Overnight AI Research (2026-07-11)

2026-07-11 Overnight AI Research

Summary

07:00 브리핑의 핵심은 tradable edge가 price direction보다 liquidity, execution, settlement, physical infrastructure의 실제 처리 능력에서 나온다는 점이다. Overnight 수집은 microstructure/execution, risk sizing, macro liquidity, crypto structure, AI infrastructure bottleneck을 중심으로 정리되었고, Henry에게는 leverage 확대보다 execution log, tail-factor sizing, liquidity indicator 구축이 우선이라는 결론이다. 특히 SOXL, MSTU, RGTI, BTC-linked exposure는 narrative상 분산되어 보여도 common liquidity shock에서는 한 덩어리 tail risk로 움직일 수 있다.

Raw Briefing

Overnight Trading Domain Research - 20260711

[01:03 KST] Robustness in Sequential Decision Making under Evolving Uncertainty: Evidence from High-Frequency Market Making

• URL: https://arxiv.org/abs/2607.08291
• Source Type: paper
• Domain: microstructure, execution, risk-sizing
• Trading Relevance: 9
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: execution, liquidity, sizing
• Why it matters: Robustness is usually treated as pure safety, but this paper splits it into uncertainty tolerance and action robustness. Henry에게 중요한 점은 defensive rule도 과하면 illiquid market에서 기회를 없애고, 부족하면 adverse fill에 노출된다는 균형이다.
• Raw excerpt / abstract: The authors study high-frequency market making under evolving uncertainty. They state that robustness has two economically meaningful dimensions: uncertainty tolerance and action robustness. Simulation and empirical evidence show action robustness has a substantially larger impact than uncertainty tolerance, and excessive robustness may reduce profitability in illiquid markets by limiting execution opportunities.
• Cold read: Trading rule의 품질은 signal accuracy만이 아니라 uncertainty에 반응하는 행동 함수의 convexity에 달려 있다. 시장이 얇을수록 보수성은 protection이면서 opportunity cost다.
• Trading insight: Henry의 큰 포지션은 entry trigger뿐 아니라 action robustness parameter를 가져야 한다. 예: liquidity가 얇으면 size를 줄이는 것과 동시에 entry patience를 늘리되, 너무 많은 confirmation 요구로 edge가 사라지는지 점검한다.
• Change sensor: bid-ask spread widening, order book depth drop, limit order non-fill ratio.
• Follow-up question: Henry의 leveraged ETF entry rule에 action robustness를 어떤 수치로 넣을 수 있는가?

[01:03 KST] Can Reinforcement Learning Efficiently Discover Price Manipulation?

• URL: https://arxiv.org/abs/2607.06121
• Source Type: paper
• Domain: microstructure, execution, ai-world-sensing
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, execution, regime
• Why it matters: RL agent가 explicit market model 없이 manipulation strategy를 발견할 수 있다면, future market structure는 사람이 보는 chart pattern보다 machine control policy에 의해 더 자주 왜곡될 수 있다. Henry는 sudden wick이나 shallow book move를 naive breakout으로 해석하면 안 된다.
• Raw excerpt / abstract: The paper compares model-based estimation and a Deep Deterministic Policy Gradient agent in an Almgren-Chriss framework with non-linear permanent impact and linear temporary impact. For intermediate volatility, the RL agent discovers profitable manipulative strategies with limited training data and outperforms the model-based approach under parameter sampling error. The authors warn about deploying learning algorithms without safeguards.
• Cold read: Market manipulation is becoming a discoverable control problem, not only a human intent problem. Regime where volatility is intermediate may be especially exploitable because price impact is readable but noise is not overwhelming.
• Trading insight: Stop placement must assume adversarial probing. Thin liquidity plus obvious stop clusters plus intermediate volatility is a red-zone condition for using market orders or tight stops.
• Change sensor: repeated wick-reversion around obvious levels, OI build without spot confirmation, abnormal impact per dollar traded.
• Follow-up question: Can we flag days when Henry’s intended stop is likely to sit in a manipulation-prone liquidity pocket?

[01:03 KST] Order Splitting and Liquidity Replenishment Are Jointly Necessary for the Square-Root Law of Market Impact

• URL: https://arxiv.org/abs/2607.04280
• Source Type: paper
• Domain: microstructure, execution
• Trading Relevance: 9
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: liquidity, execution, sizing
• Why it matters: The square-root law says impact scales non-linearly with participation rate. Henry에게 중요한 점은 large bet의 true cost가 brokerage commission이 아니라 participation, order splitting, and replenishment에 의해 결정된다는 것이다.
• Raw excerpt / abstract: The abstract tests three predictions for the square-root law of market impact using a minimal limit order book with heterogeneous interacting agents calibrated to Tokyo Stock Exchange benchmarks. It focuses on the condition that order splitting and liquidity replenishment are jointly necessary for square-root impact.
• Cold read: Impact law는 trader가 size를 늘릴 때 invisible tax처럼 작동한다. Liquidity replenishment가 충분하지 않으면 분할 주문도 impact를 제거하지 못한다.
• Trading insight: 포지션을 키울 때는 notional size보다 ADV participation과 spread-depth profile을 먼저 본다. Henry의 leveraged ETF도 open, close, macro-event window에서는 effective liquidity가 급감할 수 있다.
• Change sensor: 5-minute volume participation, top-of-book replenishment speed, realized slippage versus mid-price.
• Follow-up question: SOXL/MSTU 같은 high beta ETF에서 Henry 주문 규모별 expected implementation shortfall을 어떻게 추정할 것인가?

[01:03 KST] Is Trend Still Your Friend?: A Microstructural Account of the Demise of Short-Term Trend-Following

• URL: https://arxiv.org/abs/2607.01550
• Source Type: paper
• Domain: microstructure, execution, historical-case
• Trading Relevance: 9
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: regime, narrative, execution
• Why it matters: Short-term trend failure를 macro narrative나 crowding만으로 설명하지 않고 tick size와 HFT liquidity withdrawal로 설명한다. Henry가 breakout을 볼 때 trend signal 자체보다 그 trend가 자기강화될 microstructure가 남아 있는지 봐야 한다.
• Raw excerpt / abstract: The authors document that short-term trend returns degraded after around 2009. The central finding is that post-2008 trend PnL collapsed on small-tick contracts across signal horizons while remaining intact on large-tick contracts. They argue HFT-dominated market making withdrew liquidity in front of predictable directional flow, breaking the impact feedback loop on small-tick contracts.
• Cold read: Trend edge는 eternal chart truth가 아니라 execution ecology의 산물이다. Liquidity providers가 directional flow를 미리 피하면 trend follower의 own impact가 더 이상 trend를 sustain하지 못한다.
• Trading insight: Breakout entry는 tick size, spread, depth, and replenishment regime에 conditional해야 한다. 단순 moving average cross는 execution ecology가 바뀌면 false confidence를 준다.
• Change sensor: small-tick versus large-tick performance dispersion, post-breakout depth withdrawal, trend follow-through after first hour.
• Follow-up question: Henry의 watchlist에서 trend-following이 유효한 종목과 mean-reversion화된 종목을 어떻게 분리할 것인가?

[01:03 KST] Generating Plausible Stress Scenarios via Large Deviations

• URL: https://arxiv.org/abs/2606.31122
• Source Type: paper
• Domain: risk-sizing
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: leverage, sizing, correlation
• Why it matters: Handpicked stress scenario는 실제로 위험한 조합을 놓치거나 너무 비현실적인 shock에 attention을 낭비할 수 있다. Henry의 leveraged positions는 historical worst day만 보는 것보다 plausible path of loss를 봐야 한다.
• Raw excerpt / abstract: The paper argues that financial stress tests based on handpicked scenarios can mislead risk management. It develops a large-deviations method where, conditional on a large loss, risk factors concentrate near the most likely stress configurations. The method can generate informative stress scenarios even when historical data have few or no stressed observations.
• Cold read: Real risk is not the biggest imaginable loss; it is the most plausible path into a large loss. 큰 손실은 single shock보다 correlated moderate shocks의 sequence로 오는 경우가 많다.
• Trading insight: Henry의 pre-trade checklist에 plausible stress path를 넣는다. 예: gap down, liquidity thinning, USD strength, semis beta contraction이 동시에 오면 position heat가 어디까지 커지는지 본다.
• Change sensor: cross-asset correlation spike, volatility cluster persistence, downside gap frequency.
• Follow-up question: Henry portfolio에서 one-day shock보다 더 위험한 5-day plausible stress sequence는 무엇인가?

[01:03 KST] Fluctuations in the Treasury General Account and Their Effect on the Fed’s Balance Sheet

• URL: https://www.federalreserve.gov/econres/notes/feds-notes/fluctuations-in-the-treasury-general-account-and-their-effect-on-the-feds-balance-sheet-20250806.html
• Source Type: official
• Domain: macro-liquidity, capital-markets
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, regime, correlation
• Why it matters: TGA 변동은 reserves, ON RRP, securities, lending 중 어디선가 흡수되어야 한다. Henry가 risk-on entry를 볼 때 Fed balance sheet headline보다 reserves and money-market plumbing을 같이 봐야 한다.
• Raw excerpt / abstract: The Fed note says that when TGA fluctuates, something else must adjust on the Fed balance sheet: reserves, ON RRP balances, securities, or lending. It states reserves becoming less ample and modest ON RRP balances increase the need for active asset approaches to manage TGA fluctuations. It proposes backing the TGA with T-bills as a way to insulate reserves and policy stance.
• Cold read: Liquidity is no longer one number. With ON RRP buffer modest, TGA changes can more directly stress reserves unless offset by active asset management.
• Trading insight: Macro liquidity trigger는 WALCL minus TGA minus ON RRP만 외우는 것이 아니라, marginal buffer가 어디인지 확인해야 한다. RRP가 empty에 가까우면 same TGA rebuild가 더 큰 risk-off pressure를 만들 수 있다.
• Change sensor: TGA weekly change, ON RRP take-up, SOFR stress, reserve balances.
• Follow-up question: Henry의 leverage cap을 TGA/RRP/reserve regime에 따라 dynamic하게 줄이는 rule은 가능한가?

[01:03 KST] Do Prediction Markets Match Option Prices? Bitcoin Threshold Evidence from Binance and Polymarket

• URL: https://arxiv.org/abs/2606.19517
• Source Type: paper
• Domain: crypto-structure, microstructure
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: liquidity, narrative, execution
• Why it matters: Crypto option market and prediction market가 economically identical payoff를 다르게 price한다면, 정보 전달은 automatic하지 않다. Henry는 Polymarket odds나 options-implied probabilities를 하나만 보고 확률로 착각하면 안 된다.
• Raw excerpt / abstract: The paper compares Polymarket Yes prices with discounted risk-neutral binary values implied by Binance calls for matched Bitcoin threshold contracts. It finds a mean pricing gap of 5.6 percentage points in the main September 2023 contract and 6.3 percentage points across three compatible Bitcoin threshold markets. The gap has an AR(1) half-life of roughly four hours and is consistent with segmented venues.
• Cold read: 동일한 payoff라도 venue segmentation, participant base, and demand pressure가 persistent wedge를 만든다. Crypto market structure는 unified price discovery가 아니라 loosely connected probability islands다.
• Trading insight: 확률 기반 bet는 source diversification이 필수다. prediction market, options, perp funding, spot flow가 같은 방향일 때만 conviction을 높이고, divergence는 edge가 아니라 execution risk일 수 있다.
• Change sensor: Polymarket versus options-implied gap, funding basis, cross-venue liquidity depth.
• Follow-up question: BTC/MSTR 관련 확률 판단에서 prediction market odds를 어떤 discount로 반영해야 하는가?

[01:03 KST] The Extremity Premium: Sentiment Regimes and Adverse Selection in Cryptocurrency Markets

• URL: https://arxiv.org/abs/2602.07018
• Source Type: paper
• Domain: crypto-structure, cognition, microstructure
• Trading Relevance: 9
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: FOMO, liquidity, execution
• Why it matters: Extreme fear and extreme greed both widen spreads. Henry에게 핵심은 sentiment direction보다 intensity가 liquidity withdrawal을 만든다는 점이다. FOMO 상태와 panic 상태 모두 execution cost가 올라간다.
• Raw excerpt / abstract: Using Crypto Fear and Greed Index and Bitcoin daily data from February 2018 to January 2026, the paper documents an extremity premium: extreme fear and extreme greed regimes exhibit significantly higher spreads than neutral periods. It reports within-volatility-quintile premium with p less than 0.001, strong Granger causality from uncertainty to spreads, and replication on Ethereum and across six of seven cycles.
• Cold read: Crowd emotion is not just directional signal; it is a liquidity state variable. Extremity makes makers withdraw or charge more, increasing adverse selection for late entrants.
• Trading insight: Henry가 강한 narrative를 느낄 때는 conviction을 올리기 전에 spread, slippage, and volatility-adjusted size를 낮춰야 한다. Greed extreme에서 breakout buy는 좋은 idea라도 bad execution이 될 수 있다.
• Change sensor: Fear and Greed extremity, spread widening after volatility control, perp funding crowding.
• Follow-up question: Henry의 FOMO checklist에 sentiment intensity discount를 어떻게 수치화할까?

[01:03 KST] 2026 Crypto Market Outlook

• URL: https://www.coinbase.com/institutional/research-insights/research/market-intelligence/2026-crypto-market-outlook
• Source Type: research
• Domain: crypto-structure, capital-markets
• Trading Relevance: 7
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative, regime, correlation
• Why it matters: Coinbase frames 2026 crypto around regulation, institutional integration, tokenization, stablecoins, and market structure. 이것은 pure retail cycle보다 core financial system integration cycle에 가까운 thesis지만, 동시에 institutional narrative crowding risk가 커진다.
• Raw excerpt / abstract: Coinbase states that crypto markets are poised for transformative growth in 2026 as clearer regulation and accelerating institutional integration deepen crypto’s role in the core financial system. It highlights macro landscape, technological transformations, regulatory progress, tokenization, stablecoins, and market structure. It also says stablecoins have become the number one use case and models a total stablecoin market cap target range centered around $1.2T by the end of 2028.
• Cold read: Crypto의 next regime은 price cycle보다 rails adoption, collateral usage, and regulated access의 regime일 수 있다. 하지만 institutionalization also means correlation with liquidity and policy becomes stronger.
• Trading insight: BTC exposure 판단에서 ETF flow만 보지 말고 stablecoin supply, tokenized collateral, and regulatory access를 함께 본다. MSTR reflexivity는 BTC price뿐 아니라 capital market appetite와 연결된다.
• Change sensor: stablecoin market cap growth, tokenized equity volume, ETF net flows, Coinbase institutional product launches.
• Follow-up question: Stablecoin growth가 BTC beta보다 broader crypto liquidity를 더 잘 설명하는 구간은 언제인가?

[01:03 KST] Frontiers of Compute: The Technologies to Reduce AI Inference Costs

• URL: https://www.mckinsey.com/industries/semiconductors/our-insights/frontiers-of-compute-the-technologies-to-reduce-ai-inference-costs
• Source Type: research
• Domain: ai-world-sensing, capital-markets
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative, regime, correlation
• Why it matters: AI capex thesis의 핵심은 smarter model보다 cheaper token일 수 있다. Henry가 semis and AI infra exposure를 볼 때, demand story와 함께 cost-per-token, energy-per-token, and margin sustainability를 확인해야 한다.
• Raw excerpt / abstract: McKinsey writes that Amazon, Google, Meta, and Microsoft are collectively committing over $700B in combined capital expenditure in 2026, with a substantial majority directed at AI infrastructure. It says inference is shifting to become a major compute spending driver and that sustainable margins require inference costs to decline by several multiples. It identifies model optimization, advanced packaging, custom silicon, and co-packaged optics as high-impact levers, with top levers approaching an order-of-magnitude reduction in cost per token and combined levers potentially reducing costs by up to two orders of magnitude.
• Cold read: AI infra는 bubble or boom의 단순 이분법이 아니다. Unit economics가 내려가면 demand elasticity가 capex를 정당화할 수 있고, 비용 하락이 느리면 capex 부담이 equity narrative를 압박한다.
• Trading insight: SOXL/NXE/AI infrastructure 관련 risk는 revenue growth뿐 아니라 power and token economics다. AI trade conviction은 inference cost curve가 실제로 하락하는 evidence와 같이 키워야 한다.
• Change sensor: cost per token, energy per token, hyperscaler capex guidance, custom silicon adoption, HBM and optical interconnect bottlenecks.
• Follow-up question: AI infra winners를 chip headline이 아니라 cost-per-token bottleneck별로 map하면 Henry watchlist는 어떻게 달라지는가?

