MLB Game & Prop Picks Recap for July 31, 2026: 3-1, +2.74 units

One-minute recap

The daily card contained 1 game Pick and 3 pitcher-prop Picks. Together the settled Picks finished 3-1 and +2.74 units on their flat-risk ledgers.

Every number below comes from the saved pregame receipts. Game and prop records remain separate in the Archive, and neither the inputs nor the selections were rewritten after the result.

MLB Game Picks

1 game Pick was settled from confirmed lineups, rolling team and player values, and the exact moneyline saved before first pitch.

NYY +142 — the read and the result

The final was NYY 2, CHC 0. The saved NYY moneyline at +142 from FanDuel graded as a win and moved the flat-risk ledger +1.42 units.

The number we froze was FanDuel NYY +142. Our probability was 40.8%; that price needed 41.3% to break even. That was 0.5 percentage points under break-even, so this was not a positive price-edge read. It made the tracked daily card only as the first matchup with a complete calculation-quality screen. The strongest public push toward NYY was Elo home-win baseline, worth 0.23 probability points in the saved calculation.

On the 0–100 rolling scale, the starting-pitcher matchup was Will Warren 47.1 against Shota Imanaga 47.2.

The saved lineup values were NYY 49.6 and CHC 51.8. The top displayed hitters were Ben Rice (57.7) and Pete Crow-Armstrong (60.9).

Bullpen availability checked in at NYY 52.4 and CHC 49.8. These are the pregame values the model saw, not explanations invented from the final score.

Good result, limited lesson. A 40.8% forecast can win without proving that one starter, hitter, or bullpen input caused it. This row moves into the comparable-game review set; the live weights stay unchanged.

MLB Pitcher Prop Picks

3 pitcher strikeout Picks were settled from the confirmed opponent lineup, rolling pitcher and matchup values, workload projection, and saved prop price.

Paul Skenes Under 7.5 Ks at +120 — the read and the result

Paul Skenes finished with 7 strikeouts, so the saved under 7.5 ticket graded as a win and moved the separate prop ledger +1.20 units.

The frozen price was BetRivers +120. The model read was 71.6% against a 45.5% break-even mark, a calculated difference of +26.1 percentage points.

The saved projection was 6.0 strikeouts across 22.2 projected batters faced. On the 0–100 rolling scale, the pitcher value was 60.5 and the opponent strikeout-matchup value was 72.1; the confirmed lineup carried a 25.6% strikeout rate.

This win is one frozen observation, not a reason to rewrite the pitcher or matchup value. It enters the comparable-prop review set by workload, opponent profile, line, and price; production weights stay unchanged until a pattern survives prospective testing.

Dylan Cease Under 7.5 Ks at +112 — the read and the result

Dylan Cease finished with 7 strikeouts, so the saved under 7.5 ticket graded as a win and moved the separate prop ledger +1.12 units.

The frozen price was BetRivers +112. The model read was 69.6% against a 47.2% break-even mark, a calculated difference of +22.4 percentage points.

The saved projection was 6.1 strikeouts across 23.9 projected batters faced. On the 0–100 rolling scale, the pitcher value was 61.3 and the opponent strikeout-matchup value was 65.9; the confirmed lineup carried a 20.0% strikeout rate.

This win is one frozen observation, not a reason to rewrite the pitcher or matchup value. It enters the comparable-prop review set by workload, opponent profile, line, and price; production weights stay unchanged until a pattern survives prospective testing.

Kyle Leahy Over 3.5 Ks at -148 — the read and the result

Kyle Leahy finished with 3 strikeouts, so the saved over 3.5 ticket graded as a loss and moved the separate prop ledger -1.00 units.

The frozen price was FanDuel -148. The model read was 61.4% against a 59.7% break-even mark, a calculated difference of +1.7 percentage points.

The saved projection was 4.5 strikeouts across 22.0 projected batters faced. On the 0–100 rolling scale, the pitcher value was 54.2 and the opponent strikeout-matchup value was 38.7; the confirmed lineup carried a 19.8% strikeout rate.

This loss is one frozen observation, not a reason to rewrite the pitcher or matchup value. It enters the comparable-prop review set by workload, opponent profile, line, and price; production weights stay unchanged until a pattern survives prospective testing.

What goes into the next review

A result can move its record, but it cannot retrain either model by itself. Each game and prop stays attached to its exact pregame snapshot so comparable reads can be reviewed without inventing a story from the final result.

Any candidate adjustment must repeat across a larger frozen sample and survive a prospective challenger test. Until then, the production weights remain unchanged.