Measured, not marketed
Our track record, in full
Every prediction the committee publishes is recorded with its entry, target and invalidation, then scored against real market data when it resolves. Nothing is excluded for looking bad. The figures below are read live from that record each time this page loads.
Over 92 decided predictions the system is currently ahead on expectancy. That is a small sample and not a promise of anything; read the confidence interval below before drawing conclusions from it.
The headline numbers
How much to trust that win rate: 92 resolved trades — moderate confidence (±10pts). The 95% confidence interval runs 29.8%–49.3%, so the true rate could sit anywhere in that range.
Why win rate is the wrong number
A high win rate sells well and means very little. What decides whether a method makes money is expectancy — average win × win rate, minus average loss × loss rate. A system that wins 35% of the time while making three times more on winners than it loses on losers is profitable; one that wins 70% of the time with the reverse ratio is not.
Is the confidence figure honest?
A win rate tells you whether the calls worked. This tells you whether the confidence attached to them meant anything — of the calls made at 70% stated confidence, did roughly 70% hit? Over 61 resolved predictions the system claimed 68% on average and delivered 34% — it was over-confident by 34 points. The number shown in the product is corrected for that gap before you see it.
Bar = share of those calls that actually hit. Vertical line = the confidence claimed. A bar falling short of its line is over-confidence in that band. Bucket samples are small; read the direction, not the decimals.
Method
- Sample. 121 predictions recorded, 121 resolved, 0 still open. 29 expired without reaching either target or invalidation and are excluded from the win rate — which is why it is computed over 92 decided calls, not 121.
- Scoring. A call is a win if it reached its stated target before its invalidation level, and a loss if the reverse. Outcomes are resolved against exchange candle data, not judged after the fact.
- No survivorship.Every recorded prediction is included. Nothing is deleted, re-labelled or filtered for looking bad, and there is no separate “selected calls” set.
- Costs. Figures are gross of trading fees, funding and slippage. Real-world results would be worse, not better.
By asset
| Symbol | Decided | Expired | Win rate | Net % |
|---|---|---|---|---|
| BTCUSDT | 25 | 13 | 52.0% | +35.58% |
| BNBUSDT | 13 | 1 | 23.1% | +6.37% |
| XRPUSDT | 11 | 1 | 36.4% | +21.36% |
| SOLUSDT | 10 | 2 | 30.0% | +17.85% |
| ETHUSDT | 6 | 5 | 33.3% | +14.71% |
| PAXGUSDT | 6 | 0 | 33.3% | -2.12% |
| WLDUSDT | 4 | 0 | 75.0% | +11.88% |
| BLURUSDT | 2 | 0 | 0.0% | -32.93% |
| VANRYUSDT | 1 | 1 | 100.0% | +18.85% |
| DEXEUSDT | 2 | 0 | 0.0% | -14.92% |
| ZECUSDT | 2 | 0 | 0.0% | -2.76% |
| ETHFIUSDT | 1 | 1 | 0.0% | +23.36% |
| HEMIUSDT | 2 | 0 | 50.0% | +50.91% |
| GRAMUSDT | 1 | 0 | 0.0% | -3.25% |
| THEUSDT | 0 | 1 | — | +9.86% |
| WIFUSDT | 1 | 0 | 100.0% | +2.25% |
| SYNUSDT | 1 | 0 | 0.0% | -4.18% |
| ADAUSDT | 0 | 1 | — | -0.93% |
| AAVEUSDT | 0 | 1 | — | +6.52% |
| CLUSDT | 1 | 0 | 100.0% | +9.92% |
| ALLOUSDT | 0 | 1 | — | -0.07% |
| PLUMEUSDT | 0 | 1 | — | +6.06% |
| KIIUSDT | 1 | 0 | 100.0% | +35.80% |
| NEARUSDT | 1 | 0 | 100.0% | +29.59% |
| PRLUSDT | 1 | 0 | 0.0% | -26.28% |
Win rate here is the same measure as the headline above — targets hit over targets plus stops, expiries excluded — so the two are directly comparable. A dash means nothing on that asset reached either level yet, which is not the same as 0%. Per-asset samples are small: several of these are a handful of calls. Read them as texture, not evidence.
So what are you actually paying for?
Not a signal to follow. BlofinX compresses hours of derivatives research into a minute and shows the reasoning: what fourteen independent agents found, where they disagreed, and which one dissented. You still decide. We publish this page because a research tool that hides its own scorecard is not one worth paying for — and because you can check whether the numbers here move.
See the current forecast
The numbers above are the record. The desk is where the next call gets made.