1. Market Regime Snapshot

VIX Term Structure

Tenor Level vs. Spot
VIX Spot 15.3 —
VIX 3M 18.0 +2.7 (Contango)

Realized vs. Implied Vol: SPY 20D RV = 10.4%, VIX = 15.3. Spread = -5.0pp (realized vol below implied — calmer than feared).

Cross-Asset Momentum (1-Month)

Asset 1M Return Signal
SPY (Equity) -0.2% 📉 Negative
TLT (Bonds) -5.2% 📉 Negative
GLD (Gold) -7.3% 📉 Negative
UUP (US Dollar) +3.1% 📈 Positive

Regime: Mixed/transitional regime

Last updated: 2026-10-05 11:31 UTC

2. Factor Performance Dashboard

Source: ETF Proxies (MTUM, VLUE, QUAL, USMV, IWM vs SPY)

Note: ETF proxy returns include market beta and are not directly comparable to factor-neutral French library returns.

Factor Weekly 1M 3M Mean (52W wkly) Std (52W wkly) Z
Momentum +1.38% +8.92% +0.57% +0.49% 2.81% +0.31
Value -0.71% -0.46% +3.69% +0.93% 2.60% -0.63
Quality +0.40% +1.61% +2.96% +0.29% 1.55% +0.07
Low Volatility -0.13% -2.42% +0.66% +0.10% 1.19% -0.19
Size +0.06% -4.81% -8.06% +0.03% 1.55% +0.02

No factor stress signals this week (all within ±2σ).

Last updated: 2026-10-05 11:31 UTC

3. Macro Signal Tracker

Indicator Current 1W Change Signal
HY Spread (OAS) 310bps +17bps ⚠️ Widening
IG Spread (OAS) 85bps +4bps ⚠️ Widening
2s10s Yield Curve 0.45% +0.09% Normal
3M10Y Yield Curve 1.09% +0.16% Normal
Fed Funds Rate 3.75% N/A →
HY − IG Spread 225bps — Risk sentiment proxy

Macro Summary: Neutral macro backdrop

Last updated: 2026-10-05 11:31 UTC

4. Quant Research Digest

Three papers I found worth reading this week:

When a Correct Reward Is Not Enough: Diagnosing and Guiding PPO in an Analytically Solved Broker-Trader Game — Siu Tung Wong et al. arXiv

This paper evaluates whether Reinforcement Learning models (specifically PPO) can leverage established analytical solutions in financial optimal-control problems to guide their training. It is an excellent read for anyone building RL trading agents, highlighting that providing a correct reward function is often insufficient without structural mathematical guidance to help the model escape local optima.

Mixture-of-Experts for Cryptocurrency Order Execution: Training Stability, Tail Risk, and Failure Modes — Alexander Ardaiz et al. arXiv

The authors benchmark Double Deep Q-Learning against partitioned Mixture-of-Experts architectures using high-frequency limit order book data to optimize trade execution. This is highly practical for execution research because it emphasizes evaluating model stability across different training seeds and documenting tail-risk failure modes, rather than just reporting the one lucky backtest that achieved the highest return.

Event History Over Scale: Compact Transformers for Low-Latency Limit Order Book Forecasting — David Schaurecker et al. arXiv

This research introduces highly compact, causal transformer architectures (down to 7,203 parameters) specifically optimized for the strict latency constraints of intraday order book forecasting. It tackles a very real engineering bottleneck in deployment, proving that squeezing predictive quality out of tiny, fast architectures is often more useful in production environments than blindly scaling up parameter counts.

Last updated: 2026-10-05 11:32 UTC

5. Stat of the Week

Stat Value Context
CBOE Skew Index 145 Elevated tail risk (>130)

The Skew Index remained elevated at 145 this week, continuing to diverge from the relatively calm 10.4% realized volatility on the SPY. With high-yield spreads also widening 17 basis points to 310bps, it is interesting to observe persistent stress pricing in out-of-the-money puts and credit markets while the top-level equity index barely moves. This divergence is a clean structural reminder to monitor cross-asset metrics and implied tails rather than relying solely on rolling equity variance to evaluate market risk.

Last updated: 2026-10-05 11:32 UTC

Generated: 2026-10-05 11:32 UTC