[01:03 KST] 2026 Data Center Marketplace Report

• URL: https://www.colliers.com/en/research/nrep-usdc-data-center-marketplace-2026
• Source Type: research
• Domain: ai-world-sensing, capital-markets
• Trading Relevance: 7
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: liquidity, narrative, execution
• Why it matters: AI data center theme는 chip demand만이 아니라 power scarcity and capital structure problem이다. Henry가 AI infra narrative에 올라탈 때는 physical bottleneck과 financing bottleneck을 같이 봐야 한다.
• Raw excerpt / abstract: Colliers states AI-driven demand is rapidly reshaping data centers into a phase defined by power scarcity, rising capital intensity, and infrastructure-scale execution. The report highlights over $120B of hyperscaler debt issued to fund AI infrastructure, a 47 percent year-over-year increase in build costs driven by power intensity and complexity, 15.6 GW of North America absorption, more than 90 percent of new capacity pre-leased before delivery, and 40 to 50 percent of total project costs attributable to power infrastructure.
• Cold read: Constraint shifted from land and servers to grid access, permitting, and balance sheet strength. This makes AI infra less like software scaling and more like energy and project finance.
• Trading insight: AI capex beneficiaries can diverge sharply: companies with power access, financing capacity, and execution credibility deserve premium; narrative-only suppliers face disappointment risk.
• Change sensor: grid interconnection backlog, data center pre-lease rates, utility deposits, hyperscaler debt issuance, build cost inflation.
• Follow-up question: NXE-like energy exposure should be treated as AI derivative or separate commodity-cycle exposure under Henry risk budget?

Run Summary [01:03 KST]

• collected_count: 11
• skipped_duplicates: 0
• domain_mix: microstructure/execution 45%, risk 15%, macro-liquidity/capital-markets 20%, crypto-structure 25%, AI-history/world-sensing 20%. Some items map to multiple domains, so total exceeds 100%.
• top_theme: Execution edge is becoming state-dependent: liquidity, adversarial learning, sentiment extremity, and physical infrastructure constraints all change whether a signal can be monetized.
• sharpest_insight: The same idea can be right and still lose if the execution ecology is wrong. Robustness, trend, crypto probability, and AI capex all require checking the market mechanism that converts thesis into realized return.
• danger_of_misuse: Do not convert these papers into immediate buy or sell decisions. The misuse risk is overfitting a new academic mechanism to one trade, especially using RL manipulation or sentiment extremes as a reason to avoid all stops or chase narrative moves.
• next_probe: Build a practical Henry execution dashboard: spread, depth, ADV participation, sentiment extremity, TGA/RRP liquidity, and AI cost-per-token signals that map to position sizing rather than price prediction.

[02:00 KST] Volatility in Prediction Markets: A Structural Approach

• URL: https://arxiv.org/abs/2607.08199
• Source Type: paper
• Domain: microstructure, crypto-structure, risk-sizing
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: liquidity, narrative, execution
• Why it matters: Prediction market price is bounded probability, not a normal asset price. Henry가 Polymarket or event odds를 volatility input처럼 쓸 때, deadline resolution and informed order flow가 동시에 volatility를 만든다는 점을 분리해야 한다.
• Raw excerpt / abstract: The paper says prediction market prices are bounded probabilities, payoffs are binary, and contracts resolve at known deadlines. It develops a model combining a Wright-Fisher deadline-resolution component with a Glosten-Milgrom order-flow component that captures volatility from informed trading as reflected in spreads and volume.
• Cold read: Prediction markets are not simple crowd wisdom. Time-to-resolution mechanically forces uncertainty to collapse, while order-flow imbalance can signal informed trading or liquidity stress.
• Trading insight: Event odds should be discounted when spreads widen or volume spikes near the deadline. A high-confidence event price can still be a poor execution signal if the volatility source is adverse selection rather than new information.
• Change sensor: spread and volume around event deadlines, odds jump without external news, divergence between prediction market odds and options-implied probabilities.
• Follow-up question: BTC-related prediction market odds should be mapped to Henry’s conviction score with what spread and deadline discount?

[02:00 KST] Square-Root Price Impact Is Necessary for Endogenous Manipulation Cycles in Learning-Agent Markets

• URL: https://arxiv.org/abs/2607.05141
• Source Type: paper
• Domain: microstructure, reflexivity, cognition
• Trading Relevance: 9
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: liquidity, FOMO, narrative, execution
• Why it matters: Herding retail plus a capitalized optimizing agent can create self-sustaining cycles without a new fundamental catalyst. Henry에게 중요한 점은 chart acceleration itself can be manufactured by impact and feedback, especially when retail follows recent price.
• Raw excerpt / abstract: The authors model one evolutionary-optimized institutional agent interacting with 20,000 herding retail traders. The agent discovers a multi-cycle predatory strategy, producing 8-11 cycles over 2000 trading days, and the authors argue position-tracking feedback coupled with square-root price impact creates a self-sustained nonlinear oscillator.
• Cold read: A market can become a machine that generates its own narratives. Square-root impact makes large trades powerful enough to move price, and herding converts price movement into demand.
• Trading insight: In fast narrative names, avoid treating repeated cycles as independent confirmation. If the same feedback loop is driving price, adding size late increases exposure to the engineered reversal phase.
• Change sensor: parabolic retail volume, social mention acceleration, large impact per dollar, repeated pump-reversal cycles without fresh information.
• Follow-up question: How can Henry label a move as fundamental re-rating versus impact-herding oscillator before increasing size?

[02:00 KST] A Limit Order Market with Uncertain Informed Trading Participation

• URL: https://arxiv.org/abs/2607.04221
• Source Type: paper
• Domain: microstructure, execution
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: liquidity, execution, sizing
• Why it matters: Liquidity suppliers do not only price the average probability of informed traders; they price uncertainty about how many informed traders are actually present. Henry의 large order는 news day, macro event, or crowded theme에서 more toxic flow로 해석될 수 있다.
• Raw excerpt / abstract: The paper studies a one-period limit order market where informed trader count is random. Liquidity suppliers know only the distribution and face uncertainty about both presence and intensity of informed trading. Large order asymptotics show price impact exponents are determined jointly by asset value tail and the full distribution of informed trader count, not only the expected number.
• Cold read: Market depth is a belief about toxicity. When suppliers cannot tell whether an order is liquidity demand or informed demand, they demand a larger impact premium.
• Trading insight: Size should be cut when the market cannot distinguish Henry’s order from informed flow. This is especially relevant in thin premarket, post-news, and single-theme momentum environments.
• Change sensor: spread response to same-sized prints, depth retreat after large trades, abnormal dark or lit venue imbalance.
• Follow-up question: Can Henry’s execution checklist include a proxy for informed-flow uncertainty before placing a large order?

[02:00 KST] A Gabor—Epps Uncertainty Principle for Traders

• URL: https://arxiv.org/abs/2607.04130
• Source Type: paper
• Domain: microstructure, execution, risk-sizing
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: correlation, execution, regime
• Why it matters: High-frequency correlation is not observed in clock time alone; it depends on market activity, order-flow overlap, and finite coupling response. Henry의 correlation hedge can fail if measured at the wrong time scale.
• Raw excerpt / abstract: The paper proposes that at event scales, the more precisely one localizes market activity in time, the less well one can resolve stable cross-asset dependence; the more one resolves dependence, the more one coarse-grains away event-time structure.
• Cold read: Correlation is a measurement artifact unless the time basis is specified. Calendar time, trade time, and volume time can tell different truths.
• Trading insight: Do not size hedges or pairs by a single intraday correlation number. For SOXL/NXE/MSTU-style risk, the hedge horizon must match the liquidation horizon and the liquidity clock.
• Change sensor: correlation breakdown across calendar time versus volume time, cross-asset lead-lag shifts, hedge slippage during fast sessions.
• Follow-up question: Which time basis should Henry use for leveraged ETF correlation heat: clock time, trade count, or volume bucket?

[02:00 KST] Exact Conditional Simulation of Point Processes: Application to Pathwise Market Impact Estimation

• URL: https://arxiv.org/abs/2607.03239
• Source Type: paper
• Domain: execution, microstructure
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: execution, liquidity, sizing
• Why it matters: Execution quality requires a counterfactual: what would price have done without my order? Henry의 realized slippage만 보면 market move와 own impact를 혼동할 수 있다.
• Raw excerpt / abstract: The paper defines market impact as the difference between the observed price trajectory under an execution strategy and the counterfactual trajectory that would have prevailed without it. It proposes exact conditional simulation of point processes to reconstruct counterfactual paths on common randomness for aggressive, passive, and mixed strategies.
• Cold read: Impact is not directly visible. Good execution analytics need a controlled alternative path, not just average fill price.
• Trading insight: After meaningful orders, log arrival price, fill path, concurrent market move, and post-trade drift separately. Without this, Henry may reward bad execution in a rising market or punish good execution in a falling market.
• Change sensor: realized shortfall versus post-trade drift, non-fill opportunity cost, adverse selection after passive fills.
• Follow-up question: What minimal order log fields are needed to estimate Henry’s own implementation shortfall without institutional data?

[02:00 KST] Liquidity-Based Audit of Algorithmic Trading Strategies

• URL: https://arxiv.org/abs/2606.29018
• Source Type: paper
• Domain: execution, risk-sizing, microstructure
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: liquidity, sizing, correlation
• Why it matters: A strategy can look profitable while being a net liquidity consumer that becomes fragile when many similar strategies crowd together. Henry의 process should identify whether his entries demand liquidity at the same time as everyone else.
• Raw excerpt / abstract: The authors state net demand for liquidity by algo strategies is identifiable from trade and price history alone, without knowing the signal. A multi-period regret decomposition classifies a strategy as net liquidity consumer or provider. Aggregating across correlated strategies yields a liquidity-balance condition whose violation produces welfare loss scaling as N squared, a fire-sale externality.
• Cold read: Crowding is not just holding the same asset; it is demanding liquidity in the same direction at the same time.
• Trading insight: Momentum entries, stops, and leveraged ETF rebalancing can all synchronize liquidity demand. Henry should reduce size when his planned order aligns with crowded mechanical flows.
• Change sensor: same-direction close imbalance, ETF rebalance pressure, spike in liquidity-taking volume, correlated stop-out patterns.
• Follow-up question: Can Henry’s trade journal classify each order as liquidity-taking, liquidity-providing, or mixed?

[02:00 KST] Innovating Risk Modelling for Global Funds

• URL: https://arxiv.org/abs/2607.07465
• Source Type: paper
• Domain: risk-sizing, capital-markets
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: correlation, sizing, regime
• Why it matters: Benchmark-relative risk can hide the true risk of a concentrated, innovation-heavy book. Henry’s account is not an index fund, so risk should be measured internally by covariance and shared drivers.
• Raw excerpt / abstract: The paper argues Markowitz originally defined portfolio risk as an internal covariance property, but CAPM, style models, and benchmark deviation shifted risk into distance from an index. For a global book concentrated in a few markets and innovation sectors, active risk measures mismatch rather than true risk. It proposes recovering portfolio structure directly with PCA.
• Cold read: The relevant question is not whether Henry differs from an index. The relevant question is whether multiple holdings are secretly the same bet under stress.
• Trading insight: SOXL, RGTI, NXE, MSTU, BTC-linked exposures can share liquidity, AI narrative, and risk-on beta. Position limits should account for common factor heat, not ticker count.
• Change sensor: first principal component dominance, correlation rising during selloffs, shared sensitivity to rates and AI capex narrative.
• Follow-up question: What is Henry’s true one-factor exposure if each holding is decomposed into liquidity beta, AI beta, BTC beta, and leverage beta?

[02:00 KST] Split-Session Cluster GARCH for Overnight and Intraday Returns: The Role of Tail Heterogeneity

• URL: https://arxiv.org/abs/2607.03669
• Source Type: paper
• Domain: risk-sizing, macro-liquidity
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: leverage, correlation, sizing
• Why it matters: Overnight and intraday returns have different tail behavior. Henry의 Korean CMA account faces U.S. session risk plus overnight gap risk, so one volatility estimate can understate the danger.
• Raw excerpt / abstract: The paper proposes a Split-Session Cluster GARCH model for heavy-tailed multivariate dependence among returns decomposed into overnight and intraday components. It allows tail behavior to differ by trading session and sector, and finds pronounced tail heterogeneity between overnight and intraday returns in U.S. equities.
• Cold read: Risk is session-dependent. Overnight tails are not just intraday volatility with fewer observations.
• Trading insight: Position size should be lower when a stop cannot be executed through the actual risk window. For leveraged ETFs, gap risk can dominate normal intraday stop logic.
• Change sensor: overnight gap frequency, earnings or macro release clustering, session-specific volatility ratio, sector-level tail co-movement.
• Follow-up question: Should Henry use separate R multiples for intraday-managed risk and overnight-held risk?

[02:00 KST] Hidden Dependence and Aggregate Tail Risk

• URL: https://arxiv.org/abs/2606.30193
• Source Type: paper
• Domain: risk-sizing
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: correlation, leverage, sizing
• Why it matters: Marginal distributions can look safe while hidden dependence creates worst-case aggregate tail loss. Henry should not assume diversification from separate tickers if tail events line up.
• Raw excerpt / abstract: The paper studies risk aggregation under dependence uncertainty. It introduces hidden dependence based on risk concentration and common tail events, showing small perturbations of a reference distribution can be compatible with hidden dependence and lead to worst-case aggregate risk.
• Cold read: The portfolio can be more correlated in the tail than in normal data. Stress dependence may be invisible until it matters.
• Trading insight: Leverage cap should respond to tail-dependence risk, not only realized correlation. If AI, semis, uranium, and BTC all depend on liquidity, one macro shock can compress the entire book.
• Change sensor: downside correlation spikes, cross-asset vol-of-vol, credit spread widening, DXY and real-rate shock overlap.
• Follow-up question: What hidden common tail event would hurt every Henry position at the same time?

[02:00 KST] May 2026 Federal Reserve Balance Sheet Developments

• URL: https://www.federalreserve.gov/monetarypolicy/May-2026-Federal-Reserve-Balance-Sheet-Developments.htm
• Source Type: official
• Domain: macro-liquidity, capital-markets
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, regime, correlation
• Why it matters: The report confirms that ON RRP is very close to zero and TGA changes now matter more as a balance sheet swing variable. Henry’s risk-on exposure should distinguish flow liquidity from reserve adequacy.
• Raw excerpt / abstract: The Fed states total assets rose to $6.7TasofMarch25,2026.Liabilitiesincreasedto$6.6T. ON RRP take-up fell $28B,reachingveryclosetozero.TheTreasuryGeneralAccountgrew$79B, Federal Reserve notes rose $36B,andthesechangescontributedtoa$36B net increase in reserve balances.
• Cold read: The old RRP buffer is nearly gone. Balance sheet plumbing is now more about reserve management purchases, TGA swings, and money market alternatives than a simple net liquidity formula.
• Trading insight: Use macro liquidity as a sizing regime, not a directional price call. When RRP buffer is exhausted, TGA rebuild or funding stress can reduce tolerance for leveraged positions even if equity trend remains strong.
• Change sensor: ON RRP near zero, TGA weekly change, reserve balances, SOFR pressure, bill yields versus administered rates.
• Follow-up question: At what reserve and SOFR stress threshold should Henry automatically reduce portfolio heat?

