Week-ahead read - the rotation stalls
The early-July advance has cooled into a broad pullback, and the leadership that defined last week's rotation has thinned out. Solana and Ethereum still clear the model's multi-source gate on the upside, with constructive cross-source coverage and firm short-term momentum. But Avalanche, XRP and Dogecoin have lost their bid this week, and Bitcoin's blended read sits close to a coin-flip. With no decisive direction and no broad rotation leader, the week-ahead regime shifts from rotation to choppy.
We retired two upside leans that the model no longer supports: the Avalanche read collapsed toward even and the XRP read lost its signal entirely as this week's tape ran against both. Those calls stay in the scored record and will resolve on schedule - being removed from the live board does not remove them from the accuracy denominator.
What still clears the gate
- Solana and Ethereum hold their seven-day upside leans, both well above their base rates on bullish coverage plus momentum.
- Curve DAO keeps its lower-conviction weekly lean, though fresh coverage has thinned.
- On the quarter, both Solana and Ethereum retain a better-than-even modeled chance of outpacing Bitcoin - edges above the historically low base rate for a single large-cap.
Receipts - misses shown alongside hits
The honest read this stretch is that our defensive calls were caught by the early-July rally:
- Bitcoin, 7-day downside lean (issued 06-30, resolved 07-07): miss - price rose over the week.
- Week-ahead regime: risk-off (issued 06-30, resolved 07-07): miss - the tape turned constructive.
- Ethereum, 7-day downside lean (issued 06-27, resolved 07-04): miss.
- XRP, 7-day downside lean (issued 06-27, resolved 07-04): miss.
- Solana, 7-day upside lean (issued 06-26, resolved 07-03): hit (+12.6%).
On the claim-eligible seven-day record, our edge calls are running modestly ahead of the base rate they are measured against - a slim positive skill margin over a still-small sample. The probabilities remain provisional and uncalibrated in this phase, so the reasoning leads, not the decimal.
Bottom line
A cooling tape has narrowed the opportunity set. The model is leaning on two names it still trusts rather than filling the board, and it is carrying no new weekly calls this run - a quiet read is a valid read.
