Today's read
Regime: choppy (low confidence). The engine scores Bitcoin as sideways on every horizon: its momentum leans lower, its coverage leans slightly the other way, and the two cancel out. Across the large-cap alts, momentum still points lower. This window's coverage is thicker than last run's (60 analyzed items, up from 15). Crypto-market coverage leans bearish, with the main themes being an oil-driven macro wobble, a pending index decision on a large corporate bitcoin holder, and talk of the market chopping near support.
New calls this run (all leans lower; touch-graded)
| Call | Horizon | Engine probability | Touch base rate |
|---|---|---|---|
| XRP — leans to reach −1.03% within the next day | 1d | 76.4% | 65.9% |
| Zcash (ZEC) — leans to reach −2.36% within the next day | 1d | 64.7% | 60.1% |
| Curve DAO (CRV) — leans to reach −2.37% within the next day | 1d | 70.7% | 67.6% |
| Ethereum (ETH) — leans to reach −2.11% within the next 3 days (experimental horizon) | 3d | 64.5% | 57.7% |
| Pump.fun (PUMP) — leans to reach −5.35% within the next 3 days (experimental horizon) | 3d | 67.7% | 62.5% |
| Curve DAO (CRV) — leans to reach −4.10% within the next 3 days (experimental horizon) | 3d | 71.6% | 68.6% |
Each probability is the chance that price touches the stated % move at any point within the horizon. It is not the chance that price closes there, and the complement is "does not touch", not "goes the other way". Touch probabilities sit near ~60% because that is roughly how often a move of this size gets touched anyway. That is a property of the units, not skill. The gaps over the base rate are small and unvalidated.
Not issued, and why: Bitcoin and NEAR scored sideways (no direction). AVAX, ADA, DOGE, SUI, JUP, UNI, ENA and BNB showed strong momentum but did not clear the conviction gate. Solana already has two standing calls (the per-coin limit). Ethereum and Pump.fun 1-day calls were issued under 48 hours ago and today's numbers are not materially different. The 7th 7-day slot stays open: it is reserved for an idiosyncratic read driven by a coin's own coverage, and neither of the candidates (PUMP, CRV) leans the same way as its coverage. Standing structural and relative reads were re-scored with no material change, so nothing new was issued there.
Receipts: resolved since the last run
All four resolved on 23 Oct, 22:41 UTC, shortly after they matured.
| Call | Issued | Matured | Outcome |
|---|---|---|---|
| Week-ahead market regime: choppy | 16 Oct 15:18 | 23 Oct 15:18 | Hit: BTC moved −0.56% over the week |
| XRP — leans to reach −1.51% within the next 3 days (experimental) | 20 Oct 21:20 | 23 Oct 21:20 | Hit: level reached; −1.90% at expiry (we had it at 80.7%) |
| Pump.fun (PUMP) — leans to reach −2.96% within the next day | 22 Oct 22:39 | 23 Oct 22:39 | Hit: level reached; −4.48% at expiry (66.2%) |
| Ethereum (ETH) — leans to reach −1.23% within the next day | 22 Oct 22:39 | 23 Oct 22:39 | Miss: right direction, level not reached; −0.08% at expiry (60.8%) |
That is 3 hits and 1 miss. The ETH miss closed slightly lower, but a touch call that does not reach its level is a miss.
How the record stands (trailing 30 days to 23 Oct)
- Headline, the 7-day claim pool (Brier Skill Score vs the base rate): −0.0034 (n = 10). Our probabilities are still roughly level with, and slightly worse than, simply quoting the historical base rate. They have not yet beaten it. Reliability 0.0375 (where overconfidence shows up); resolution 0.045 (how well our probabilities tell outcomes apart). With n = 10 against a floor of 25, the verdict is insufficient data.
- 1-day claim pool: Brier Skill Score −0.148 (n = 31), clearly worse than the base rate.
- Directional-accuracy note (this measures direction picking and ignores the probabilities): −1.1 points vs the base rate on the 7-day pool.
- The experimental 3-day lane is excluded from the accuracy record.