[02:00 KST] Settlement Manipulation in Prediction Markets

• URL: https://arxiv.org/abs/2606.31675
• Source Type: paper
• Domain: microstructure, crypto-structure, reflexivity
• Trading Relevance: 9
• Time Horizon: intraday
• Actionability: monitor
• Henry Risk Link: liquidity, execution, narrative
• Why it matters: If a prediction market contract settles on an asset price that traders can move, the derivative can contaminate the underlying. Henry should treat five-minute BTC event contracts as possible order-flow distortion, not clean information.
• Raw excerpt / abstract: The paper models prediction contracts settling on an asset price that holders can move by trading the underlying. After Polymarket launched five-minute Bitcoin contracts, settlement-time spot order flow spiked and caused large price reversals after settlement. The authors find manipulation largely absent in fifteen-minute contracts, suggesting longer horizons as a market-design remedy.
• Cold read: A contract designed to predict price can become an incentive to move price. Market structure turns information venue into manipulation venue when settlement is too short and underlying liquidity is movable.
• Trading insight: Avoid reading spot wicks around prediction-market settlement as normal BTC price discovery. Execution should pause or use wider controls around known settlement windows.
• Change sensor: settlement-time spot volume spikes, rapid post-settlement reversal, open interest in short-horizon event contracts, venue-specific BTC threshold clustering.
• Follow-up question: Which BTC-linked products Henry watches are exposed to event-settlement distortion rather than organic flow?

[02:00 KST] The Inference-Compute Frontier and a Latency-Efficient Architecture for Limit Order Book Prediction

• URL: https://arxiv.org/abs/2606.25986
• Source Type: paper
• Domain: microstructure, execution, ai-world-sensing
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: study
• Henry Risk Link: execution, regime, liquidity
• Why it matters: AI performance in order book prediction is constrained by compute and latency, not only model accuracy. Henry should expect professional execution systems to improve fastest where latency-friendly architectures convert prediction into fill quality.
• Raw excerpt / abstract: The paper tests whether a scaling-law-style inference-compute frontier appears in LOB prediction. It finds predictive loss versus structural forward work is summarized by a power law with R squared of 0.941 on an excluded high-compute target frontier. The latency-space relationship is weaker, showing latency is not merely noisy compute, and motivating a hardware-friendly FastBiNLOB architecture.
• Cold read: Better market AI is becoming an engineering frontier. The edge belongs to systems that jointly optimize prediction loss, latency, and hardware path.
• Trading insight: Retail chart signals may decay faster in liquid instruments as professional systems improve book-level prediction. Henry’s edge must shift toward timeframe selection, risk control, and waiting for situations where execution quality can be managed.
• Change sensor: latency-sensitive LOB model releases, exchange colocation upgrades, spread compression in liquid names, worse retail fill quality during high-speed regimes.
• Follow-up question: Which Henry trades rely on patterns that professional LOB models are most likely to arbitrage away?

Run Summary [02:00 KST]

• collected_count: 12
• skipped_duplicates: 11
• domain_mix: microstructure/execution 58%, risk-sizing 33%, macro-liquidity/capital-markets 8%, crypto-structure/reflexivity 17%, AI-world-sensing 8%. Some items map to multiple domains, so total exceeds 100%.
• top_theme: The same signal can become dangerous when the mechanism that converts it into price is adversarial, crowded, or settlement-driven.
• sharpest_insight: Risk is increasingly hidden in the conversion layer: event odds to volatility, orders to impact, benchmark risk to true covariance, and prediction markets to underlying spot manipulation.
• danger_of_misuse: Do not use these papers to avoid all trading or to assume every wick is manipulation. The correct use is to add sizing, liquidity, and execution filters before acting on a thesis.
• next_probe: Convert today’s mechanisms into a practical pre-trade risk filter: liquidity-taking classification, hidden common factor heat, session-specific tail risk, and settlement-window distortion.

[03:01 KST] Outcome-Classified Precision Auditing of Filter Rules in Algorithmic DEX Trading

• URL: https://arxiv.org/abs/2607.02830
• Source Type: paper
• Domain: execution, crypto-structure, risk-sizing
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: execution, sizing, FOMO
• Why it matters: 이 논문은 DEX filter가 실제로 손실을 막았는지 사후 관찰로 audit한다. Henry에게 핵심은 filter의 목적이 missed winner를 줄이는 것이 아니라, 나쁜 trade를 얼마나 검증 가능하게 걸렀는지 측정하는 데 있다.
• Raw excerpt / abstract: The audit yielded 99,510 follow-up samples across 2,402 unique rejection events spanning eight active filter rules. The conservative save-to-miss ratio from measured-drawdown saves is 3.7 : 1, while a wider 14.8 : 1 interpretation did not survive matched-comparison testing.
• Cold read: Risk filter는 느낌이 아니라 rejection ledger로 검증해야 한다. 넓게 정의한 success tier는 보기 좋지만, matched lifecycle data가 받쳐주지 않으면 false comfort다.
• Trading insight: Henry의 pre-trade checklist도 rejected trade를 기록해야 한다. 실행하지 않은 trade의 이후 경로를 추적해야 FOMO와 over-filtering 중 어느 쪽이 문제인지 알 수 있다.
• Change sensor: rejected trade follow-up return, drawdown saved versus winner missed, filter별 false reject rate.
• Follow-up question: Henry의 손절, liquidity, narrative filter 각각에 대해 save-to-miss ratio를 어떻게 산출할까?

[03:01 KST] Pump.fun Graduation Regime Windows: Survival Analysis of 832,941 Token Launches

• URL: https://arxiv.org/abs/2607.02823
• Source Type: paper
• Domain: crypto-structure, microstructure, cognition
• Trading Relevance: 7
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: FOMO, narrative, liquidity
• Why it matters: Meme-token microstructure는 extreme survivorship bias의 실험실이다. Henry가 high-volatility theme를 볼 때 visible winners가 전체 launch population의 극소수라는 base rate를 잊으면 안 된다.
• Raw excerpt / abstract: The study observes 832,941 Solana token launches and reports a pooled 24-hour graduation rate of 0.198%, down 3.18x from a prior 0.63% rate. Telegram-advertised launches graduated at 1.485% versus 0.166% without, and all-three-social-channel launches graduated at 1.919% versus 0.110% with none.
• Cold read: Social presence raises survival odds but the absolute base rate remains tiny. Narrative visibility is a conditional filter, not proof of durable value.
• Trading insight: In reflexive assets, do not confuse social proof with edge. Social signal can improve odds but also crowds entry and worsens execution when everyone sees the same cue.
• Change sensor: launch survival rate, social-channel lift decay, creator self-buy proxy, first-hour liquidity concentration.
• Follow-up question: Equity and crypto narrative trades에서 social presence가 absolute base rate를 얼마나 바꾸는지 어떻게 측정할까?

[03:01 KST] Coordinated Sniper Cohorts on Pump.fun

• URL: https://arxiv.org/abs/2607.02795
• Source Type: paper
• Domain: crypto-structure, microstructure, cognition
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: FOMO, liquidity, execution
• Why it matters: 초기 wallet flow를 smart money로 해석하는 습관에 제동을 건다. 논문은 coordinated cohort가 보이더라도 placebo가 더 큰 lift를 보였기 때문에, naive causal inference가 위험하다고 말한다.
• Raw excerpt / abstract: A two-stage pipeline identifies 1,012 persistent wallet cohorts across 166,098 launches. Cohort-touched launches show +132.3% first-30-minute buyer count lift, but an activity-matched placebo shows a larger +216.3% lift, refuting a strong cohort-specific causal interpretation.
• Cold read: Early buyer clustering may be selection, not causation. Active wallets appear where flow was already likely to concentrate.
• Trading insight: Whale or smart-wallet tracking은 causal proof가 아니라 hypothesis다. Henry는 큰 주소 움직임을 보고 진입하기 전에 launch quality, liquidity, and independent catalyst를 같이 확인해야 한다.
• Change sensor: wallet cohort overlap, activity-matched placebo performance, first-hour inflow persistence after initial spike.
• Follow-up question: MSTR/BTC or small-cap AI names에서 smart money narrative를 placebo control 없이 믿고 있지는 않은가?

[03:01 KST] Liquidity Premium and Investment Horizons

• URL: https://arxiv.org/abs/2607.01377
• Source Type: paper
• Domain: microstructure, execution, capital-markets
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: liquidity, execution, sizing
• Why it matters: Liquidity premium을 risk compensation이 아니라 adverse-selection mechanism으로 설명한다. Henry에게 중요한 점은 낮은 order flow가 가격을 싸게 만드는 동시에 execution risk도 키운다는 양면성이다.
• Raw excerpt / abstract: The paper estimates Kyle’s price-impact coefficient from daily equity order flow. Signed order flow predicts contemporaneous and one-month-ahead returns, while volume volatility predicts lower subsequent returns, consistent with widening price impact degrading price discovery.
• Cold read: Illiquidity can create return opportunity but it is not free alpha. Price discovery가 느려지고 Kyle lambda가 커지면 진입과 청산 모두 더 비싸진다.
• Trading insight: Cheap-looking pullback은 liquidity premium인지 falling knife인지 분리해야 한다. Low order flow normalization이 보이지 않으면 size를 줄이고 confirmation을 더 요구한다.
• Change sensor: Kyle lambda proxy, signed order flow persistence, volume volatility, next-month reversal after liquidity normalization.
• Follow-up question: Henry watchlist에서 illiquidity discount와 genuine deterioration을 구분하는 최소 지표 세트는 무엇인가?

[03:01 KST] When Large Trades Are Not News: Liquidity Tail Risk and Price Discovery

• URL: https://arxiv.org/abs/2607.01198
• Source Type: paper
• Domain: microstructure, execution, risk-sizing
• Trading Relevance: 9
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: liquidity, execution, sizing
• Why it matters: 큰 trade를 정보로 볼지 liquidity shock으로 볼지 구분하는 문제다. Henry가 sudden block, wick, volume spike를 볼 때 즉시 news로 해석하면 잘못된 방향으로 추격할 수 있다.
• Raw excerpt / abstract: Liquidity suppliers observe aggregate order flow but not its decomposition into informed demand and uninformed liquidity demand. With heavy-tailed liquidity demand, large imbalances remain plausibly liquidity-driven, flattening and concavifying price impact, slowing learning from order flow, and delaying the decline of adverse-selection premia.
• Cold read: Large order impact의 의미는 tail regime에 따라 바뀐다. Heavy-tailed liquidity demand에서는 큰 print가 정보보다 forced flow일 가능성이 오래 남는다.
• Trading insight: Henry는 large print 후 entry 전에 follow-through, spread resilience, and replenishment를 본다. Forced liquidity move라면 mean reversion 가능성이 크고, informed trade라면 adverse selection이 지속된다.
• Change sensor: large trade follow-through, post-print spread persistence, depth replenishment, price impact convexity.
• Follow-up question: SOXL/MSTU에서 큰 거래량 spike 후 15분 내 follow-through rule은 어떻게 만들까?

[03:01 KST] End-to-End Parametric Portfolio Policies for Cross-Asset Futures Timing

• URL: https://arxiv.org/abs/2607.00475
• Source Type: paper
• Domain: risk-sizing, execution, ai-world-sensing
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing, regime, execution
• Why it matters: AI model이 simple rules를 이기는 조건이 transaction cost까지 포함하면 달라진다. Henry의 자동화나 quantitative trigger도 gross Sharpe보다 turnover-adjusted survival을 먼저 봐야 한다.
• Raw excerpt / abstract: The authors train end-to-end AI policies on sixteen liquid CME futures with differentiable Sharpe loss. LSTM and transformer perform comparably gross, but diverge after transaction costs; the transformer trades far less and matches or exceeds equal weighting through moderate cost.
• Cold read: Model sophistication보다 trading frequency and cost discipline이 성과를 가른다. 좋은 forecast가 너무 많이 거래하면 edge가 cost로 사라진다.
• Trading insight: Henry의 rule은 signal count가 아니라 net execution quality로 평가해야 한다. High-confidence but low-turnover policy가 low-confidence high-turnover policy보다 계좌 생존에 유리할 수 있다.
• Change sensor: turnover, implementation shortfall, gross-to-net Sharpe decay, signal persistence by asset class.
• Follow-up question: Henry의 discretionary entries를 turnover-adjusted rule로 환산하면 어떤 trade가 제거될까?

[03:01 KST] Heads, Not Backbones: Output Heads Dominate Architectures on Fat-Tailed Returns

• URL: https://arxiv.org/abs/2606.30037
• Source Type: paper
• Domain: risk-sizing, ai-world-sensing
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing, leverage, correlation
• Why it matters: Financial return forecasting에서 backbone보다 output distribution이 tail risk를 더 잘 잡을 수 있다. Henry에게 핵심은 point forecast보다 distribution forecast가 position sizing에 더 중요하다는 점이다.
• Raw excerpt / abstract: On S and P 500 monthly log returns from 1871 to 2023, switching from point to Gaussian improves CRPS by about 1.3%, and switching from Gaussian to mixture adds about 2.4%. Mixture head value is largest in high-volatility regimes, with 13.9% improvement in 1970s stagflation at horizon 12.
• Cold read: Crisis risk는 average error metric에 잘 안 보인다. Distribution head가 tail and coverage를 개선할 때 risk management value가 생긴다.
• Trading insight: Henry의 sizing model은 target price 하나보다 downside distribution을 입력받아야 한다. 특히 leverage ETF는 tail mixture를 무시하면 손절 gap과 volatility decay를 과소평가한다.
• Change sensor: forecast coverage, pinball loss, volatility regime별 tail calibration, Gaussian versus mixture downside estimate.
• Follow-up question: Henry 포트폴리오의 worst 5% path를 single Gaussian이 얼마나 과소평가하는가?

[03:01 KST] Robust Hedging Valuation Adjustment under Liquidity-Demand Stress

• URL: https://arxiv.org/abs/2606.26731
• Source Type: paper
• Domain: risk-sizing, execution
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: execution, sizing, liquidity
• Why it matters: Rebalancing을 자주 하면 cost가 늘고, 너무 드물게 하면 hedge error가 커진다. Henry의 stop adjustment, tranche entry, and rebalance rule도 같은 trade-off를 가진다.
• Raw excerpt / abstract: The paper defines robust HVA as worst-case expected loss over a relative-entropy neighborhood of simulated rebalancing and maturity-unwind loss distributions. It distinguishes fixed-radius from fixed benchmark-stress conventions and shows wider no-trade bands lower rebalancing costs but raise hedge-error risk.
• Cold read: No-trade band는 게으름이 아니라 cost와 hedge error 사이의 명시적 선택이다. Stress convention을 고정하지 않으면 서로 다른 band의 risk를 공정하게 비교하기 어렵다.
• Trading insight: Henry가 분할 매수 후 리밸런싱할 때 every tick 반응은 cost trap이고, 너무 넓은 band는 gap risk를 키운다. Band width는 volatility, liquidity, and account heat에 따라 정해야 한다.
• Change sensor: rebalance frequency, hedge error, transaction cost per adjustment, liquidity stress around rebalance windows.
• Follow-up question: SOXL/MSTU tranche management에서 no-trade band를 ATR 몇 배로 둘 때 net R이 가장 안정적인가?

[03:01 KST] High Public Debt and Shifting Financial Markets: Challenges for Central Banks

• URL: https://www.bis.org/publ/arpdf/ar2026e2.htm
• Source Type: official
• Domain: macro-liquidity, capital-markets, risk-sizing
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, regime, correlation
• Why it matters: BIS는 high public debt와 NBFI sovereign bond intermediation이 새로운 fiscal-financial stability nexus를 만든다고 본다. Henry의 leveraged risk-on book은 equity story만이 아니라 sovereign liquidity shock에 의해 동시에 de-risk될 수 있다.
• Raw excerpt / abstract: BIS says central banks face mounting challenges from near record-high public debt and the growing role of non-banks in sovereign debt markets. NBFIs’ share of AE sovereign debt holdings rose from 44% in 2021 to 53% in 2025, and around 70% of bilateral USD repos with hedge funds are at zero haircuts. BIS estimates the probability of a GFC-like stress event within three months is about 10 times higher when public debt to GDP is high, approximately 3.8% versus 0.3%.
• Cold read: Fiscal space is now partly a market microstructure variable. Debt sustainability can deteriorate through dealer balance sheet constraints, repo haircuts, volatility, and NBFI deleveraging before fundamentals visibly break.
• Trading insight: Macro liquidity should cap leverage before price breaks. If repo funding, Treasury liquidity, or sovereign stress worsens, Henry should reduce portfolio heat even if AI or BTC narratives remain bullish.
• Change sensor: repo haircuts, Treasury market depth, NBFI leverage, dealer balance sheet usage, fiscal surprise impact in illiquid sessions.
• Follow-up question: Henry의 leverage cap을 Treasury liquidity and repo stress regime에 연결하면 threshold는 무엇인가?

