Track record
Score track record
A quality score is only useful if it predicts the future. So we freeze the scores from a past date and follow every leader scored that day.
Leaders who disappeared from the ranking in the meantime (portfolio closed, made private, stopped) are not dropped. For a copier the copy simply ended, so we count it as a failure. If you keep only the survivors, any score looks good.
Cohort scored on 2026-07-23 · 176 leaders — 60 days later
The chance that grade A's lower failure rate versus grade F is luck is about 1.7% (one-sided Fisher exact test).
What the score got right
The risk of a large loss or disappearing moved with the grade — grade A 19% vs grade F 54%.
What it did not predict
Among survivors, returns did not follow the grade order — the highest survivor median was grade D (+13.4%). Read the score as a filter for blow-up and exit risk, not as a forecast of how much a leader will make.
30 days later (frozen, computed 2026-10-03)
| Grade | Leaders | Ended positive | Failure rate | Survivor median (survivors) | Exited / closed |
|---|---|---|---|---|---|
| A | 16 | 69% | 13% | +8.4% (14) | 2 |
| B | 32 | 50% | 28% | +4.3% (24) | 8 |
| C | 54 | 30% | 50% | +0.7% (31) | 23 |
| D | 39 | 38% | 49% | +6.3% (22) | 17 |
| F | 35 | 20% | 51% | 0.0% (18) | 17 |
60 days later (frozen, computed 2026-10-03)
| Grade | Leaders | Ended positive | Failure rate | Survivor median (survivors) | Exited / closed |
|---|---|---|---|---|---|
| A | 16 | 56% | 19% | +7.8% (14) | 2 |
| B | 32 | 47% | 28% | +9.4% (24) | 8 |
| C | 54 | 33% | 54% | +3.0% (31) | 23 |
| D | 39 | 36% | 49% | +13.4% (22) | 17 |
| F | 35 | 26% | 54% | +3.2% (18) | 17 |
So far: 71 days (in progress — for reference)
| Grade | Leaders | Ended positive | Failure rate | Survivor median (survivors) | Exited / closed |
|---|---|---|---|---|---|
| A | 16 | 56% | 25% | +9.9% (14) | 2 |
| B | 32 | 50% | 31% | +8.0% (24) | 8 |
| C | 54 | 33% | 52% | +6.7% (31) | 23 |
| D | 39 | 36% | 54% | +7.8% (22) | 17 |
| F | 35 | 29% | 54% | +3.7% (18) | 17 |
Leaders scored 95+ at the time
| Leader | Score then | 60-day return | Max drawdown | Score now |
|---|---|---|---|---|
| SaGoCrypto | 100 (A) | +23.2% | -6.7% | 100 (A) |
| CuanBot | 100 (A) | +4.6% | -2.6% | 100 (A) |
| 一亏三千里 | 100 (A) | +45.0% | -21.6% | 100 (A) |
| ALL KILL OUT | 100 (A) | -52.5% | -54.2% | 65 (D) |
| 求其_ | 100 (A) | Portfolio closed | — | No longer listed |
| 富一次就足够 | 100 (A) | +130.1% | -12.5% | 91 (B) |
| Operation Last Raven | 99 (A) | -7.5% | -48.5% | 62 (D) |
| MoneyMakersTM | 99 (A) | Left the ranking | — | No longer listed |
| GridMaster | 98 (A) | +11.0% | -1.8% | 74 (C) |
| RIOSE_CPS | 98 (A) | -17.6% | -18.8% | 61 (D) |
| mumudy | 98 (A) | -20.9% | -50.6% | 84 (C) |
| 交易员大侠 | 96 (A) | +4.4% | -8.0% | 71 (D) |
| 人生到处知何似 | 96 (A) | +17.5% | -17.6% | 76 (C) |
| 无秋⁶⁶⁶ | 96 (A) | +23.8% | -26.0% | 96 (A) |
| iiTrader | 96 (A) | +20.9% | -6.1% | 96 (A) |
| Sky株式会社 | 95 (A) | -16.8% | -18.4% | 90 (B) |
How it is computed
- Failure = a loss of 30% or more over the period, or leaving the ranking (including closure).
- Returns use the leader-account ROI curve published by the exchange. The copier's profit share (usually 10%), entry delay and slippage are not deducted, so real copy results can be worse.
- The exchange's 90-day ROI curve is re-based as the window moves, so stitching days across snapshots is wrong. Each period is computed from a single curve that contains both its start and end date, and frozen on the day it is computed.
- Score snapshots are saved daily. New cohorts appear here as they pass 30, 60 and 90 days.
Limitations
- There are only 1 cohort(s) so far, and grade A has only 16 leaders. One market period may drive the result.
- The first cohort (2026-07-23) was scored with the site's initial formula. The formula has changed since, so the current score's record will come from the daily cohorts now accumulating.
1 score snapshot(s) are waiting to reach 30 days. The next results start around 2026-11-02.