The read
The engine's slate still leans lower across every coin it scored this run, but only three clear the conviction gate: Solana, Ethereum and Bitcoin. Solana already carries two standing calls, and each of the three already carries a 7-day call, so the only new reads are two 3-day experimental leans:
- Ethereum (ETH) — leans to reach −1.43% within the next 3 days (experimental horizon). Downside momentum leads, and Ethereum's own coverage over the past week ran less positive than its usual mix. Roughly 71% to touch −1.43% within 3 days, against a ~59% base rate for a move that size; uncalibrated, experimental horizon excluded from the public accuracy claim.
- Bitcoin (BTC) — leans to reach −1.00% within the next 3 days (experimental horizon). Momentum alone carries this lean: our corpus input on Bitcoin sits essentially at its usual baseline. Roughly 69% to touch −1.00% within 3 days, against a ~64% base rate, so the gap is small; uncalibrated, experimental horizon excluded from the public accuracy claim.
The standing risk-off regime read runs to 16 October and is unchanged. Every structural question (new BTC record, altseason, majors drawdown, Ethereum- and Solana-vs-BTC) already holds its two-row maximum, and the XRP-vs-BTC read did not clear the conviction gate, so no structural row was added.
Corpus note. At the time of this run, no stories from the 14–15 October window had yet been analyzed; the corpus held nothing newer than late on 13 October. The engine's corpus input uses a trailing 7-day window, so it still reads coverage from earlier in the week, but nothing in this article draws on coverage from the last two days.
Receipts
Four calls matured since the last read; all four touched their stated level (resolved 15 October, 15:07 UTC):
| Call | Issued | Matured | Outcome |
|---|---|---|---|
| Uniswap (UNI) — leans to reach −6.87% within the next 7 days | 8 Oct 14:01 | 15 Oct 14:01 | Hit — level reached; −7.73% at expiry |
| Ethereum (ETH) — leans to reach −2.08% within 3 days (experimental) | 12 Oct 14:36 | 15 Oct 14:36 | Hit — level reached; −6.18% at expiry |
| Uniswap (UNI) — leans to reach −4.31% within 3 days (experimental) | 12 Oct 14:36 | 15 Oct 14:36 | Hit — level reached; −4.10% at expiry |
| Bitcoin (BTC) — leans to reach −0.55% within the next day | 14 Oct 14:49 | 15 Oct 14:49 | Hit — level reached; −0.08% at expiry |
A run of hits is not evidence of skill on touch calls: touch base rates sit near 60% at these levels, so an always-lower forecaster would have scored the same four.
Accuracy
Our probabilities have not yet beaten the base rate. On the 7-day claim pool (current engine, trailing 30 days to this run on 15 October), the Brier Skill Score against the base-rate reference is −0.135 (n=6) — worse than quoting the historical frequency and ignoring our signals. Reliability is 0.202 (the over-confidence term) against resolution of 0.111, so the stated probabilities run higher than outcomes support. The 1-day lane reads −0.171 (n=23). Both remain below the 25-call floor, so the verdict is insufficient data.
As a directional-accuracy note only (it does not read the probability): the 7-day hit rate is 33% (2 of 6) against the base rates those calls were issued at.
