Today's read
The market regime call stays risk-off (issued 9 Oct, runs to 16 Oct) and is not re-issued: nothing in today's inputs changes it. Coverage over 12-13 October (62 analyzed items, up from 52) tilts cautious on the broad market, with commentary on an extended altcoin correction, positioning risk on leveraged ETH exposure and an inflation print due this week; Bitcoin-specific coverage reads bullish while ETH coverage reads bearish.
The engine's conviction gate passed three coins, all leaning lower: Ethereum, Solana and Uniswap (Bitcoin, which cleared yesterday, did not). Ethereum and Uniswap already have open calls on the slots the engine would fill, Solana's free slots were held back by the cap on concurrent 3-day calls and an unchanged 1-day reading, and today's numbers sit within a few points of the standing ones, so the book is left to run rather than restated. Records are never edited; each open call is graded on its own expiry.
New today: Solana versus Bitcoin over 90 days
The engine puts the chance that Solana's 90-day return finishes at or below Bitcoin's at roughly 63%, against a historical frequency of about 59% for that outcome. The gap is small, and the two signals behind it point in different directions, so this is a modest lean, not a strong one. It sits beside the July read on the same question, which matures on 16 October and is graded separately.
Receipts
Resolved since the last run:
- Solana, 1-day lean to reach -0.96% (issued 12 Oct 14:36 UTC, matured 13 Oct 14:36, resolved 14:38): Hit. The level was reached; the close was 1.85% lower.
The six calls that resolved at the start of yesterday's run (one hit, five misses, including two that moved the right way without reaching their level) were reported in yesterday's read.
How the record stands
On the current grading rule, our probabilities have not beaten the base rate. The Brier Skill Score against the base-rate reference is -0.18 on the 7-day claim pool (n=5) and -0.18 on the 1-day pool (n=22): quoting the historical frequency alone would have scored better than our stated numbers. The Murphy split on the 7-day pool shows why: reliability 0.26 (our stated probabilities run well above how often the calls land) against resolution 0.06 (they barely separate hits from misses). The sample is far below the 25-call minimum, so no skill claim is made in either direction.