[03:01 KST] The Evolution of Central Banks’ Lending Operations

• URL: https://www.bis.org/publ/qtrpdf/r_qt2606b.htm
• Source Type: official
• Domain: macro-liquidity, capital-markets, execution
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, regime, execution
• Why it matters: 중앙은행 liquidity toolkit이 smaller balance sheets and NBFI footprint에 맞춰 재설계되고 있다. Henry는 central bank backstop을 단순 bullish put으로 보지 말고, who has access, what collateral, what stigma, and what price를 봐야 한다.
• Raw excerpt / abstract: BIS notes that as major central banks reduce balance sheets, lending operations become more important for meeting reserve demand and curbing money market volatility. The updated Compendium highlights different approaches to NBFI access, collateral frameworks, pricing, and disclosure. The Bank of England introduced a contingent NBFI repo facility, while the Federal Reserve removed the aggregate operational limit for standing repo operations in 2025.
• Cold read: Liquidity backstop quality is operational, not rhetorical. A facility that is stigmatized, expensive, or inaccessible to the stressed holder may not stabilize markets in time.
• Trading insight: During stress, do not assume all liquidity facilities transmit equally to risk assets. Henry should track whether stressed entities can actually convert collateral into cash without fire sales.
• Change sensor: standing repo usage, discount window stigma indicators, eligible collateral changes, NBFI access expansion, money-market rate volatility.
• Follow-up question: Which macro liquidity indicators prove the backstop is being used early enough rather than after forced selling?

[03:01 KST] Economy - The 2026 AI Index Report

• URL: https://hai.stanford.edu/ai-index/2026-ai-index-report/economy
• Source Type: research
• Domain: ai-world-sensing, capital-markets, historical-case
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative, regime, correlation
• Why it matters: AI investment, adoption, labor displacement, and robotics are moving at different speeds. Henry의 AI exposure는 capex headline만이 아니라 revenue scale, consumer surplus, adoption saturation, and labor-market friction까지 봐야 한다.
• Raw excerpt / abstract: Stanford HAI reports global corporate AI investment more than doubled in 2025, private investment grew 127.5%, generative AI funding grew more than 200%, and U.S. private AI investment was 23 times China’s. It also notes frontier company revenue is rising fast while compute spend and cloud capex are at record levels, organizational AI adoption reached 88%, and employment for software developers ages 22 to 25 fell nearly 20% from 2024.
• Cold read: AI는 single stock theme가 아니라 capital allocation and labor-market reallocation shock다. Benefits are real, but capex intensity and worker displacement create second-order macro and political feedback.
• Trading insight: AI trade는 demand boom and cost burden을 동시에 가격에 반영한다. Henry는 SOXL-style exposure를 늘릴 때 adoption data와 capex financing stress를 같이 확인해야 한다.
• Change sensor: cloud capex guidance, frontier AI revenue run-rate, AI adoption by function, young developer employment, industrial robot installations.
• Follow-up question: AI adoption이 기업 margin expansion으로 연결되는 구간과 capex burden으로 equity multiple을 압박하는 구간을 어떻게 구분할까?

Run Summary [03:01 KST]

• collected_count: 11
• skipped_duplicates: 23
• domain_mix: microstructure/execution 55%, risk-sizing 45%, macro-liquidity/capital-markets 27%, crypto-structure 27%, AI-history/world-sensing 18%. Some items map to multiple domains, so total exceeds 100%.
• top_theme: Edge must be audited at the rejection, execution, and tail-distribution layers; macro backstops and AI narratives only matter after checking who bears liquidity and funding stress.
• sharpest_insight: A signal is not enough. The decisive question is whether the mechanism behind the signal is causal, liquid, and executable under stress rather than merely visible after the fact.
• danger_of_misuse: Do not use DEX base-rate studies to dismiss all high-volatility opportunities or use BIS backstop discussion as a reason to lever risk-on. The correct use is to tighten filters, log rejected trades, and reduce size when liquidity mechanisms are fragile.
• next_probe: Design a Henry rejection ledger and execution audit: every skipped trade, every liquidity-taking entry, every stop adjustment, and each post-trade path should be classified for save-to-miss and gross-to-net R.

[04:00 KST] Reaction-Boundary Variance and Adjoint-Consistent Local-Volatility Projection

• URL: https://arxiv.org/abs/2607.05011
• Source Type: paper
• Domain: microstructure, execution
• Trading Relevance: 7
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: liquidity, execution, sizing
• Why it matters: 이 논문은 bid-ask imbalance field의 zero boundary를 변동성의 구조적 원천으로 본다. Henry에게 중요한 점은 volatility가 price candle에서만 나오는 것이 아니라 order book imbalance가 반응하는 방식에서 생긴다는 것이다.
• Raw excerpt / abstract: The paper derives an operational-time variance kernel for a latent-order-book reaction boundary. It separates a structural boundary cumulant, a clock projection, and a pricing-measure choice, where signed order-flow perturbations displace the bid-ask imbalance zero through a damped Abel response kernel.
• Cold read: Book imbalance는 단순 보조지표가 아니라 가격 경계가 움직이는 물리적 장치다. 같은 volatility라도 order-flow clock과 calendar clock이 다르면 체감 execution risk가 달라진다.
• Trading insight: 시장가 진입 전에는 price volatility보다 imbalance response와 depth recovery를 봐야 한다. 특히 open 직후와 macro release 직후에는 clock-time ATR이 실제 book risk를 늦게 반영할 수 있다.
• Change sensor: top-of-book imbalance persistence, order-flow burst after macro data, spread recovery time.
• Follow-up question: Henry의 진입 rule에 book imbalance 기반 pause condition을 어떤 간단한 proxy로 넣을 수 있을까?

[04:00 KST] Adapted Law Invariance and Time-Consistent Dynamic Risk Measures

• URL: https://arxiv.org/abs/2607.04392
• Source Type: paper
• Domain: risk-sizing
• Trading Relevance: 7
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing, regime, leverage
• Why it matters: 정적 risk measure는 최종 손익 분포만 보지만, 실제 트레이딩은 정보가 시간에 따라 공개되는 경로를 따라간다. Henry의 position sizing은 최종 손실 확률뿐 아니라 중간에 어떤 정보가 드러나고 어떤 행동 여지가 남는지를 반영해야 한다.
• Raw excerpt / abstract: The paper says law invariance in static risk measurement means risk depends only on distribution. In dynamic settings it becomes adapted law invariance: risk assessment should depend on the probabilistic structure of the position together with how information about it is revealed over time.
• Cold read: 같은 최종 payoff라도 정보 공개 경로가 다르면 risk는 다르다. Stop을 실행할 수 있는 경로와 gap으로 정보가 한 번에 반영되는 경로는 같은 분포가 아니다.
• Trading insight: Henry는 overnight hold와 intraday managed trade를 같은 R로 취급하면 안 된다. Dynamic risk measure 관점에서는 actionable information timing이 sizing의 핵심이다.
• Change sensor: scheduled event density, overnight news clustering, stop executable window, liquidity during information release.
• Follow-up question: Henry의 R multiple을 최종 손실폭이 아니라 정보 공개 경로별로 나누면 어떤 trade가 축소될까?

[04:00 KST] Governing Generative AI Across Financial Institutions

• URL: https://arxiv.org/abs/2607.04103
• Source Type: paper
• Domain: risk-sizing, ai-world-sensing, capital-markets
• Trading Relevance: 7
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: regime, narrative, execution
• Why it matters: Generative AI가 직접 신용모델이 아니어도 control environment를 바꾸면 금융기관의 operational risk가 된다. Henry에게는 AI adoption narrative가 생산성 상승만이 아니라 model governance, audit trail, authority boundary 비용을 동반한다는 냉정한 렌즈가 필요하다.
• Raw excerpt / abstract: The paper argues SR 26-2 modernizes U.S. model risk management but excludes generative and agentic AI. It says generative AI may not estimate credit risk directly, yet its outputs can materially affect the surrounding control environment through documentation, decision support, and operational workflows.
• Cold read: AI 금융 도입은 속도와 통제의 trade-off다. 규제권이 agentic AI를 직접 다루지 않아도 사고는 workflow boundary에서 생긴다.
• Trading insight: AI 금융주나 fintech narrative를 볼 때 adoption headline보다 auditability, human override, and error containment를 점검한다. 규제 마찰은 multiple 압박이자 장기 moat가 될 수 있다.
• Change sensor: SR 26-2 적용 가이드, bank AI incident disclosure, audit vendor demand, human-in-the-loop policy changes.
• Follow-up question: Henry의 AI exposure 중 governance cost가 가장 늦게 가격에 반영될 후보는 무엇인가?

[04:00 KST] Portfolio Optimization and Tail-Risk Analytics of Actively Managed ETFs

• URL: https://arxiv.org/abs/2607.03082
• Source Type: paper
• Domain: risk-sizing, capital-markets
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: leverage, correlation, sizing
• Why it matters: ETF는 wrapper가 같아도 tail risk와 turnover 구조가 다르다. Henry가 leveraged ETF와 thematic ETF를 볼 때, 이름과 headline holdings보다 CVaR, drawdown concentration, and rebalance behavior를 먼저 봐야 한다.
• Raw excerpt / abstract: The paper evaluates buy-and-hold, mean-variance, CVaR-based, and tangency-type strategies for 30 actively managed funds from December 2020 to December 2025, with long-only and long-short constraints, focusing on portfolio optimization and tail-risk analytics.
• Cold read: ETF diversification은 구조적 safety를 보장하지 않는다. Active management, turnover, and factor concentration can create hidden tail exposure.
• Trading insight: SOXL/MSTU 같은 leveraged wrapper는 단순 ETF가 아니라 daily rebalanced convex instrument다. Henry는 CAGR보다 path risk and CVaR를 pre-trade sheet에 넣어야 한다.
• Change sensor: ETF holdings turnover, factor concentration, CVaR drift, daily rebalance drag under high volatility.
• Follow-up question: Henry 보유 ETF들의 95% CVaR와 max 5-day drawdown을 같은 화면에서 보면 position size가 어떻게 달라질까?

[04:00 KST] Strategic Risk Reduction: Self-Protection and Self-Insurance

• URL: https://arxiv.org/abs/2606.30363
• Source Type: paper
• Domain: risk-sizing, cognition
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing, leverage, FOMO
• Why it matters: 손실 빈도를 줄이는 self-protection과 손실 크기를 줄이는 self-insurance는 다르다. Henry의 전략도 entry filter는 빈도를 낮추고 stop, hedge, size cap은 severity를 낮춘다는 식으로 분리해야 한다.
• Raw excerpt / abstract: The paper studies how a risk holder combines self-protection and self-insurance when market insurance is absent. Self-protection reduces loss frequency, self-insurance reduces loss severity, and Value-at-Risk creates threshold-driven solutions while Tail Value-at-Risk remains sensitive to the loss magnitude.
• Cold read: 좋은 리스크 관리는 한 개의 방어막이 아니다. 빈도와 크기를 각각 줄이는 장치가 같이 있어야 risk of ruin이 낮아진다.
• Trading insight: Henry가 좋은 setup만 고르는 것은 self-protection이고, fractional Kelly, stop, no-averaging rule은 self-insurance다. 하나가 있다고 다른 하나를 생략하면 levered portfolio에는 구멍이 생긴다.
• Change sensor: rejected setup rate, average loss size, tail loss contribution, stop compliance rate.
• Follow-up question: Henry의 현재 checklist에서 self-protection 항목과 self-insurance 항목은 각각 몇 개인가?

[04:00 KST] Adaptive AI Delegation under Uncertainty

• URL: https://arxiv.org/abs/2606.29406
• Source Type: paper
• Domain: ai-world-sensing, cognition, risk-sizing
• Trading Relevance: 7
• Time Horizon: secular
• Actionability: study
• Henry Risk Link: execution, sizing, regime
• Why it matters: AI recommendation을 언제 사람에게 넘기고 언제 자동화할지의 문제는 prediction 문제가 아니라 decision authority 문제다. Henry의 trading AI stack도 신호 정확도보다 불확실성에 따라 권한을 조절하는 governance가 필요하다.
• Raw excerpt / abstract: The paper frames AI delegation as a sequential decision-authority problem under uncertainty. It asks how organizations should allocate authority to AI-generated recommendations as evidence quality, uncertainty, and objectives evolve over time.
• Cold read: AI가 맞힐 확률이 높아도 권한을 넘겨도 되는 것은 아니다. Error cost, reversibility, and uncertainty state가 authority boundary를 정한다.
• Trading insight: Henry의 자동화는 order placement까지 바로 가지 말고 alert, draft, size suggestion, execution hold 같은 단계로 나눠야 한다. High uncertainty regime에서는 AI output을 conviction이 아니라 checklist input으로만 써야 한다.
• Change sensor: AI signal confidence calibration, human override frequency, error severity by automation level, uncertainty regime shift.
• Follow-up question: Henry의 ATLAS or agent output은 어떤 조건에서 자동 실행이 아니라 사람 승인 대기로 강등되어야 하는가?

[04:00 KST] Methods for Uncertainty Representation in Risk Management

• URL: https://arxiv.org/abs/2606.27804
• Source Type: paper
• Domain: risk-sizing, cognition
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing, regime, narrative
• Why it matters: 리스크 관리에서 uncertainty를 한 숫자로 줄이면 traceability가 사라진다. Henry에게 필요한 것은 bullish or bearish 결론보다 어떤 불확실성이 aleatory randomness이고 어떤 것이 epistemic ignorance인지 구분하는 습관이다.
• Raw excerpt / abstract: The review says uncertainty is a central but often inadequately addressed component of risk management. It compares how established approaches conceptualize and represent uncertainty and proposes a decision-oriented framework for quality and traceability of decision processes.
• Cold read: 모르는 것에도 종류가 있다. 측정 가능한 변동성과 모델이 틀렸을 가능성을 섞으면 false precision이 생긴다.
• Trading insight: Henry의 pre-trade note에는 confidence number 하나보다 data uncertainty, model uncertainty, liquidity uncertainty를 따로 적는 방식이 더 안전하다. 특히 narrative trade에서는 epistemic uncertainty가 가장 크다.
• Change sensor: data revision risk, model disagreement, liquidity regime breaks, forecast dispersion.
• Follow-up question: Henry의 다음 큰 베팅에서 모르는 것을 세 칸으로 나누면 size가 얼마나 줄어드는가?

[04:00 KST] Mitigating Adverse Selection in Concentrated Liquidity AMMs with Dynamic Fees

• URL: https://arxiv.org/abs/2606.23070
• Source Type: paper
• Domain: crypto-structure, microstructure, execution
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: liquidity, execution, narrative
• Why it matters: Concentrated liquidity AMM은 capital efficiency를 높이지만 LP에게 Loss-Versus-Rebalancing 형태의 adverse selection cost를 만든다. Henry가 DeFi liquidity나 crypto market structure를 볼 때, 높은 fee yield는 toxic flow 보상인지 확인해야 한다.
• Raw excerpt / abstract: The paper develops an agent-based model of a Uniswap v3 pool interacting with a stochastic reference market. It studies how dynamic fees can mitigate adverse selection costs, formalized as Loss-Versus-Rebalancing, in concentrated liquidity AMMs.
• Cold read: DeFi liquidity는 passive yield가 아니라 market making inventory risk다. Dynamic fee가 없다면 informed flow가 LP에게 비용을 이전한다.
• Trading insight: Crypto liquidity depth를 볼 때 TVL만으로 안전하다고 판단하지 않는다. LP가 toxic flow에 계속 당하면 depth는 stress 때 사라지고, token price는 더 얇은 book에서 움직인다.
• Change sensor: LVR estimates, dynamic fee adoption, LP withdrawal during volatility, DEX-CEX price lead-lag.
• Follow-up question: BTC and crypto beta 판단에서 DEX liquidity withdrawal이 spot exchange depth보다 먼저 경고하는 구간은 언제인가?

