Research
What was actually tested, what survived, and the six ways the measuring instrument turned out to be broken.
- In useValidated and published as a live number in the project.
- PreliminaryPasses its tests but rests on few independent observations.
- Finding, not usedReal, verified information that no simple construction manages to harvest.
- Null resultSeriously tested with no defensible edge — or real but impossible to collect after costs.
- Method lessonNot a finding about the market: a finding about how to measure it.
Findings
Each page is one hypothesis with an economic mechanism behind it, its measured numbers and a verdict. The null results are written up in the same detail as the positives, on purpose.
In useThe basis carry
Collecting the interest leveraged longs pay, with no exposure to price direction at all.
PreliminaryTimed variance selling
Selling 30-day volatility only into the months the regime model calls calm — the one result that brings together the project's two best pieces.
Finding, not usedThe variance risk premium
The volatility the options market charges systematically exceeds the one that actually happens — and not only in crypto.
In usePredicting risk, not direction
With the same data and the same method, price direction can't be predicted and volatility can — but turning that into an exposure rule didn't work.
Finding, not usedLeverage crowding
Assets whose leveraged longs are most crowded do worse over the following days — real information that no simple portfolio manages to collect.
Finding, not usedOrder-flow reversal
An hour of unusually aggressive buying is followed by reversal over the next two hours, in 14 of 14 assets — and the entire edge lives inside the spread.
Finding, not usedThe compression of the premia
Four independent fee streams, measured separately, falling by the same order of magnitude on the same dates.
Null resultThe overlay that was retired
A rule that raised exposure to 1.8x the further price sat below trend, retired in 2026 after it was shown to subtract value at every scale.
Preliminary2026, actually applied
What each validated piece would have returned since 1 January 2026 — including the uncomfortable result.
Null resultPredicting price direction
6.68 million hypotheses, six independent methodologies and no defensible edge. The project's best-established null result.
Null resultClassic patterns, Elliott and Wyckoff
Head and shoulders, double bottoms, triangles, flags, Wyckoff, Elliott waves, SMC/ICT and divergences. All null — and one instructive bug.
Null resultThe other null families
Cross-sectional momentum, seasonality, macro, stablecoins, ETF flows, futures positioning and options skew — tested with the corrected method.
Method
The six ways the measuring instrument failed, and the calculation that would have prevented each one. These pages carry the most useful widgets on the site.
Method lessonPower: what the experiment can see
The thirty-second calculation that would have saved weeks of work: before running a test, knowing the smallest effect it could possibly detect.
Method lessonMultiplicity: the best of many tries
With 432,053 hypotheses tested, the best result from pure chance already sits at 4.6 standard deviations. Anything below that is not a finding.
Method lessonThe test that wasn't measuring skill
For two weeks, the project's "mandatory" acceptance criterion scored a signal with a real +40% per trade exactly the same as one with no edge at all.
Method lessonWhat it's compared against: the null
A signal isn't compared against zero, but against what chance would have given in the same era and with the same duration. Getting the era wrong multiplies the bar by seven.
Method lessonThe holdout and pre-registration
Freezing 2023 onward, never looking at it during exploration, and opening that door ONCE per candidate.
Method lessonLooking ahead without noticing
Seven confirmed look-ahead and misalignment bugs, and why the most famous of them — repainting — produces spectacular results.