Every engine races.
Nothing gets hidden.
We run 11 isolated engines against the same market in paper. The board ranks them on their whole life, so the leader is the one that is genuinely ahead since it started — switch the window and they re-rank in front of you, with age and drawdown printed on every lane so no good session gets to pose as a track record.
Pick a window. Watch them re-rank.
Top 6 of 11 live engines · ranked by return over each engine's whole life.
- +4.71%worst drop −8.18% · +1.21% last 24h · 28d live
- +0.38%worst drop −2.00% · +0.49% last 24h · 18d live
- +0.36%worst drop −2.63% · +0.61% last 24h · 18d live
- +0.14%worst drop −1.56% · +0.54% last 24h · 18d live
- +0.04%worst drop −1.99% · +0.62% last 24h · 18d live
- -0.15%worst drop −1.63% · +0.51% last 24h · 24d live
Bar length = return since that engine started · each lane on its own life, age printed beside it
Every engine trades paper money against live prices and carries its own modeled costs. Engines are ranked on their whole life, so a strong day can never stand in for a record — but they did not all start on the same day, which is why each lane prints its age, its worst drawdown and its last 24 hours. Forward test engines are candidates still earning their mandate — they race, but they cannot be mirrored.
Paper performance is hypothetical and does not guarantee future results. Returns are normalized and include each engine’s modeled trading costs; account size, leverage and individual positions are not part of the ranking.
A leader that cannot be grandfathered in.
Four mechanical rules, applied on every refresh. No editorial thumb on the scale.
The lifetime leads
The board opens on each engine's whole life, so the headline goes to whoever is actually ahead since it started — not to whoever won the last session while deep in the red.
You pick the window
Switch to 24 hours or 7 days and the whole field re-ranks. The window is named on every number it produced.
One clock, or one whole life
Inside a rolling window every engine is measured between the same two timestamps. On the lifetime view each engine runs from its own first mark — so every lane prints its age.
Two numbers, always
No result stands alone: a short window carries the whole-life number, the lifetime view carries the age, the worst drawdown and the last 24 hours.
These engines are young.
The fleet was reset to a clean slate, so the oldest engine here has 39d of live history and the newest 18d. That is nowhere near enough to judge a strategy — it is why the window switcher exists and why every lane shows its age. Come back as the record lengthens; we are not going to reset it away.
Each engine against its own mandate.
These references have different provenance and are context — not a second leaderboard.
Curated Coins · staggered
The curated list with laddered rebalancing.
- Last 24h
- +0.51%
- Since it started
- -0.15% · 24d
- Backtest CAGR
- +21%/yr
- Backtest max DD
- −4.5%
Curated Coins
The same books on a hand-curated list of currently-liquid majors.
- Last 24h
- +0.55%
- Since it started
- -2.07% · 24d
- Backtest CAGR
- +21%/yr
- Backtest max DD
- −4.5%
Staggered rebalancing
The same books, laddered into phase-offset tranches — the one overlay that beat hold-to-boundary.
- Last 24h
- -0.07%
- Since it started
- -3.22% · 24d
- Backtest CAGR
- +28%/yr
- Backtest max DD
- −6.4%
Flagship · locked universe
The proven four-book market-neutral engine on its frozen 62-coin research universe.
- Last 24h
- +0.03%
- Since it started
- -4.15% · 24d
- Backtest CAGR
- +20%/yr
- Backtest max DD
- −12.0%
Trend sleeve · long-only
Directional structural-breakout engine — active only in a BTC uptrend.
- Last 24h
- +0.00%
- Since it started
- -6.06% · 39d
- Backtest CAGR
- +55%/yr
- Backtest max DD
- −39.9%
Adaptive liquid universe
The same books ranked over the self-maintaining liquid universe.
- Last 24h
- -0.83%
- Since it started
- -16.86% · 24d
- Backtest CAGR
- +26%/yr
- Backtest max DD
- −6.5%
5 candidates with nothing to show yet.
No published backtest, no executor bridge — they race in public from day one and have to earn a mandate on live evidence alone.
- 01WorkbenchThe original nomination engine, still running as a long-lived control.+1.21%24h+4.71%28d
- 02Curated II · no carry, staggeredNo-carry books on the second curated universe, laddered.+0.49%24h+0.38%18d
- 03Curated II · no carryThe second curated universe with the carry book removed.+0.61%24h+0.36%18d
- 04Curated II · staggeredThe second curated universe with laddered rebalancing.+0.54%24h+0.14%18d
- 05Curated IIA second curated universe, forward-testing in public from its first day.+0.62%24h+0.04%18d
How to read this page
Live ranking uses only marked paper equity inside the window you selected. Backtest references use each engine’s own documented methodology — holdout, point-in-time, overlay, scouting or survivorship-adjusted — and therefore must not be ranked against one another. Engines marked forward test are candidates with no published backtest and no executor bridge: they race in public from day one, but they are not available to mirror. Individual positions and coin lists remain a paid view.