[04:00 KST] Endogenous Randomness from Adversarial Market Learning

• URL: https://arxiv.org/abs/2606.22743
• Source Type: paper
• Domain: microstructure, reflexivity, ai-world-sensing
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: regime, execution, narrative
• Why it matters: 이 논문은 외부 noise 없이도 predictive traders와 market mechanism 상호작용만으로 apparent randomness가 생길 수 있다고 본다. Henry에게 중요한 점은 시장이 복잡해 보이는 이유가 정보 부족뿐 아니라 참가자들이 서로의 예측 가능성을 제거하기 때문일 수 있다는 것이다.
• Raw excerpt / abstract: The paper proposes a deterministic adversarial market model in which apparent randomness emerges endogenously from the interaction between a market mechanism and predictive traders, without injecting random noise or imitating an external empirical data distribution.
• Cold read: Market randomness is partly a defense system. Predictability attracts exploitation, and exploitation changes the path until the pattern degrades.
• Trading insight: Backtest edge가 공개적이고 단순할수록 반감기가 짧다. Henry는 signal discovery보다 edge decay monitoring, position size throttling, and regime revalidation을 중시해야 한다.
• Change sensor: strategy crowding, declining post-signal drift, faster mean reversion after popular triggers, social sharing of same setup.
• Follow-up question: Henry가 믿는 반복 setup 중 시장 참가자가 쉽게 역이용할 수 있는 것은 무엇인가?

[04:00 KST] Energy Markets Race to Solve the AI Power Bottleneck

• URL: https://www.morganstanley.com/insights/articles/powering-ai-energy-market-outlook-2026
• Source Type: research
• Domain: ai-world-sensing, capital-markets, macro-liquidity
• Trading Relevance: 9
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative, dilution, regime, correlation
• Why it matters: AI trade의 병목은 GPU만이 아니라 power, grid, credit markets다. Henry의 SOXL and NXE-linked thinking에는 compute demand와 에너지 인프라 financing이 함께 들어가야 한다.
• Raw excerpt / abstract: Morgan Stanley says developers expect power constraints by 2027-2028. AI-driven data centers contribute nearly one-fifth of global electricity demand growth, data center power consumption is expected to increase nearly 126 GW annually through 2028, U.S. data center demand could reach 74 GW by 2028 with about 49 GW shortfall in available power access, and hyperscalers could spend more than $1T in 2025-2026, with debt markets covering roughly half.
• Cold read: AI is becoming a project finance and energy bottleneck trade. The winners may be those with power access, balance sheet, permitting, and supply chain control rather than only model quality.
• Trading insight: AI exposure sizing should react to credit spread pressure, utility politics, grid equipment delays, and hyperscaler capex financing. NXE-like energy exposure may be an AI derivative, but it also carries commodity-cycle and regulatory risk.
• Change sensor: hyperscaler debt issuance, grid interconnection queue, power spread, transformer and switchgear lead times, political backlash on electricity bills.
• Follow-up question: Henry의 AI basket을 chip, power, grid equipment, credit financing으로 나누면 어느 bucket이 가장 fragile한가?

Run Summary [04:00 KST]

• collected_count: 10
• skipped_duplicates: 24
• domain_mix: microstructure/execution 40%, risk-sizing/cognition 60%, macro-liquidity/capital-markets 30%, crypto-structure/reflexivity 30%, AI-history/world-sensing 40%. Some items map to multiple domains, so total exceeds 100%.
• top_theme: Risk quality is shifting from static prediction to path-aware control: order-book reaction, information timing, AI authority boundaries, adverse selection, and power-financing bottlenecks all decide whether a thesis survives contact with execution.
• sharpest_insight: The trade is not just the asset. It is the path of information, the venue’s adverse-selection mechanics, the authority granted to models, and the physical-credit infrastructure behind the narrative.
• danger_of_misuse: Do not treat these papers as reasons to overcomplicate every trade or to chase AI power names immediately. The correct use is to separate uncertainty types, cap size when information or liquidity is non-actionable, and monitor bottlenecks before adding leverage.
• next_probe: Build a Henry risk taxonomy with three columns: path risk, liquidity and adverse-selection risk, and narrative infrastructure risk. Map each current holding and watchlist ticker to those columns before any large add.

[05:00 KST] SHARC: SHAP-Based Interpretability in ML Risk Models for Regulatory Capital

• URL: https://arxiv.org/abs/2607.05484v1
• Source Type: paper
• Domain: risk-sizing | capital-markets | ai-world-sensing
• Trading Relevance: 7
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing | regime | narrative
• Why it matters: Henry의 큰 베팅은 모델 출력 자체보다 그 모델이 어떤 stress driver에 의해 움직이는지 설명 가능해야 한다. 이 논문은 ML risk model을 SHAP으로 baseline, mean-driven, volatility-driven component로 분해해 stress capital의 원인을 추적한다. 개인 트레이더에게도 “내 리스크 숫자가 왜 커졌는가”를 설명하지 못하면 sizing을 줄여야 한다는 교훈이다.
• Raw excerpt / abstract: “SHARC decomposes SVaR into baseline, mean-driven, and volatility-driven components… under stress conditions, the mean return component (directional loss magnitude) dominates the variance component (volatility baseline) in determining capital levels.”
• Cold read: risk model의 다음 경쟁력은 더 복잡한 예측이 아니라 auditable decomposition이다. 방향성 손실 magnitude가 volatility baseline보다 capital을 더 지배한다는 결과는 “변동성만 보면 된다”는 단순화를 깨뜨린다.
• Trading insight: SOXL, MSTU 같은 convex exposure는 volatility 확대보다 directional gap loss가 계좌 capital requirement를 지배할 수 있다. 진입 전 stress scenario별 손실 기여도를 적어야 하며, 설명 불가능한 모델 confidence는 position size로 번역하지 않는다.
• Change sensor: stress VaR decomposition, mean-loss vs volatility contribution, AI bubble/regulatory burden scenario의 capital driver.
• Follow-up question: Henry 계좌의 각 보유 후보에 대해 SHAP식은 아니더라도 loss driver를 direction, volatility, liquidity, FX로 분해하면 어떤 포지션이 가장 불투명한가?

[05:00 KST] Strategic Risk Reduction: Self-Protection and Self-Insurance

• URL: https://arxiv.org/abs/2606.30363v2
• Source Type: paper
• Domain: risk-sizing | cognition
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: leverage | sizing | execution
• Why it matters: 이 논문은 risk holder가 loss frequency를 줄이는 self-protection과 loss severity를 줄이는 self-insurance를 어떻게 조합해야 하는지 다룬다. 트레이딩 언어로 번역하면 self-protection은 “나쁜 진입을 줄이는 필터”, self-insurance는 “손절, hedge, cash buffer로 손실 크기를 제한”하는 행위다. Henry에게 중요한 점은 두 장치가 항상 보완재가 아니며 VaR와 Tail VaR 기준에 따라 대체재가 될 수도 있다는 점이다.
• Raw excerpt / abstract: “Self-protection reduces loss frequency, while self-insurance reduces loss severity… Value-at-Risk leads to a threshold-driven solution… By contrast, Tail Value-at-Risk… creates a direct residual frequency-severity interaction.”
• Cold read: 리스크 관리에는 두 축이 있다. 첫째는 거래를 덜 하게 만드는 구조, 둘째는 틀렸을 때 덜 죽게 만드는 구조다. Tail risk 기준에서는 두 축의 상호작용이 비선형이어서 단순한 stop loss 하나로 충분하지 않다.
• Trading insight: FOMO 상태에서는 self-protection이 먼저다. 이미 진입했다면 self-insurance가 우선이다. 대형 베팅은 필터와 손실 제한 장치가 동시에 작동할 때만 가능하며, 둘 중 하나가 약하면 fractional Kelly를 더 낮춰야 한다.
• Change sensor: 거래 필터의 miss/save ratio, 손절 실행률, hedge 비용, Tail VaR 기준 손실 빈도와 손실 크기의 동시 악화.
• Follow-up question: Henry의 pre-trade checklist에서 self-protection 항목과 self-insurance 항목을 분리하면 어떤 항목이 비어 있는가?

[05:00 KST] Portfolio Optimization for Commodity ETFs under Heavy-Tailed Returns

• URL: https://arxiv.org/abs/2606.26625v1
• Source Type: paper
• Domain: risk-sizing | execution | capital-markets
• Trading Relevance: 7
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: correlation | sizing | execution
• Why it matters: commodity ETF 연구지만 핵심은 ETF allocation에서 heavy tail, skewness, kurtosis, transaction cost가 최적화를 어떻게 망가뜨리는지다. Henry의 레버리지 ETF와 thematic ETF 접근에도 같은 구조가 적용된다. mean-variance tangent portfolio가 expected-return estimation error에 민감하다는 결론은 “기대수익률 추정으로 사이즈를 키우는 것”의 위험을 다시 확인한다.
• Raw excerpt / abstract: “CVaR-based portfolios provided more stable cumulative performance… optimized portfolios remained exposed to heavy downside tails… practical value of dynamic optimization depended on turnover control.”
• Cold read: 최적화는 tail을 제거하지 못한다. 단지 손실 분포를 덜 나쁘게 만들 뿐이며, dynamic model은 turnover와 implementation cost 앞에서 쉽게 무너진다.
• Trading insight: SOXL 같은 상품에서 expected return 기반 tangent sizing은 과최적화 위험이 크다. CVaR, Calmar, STARR 같은 downside-aware metrics를 우선하고, 리밸런싱 빈도는 수익이 아니라 transaction-cost budget으로 제한해야 한다.
• Change sensor: ETF tail kurtosis, rolling CVaR, turnover-adjusted performance, sector별 commodity 또는 thematic ETF volatility clustering.
• Follow-up question: Henry의 현재 watchlist를 mean-variance가 아니라 CVaR minimization과 turnover cap으로 보면 어떤 종목의 size가 가장 크게 줄어드는가?

[05:00 KST] MPFlow: Budgeted Max-Flow Optimization on the Bitcoin Lightning Network

• URL: https://arxiv.org/abs/2607.08703v1
• Source Type: paper
• Domain: crypto-structure | microstructure | ai-world-sensing
• Trading Relevance: 6
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: liquidity | regime
• Why it matters: Lightning Network liquidity placement는 crypto가 단순 가격 자산이 아니라 payment routing infrastructure로 진화하는지를 보여주는 구조 신호다. 논문은 제한된 budget으로 어떤 channel을 열어 routing capacity를 최대화할지 deep graph RL로 풀고, 실제 4,640개 channel-open decision과 267.3 BTC 배치 사례를 제시한다. BTC market structure 관찰에서 ETF flow만 보지 말고 network utility와 liquidity routing도 봐야 한다.
• Raw excerpt / abstract: “The agent has been deployed in production for peer recommendations, executing 4640 channel-open decisions that cumulatively allocate 267.3 BTC over $16 million across 30 managed nodes.”
• Cold read: crypto liquidity는 exchange order book 밖에서도 최적화되고 있다. 결제 네트워크의 routing capital allocation이 자동화되면 BTC의 utility narrative와 liquidity moat가 가격 narrative와 분리되어 성장할 수 있다.
• Trading insight: BTC 관련 포지션은 ETF inflow, perp funding, spot reserve만으로 구조를 판단하면 좁다. Lightning capacity, routing revenue, node centralization 변화가 장기 narrative의 hard data가 될 수 있다.
• Change sensor: Lightning capacity, channel concentration, routing fee market, production RL peer recommendation adoption.
• Follow-up question: BTC bull thesis에서 “store of value” 외에 payment-routing utility가 실제 capital allocation으로 확인되는 지표는 무엇인가?

[05:00 KST] Formal Mechanisms for Market Stability in Self-Interested Agent Societies

• URL: https://arxiv.org/abs/2607.08652v1
• Source Type: paper
• Domain: ai-world-sensing | microstructure | cognition
• Trading Relevance: 6
• Time Horizon: secular
• Actionability: study
• Henry Risk Link: regime | narrative | execution
• Why it matters: LLM agents가 반복 거래하는 marketplace simulation에서 self-interested agents는 제약이 없으면 defection으로 흐르고 gains from trade가 붕괴한다. Mediation mechanism이 optimized troll attack에서도 honest-agent utility를 13.3%만 낮추고 붕괴는 막았다는 결과는, 미래 agent market의 안정성이 가격 예측보다 mechanism design에 달려 있음을 보여준다.
• Raw excerpt / abstract: “Self-interested agents, left unconstrained, tend toward defection… the best attack (v6) reduces honest-agent utility by 13.3% but cannot collapse the market. Mediation enables recovery even under sustained adversarial pressure.”
• Cold read: AI agent economy는 선의로 굴러가지 않는다. 시장이 유지되려면 정보, 평판, 중재, penalty 같은 formal mechanism이 필요하다.
• Trading insight: agentic finance가 확산될수록 edge는 더 빠르게 arbitraged 되고, manipulation과 defense가 함께 발전한다. Henry는 AI narrative를 매수 이유로 쓰기보다 “어떤 시장 설계가 adversarial pressure를 견디는가”를 봐야 한다.
• Change sensor: agent marketplace protocol, mediation mechanism, adversarial audit result, automated trading venue의 rule-change frequency.
• Follow-up question: ATLAS 또는 개인 trading agent를 만들 때 defection, overtrading, hallucinated confidence를 막는 mediation layer는 무엇이어야 하는가?

[05:00 KST] Workflow as Knowledge: Semantic Persistence for LLM-Mediated Workflows

• URL: https://arxiv.org/abs/2607.08740v1
• Source Type: paper
• Domain: ai-world-sensing | cognition
• Trading Relevance: 6
• Time Horizon: secular
• Actionability: study
• Henry Risk Link: execution | FOMO | regime
• Why it matters: 트레이딩 연구와 실행 품질은 단발 prompt가 아니라 workflow, inference record, context snapshot, dependency relation이 남는 persistent system이 되어야 한다. 이 논문은 derive와 infer를 구분하며 LLM judgment를 declared context와 capability policy 아래에 두는 개념을 제시한다. Henry에게는 매매 판단의 흔적을 남겨 다음 날 냉정하게 복기하는 구조로 연결된다.
• Raw excerpt / abstract: “workflow definitions, workflow instances, inference records, context snapshots, and dependency relations are represented as persistent knowledge objects… derive is deterministic computation… infer is mediated LLM judgment under declared context.”
• Cold read: AI workflow의 성숙은 더 긴 context가 아니라 검토 가능한 state object다. 매매도 동일하게, 결론보다 결론이 만들어진 경로가 기록되어야 한다.
• Trading insight: pre-trade checklist, position sizing, scenario matrix, execution result를 같은 object graph로 남기면 FOMO 매매와 사후 합리화를 줄일 수 있다. inference와 deterministic calculation을 분리하는 것이 중요하다.
• Change sensor: workflow persistence tooling, audit trail schema, deterministic calculation vs LLM inference separation, human approval checkpoint.
• Follow-up question: Henry의 매매 기록에서 deterministic fact, quantitative calculation, LLM interpretation, final human decision을 분리해 저장할 수 있는가?

[05:00 KST] New York Fed Markets Data Dashboard

• URL: https://www.newyorkfed.org/markets/data-hub
• Source Type: official
• Domain: macro-liquidity | capital-markets | execution
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: liquidity | regime | correlation
• Why it matters: New York Fed dashboard는 SOFR, TGCR, repo, reverse repo, Treasury operations, securities lending, primary dealer statistics 같은 liquidity plumbing을 한 곳에 모은다. 가격 브리핑보다 중요한 것은 funding stress가 risk assets의 correlation regime을 바꾸기 전에 감지하는 일이다. 특히 repo와 dealer financing data는 레버리지 asset의 체감 risk를 바꿀 수 있다.
• Raw excerpt / abstract: 원문은 “Tri-Party General Collateral Rate”, “Treasury Securities”, “Repos”, “Reverse Repos”, “System Open Market Account Holdings”와 “Primary Dealer Statistics… positions, transactions, financing and fails”를 제공한다고 설명한다.
• Cold read: macro liquidity는 하나의 headline이 아니라 plumbing dashboard다. funding, collateral, dealer balance sheet, settlement fail이 함께 악화될 때 가격 움직임은 narrative가 아니라 balance sheet constraint가 된다.
• Trading insight: 레버리지 ETF size를 키우기 전에는 VIX만 볼 것이 아니라 SOFR spread, repo usage, dealer fails, primary dealer financing 흐름을 확인해야 한다. liquidity가 얇아지는 regime에서는 technical breakout의 신뢰도가 낮아진다.
• Change sensor: SOFR-TGCR spread, repo and reverse repo operations, primary dealer financing volume, Treasury fails, SOMA holdings 변화.
• Follow-up question: Henry의 weekly risk dashboard에 New York Fed의 어떤 3개 liquidity series를 고정으로 넣을 것인가?

