The Side Cluster Lost, the Tiny NBA Overs Did Not

The pick was not a single pick. The final logged action kept the top-five raw-percentage trim, which meant three MLB +1.5 sides and two NBA rebounds+assists overs. That is a neat ranking system and an ugly diversification profile. Kornet’s price drift from -139 on the card to -154 at placement was the loudest warning, but the actual ticket shape mattered more.

The signal was narrow. There were no SportsLine, Perplexity, or Claude recommendations in the reconciled note, so the board lived entirely inside Rithmm plus the settlement reconstruction. The useful split was not between models. It was between market families. The three MLB +1.5 sides were the bulk exposure, while the two tiny NBA combo overs were the cleaner threshold bets.

The result was good enough to avoid a full wipeout and bad enough to keep the lesson obvious. The full evaluated board finished 3-4. The actual placed five-leg ticket finished 3-2. Angels +1.5 survived a 1-0 loss, both NBA rebounds+assists overs cashed with 4 combined each, and Marlins +1.5 plus White Sox +1.5 lost cleanly. In martingale terms, this is not a reset story so much as a reminder that portfolio shape can rescue or sink the day before any single read does.

The learning is structural. The tracked NBA rebounds+assists over bucket moved to 3-1 overall in the local log through June 5, and today’s share of that bucket was 2 of 2. Meanwhile, the placed MLB +1.5 cluster went 1 of 3 on this card. That is not proof that tiny combo overs are secretly safe. It is proof that trimming by raw win probability can still select for one weak board archetype. Here, the side cluster was the leak and the supposedly smaller props carried the card.

The verdict is plain. The placed bets followed the final top-five trim, and that alignment worked only where it escaped the MLB side cluster.