2. Speed of formation
Is a consensus that formed fast more informed than one that built up over days?
Definition
Hours between the first tracked holder's first fill in the market and the fifth holder's first fill, over the holders counted at the hour the consensus formed. Variants: within 1 h, within 6 h, within 24 h, slower.
Baseline for every comparison: the plain consensus on the same markets (5+ tracked holders, 60 %+ on one side, price 0.05–0.95, first signal, held to resolution), 1 ¢ half-spread plus Polymarket's taker fee; $1 a trade.
Mean return per trade, with 95 % intervals
One marker per window: circle = A (curated wallets, Mar–Apr 2026, clean), square = B (all tracked wallets, Oct 2026, provisional, drift ranges dropped), diamond = pooled. Variants with fewer than 30 trades in a window are left out of the chart; the table below has every number.
Every variant
| Variant | Expected | Window A | Window B | Pooled A + B | IS / OOS (pooled) | vs its baseline (pooled) | Guard | Verdict |
|---|---|---|---|---|---|---|---|---|
| 2a fifth holder within 1 h of the first entry filter, decided at the signal | unknown | −17.9 % [−39.7, +4.7] n = 42, hit 60 %, price 0.59 | −5.4 % [−14.9, +4.2] n = 320, hit 63 %, price 0.62 | −6.8 % [−15.7, +1.5] n = 362, hit 62 %, price 0.62 | IS −2.6 % (152) OOS −8.8 % (193) | −2.6 [−10.9, +6.2] points vs the dropped trades, p = 0.7178 | no | nothing |
| 2b within 6 h entry filter, decided at the signal | unknown | −15.2 % [−26.1, −4.7] n = 176, hit 61 %, price 0.62 | −5.1 % [−10.5, +0.4] n = 868, hit 63 %, price 0.63 | −6.8 % [−11.7, −2.0] n = 1,044, hit 63 %, price 0.63 | IS −2.8 % (429) OOS −8.7 % (535) | −2.9 [−8.6, +2.5] points vs the dropped trades, p = 0.8556 | no | nothing |
| 2c within 24 h entry filter, decided at the signal | unknown | −8.4 % [−13.7, −2.9] n = 989, hit 62 %, price 0.62 | −4.7 % [−7.7, −1.3] n = 2,435, hit 65 %, price 0.64 | −5.8 % [−8.4, −2.8] n = 3,424, hit 64 %, price 0.64 | IS −3.2 % (1,553) OOS −7.7 % (1,597) | −2.8 [−6.9, +1.7] points vs the dropped trades, p = 0.8966 | no | negative |
| 2d slower than 24 h entry filter, decided at the signal | worse | −1.5 % [−6.0, +3.1] n = 2,074, hit 64 %, price 0.62 | −5.2 % [−9.6, −1.1] n = 1,321, hit 65 %, price 0.65 | −2.9 % [−6.3, +0.0] n = 3,395, hit 64 %, price 0.63 | IS −2.3 % (2,029) OOS −3.3 % (1,145) | +2.8 [−1.4, +7.4] points vs the dropped trades, p = 0.0914 | no | nothing |
What this data can and cannot show
Can: Whether the pace at which tracked wallets piled in predicts the outcome.
Cannot: Why they were fast: the fast markets are also the short-dated ones (in-play games, crypto windows), so speed is confounded with market type. The report shows the mix.
Caveats that apply to every signal
- Only tracked wallets' fills exist, from when the site started tracking them; the replay is closer to the chain than to what the site displayed (no Polymarket position overrides, merges missing May–Oct 2026).
- Window B's fill times run early in parts of October 2026; the affected ranges are dropped and the rest is not proven clean. Its numbers are provisional until the repaired export is in.
- Prices are the last tracked fill, not the order book (median error 0.5 ¢, p90 up to 16 ¢ on fast markets); entries happen on the hour (5-minute steps in play).
- The windows are 7 and 3 weeks long; effects of a few points cannot pass a guard over 40 variants there. A failed guard is not evidence of no effect — it is the absence of evidence.
- Nothing here is trading advice. Past replays, even clean ones, do not pay future trades.