[05:00 KST] Data Center Investment in 2026: AI Demand, Power Constraints, and Private Equity Trends

• URL: https://www.ropesgray.com/en/insights/viewpoints/102mvfl/data-center-investment-in-2026-ai-demand-power-constraints-and-private-equity
• Source Type: research
• Domain: ai-world-sensing | capital-markets | macro-liquidity
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative | dilution | regime
• Why it matters: AI infrastructure cycle의 bottleneck이 GPU에서 power, interconnection, financing structure로 확장되고 있다. Ropes & Gray는 power availability가 capital보다 primary constraint이고 grid interconnection이 up to four years 걸릴 수 있으며, GPU financing, land-cost facilities, BYOP financing, ABS, CMBS, 144A까지 자금 조달 구조가 진화한다고 정리한다. AI narrative 투자는 compute demand뿐 아니라 debt market absorption과 power delivery risk를 봐야 한다.
• Raw excerpt / abstract: “Power availability—not capital—is the primary constraint on data center development… Electrical grid interconnections are often taking up to four years… Outside capital now arrives earlier in the life cycle through GPU financings and land-cost facilities… investors now require firmly committed, transparent hyperscale credit.”
• Cold read: AI boom은 software multiple story에서 infrastructure finance story로 이동하고 있다. power access와 tenant credit이 alpha와 blow-up을 가르는 underwriting 변수다.
• Trading insight: AI 관련 종목을 볼 때 revenue growth만으로 과감히 베팅하면 위험하다. capex financing, power availability, permitting, hyperscaler credit commitment, exit market depth가 cycle risk를 만든다.
• Change sensor: grid interconnection queue, BYOP deal count, GPU-backed debt spread, datacenter ABS or CMBS issuance, hyperscaler take-or-pay contract quality.
• Follow-up question: AI infra exposure를 반도체, 전력, data center REIT, private credit risk로 나누면 Henry의 현재 narrative exposure는 어디에 과집중되어 있는가?

Run Summary [05:00 KST]

• collected_count: 8
• skipped_duplicates: 16
• domain_mix: microstructure/execution 2개, risk-sizing 3개, macro/capital-markets 3개, crypto-structure 1개, AI-history/world-sensing 4개로 편중은 AI와 risk 쪽이지만 liquidity plumbing을 함께 보강함.
• top_theme: 좋은 트레이딩 시스템은 예측 모델보다 설명 가능한 risk decomposition, self-protection와 self-insurance의 분리, persistent workflow audit trail, liquidity plumbing monitor로 구성된다.
• sharpest_insight: Tail risk에서 “나쁜 거래를 피하는 장치”와 “틀렸을 때 덜 죽는 장치”는 단순 보완재가 아니며, regime에 따라 대체재 또는 비선형 조합이 된다.
• danger_of_misuse: 오늘 자료를 AI, BTC, data center narrative의 즉시 매수 근거로 쓰면 안 된다. 대부분은 구조와 process 개선 자료이며, 가격 entry, liquidity, sizing, stop plan이 없는 베팅은 FOMO다.
• next_probe: Henry의 실제 pre-trade checklist를 self-protection, self-insurance, liquidity plumbing, audit trail 네 구역으로 재분류하면 어떤 구역이 가장 약한가?

[06:00 KST] Signature-Based Optimal Execution for Statistical Arbitrage with Path-Dependent Trading Signals

• URL: https://arxiv.org/abs/2606.31387
• Source Type: paper
• Domain: execution, microstructure, risk-sizing
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: execution, sizing, liquidity
• Why it matters: 이 논문은 alpha signal과 execution speed를 같은 path-dependent representation 안에서 다룬다. Henry에게 핵심은 좋은 signal도 inventory risk, temporary impact, terminal liquidation cost를 동시에 고려하지 않으면 실제 수익으로 변환되지 않는다는 점이다.
• Raw excerpt / abstract: The authors develop a signature-based framework for optimal execution in statistical arbitrage strategies with path-dependent predictive signals. Both alpha process and trading speed are modelled as linear functionals of the truncated signature of a time-augmented market path, allowing the trading rule to react to realised signal history while accounting for temporary impact, inventory exposure, and terminal liquidation.
• Cold read: Execution은 signal 뒤에 붙는 사후 작업이 아니라 signal과 같은 모델 안에서 풀어야 하는 제어 문제다. Path history가 중요할수록 단순 entry price와 stop price만으로는 부족하다.
• Trading insight: Henry가 momentum or mean-reversion setup을 볼 때, signal strength가 커도 남은 liquidity window와 inventory unwind plan이 없으면 size를 키우면 안 된다. 진입 전부터 exit path를 같이 계산해야 한다.
• Change sensor: signal half-life, temporary impact estimate, inventory holding time, terminal liquidation cost.
• Follow-up question: Henry의 실제 주문 계획에서 alpha signal과 execution speed를 분리하지 않고 같이 조정하는 최소 rule은 무엇인가?

[06:00 KST] The Bounce Has No Direction: Sign, Magnitude, and the Microstructure of Equity Return Predictability

• URL: https://arxiv.org/abs/2606.29591
• Source Type: paper
• Domain: microstructure, execution, cognition
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: execution, FOMO, regime
• Why it matters: SPY의 lag-1 return autocorrelation이 mean reversion인지 단순 magnitude shrinkage인지 구분한다. Henry가 candle bounce를 방향 신호로 착각하면, 실제로는 방향 edge가 아니라 변동폭이 줄어드는 현상을 잘못 매매할 수 있다.
• Raw excerpt / abstract: The paper states that SPY’s lag-1 return autocorrelation is highly significant, but standard variance-ratio tests cannot determine whether it reflects directional reversal or magnitude shrinkage. The authors introduce a Fourier-Residue Identity to decompose autocorrelation into sign and magnitude channels with different trading implications.
• Cold read: 모든 bounce가 방향성을 의미하지 않는다. 시장이 다음 bar에서 덜 움직인다는 사실과 반대로 움직인다는 사실은 전혀 다른 edge다.
• Trading insight: Henry는 oversold bounce에서 방향 reversal과 volatility compression을 분리해야 한다. 방향 edge가 없으면 position size를 줄이고, volatility compression만 기대하는 trade는 option or mean-reversion 구조로 따로 설계해야 한다.
• Change sensor: post-shock sign reversal rate, realized range shrinkage, variance ratio decomposition, bounce follow-through.
• Follow-up question: SOXL 급락 후 반등 setup에서 Henry가 보는 것은 방향 reversal인가, 단순 변동폭 축소인가?

[06:00 KST] Hierarchical Graph Learning for Calendar Spread Strategies in Commodity Futures Markets

• URL: https://arxiv.org/abs/2606.25811
• Source Type: paper
• Domain: execution, capital-markets, ai-world-sensing
• Trading Relevance: 7
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: correlation, sizing, regime
• Why it matters: Commodity futures는 underlying asset과 개별 contract가 계층적으로 연결된 구조다. Henry의 NXE or energy-linked thinking에도 front-end physical constraint, long-end narrative, curve shape가 서로 다르게 움직일 수 있다는 교훈을 준다.
• Raw excerpt / abstract: The paper represents commodity futures hierarchically, with underlying assets at the upper level and individual futures contracts at the lower level. It proposes hierarchical graph learning for calendar spread strategies, using edges for within-level correlations and cross-level contract-to-underlying connections.
• Cold read: 같은 commodity theme라도 curve location이 다르면 다른 bet다. Spot shortage, carry, storage, funding, and macro narrative가 contract별로 다른 risk를 만든다.
• Trading insight: Henry가 uranium or energy derivative exposure를 볼 때 single ticker narrative 대신 curve, inventory, and financing structure를 확인해야 한다. Correlation은 asset class가 아니라 maturity bucket에서 깨질 수 있다.
• Change sensor: calendar spread slope, inventory data, contract-level liquidity, curve inversion or contango persistence.
• Follow-up question: NXE 같은 equity exposure를 commodity curve signal과 어떻게 연결하면 narrative 과신을 줄일 수 있을까?

[06:00 KST] Time-Dependent Weighted Directed Networks of Cryptocurrency Interaction from High-Frequency Returns

• URL: https://arxiv.org/abs/2606.25466
• Source Type: paper
• Domain: crypto-structure, microstructure, risk-sizing
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: monitor
• Henry Risk Link: correlation, liquidity, regime
• Why it matters: Crypto assets are not independent bets; influence flows across assets and changes over time. Henry가 BTC, MSTR, crypto beta를 볼 때 단순 상관계수보다 directed influence network가 regime shift를 더 빨리 보여줄 수 있다.
• Raw excerpt / abstract: The authors build directed and weighted cryptocurrency interaction networks from high-frequency returns between 2020 and 2025 using statistically significant Granger causal relationships. They report heavy-tailed normalized returns and evolving influence across crypto assets.
• Cold read: Crypto market risk는 network contagion risk다. A leading asset can become a transmitter of stress before portfolio correlation statistics fully catch up.
• Trading insight: BTC exposure를 늘릴 때 altcoin weakness, stablecoin liquidity, and exchange-specific flow가 BTC로 전염되는지 확인한다. MSTR도 BTC 자체보다 crypto liquidity network shock에 더 민감해질 수 있다.
• Change sensor: directed lead-lag edges, centrality of BTC and ETH, stablecoin pair liquidity, network density during selloffs.
• Follow-up question: Henry의 crypto risk dashboard는 simple BTC chart 대신 어떤 lead-lag network indicators를 봐야 하는가?

[06:00 KST] Empirical Confirmation of the Square-Root Law of Market Impact in a U.S. Large-Cap Equity

• URL: https://arxiv.org/abs/2606.24019
• Source Type: paper
• Domain: microstructure, execution
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: study
• Henry Risk Link: liquidity, execution, sizing
• Why it matters: AAPL 같은 대형주에서도 metaorder impact가 square-root law에 가깝다면, liquidity가 풍부한 종목에서도 size는 공짜가 아니다. Henry의 ETF 주문도 개별 주문이 작아 보여도 participation이 커지면 non-linear cost를 낸다.
• Raw excerpt / abstract: The paper tests the square-root law on Apple using Nasdaq TotalView-ITCH market-by-order data over 178 trading days and roughly 0.5 billion events. Without broker-tagged parent orders, it reconstructs metaorders from anonymous tape and calibrates impact as a function of Q over daily volume.
• Cold read: Large-cap liquidity is deep but not infinite. Impact is hidden until order participation, timing, and liquidity replenishment make it visible.
• Trading insight: Henry의 order ticket에는 dollar amount뿐 아니라 expected participation rate가 있어야 한다. Open, close, macro release window에서는 apparent liquidity가 커 보여도 impact coefficient가 악화될 수 있다.
• Change sensor: ADV participation, impact per dollar, top-of-book replenishment, post-order drift.
• Follow-up question: Henry의 평균 주문 규모가 SOXL, MSTU, RGTI의 5-minute volume 대비 몇 percent인지 기록하고 있는가?

[06:00 KST] Transaction Costs and Speed in the Ethereum Ecosystem: Scalability of the Mainnet and Layer 2s

• URL: https://arxiv.org/abs/2606.22206
• Source Type: paper
• Domain: crypto-structure, execution, capital-markets
• Trading Relevance: 7
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: liquidity, regime, narrative
• Why it matters: Ethereum scaling is market structure, not just technology. Lower fees and higher throughput can change DeFi liquidity, stablecoin settlement, and on-chain arbitrage speed, which feeds back into crypto beta and institutional usage.
• Raw excerpt / abstract: The paper compares Ethereum Mainnet, Layer 2 networks, Solana, and Polygon from January 2024 through March 2026. It states Ethereum upgrades doubled transactions per second on both mainnet and L2s, mainnet median fees fell from over $2tounder$0.02, and L2 median fees fell more than 95% from $0.05to$0.0015.
• Cold read: Crypto rails are becoming cheaper and faster, but that can compress some edges while increasing volume and composability. Lower transaction cost can also increase high-frequency adversarial activity.
• Trading insight: Crypto infrastructure progress is not automatically bullish for every token. It changes where liquidity forms, where MEV and arbitrage concentrate, and which venues get adverse-selected.
• Change sensor: L2 fee dispersion, TPS, bridge volume, stablecoin settlement share, MEV and sequencer revenue.
• Follow-up question: BTC and ETH exposure should be sized differently when on-chain execution costs fall but adversarial speed rises?

[06:00 KST] Optimal Dynamic Fees for Automated Market Makers: A Stochastic Control Approach to Loss-Versus-Rebalancing

• URL: https://arxiv.org/abs/2606.21769
• Source Type: paper
• Domain: crypto-structure, microstructure, execution
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: liquidity, execution, narrative
• Why it matters: AMM liquidity is not passive depth; fee policy decides whether LPs are compensated for toxic flow. Henry가 DeFi liquidity or crypto market depth를 볼 때, TVL보다 adverse selection compensation이 더 중요할 수 있다.
• Raw excerpt / abstract: The paper studies an AMM liquidity provider whose fee can be continuously adjusted through programmable hooks. In a constant-product AMM, the fee controls two opposing forces: higher revenue per uninformed trade and worse selection against informed flow, measured relative to a continuously rebalanced benchmark.
• Cold read: Dynamic fee is an immune system for on-chain liquidity. If fees cannot adapt, informed flow extracts value and liquidity disappears when it is most needed.
• Trading insight: Crypto slippage risk is a function of fee regime, LP inventory, and volatility. During high volatility, static low-fee pools may look cheap but become fragile.
• Change sensor: dynamic fee adoption, LP withdrawal, LVR, pool depth under volatility, CEX-DEX lead-lag.
• Follow-up question: Henry가 crypto liquidity를 볼 때 TVL 대신 어떤 toxic-flow and LVR proxy를 써야 하는가?

[06:00 KST] Beyond Reserves: The Federal Reserve’s Balance Sheet and the Repo Market

• URL: https://www.federalreserve.gov/econres/feds/beyond-reserves-the-federal-reserves-balance-sheet-and-the-repo-market.htm
• Source Type: official
• Domain: macro-liquidity, capital-markets, risk-sizing
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, regime, correlation
• Why it matters: 이 Fed paper는 QT의 제약을 bank reserves가 아니라 repo market capacity로 본다. Henry에게 중요한 점은 liquidity regime을 reserves 한 숫자로 단순화하면 hedge fund demand, money market fund supply, policy rate interaction을 놓친다는 것이다.
• Raw excerpt / abstract: The Fed abstract states: “We present a new constraint on the size of the Fed’s balance sheet: repo market capacity.” It says repo market capacity, driven by money market fund liquidity supply, is the binding constraint on the Fed balance sheet, not bank reserve demand, and higher policy rates can expand repo capacity.
• Cold read: Balance sheet liquidity is now a shadow-bank plumbing problem. Reserves matter, but the binding constraint can sit in repo capacity and money fund willingness to lend.
• Trading insight: Leveraged risk-on sizing should watch secured funding stress and repo capacity, not only Fed balance sheet totals. If repo capacity tightens, high beta equity and BTC can sell off through funding channels even before economic news changes.
• Change sensor: TGCR-IORB spread, SOFR pressure, ON RRP usage, money market fund repo lending, hedge fund Treasury basis leverage.
• Follow-up question: Henry의 leverage cap should be reduced when repo capacity indicators deteriorate by how much?

