The read, 6 October
Regime: the standing choppy read (issued 2 October, matures 9 October) is unchanged and stays in place; nothing in this window's coverage or breadth amounts to a material change, so no new regime row is appended.
New calls (all near-term horizons — the 7-day book is full at seven standing calls):
| Call | Engine probability | Base rate |
|---|---|---|
| PUMP reaches +5.89% within 3 days | ~75% | ~67% |
| NEAR reaches +6.15% within 3 days | ~71% | ~63% |
| BTC reaches +1.49% within 3 days | ~66% | ~58% |
| SOL reaches +1.37% within 1 day | ~68% | ~61% |
| ETH reaches +1.13% within 1 day | ~58% | ~54% |
Each is a touch read: it scores as a hit if price reaches the stated percentage in the called direction at any point inside the horizon, wherever it closes. The complement is "does not reach the level", not "falls". Touch base rates sit near 54–67% for these sizes, so probabilities in that range are a units effect, not a sign of skill — the gap to the base rate is the number to watch, and it is unvalidated.
Why these five: they are the only names in today's scan that clear the conviction gate (two independent signals — our corpus and price momentum — agreeing). Other names in the scan (ADA, AVAX, BNB, XRP, DOGE) carried a momentum read only, with no usable corpus leg, and failed the gate by design. This window's coverage (79 analyzed items) skews constructive on Bitcoin and Solana: exchange-outflow and options-skew reads on BTC, and a Solana upgrade that cuts finality time. Ethereum is the exception — its handful of items this window tilt bearish, and its call is the weakest on the slate.
Receipts — resolved since the last read
All six were matured and scored by the deterministic resolver at 04:13 UTC on 6 October.
| Call | Issued | Matured | Outcome |
|---|---|---|---|
| PUMP reaches +5.90% within 3 days (p 0.764) | 3 Oct 03:48 UTC | 6 Oct 03:48 UTC | Hit — level reached; +13.97% at expiry |
| NEAR reaches +6.46% within 3 days (p 0.712) | 3 Oct 03:48 UTC | 6 Oct 03:48 UTC | Hit — level reached; +11.84% at expiry |
| SOL reaches +2.71% within 3 days (p 0.688) | 3 Oct 03:48 UTC | 6 Oct 03:48 UTC | Miss — right direction, target not reached (+0.54% at expiry) |
| BTC reaches +0.86% within 1 day (p 0.637) | 5 Oct 04:05 UTC | 6 Oct 04:05 UTC | Miss — level not reached; −0.57% at expiry |
| ETH reaches +1.17% within 1 day (p 0.590) | 5 Oct 04:05 UTC | 6 Oct 04:05 UTC | Miss — level not reached; −0.52% at expiry |
| Regime: rotation (week of 29 Sep) | 29 Sep 02:57 UTC | 6 Oct 02:57 UTC | Miss — BTC +2.34% with mixed breadth, not a rotation |
Two hits, four misses.
Where the track record stands
The headline measure — the Brier Skill Score of our published probabilities against simply quoting the historical base rate — cannot be computed yet: the current engine's seven-day claim pool has zero resolved calls (the first matures on 9 October), against a minimum of 25 before any claim is made. The verdict is insufficient data, and nothing here should be read as skill.
The one-day lane, a small segment, has six resolved calls with a Brier Skill Score of −0.10: on those six, our probabilities did slightly worse than the base rate. That is a tiny sample and is not the headline, but it is reported because it is negative. The three-day lane is a pre-registered experiment and is excluded from the accuracy claim.
The prediction-market input was not admitted on any of today's calls; all five rest on our corpus and price momentum.
