Methodology,
shown in full.
The research discipline behind every number here: survivorship-clean data, point-in-time universes, walk-forward validation, overfitting tests, and the audits we run against our own work first.
13 notes · Methodology
Probability of backtest overfitting (PBO): how much of the backtest is the act of choosing it?
Given every variant you tried, how likely is your winner just the luckiest? The CSCV method, our equity audit (128 trials, PBO 2.6%), and the crypto gate.
Backtest the point-in-time universe. Not the one in your download.
Which assets qualified on each date, using only what was knowable then: why current-list backtests simulate a fiction, and what point-in-time requires.
Proof of process, never proof of profit.
The evidence ladder from backtest to audited live, what attested paper trading proves about a track record, the fills and capital gaps, and the 8-week gate.
Secret sauce protects the pricing, rarely the edge.
What secrecy protects, why a published trading method doesn't decay with adoption, the index-provider precedent, and the one boundary we keep, with an expiry.
Same daily behavior, different year. Check the calendar.
The calendar hiding inside volatility, Sharpe and drawdown figures, three traps when comparing crypto and equity statistics, and the conventions we pin.
The curve looks magnificent. That was the selection criterion.
Memory versus understanding, the five amplifiers that make crypto the easiest market to fool yourself in, the trial-count arithmetic, and the plateau principle.
Tune on one era. Grade on one it never saw.
Why full-window optimization takes the exam with the answer key, the holdout and walk-forward designs, what validation costs, and our pre-registered split.
Every ranking hides its biggest call in the guest list.
The monitor and research universes, every inclusion rule with its reason, the ledger for rebrands and delistings, and why public criteria resist gaming.
Price needs capital to move. Volume just needs a printer.
Why reported crypto volume lies more than anywhere, the landmark evidence, four detection heuristics anyone can apply, and how our pipeline mitigates it.
Half the audit came back clean. We published the other half too.
PBO, deflated Sharpe, and a survivorship bound aimed at our own flagship: selection came back clean, the universe did not, and the report became our rules.
We ran our stock engine on crypto. It lost. Here is the anatomy.
We ran a validated US-equity breakout strategy on five years of crypto bars, zero re-tuning, and published the losing backtest: what failed and what carried.
The coins that aren't there are the whole story.
Exchange APIs only remember the living. How many coins die, what the graveyard does to crypto backtests, and the point-in-time universe rules we commit to.
Screenshots ask for trust. Hashes make it unnecessary.
How cryptographic commitments make a published record tamper-evident: the hash chain, Bitcoin timestamps, what it proves, and how to verify a day yourself.