deadlist.run Get early access

Survivorship-free dataset + strategy audits · Binance perpetuals

Test your strategy against the coins that died.

Deadlist is two things: a dataset of all 741 Binance perps ever listed — including the 102 delisted — and an audit that runs your strategy against it. Ours showed +$299,099 in-sample. The honest test showed −$138,419. We found out for $0. Most traders find out with their deposit.

The autopsy

Four strategies. Four green backtests. Four honest kills.

These are real results from our own research — every one looked deployable until it met a point-in-time universe, dead contracts, and gem-stripping. Total real money lost: $0. That's the product.

alt-perp trail convexity

in-sample+$299,099
honest OOS−$138,419

The mega-gems that carried it never appear under point-in-time selection. The universe was today's survivors, chosen by today's liquidity — doubly biased.

SELECTION ARTIFACT

memecoin discovery scanner

headline net+$5,642
base ex-gems−$1,488

Two tokens carried 100% of the profit. Only 5% of spikes were liquid enough to enter — and 13 delisted rugs were silently excluded from the math.

RUG FUNNEL

1h alts, uncapped winners

backtest net+$80,262
ex-top-10 trades−$1,448

Max drawdown ($112k) exceeded the entire profit. Strip the ten best trades of 98,380 and the edge is gone.

10-TRADE MIRAGE

energy trend-following

26-yr edgereal ✓
time underwater8.5 years

The only edge that survived OOS — and drawdown stress showed 43% peak-to-trough and nearly a decade below high-water. Real, and untradeable at size.

UNTRADEABLE

Why backtests lie

Three biases no chart platform shows you

The dead are missing

Delisted contracts vanish from standard data feeds. 102 of 741 Binance perps ever listed are dead — and if your strategy would have entered them, those are losses your backtest never saw.

The frozen universe

Testing on "the top 40 coins by volume" means today's winners, selected by liquidity they earned by pumping. In our flagship audit this bias was bigger than survivorship itself — it flipped the sign even restricted to survivors.

Σ

Gem concentration

Every strategy we tested had a negative median trade — all profit came from rare outliers. Strip the top ten trades and ask what's left. If the answer is a loss, you're betting on lightning striking on schedule.

The actual console

Not a mockup — the pipeline that ran every number on this page

Live dryrun dashboards across 26 instances, a cross-strategy verdict grid, walk-forward OOS tables, order-book capacity scans, and a hard-guard executor with every rejection tested. These are the screens your audit report comes from.

What you get

The dataset and the audit

Product 01 · Data

The Deadlist dataset

One-time purchase · early-access pricing for the waitlist

  • Every Binance USDⓈ-M perp ever listed — 741 contracts, reconstructed from exchange archives
  • Including all 102 delisted contracts with full 1d + 4h klines to their final candle
  • Per-contract listing spans and point-in-time volume, so you can build the universe as it existed on any date
  • Clean CSV/JSON — drops into any backtester

Product 02 · Audit

The honest audit report

Per-strategy · send your trade log or rules, get the verdict report

  • Walk-forward OOS table — is your edge broad, or clustered in two lucky windows?
  • Survivor vs. corrected contrast — exactly how much the dead names change your number
  • Gem-stripping — your net with the top 5 / top 10 trades removed
  • Capacity ceiling — the position size where slippage eats your edge, from live order books
  • Drawdown & time-underwater — the number that decides whether you could actually hold it

The story

Built by a trader it saved four times

This pipeline wasn't built as a product. It was built out of self-defense — by a founder about to deploy real money into a strategy showing +$299,099 in backtest.

Before wiring the wallet, one question: would this have worked on the coins that no longer exist? Reconstructing the full point-in-time universe — every contract ever listed, dead ones included — took days. Running the honest test took minutes. The sign flipped.

Then it flipped three more strategies. A memecoin scanner whose profit was two lucky tokens. A 1-hour system whose entire edge was ten trades. An energy edge that was real — and spent 8.5 years underwater.

The machine's track record is that it stopped its own builder from making four expensive mistakes. Total deployed into fake edges: $0.

Now it's being packaged so it can do the same for you — before the market runs the test with your money.

Early access

Audit your backtest before the market does