Most strategies die here.
The survivors are judged in public.
Bred on real futures history, culled at four gates, frozen into cards — then graded every day on candles they've never seen. Winners hang on the wall. Losers stay on it.
Strategies here are guilty until proven otherwise.
Line four of our validator's own documentation reads: "without this module, the Factory becomes a backtest fraud factory." The engine names the industry's sin before it runs a single test — then spends its whole life refusing to commit it.
So the burden of proof sits on the strategy. Too few trades to judge? That's a fail, not an "inconclusive" — if you didn't take the test, you get zero, not fifty. A flawless record isn't waved through either; here, a perfect score is treated as a red flag, not genius.
The engine's standing rules →Nothing cherry-picked, nothing deleted. Futures are risky — this is research, not financial advice.
A candle is a photo.
This engine watches the video.
Inside one candle, a stop-loss and a take-profit can both look true — and a lazy engine gets to assume which came first. It will assume the one that flatters the result. This engine isn't allowed to assume anything: it replays every trade against every second of the day — so when a stop and a liquidation land on the same second, liquidation wins — because that's what the exchange would do to you.
Fees on both sides, slippage on every fill, funding charged like a toll every eight hours. The score you see is what's left after the bills, not before.
The numbers below aren't features. They're the cost of refusing to assume.
Half the market
is a fall.
Before building any of this, we counted ten years of candles, timeframe by timeframe — every close, up or down. The split isn't what the ads assume: half of everything the market has ever done is a fall. Yet nearly every strategy sold anywhere is a long, built to earn only in the half where price rises.
So this lab breeds short strategies with the same seriousness as longs — separate champions for each direction, graded at the same gates, hung on the same public wall. Half the Boutique's seats are reserved for shorts, by design.
A shelf that can earn in both halves of the market is dressed for more weather than one that only works while the sun is out.
Four steps,
start to signal.
Draw a strategy from the vault, read its grade, keep what earns your trust — and let it message you the moment it moves.
Survive all four,
or die in testing.
History is cut into windows and judged window by window — at least half must come up profitable, so one lucky stretch can't carry the grade.
The most recent stretch is graded separately against the earlier one — an edge that faded late, the signature of memorized history, fails here.
Parameters get nudged up and down — a real edge bends without breaking; a fragile one-off fit shatters.
Trades are reshuffled a thousand times — a smooth curve that only exists in one lucky ordering of wins and losses gets exposed.
Rarity is earned,
not rolled.
Every survivor comes back as a card with a seal on the front. The tier isn't a dice roll — it's the strategy's test score, sealed the moment it passes the gates.
Saved cards can't
rewrite their past.
A saved strategy is frozen at creation — the recipe can't be quietly fixed after the fact. From that day forward its record grows daily, wins and losses alike. What you read is what happened.
Don't trust — verify. Save a strategy and watch it with your own eyes: the record updates every day it stays live, and if it fades, you'll see that too.
That's the whole trick, and it isn't one: keep the losers on the board, and the survivors mean something.
The board is public and refreshes every day — losers kept, nothing quietly deleted. The first champions are proving themselves on live candles right now.
Pull your first card.
Nothing cherry-picked, nothing deleted. Futures are risky — research, not financial advice.