Papers With Backtest

[Papers With Backtest @paperswbacktest]

@JoeTigay @SchwabNetwork @NPetallides Long VIX futures fight contango ~80% of the time, so the overlay bleeds via negative roll. 'Buy low/sell high' only nets positive when rebalancing harvests enough spike convexity to beat that carry drag. Sizing is the edge, not just being long vol.

[Papers With Backtest @paperswbacktest]

@lopezdeprado A big driver of that 80% is multiple testing. Mine hundreds of factors and a t-stat of 2 is far too lax. Harvey, Liu & Zhu put the hurdle near 3. Your deflated Sharpe nails it: most 'discoveries' don't survive once you count the trials.

[Papers With Backtest @paperswbacktest]

@profplum99 The low-vol anomaly is the cleanest counterexample: high-beta stocks have historically underperformed low-beta on a risk-adjusted and often absolute basis. Frazzini-Pedersen tie it to leverage constraints forcing demand into high-beta names.

[Papers With Backtest @paperswbacktest]

@JohnHCochrane Agree. This is Asness's 'Fight the Fed Model' point: pairing E/P with nominal yields only worked over 1965-2000 when inflation drove both. But even vs TIPS, the yield gap predicts 10y returns worse than CAPE alone.

[Papers With Backtest @paperswbacktest]

@quant_arb Half the story. Even with zero alpha, the impact-vs-timing-risk tradeoff sets the baseline schedule. Short-term alpha just tilts it front- or back-loaded. Treating execution as pure alpha is how you overtrade into your own impact.

[Papers With Backtest @paperswbacktest]

Another fund raising seven figures to throw foundation models at markets. But financial time series are mostly noise, not language with rich structure. Bigger models overfit faster. The bottleneck was never compute. It is a signal that survives costs and capacity.

[Papers With Backtest @paperswbacktest]

@SSRN Watch the branching ratio in these Hawkes fits, it often creeps near 1: reads as pure self-excitation when it is really a latent common factor moving both sides. Separating endogenous feedback from exogenous news flow is where most trade-intensity models quietly break.

[Papers With Backtest @paperswbacktest]

@quant_arb HPCA clusters on linear correlation, so it can lump features that are linearly redundant yet carry distinct nonlinear signal. Worth screening the within-cluster survivor with mutual information or distance correlation before dropping the rest.

[Papers With Backtest @paperswbacktest]

'React, don't anticipate' sounds wise but it's untestable as stated. Make it a rule: enter on confirmation, not forecast, then backtest both and compare Sharpe net of costs. If feeling the market's flow beats a coded signal, prove it. Usually it's just hindsight talking.

[Papers With Backtest @paperswbacktest]

Oldies but Goodies

Machine Learning Risk Models build equity covariances from averaged k-means clusterings of recent returns, not a factor model, giving an invertible matrix for portfolio optimization.

https://t.co/lwNPlIk3mw

[Papers With Backtest @paperswbacktest]

Backtests treat every fill as real and final. But exchanges keep an "obvious error" rule: clearly erroneous trades get busted hours later. Your best print of the day can simply vanish. Edge that only lives in fills no human would honor was never edge.

[Papers With Backtest @paperswbacktest]

@6_Figure_Invest RV vol books are implicitly short correlation. When several structural pairs decouple at once, it is rarely independent: one positioning/liquidity shock hits them together, so the diversifying legs realize the same move and tail correlation goes to 1.

[Papers With Backtest @paperswbacktest]

The biggest tail risk in a trading system isn't a bad signal. It's wiring untested, high-blast-radius tooling into your live execution path. Operational risk earns no Sharpe, but its drawdown is uncapped. It isn't backtestable, which is exactly why people skip it.

[Papers With Backtest @paperswbacktest]

@TheMintingM The fatal flaw is momentum crashes: severe negative skew when beaten-down shorts rip in a sharp rebound (1932, 2009). The documented fix is scaling exposure by recent realized volatility, which roughly doubles the Sharpe and tames the left tail.

[Papers With Backtest @paperswbacktest]

A 7-year streak of positive annual returns feels like proof of edge. It usually isn't. Shift the start date and most low-vol strategies show one. The testable question is never the streak. It's Sharpe net of costs, worst drawdown, and out-of-sample stability.

[Papers With Backtest @paperswbacktest]

Oldies but Goodies

Build a global expected business condition factor from lagged OECD leading indicators, then time international equity markets monthly. It predicts aggregate returns in and out of sample, working through the cash-flow news channel.

https://t.co/XF2KlfRXKh

[Papers With Backtest @paperswbacktest]

A stretched pairs spread looks like free money. Before you fade it, ask: did the relationship break, or just deviate? Most stat arb mispricings are regime changes wearing a mean-reversion costume. The edge isn't spotting the gap, it's knowing it closes.

[Papers With Backtest @paperswbacktest]

Full breakdown + Python code and backtest results → https://t.co/0tKKqYivOw
Explore all strategies and datasets on https://t.co/9KTq6MzFVQ

[Papers With Backtest @paperswbacktest]

Bitcoin trades 24/7 — and that never-sleeping market leaks a pattern. From Oct 2015 to Feb 2022, simply buying BTC at 21:00 UTC and selling at 23:00 UTC each day returned ~33% annualized. Two hours a day. 🧵

[Papers With Backtest @paperswbacktest]

The insight that ties it together: a new high and a new low can't happen at the same time. So the momentum and reversion signals are mutually exclusive — you can fuse them into one strategy. The result: the smoothest equity curve of all.