Your chart shows one coin at a time. This shows all 81.
Ten indicator families across nine time scales, recomputed for every one of 81 coins, every minute. Build a rule against years of real history, then watch the live market for the moment it fires.
The rule you test is the rule that watches.
On most platforms the backtester and the live screener are separate systems with separate maths, so a condition can quietly mean two different things. Here it is one system.
Exchanges typically charge 0.015-0.045% per side depending on your fee tier and whether you take or make. This test assumes 0.05% on both sides — the worst case.
No coin satisfies all four conditions right now. This rule fired 721 times in eight months — roughly three times a day. Being rare is the rule doing its job.
Take-profit is twice stop-loss, so a 40% win rate still finishes ahead. Judge a result by trade count, not by win rate.
Commission is charged at the worst case: 0.05% in, 0.05% out, taker on both sides. Your own exchange tier and maker orders will cost less.
1,802,500 USDT of volume, from 100-dollar positions. On a real exchange that volume raises your fee tier, which lowers every trade that follows.
Leverage multiplies losses as well as gains.
Measurements you will not find ready-made elsewhere.
Taker buy ratio
Aggressive buying as a share of total volume
Computed across all nine time scales for every coin. Outside professional order-flow terminals this is rarely offered as a ready value.
Most coins sit near the middle. The extremes are where the information is.
Read more →Volume against a coin's own normal VOAR
Volume measured against that coin's own seven-day baseline, not against the market. A volume that is noise for BTC is an event for a small token. This is what lets 81 coins be ranked against each other on the same scale.
Equal distance means equal ratio.
Read more →Momentum histogram delta MHD
The histogram measures the speed of momentum. This measures the change in that speed. "Strong but fading" and "weak but building" produce the same histogram value; this is the only thing that separates them.
Same histogram value. Opposite direction.
Twenty-one price lookbacks
Most tools give you 1h, 4h and 24h. Here there are twenty-one. If the 10-minute change is large but the 20-minute change is not, the move began in the last ten minutes. Three lookbacks cannot tell you that.
Most tools give you three of these.
Read more →Most backtest results are wrong, and the reason is invisible.
A simulation that uses information which did not exist yet at the moment of the decision produces excellent results that mean nothing.
The future never leaks into the past
A minute's indicators are computed only from data finalised through the previous minute. Entry is taken at that minute's open. Stop and target are checked against the previous minute's finalised high and low.
Ambiguity resolves against you
If stop and target both appear within the same minute, which came first is unknowable. The engine takes the stop. Results are deliberately pessimistic.
Empty is not zero
A value that cannot be computed stays empty. It is never written as 0, never carried forward from the previous minute, never interpolated. Every 0 you see was measured.
Live and historical are forced to agree
Two independent pipelines — one running against the live market, one that built the archive — are compared value by value and required to produce the same number.
Look before you sign up.
Analyze, Discover and Backtest are all open to visitors. An account is only needed to run a backtest of your own.
Not every coin is listed. The selection is the highest-volume, most liquid pairs — the ones where technical analysis produces reliable results and a backtest reflects real market conditions.
See every supported coin →Is it an idea, or a habit? A number tells you the difference.
Start free — no card required