What traders actually predicted in September 2026
Every figure below comes from calls published on TradeLiveX and graded under our published methodology. This describes the traders who post here — a self-selected group — during one month. It is not a description of the market and not a forecast.
How far calls got
The question almost no published track record answers.
Breakdowns
Each with its own sample size, so a thin slice is visible as one.
By market
| n | Above BE | Mean R | Median R | |
|---|---|---|---|---|
| crypto | 424 | 41% | -0.07R | -0.26R |
| forex | 175 | 45% | +0.00R | -0.04R |
| metal | 30 | 43% | +0.07R | -0.14R |
By timeframe
| n | Above BE | Mean R | Median R | |
|---|---|---|---|---|
| 1h | 207 | 48% | +0.07R | -0.02R |
| 4h | 152 | 40% | -0.07R | -0.63R |
| 15m | 137 | 45% | -0.00R | -0.04R |
| 1d | 132 | 34% | -0.25R | -1.00R |
| 5m | 1 | 0% | -0.08R | -0.08R |
By stated conviction
| n | Above BE | Mean R | Median R | |
|---|---|---|---|---|
| medium | 271 | 44% | +0.04R | -0.08R |
| high | 179 | 45% | -0.05R | -0.08R |
| low | 179 | 39% | -0.16R | -0.30R |
The conviction table is published whichever way it points. If high-conviction calls did not outperform this month, that is a finding about this community worth reporting.
What the headline figures leave out
Reported separately rather than quietly dropped.
668 calls were posted during September 2026. Unentered and cancelled calls carry no R and are excluded from every average above; calls held for review were not graded at all, because the candle series covering them had gaps inside an open session. 29% of graded calls had at least one fill taken at a candle open because price gapped through a level.
Methodology and limitations
- Window. Calls that resolved between 2026-09-01 and 2026-10-01 UTC. A call posted in a previous month appears here if it resolved in this one.
- Exclusions. Calls marked invalid after moderation, calls that never entered, calls cancelled before entry, and calls held for review.
- Modelled fills. No slippage, no spread on the exit, no financing, no position sizing. A live account trading identical calls would do slightly worse.
- R is not a return. It is a ratio to each trader's own defined risk, and those risks are not comparable in currency terms.
- Self-selection. Contributors chose to publish here. Nothing in this report generalises to traders who did not.
- Sample size. 629 graded calls from 254 traders can support an interesting description of one month. It cannot support a predictive claim, and none is made.