Out of sample · 2026-09
The month, reconciled
2026-09: 28 resolved calls averaging -0.2616R. That is below the 30 this product treats as meaningful, so it is published for transparency and must not be read as evidence either way.
- Published
- 31
- Resolved
- 28
- Win rate
- 42.9%
- Total R
- -7.32
28 resolved calls is below the 30 this product treats as meaningful. These numbers are published because hiding a month until it flatters us is the habit this whole page exists to refuse — not because one month settles anything.
By style
| Style | Resolved | Win rate | Mean R | Total R | Measured verdict |
|---|---|---|---|---|---|
| scalping | 19 | 52.6% | -0.1059 | -2.01 | no demonstrated edge |
| swing | 1 | 0.0% | -1.0000 | -1.00 | measured no edge |
| intraday | 8 | 25.0% | -0.5390 | -4.31 | no demonstrated edge |
What this month does and does not prove
The reference this is read against is data/measurement.json, which currently records no demonstrated edge for any style. A month that runs positive is therefore not confirmation of an edge, and a month that runs negative is not evidence of decay -- both are single samples, and the control chart on the record page is what actually watches for decay.
The whole record, call by call, is on the verify page, and every individual call has its own page with the readings that produced it.