[06:00 KST] Rising Hedge Fund Leverage Affects Monetary Policy Implementation

• URL: https://www.dallasfed.org/research/economics/2026/0528
• Source Type: official
• Domain: macro-liquidity, capital-markets, risk-sizing
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: leverage, liquidity, regime
• Why it matters: Dallas Fed는 hedge fund net repo borrowing이 2025년 말 약 $1.8T, marketable notes and bonds의 6%에 도달했다고 말한다. Treasury supply를 levered buyers가 흡수하는 구조는 평상시에는 수요지만 stress 때는 forced deleveraging risk다.
• Raw excerpt / abstract: The article says hedge fund net repo borrowing reached roughly $1.8T by year-end 2025, more than doubling since the start of 2024. It estimates expansion of leveraged Treasury relative value activity over the past decade is associated with 10-20 basis points of widening in repo-market spreads, and says net-funding-demand trades transmit specifically through secured rates.
• Cold read: Treasury market 안정성은 increasingly levered relative-value balance sheet에 의존한다. 공급을 흡수하는 주체가 leverage를 쓸수록 funding spread가 risk asset signal이 된다.
• Trading insight: Henry의 high beta position은 equity-only story가 아니다. Basis trade unwind or repo spread widening can force cross-asset deleveraging and correlation spike.
• Change sensor: hedge fund net repo borrowing, leveraged fund short Treasury futures, TGCR-IORB and GCF-IORB spreads, dealer balance sheet constraints.
• Follow-up question: Treasury basis stress가 SOXL or MSTR drawdown으로 전염되는 path를 Henry scenario matrix에 어떻게 넣을까?

[06:00 KST] Domestic Banks Are Inelastic Providers of Marginal Funding to Repo Markets

• URL: https://www.dallasfed.org/research/economics/2026/0212-levymccormick-repotiming
• Source Type: official
• Domain: macro-liquidity, execution, capital-markets
• Trading Relevance: 8
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, execution, regime
• Why it matters: Repo liquidity has an intraday clock. Dallas Fed는 repo trading이 early morning에 집중되고 banks는 그 시간에 reserve needs를 확정하지 못해 marginal lending이 inelastic하다고 설명한다. Henry에게는 macro liquidity도 시간대별 execution risk라는 교훈이다.
• Raw excerpt / abstract: The article says 64% of cleared repo volume traded by 8:30 a.m. and about three quarters of tri-party repo volume by 9 a.m. as of 2021. It argues domestic banks are highly inelastic sources of unplanned day-of liquidity provision because repo lending drains their intraday liquidity before payment needs are known.
• Cold read: Liquidity shortage is not only about quantity; it is about timing. Cash that arrives after the market-clearing window may not prevent a funding spike.
• Trading insight: Funding stress can create overnight or premarket risk that normal equity-session indicators miss. Henry should be cautious holding oversized leverage through known funding-sensitive dates.
• Change sensor: month-end and quarter-end repo spikes, SOFR over IORB, early-morning funding volatility, term funding demand before statement dates.
• Follow-up question: Henry의 overnight risk rule에 statement date and funding calendar를 어떻게 넣을까?

[06:00 KST] 2026 Data Center Power Report: Mid-Year Pulse

• URL: https://www.bloomenergy.com/wp-content/uploads/power-report-2026-mid-year-pulse-final.pdf
• Source Type: research
• Domain: ai-world-sensing, capital-markets, macro-liquidity
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative, regime, dilution
• Why it matters: AI infra story는 GPU scarcity에서 power, permitting, community scrutiny, construction cost, architecture transition으로 넓어지고 있다. Henry의 SOXL and NXE framing에는 demand뿐 아니라 physical execution bottleneck and local political friction이 들어가야 한다.
• Raw excerpt / abstract: The report says power availability remains the defining constraint on data center growth, while construction costs and community scrutiny are worsening. It states U.S. data center electricity demand is projected to more than double by 2030, forecasts point to 700-1,200 TWh demand, inference already accounts for more than half of AI compute, and 51% of respondents rank availability of power among the top three site-selection factors.
• Cold read: AI buildout is entering a bottleneck diversification phase. Power remains central, but community acceptance, water, grid reliability, emissions, and AC-to-DC architecture readiness now decide who can actually deploy.
• Trading insight: AI exposure should not be sized only from chip demand. Monitor whether the physical buildout can keep pace without credit stress, local backlash, or architecture mismatch.
• Change sensor: data center power queue, construction cost inflation, community opposition, onsite power and CCUS adoption, AC-to-DC architecture timing.
• Follow-up question: Henry의 AI basket에서 which holdings benefit from actual power deployment versus only headline capex narrative?

Run Summary [06:00 KST]

• collected_count: 11
• skipped_duplicates: 18
• domain_mix: microstructure/execution 8개, risk-sizing 4개, macro-liquidity/capital-markets 6개, crypto-structure 3개, AI-world-sensing 2개. Some items map to multiple domains, so total exceeds 100%.
• top_theme: Edge is moving from forecast accuracy to conversion quality: path-dependent execution, market impact, funding plumbing, crypto liquidity design, and AI power delivery decide whether a thesis becomes realized return.
• sharpest_insight: Liquidity has both a mechanism and a clock. Repo funding, AMM fees, order-book impact, and data-center power access all show that capacity that arrives too late or cannot adapt under stress is not real capacity.
• danger_of_misuse: Do not turn these notes into immediate bearishness on AI or crypto, or into paralysis about every trade. The correct use is to lower size when conversion mechanisms are fragile and to demand evidence that liquidity, funding, and execution path can support the thesis.
• next_probe: Convert the liquidity-clock idea into Henry’s checklist: order participation clock, repo funding calendar, crypto settlement or fee adaptation, and AI power deployment timeline.

[07:00 KST] The Anatomy of Stablecoin Transactions

• URL: https://www.bis.org/publ/work1359.htm
• Source Type: research
• Domain: crypto-structure, microstructure, capital-markets
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, narrative, regime
• Why it matters: BIS는 stablecoin activity를 단순 payment count로 보면 구조를 크게 오해한다고 말한다. Henry가 crypto liquidity를 볼 때 transfer volume headline만 보면 trading, lending, arbitrage, liquidity provision, settlement가 한 transaction bundle 안에 섞인 것을 놓친다.
• Raw excerpt / abstract: The BIS paper analyzes 593 million event logs from 141 million Ethereum transactions involving USDT, USDC, and PYUSD. It finds that roughly one third of stablecoin transactions generate multiple steps and nearly 60% of transfer events occur within complex transactions; treating each transfer as a standalone payment misclassifies almost six in 10 transfer events.
• Cold read: Stablecoin volume is not one homogeneous demand signal. It is programmable financial plumbing where the same token can be payment, collateral, arbitrage leg, settlement asset, or liquidity-routing device.
• Trading insight: BTC and crypto beta exposure should not be sized from stablecoin transfer growth alone. Separate organic settlement demand from exchange arbitrage, DeFi leverage loops, and liquidity migration across protocols.
• Change sensor: complex transaction share, USDT versus USDC versus PYUSD timing and urgency, stablecoin co-usage with lending and DEX contracts.
• Follow-up question: Henry의 crypto liquidity dashboard에서 stablecoin transfer volume을 transaction complexity로 어떻게 haircut할까?

[07:00 KST] Anchoring Trust in Money: Innovation Beyond Stablecoins

• URL: https://www.bis.org/publ/arpdf/ar2026e3.htm
• Source Type: official
• Domain: macro-liquidity, capital-markets, crypto-structure
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: study
• Henry Risk Link: liquidity, regime, narrative
• Why it matters: BIS의 핵심은 money가 technology가 아니라 institutional trust라는 점이다. Stablecoin과 tokenization narrative가 강해질수록 Henry는 speed와 programmability보다 par settlement, liquidity elasticity, interoperability, integrity를 먼저 봐야 한다.
• Raw excerpt / abstract: BIS says stablecoins display tokenisation potential for faster and programmable payments, but current designs fall short on foundational properties of money. It argues monetary arrangements depend on a common unit of account, singleness of money, elastic liquidity supply, interoperability, and financial integrity.
• Cold read: Tokenized finance가 커져도 ultimate edge는 code novelty가 아니라 stress 때 par를 지키는 balance sheet와 backstop이다.
• Trading insight: Crypto infrastructure bullish thesis는 regulatory and institutional settlement design과 함께 읽어야 한다. Par가 흔들리는 private money는 bull market에서는 liquidity처럼 보이지만 stress에서는 redemption and funding shock가 된다.
• Change sensor: stablecoin reserve composition, secondary-market depeg frequency, central bank tokenized reserve pilots, unified ledger experiments, bridge failure incidents.
• Follow-up question: Stablecoin reserve composition change가 Treasury bill liquidity and BTC beta에 전염되는 channel은 무엇인가?

[07:00 KST] Customer Order Handling: Best Execution and Order Routing Disclosures

• URL: https://www.finra.org/index.php/rules-guidance/guidance/reports/2026-finra-annual-regulatory-oversight-report/best-execution
• Source Type: official
• Domain: execution, microstructure
• Trading Relevance: 8
• Time Horizon: intraday
• Actionability: now
• Henry Risk Link: execution, liquidity, FOMO
• Why it matters: FINRA의 best execution section은 broker가 order routing duty를 남에게 넘길 수 없고, order type별 execution quality review를 해야 한다고 강조한다. Henry에게도 broker app에서 fill price만 보는 수준이 아니라 order type, venue behavior, PFOF conflict, odd lot handling을 점검하라는 의미다.
• Raw excerpt / abstract: FINRA says firms must use reasonable diligence to ascertain the best market and cannot transfer their duty of best execution. Findings include no assessment of competing markets, unreasonable regular and rigorous reviews, no review of certain order types, inaccurate Rule 606 reports, and incomplete PFOF or rebate disclosures.
• Cold read: Retail execution is a process-quality problem, not only a commission problem. Free trading can hide routing incentives and delayed execution-quality decay.
• Trading insight: Henry should compare limit order discipline, market order slippage, stop activation fills, and odd-lot treatment across brokers or order types before scaling leveraged ETF trades.
• Change sensor: realized slippage by order type, price improvement statistics, fill speed, Rule 606 disclosures, stop order activation quality.
• Follow-up question: Henry의 한국 CMA 계좌 주문 로그에서 market order와 limit order의 realized slippage를 월별로 측정하고 있는가?

[07:00 KST] A New Era for Equity Market Structure: SEC Proposes Rescinding Rule 611 and Rule 610(e)

• URL: https://www.skadden.com/insights/publications/2026/06/a-new-era-for-equity-market-structure
• Source Type: research
• Domain: microstructure, execution, capital-markets
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: execution, liquidity, regime
• Why it matters: Rule 611 trade-through protection은 미국 equity market structure의 핵심 장치였다. 제안이 현실화되면 displayed best price 보호보다 broker best execution process와 routing 판단의 질이 더 중요해질 수 있다.
• Raw excerpt / abstract: Skadden summarizes that the SEC proposed amendments on June 11, 2026 to rescind Rule 611 and Rule 610(e). Trading centers may no longer be required to access all better-priced protected quotations and could execute at prices inferior to quotations displayed elsewhere, while scrutiny may shift toward routing practices and execution quality.
• Cold read: National best bid and offer 중심의 단일 가격 보호 체계가 약해지면, liquidity access는 더 local and broker-specific해질 수 있다.
• Trading insight: Henry의 execution checklist should move from displayed price only to fill certainty, routing quality, market impact, and post-trade drift. 특히 high beta ETF와 small-cap quantum names에서 venue fragmentation risk가 커질 수 있다.
• Change sensor: SEC final rule status, broker routing policy updates, locked and crossed quote frequency, retail execution price dispersion.
• Follow-up question: Rule 611이 약해지는 market에서 Henry의 pre-trade checklist는 어떤 broker and order-type tests를 추가해야 하는가?

[07:00 KST] Powering Intelligence 2026 Executive Summary

• URL: https://powering-intelligence.epri.com/executive-summary.html
• Source Type: research
• Domain: ai-world-sensing, macro-liquidity, capital-markets
• Trading Relevance: 8
• Time Horizon: secular
• Actionability: monitor
• Henry Risk Link: narrative, regime, dilution
• Why it matters: EPRI는 data center electricity demand가 2030년에 미국 전력의 9% to 17%가 될 수 있다고 추정한다. AI trade는 chip shipment만으로 결정되지 않고 grid, gas, nuclear, storage, permitting, state-level load concentration이 capex realization을 제한한다.
• Raw excerpt / abstract: EPRI estimates AI workloads consume 15% to 25% of data center electricity today and that share is rising. It projects U.S. data centers to consume 9% to 17% of U.S. electricity by 2030, up from 4% to 5% today, with Virginia potentially reaching 39% to 57% and seven additional states exceeding 20% in the Medium scenario.
• Cold read: AI infrastructure bottleneck is becoming regional power-market structure. The winners may be those with executable interconnection and flexible load strategy, not only better chips.
• Trading insight: SOXL upside narrative needs a physical deployment check. If power access, transformers, gas buildout, and community acceptance lag, semiconductor revenue timing and valuation multiple can diverge.
• Change sensor: state-level interconnection queues, gas turbine lead times, transformer supply, data center flexible-load programs, electricity price backlash.
• Follow-up question: Henry의 AI exposure에서 revenue recognized from deployable capacity와 order backlog narrative를 어떻게 구분할까?

[07:00 KST] Stablecoins and Safe Asset Prices

• URL: https://www.bis.org/publ/work1270.htm
• Source Type: research
• Domain: macro-liquidity, crypto-structure, capital-markets
• Trading Relevance: 9
• Time Horizon: cycle
• Actionability: monitor
• Henry Risk Link: liquidity, correlation, regime
• Why it matters: Stablecoin issuers are now meaningful short-term Treasury buyers. Henry가 BTC liquidity를 볼 때 stablecoin inflow를 crypto-native cash로만 보면 안 되고, Treasury bill yield, money-market stress, reserve transparency까지 연결해야 한다.
• Raw excerpt / abstract: BIS says dollar-backed stablecoin AUM exceeded $270billionasofDecember2025andstablecoinspurchasednearly$35 billion of U.S. Treasury bills in 2025. A $3.5 billion inflow lowers 3-month Treasury bill yields by 0.71 bps on impact and up to 4 bps within 10 days, with effects amplified during Treasury-market intermediary stress.
• Cold read: Crypto liquidity is now a safe-asset flow. Stablecoin growth can ease bill yields in calm periods but create redemption pressure in stress.
• Trading insight: BTC and MSTR risk should include stablecoin reserve-flow shock. A stablecoin redemption wave can become both crypto liquidity drain and Treasury bill market flow event.
• Change sensor: stablecoin AUM growth, T-bill reserve share, redemption episodes, 3-month bill yield sensitivity, MMF and stablecoin relative flows.
• Follow-up question: Stablecoin AUM drawdown threshold should trigger what reduction in Henry’s crypto-linked position sizing?

[07:00 KST] Continuous Hidden Markov Models for Equity Returns

• URL: https://arxiv.org/abs/2606.23492
• Source Type: paper
• Domain: risk-sizing, macro-liquidity
• Trading Relevance: 8
• Time Horizon: swing
• Actionability: study
• Henry Risk Link: sizing, regime, correlation
• Why it matters: 이 paper는 simple Markov model도 heavy-tailed emissions를 쓰면 equity return의 volatility clustering and regime-conditional VaR를 재현할 수 있다고 주장한다. Henry의 regime filter는 복잡한 AI forecast보다 heavy-tail distribution과 state transition을 분리하는 것부터 시작해도 된다.
• Raw excerpt / abstract: The abstract says the original failure of few-state HMMs is distributional, not temporal. Heavy-tailed marginals, not additional decay modes, closed most of the fit gap, recovering volatility clustering above the i.i.d. baseline and yielding regime-conditional Value-at-Risk that passes a joint conditional-coverage test.
• Cold read: Regime risk는 state count보다 tail modeling에서 먼저 깨진다. Gaussian assumption이 leverage sizing을 과감하게 만들 수 있다.
• Trading insight: Henry의 SOXL or MSTU sizing should use regime-conditional VaR with Student-t or similar heavy-tail emissions, not a fixed percent stop from recent average volatility only.
• Change sensor: inferred high-vol regime probability, regime transition persistence, VaR exception clustering, cross-asset copula stress.
• Follow-up question: Henry portfolio의 daily return history에 3-state heavy-tail HMM을 적용하면 leverage cap이 언제 낮아지는가?

