May 24 — 0-1, moneyline loss after model retrain

Yesterday

0-1 · -0.50u · ROI -100.0% · CLV -0.96%

Night 1 Recap: 0-1, -0.50u

The Rangers fell 2-1 to the Angels despite carrying significant advantages on paper. Texas entered with a superior bullpen (44 vs 8 rating) and MacKenzie Gore's edge over Reid Detmers (4.05 vs 4.39 xFIP), but couldn't capitalize in the low-scoring affair. The Rangers' recent .50 win rate compared to LAA's .31 suggested better form, yet execution faltered in Anaheim. With -0.96% CLV, the market moved against us post-bet, indicating we may have been early rather than wrong on a fundamentally sound thesis.

What missed

Worst result. TEX ML (TEX @ LAA) at +102, -0.50u — final 2-1, CLV -1.0%. Graded bad bet.

The model catastrophically misjudged the total runs environment, predicting 7.37 runs versus 3 actual in TEX@LAA. Both starters massively outperformed xFIP projections - Gore allowed just 1 ER over 6 IP against a 4.05 xFIP, while Detmers dominated with 1 ER in 8 IP versus 4.39 xFIP predicted. The away bullpen score of 44.0 suggested strength that held true. However, the -0.96% CLV indicates this was a correctly identified edge that simply ran into exceptional pitching variance.

Under the hood

Yesterday's development work focused on three technical improvements to the betting analysis platform. The team fixed how the system logs information about favorite and underdog teams in run line predictions, ensuring more accurate tracking. They also updated database records to reflect recent structural changes and completed a comprehensive renaming project across all system components, replacing old terminology with the new "closerbets" branding throughout the codebase.

This week

Last 7 days: 1-2 · -0.30u · ROI -20.0%