Daily read, 25 October 2026
The standing regime read is still choppy (issued 23 October, runs to 30 October). Corpus coverage for 24 and 25 October was thinner than in the previous window: 85 analyzed items, against 134. Coverage of the broad crypto market and of Bitcoin leaned bearish, while stablecoin, derivatives and lending coverage leaned bullish. Bearish items this window covered reported ETF outflows and long liquidations, an unverified DeFi governance-exploit report and a stablecoin loss at an exchange.
New calls this run
Each figure is P(price touches the stated level, in the stated direction, at any point within the horizon). The base rate is how often such a touch happens anyway.
| Call | Engine probability | Base rate |
|---|---|---|
| Ethereum (ETH) — leans to reach −3.35% within the next 7 days | 67.2% | ~57% |
| Solana (SOL) — leans to reach −2.49% within the next 7 days | 75.2% | ~68% |
| Bitcoin (BTC) — leans to reach −2.41% within the next 7 days | 69.2% | ~63% |
| Pump.fun (PUMP) — leans to reach −6.82% within the next 7 days | 69.4% | ~65% |
| XRP — leans to reach −0.98% within the next day | 71.7% | ~66% |
| Jupiter (JUP) — leans to reach −2.11% within the next day | 67.1% | ~64% |
| Ethena (ENA) — leans to reach +2.58% within the next day | 61.3% | ~58% |
Most of these leans come from price momentum. For Pump.fun and Jupiter, our coverage points the opposite way. Ethena is the exception: its read comes from coverage while momentum is flat. The engine has not shown a directional edge, so read these as provisional.
Receipts: resolved since the last run
| Call | Stated | Issued | Matured | Resolved | Outcome |
|---|---|---|---|---|---|
| Ethereum (ETH) — leans to reach −3.35% within the next 7 days | 66.2% | 18 Oct 17:45Z | 25 Oct 17:45Z | 25 Oct 23:06Z | Hit |
| Solana (SOL) — leans to reach −2.46% within the next 7 days | 75.8% | 18 Oct 17:45Z | 25 Oct 17:45Z | 25 Oct 23:06Z | Hit |
| Bitcoin (BTC) — leans to reach −2.39% within the next 7 days | 66.8% | 18 Oct 17:45Z | 25 Oct 17:45Z | 25 Oct 23:06Z | Hit |
| Zcash (ZEC) — leans to reach −5.73% within the next 7 days | 64.7% | 18 Oct 17:45Z | 25 Oct 17:45Z | 25 Oct 23:06Z | Miss: ZEC closed the window up about 13% |
| Uniswap (UNI) — leans to reach −2.17% within the next day | 73.7% | 24 Oct 23:05Z | 25 Oct 23:05Z | 25 Oct 23:08Z | Hit |
| Solana (SOL) — leans to reach −0.97% within the next day | 72.5% | 24 Oct 23:05Z | 25 Oct 23:05Z | 25 Oct 23:08Z | Hit |
| Bitcoin (BTC) — leans to reach −0.91% within the next day | 64.0% | 24 Oct 23:05Z | 25 Oct 23:05Z | 25 Oct 23:08Z | Miss: right direction, but the level was not reached |
That is 5 hits and 2 misses. These are mostly correlated down-calls on co-moving majors, so they are not 7 independent results.
Accuracy, trailing 30 days to 25 October 2026
- 7-day claim pool: Brier Skill Score versus the base rate is +0.012 (n=14). That is essentially level with simply quoting the historical base rate, and it is far below the 25-call floor needed before any skill claim. Murphy split: reliability 0.027 (lower is better; this is where overconfidence shows), resolution 0.039.
- 1-day claim pool: Brier Skill Score is −0.119 (n=37). At one day, our probabilities have scored worse than ignoring every signal and quoting the base rate.
- Directional-accuracy diagnostic only (this does not read our probabilities): the 7-day hit rate is 3.2 points above its mean base rate.
- The 3-day horizon remains an experiment and sits outside the accuracy claim.
The verdict is still insufficient data. We report these figures as they are.