[07:00 KST] Universal Value-at-Risk Superadditivity

• URL: https://arxiv.org/abs/2606.22884
• Source Type: paper
• Domain: risk-sizing, cognition
• Trading Relevance: 7
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: sizing, correlation, leverage
• Why it matters: VaR가 diversification benefit을 과대평가할 수 있다는 것은 이미 알려져 있지만, 이 paper는 heavy-tailed losses에서 VaR가 모든 probability level에서 superadditive가 될 수 있다고 말한다. Henry가 여러 high-beta 종목을 나누어 들고도 실제로는 같은 tail에 노출될 수 있다는 경고다.
• Raw excerpt / abstract: The abstract states that for sufficiently heavy-tailed losses, VaR can be superadditive uniformly across all probability levels. For portfolios satisfying weighted universal VaR superadditivity, every distortion risk measure is superadditive, so an optimal allocation concentrates on a single asset and diversification is never beneficial.
• Cold read: Bad diversification is possible when tails are too heavy and dependence is pathological. More tickers do not automatically mean less ruin risk.
• Trading insight: Henry should treat SOXL, RGTI, MSTU, BTC-linked assets as separate symbols but not necessarily separate tail risks. Portfolio heat must be computed under joint stress, not line-item position count.
• Change sensor: tail dependence between holdings, common liquidity shock, crash beta, VaR exceptions across positions on the same day.
• Follow-up question: Henry portfolio에서 position count가 아니라 common tail factor 기준으로 true diversification score를 어떻게 만들까?

[07:00 KST] Temporal Coarse-Graining of Multi-Sector Default Count Data Generates Posterior-Implied Copulas

• URL: https://arxiv.org/abs/2606.22162
• Source Type: paper
• Domain: risk-sizing, macro-liquidity, capital-markets
• Trading Relevance: 7
• Time Horizon: cycle
• Actionability: study
• Henry Risk Link: correlation, regime, sizing
• Why it matters: Credit stress dependence is horizon-dependent. Monthly sector default data can look different after annual aggregation, and this matters because equity drawdown risk often follows credit regime deterioration with lag and clustering.
• Raw excerpt / abstract: The paper proposes a dynamic low-rank state-space model for monthly multi-sector default-count data. Applied to S&P monthly sector-level default-count data from 1981-01 to 2021-09, a two-factor specification captures market-wide and sector-rotation modes, reproduces annual amplification of leading eigenvalues, and generates heterogeneous copula structures across sector pairs.
• Cold read: Correlation is not a static matrix; it changes with observation horizon and latent credit state. Annual stress can amplify dependence hidden in monthly data.
• Trading insight: Henry should not size cyclical or high-beta exposure using only short-window equity correlation. Credit spreads, default counts, and sector-level stress can reveal when diversification will fail over the holding horizon.
• Change sensor: HY spread dispersion, sector default count trends, leading eigenvalue of credit stress, monthly to quarterly correlation amplification.
• Follow-up question: Henry의 cycle-level risk dashboard에 credit-state factor를 넣으면 SOXL and NXE exposure cap이 어떻게 달라질까?

Run Summary [07:00 KST]

• collected_count: 9
• skipped_duplicates: 63 existing same-day URLs plus repeated q-fin category items were avoided.
• domain_mix: microstructure/execution 3개, risk-sizing 3개, macro-liquidity/capital-markets 6개, crypto-structure 3개, AI-world-sensing 1개. Multi-domain mapping means totals exceed collected_count.
• top_theme: Liquidity is becoming programmable, fragmented, and institution-dependent; execution quality now depends on plumbing design, not only price direction.
• sharpest_insight: Stablecoins, broker routing, Rule 611, heavy-tail VaR, and AI power all point to the same lesson: headline capacity is not real capacity unless it survives stress, routing, settlement, and physical delivery.
• danger_of_misuse: Do not convert these notes into an immediate trade against AI, crypto, or high-beta equities. The right use is to demand better sizing, tail-risk aggregation, execution logs, and liquidity indicators before increasing leverage.
• next_probe: Build a Henry-specific liquidity and execution checklist that combines order-routing quality, stablecoin reserve-flow stress, repo calendar, and AI power deployment constraints.

Daily Cold Synthesis

• Overnight Thesis: Today’s core update is that tradable edge increasingly comes from knowing whether liquidity, execution, settlement, and physical infrastructure can actually convert a narrative into realized return.
• World model update:
  • Stablecoin volume is not a clean payment or cash-demand signal; it is bundled programmable finance where transfer counts can misclassify activity.
  • U.S. equity execution rules may shift away from uniform trade-through protection toward broker-specific routing quality and best execution evidence.
  • AI buildout is migrating from chip scarcity to regional power, permitting, gas, nuclear, storage, grid flexibility, and community acceptance constraints.
• Trading-process lessons:
  • Every large trade needs an execution-quality record: order type, expected participation, fill slippage, venue or broker behavior, and post-fill drift.
  • Position sizing must be tail-factor based, not ticker-count based; multiple high-beta names can be one common crash exposure.
  • Regime filters should separate temporal state transition from heavy-tail distribution; Gaussian shortcuts are dangerous for leveraged ETFs.
• Henry-specific caution:
  • FOMO risk: AI or crypto headlines can create urgency, but infrastructure and settlement bottlenecks decide conversion speed.
  • Leverage risk: SOXL, MSTU, RGTI, and BTC-linked exposures may diversify narratives while concentrating tail liquidity risk.
  • Execution risk: market order convenience can hide routing conflicts and adverse fill quality, especially around volatile windows.
• Signals to monitor:
  • Broker fill slippage by order type and stop activation quality.
  • Stablecoin AUM, reserve composition, redemption spikes, and T-bill flow impact.
  • Rule 611 and Rule 610(e) proposal status plus broker routing policy changes.
  • EPRI-style state-level data center load concentration, interconnection queues, and transformer or gas turbine bottlenecks.
  • Regime-conditional VaR exceptions and cross-holding tail dependence during drawdowns.
• Do not trade directly from this: These clippings are process inputs, not buy or sell signals. They identify where Henry should reduce overconfidence, improve sizing, and demand execution evidence before taking risk.

Sources

• https://arxiv.org/abs/2607.08291 — Robustness in Sequential Decision Making under Evolving Uncertainty: Evidence from High-Frequency Market Making
• https://arxiv.org/abs/2607.06121 — Can Reinforcement Learning Efficiently Discover Price Manipulation?
• https://arxiv.org/abs/2607.04280 — Order Splitting and Liquidity Replenishment Are Jointly Necessary for the Square-Root Law of Market Impact
• https://arxiv.org/abs/2607.01550 — Is Trend Still Your Friend?: A Microstructural Account of the Demise of Short-Term Trend-Following
• https://arxiv.org/abs/2606.31122 — Generating Plausible Stress Scenarios via Large Deviations
• https://www.federalreserve.gov/econres/notes/feds-notes/fluctuations-in-the-treasury-general-account-and-their-effect-on-the-feds-balance-sheet-20250806.html — Fluctuations in the Treasury General Account and Their Effect on the Fed’s Balance Sheet
• https://arxiv.org/abs/2606.19517 — Do Prediction Markets Match Option Prices? Bitcoin Threshold Evidence from Binance and Polymarket
• https://arxiv.org/abs/2602.07018 — The Extremity Premium: Sentiment Regimes and Adverse Selection in Cryptocurrency Markets
• https://www.coinbase.com/institutional/research-insights/research/market-intelligence/2026-crypto-market-outlook — 2026 Crypto Market Outlook
• https://www.mckinsey.com/industries/semiconductors/our-insights/frontiers-of-compute-the-technologies-to-reduce-ai-inference-costs — Frontiers of Compute: The Technologies to Reduce AI Inference Costs
• https://www.colliers.com/en/research/nrep-usdc-data-center-marketplace-2026 — 2026 Data Center Marketplace Report
• https://arxiv.org/abs/2607.08199 — Volatility in Prediction Markets: A Structural Approach
• https://arxiv.org/abs/2607.05141 — Square-Root Price Impact Is Necessary for Endogenous Manipulation Cycles in Learning-Agent Markets
• https://arxiv.org/abs/2607.04221 — A Limit Order Market with Uncertain Informed Trading Participation
• https://arxiv.org/abs/2607.04130 — A Gabor—Epps Uncertainty Principle for Traders
• https://arxiv.org/abs/2607.03239 — Exact Conditional Simulation of Point Processes: Application to Pathwise Market Impact Estimation
• https://arxiv.org/abs/2606.29018 — Liquidity-Based Audit of Algorithmic Trading Strategies
• https://arxiv.org/abs/2607.07465 — Innovating Risk Modelling for Global Funds
• https://arxiv.org/abs/2607.03669 — Split-Session Cluster GARCH for Overnight and Intraday Returns: The Role of Tail Heterogeneity
• https://arxiv.org/abs/2606.30193 — Hidden Dependence and Aggregate Tail Risk
• https://www.federalreserve.gov/monetarypolicy/May-2026-Federal-Reserve-Balance-Sheet-Developments.htm — May 2026 Federal Reserve Balance Sheet Developments
• https://arxiv.org/abs/2606.31675 — Settlement Manipulation in Prediction Markets
• https://arxiv.org/abs/2606.25986 — The Inference-Compute Frontier and a Latency-Efficient Architecture for Limit Order Book Prediction
• https://arxiv.org/abs/2607.02830 — Outcome-Classified Precision Auditing of Filter Rules in Algorithmic DEX Trading
• https://arxiv.org/abs/2607.02823 — Pump.fun Graduation Regime Windows: Survival Analysis of 832,941 Token Launches
• https://arxiv.org/abs/2607.02795 — Coordinated Sniper Cohorts on Pump.fun
• https://arxiv.org/abs/2607.01377 — Liquidity Premium and Investment Horizons
• https://arxiv.org/abs/2607.01198 — When Large Trades Are Not News: Liquidity Tail Risk and Price Discovery
• https://arxiv.org/abs/2607.00475 — End-to-End Parametric Portfolio Policies for Cross-Asset Futures Timing
• https://arxiv.org/abs/2606.30037 — Heads, Not Backbones: Output Heads Dominate Architectures on Fat-Tailed Returns
• https://arxiv.org/abs/2606.26731 — Robust Hedging Valuation Adjustment under Liquidity-Demand Stress
• https://www.bis.org/publ/arpdf/ar2026e2.htm — High Public Debt and Shifting Financial Markets: Challenges for Central Banks
• https://www.bis.org/publ/qtrpdf/r_qt2606b.htm — The Evolution of Central Banks’ Lending Operations
• https://hai.stanford.edu/ai-index/2026-ai-index-report/economy — Economy - The 2026 AI Index Report
• https://arxiv.org/abs/2607.05011 — Reaction-Boundary Variance and Adjoint-Consistent Local-Volatility Projection
• https://arxiv.org/abs/2607.04392 — Adapted Law Invariance and Time-Consistent Dynamic Risk Measures
• https://arxiv.org/abs/2607.04103 — Governing Generative AI Across Financial Institutions
• https://arxiv.org/abs/2607.03082 — Portfolio Optimization and Tail-Risk Analytics of Actively Managed ETFs
• https://arxiv.org/abs/2606.30363 — Strategic Risk Reduction: Self-Protection and Self-Insurance
• https://arxiv.org/abs/2606.29406 — Adaptive AI Delegation under Uncertainty
• https://arxiv.org/abs/2606.27804 — Methods for Uncertainty Representation in Risk Management
• https://arxiv.org/abs/2606.23070 — Mitigating Adverse Selection in Concentrated Liquidity AMMs with Dynamic Fees
• https://arxiv.org/abs/2606.22743 — Endogenous Randomness from Adversarial Market Learning
• https://www.morganstanley.com/insights/articles/powering-ai-energy-market-outlook-2026 — Energy Markets Race to Solve the AI Power Bottleneck
• https://arxiv.org/abs/2607.05484v1 — SHARC: SHAP-Based Interpretability in ML Risk Models for Regulatory Capital
• https://arxiv.org/abs/2606.30363v2 — Strategic Risk Reduction: Self-Protection and Self-Insurance
• https://arxiv.org/abs/2606.26625v1 — Portfolio Optimization for Commodity ETFs under Heavy-Tailed Returns
• https://arxiv.org/abs/2607.08703v1 — MPFlow: Budgeted Max-Flow Optimization on the Bitcoin Lightning Network
• https://arxiv.org/abs/2607.08652v1 — Formal Mechanisms for Market Stability in Self-Interested Agent Societies
• https://arxiv.org/abs/2607.08740v1 — Workflow as Knowledge: Semantic Persistence for LLM-Mediated Workflows
• https://www.newyorkfed.org/markets/data-hub — New York Fed Markets Data Dashboard
• https://www.ropesgray.com/en/insights/viewpoints/102mvfl/data-center-investment-in-2026-ai-demand-power-constraints-and-private-equity — Data Center Investment in 2026: AI Demand, Power Constraints, and Private Equity Trends
• https://arxiv.org/abs/2606.31387 — Signature-Based Optimal Execution for Statistical Arbitrage with Path-Dependent Trading Signals
• https://arxiv.org/abs/2606.29591 — The Bounce Has No Direction: Sign, Magnitude, and the Microstructure of Equity Return Predictability
• https://arxiv.org/abs/2606.25811 — Hierarchical Graph Learning for Calendar Spread Strategies in Commodity Futures Markets
• https://arxiv.org/abs/2606.25466 — Time-Dependent Weighted Directed Networks of Cryptocurrency Interaction from High-Frequency Returns
• https://arxiv.org/abs/2606.24019 — Empirical Confirmation of the Square-Root Law of Market Impact in a U.S. Large-Cap Equity
• https://arxiv.org/abs/2606.22206 — Transaction Costs and Speed in the Ethereum Ecosystem: Scalability of the Mainnet and Layer 2s
• https://arxiv.org/abs/2606.21769 — Optimal Dynamic Fees for Automated Market Makers: A Stochastic Control Approach to Loss-Versus-Rebalancing
• https://www.federalreserve.gov/econres/feds/beyond-reserves-the-federal-reserves-balance-sheet-and-the-repo-market.htm — Beyond Reserves: The Federal Reserve’s Balance Sheet and the Repo Market
• https://www.dallasfed.org/research/economics/2026/0528 — Rising Hedge Fund Leverage Affects Monetary Policy Implementation
• https://www.dallasfed.org/research/economics/2026/0212-levymccormick-repotiming — Domestic Banks Are Inelastic Providers of Marginal Funding to Repo Markets
• https://www.bloomenergy.com/wp-content/uploads/power-report-2026-mid-year-pulse-final.pdf — 2026 Data Center Power Report: Mid-Year Pulse
• https://www.bis.org/publ/work1359.htm — The Anatomy of Stablecoin Transactions
• https://www.bis.org/publ/arpdf/ar2026e3.htm — Anchoring Trust in Money: Innovation Beyond Stablecoins
• https://www.finra.org/index.php/rules-guidance/guidance/reports/2026-finra-annual-regulatory-oversight-report/best-execution — Customer Order Handling: Best Execution and Order Routing Disclosures
• https://www.skadden.com/insights/publications/2026/06/a-new-era-for-equity-market-structure — A New Era for Equity Market Structure: SEC Proposes Rescinding Rule 611 and Rule 610(e)
• https://powering-intelligence.epri.com/executive-summary.html — Powering Intelligence 2026 Executive Summary
• https://www.bis.org/publ/work1270.htm — Stablecoins and Safe Asset Prices
• https://arxiv.org/abs/2606.23492 — Continuous Hidden Markov Models for Equity Returns
• https://arxiv.org/abs/2606.22884 — Universal Value-at-Risk Superadditivity
• https://arxiv.org/abs/2606.22162 — Temporal Coarse-Graining of Multi-Sector Default Count Data Generates Posterior-Implied Copulas

Keywords

• AI research
• overnight briefing
• trading process
• market microstructure
• execution quality
• order splitting
• square-root market impact
• liquidity replenishment
• high-frequency market making
• reinforcement learning market manipulation
• trend following
• stress scenarios
• large deviations
• tail risk
• tail-factor sizing
• portfolio heat
• macro liquidity
• Fed balance sheet
• Treasury General Account
• ON RRP
• stablecoins
• crypto market structure
• prediction markets
• Bitcoin options
• sentiment regimes
• adverse selection
• AI infrastructure
• data center power
• grid bottlenecks
• Rule 611
• broker routing
• best execution
• Henry risk dashboard
• SOXL
• MSTU
• RGTI
• BTC-linked exposure