Yesterday's picks: how the model did (MLB, July 7)

Tuesday graded as a mixed day for Verdexed's board, and the honest read sits right in the middle: some of the model's favorites held, but the single side it liked most on the slate did not. The Dodgers, the best team in baseball at home behind a 10-2 starter against a last-place club, lost 4-3 to Colorado on a pair of eighth-inning errors, and the loss spoiled Shohei Ohtani's 300th career home run in the same game. This column grades Verdexed's own model against what actually happened, win or lose, and July 7 grades as a day where the top of the board took a real hit while the rest of the card mostly cooperated.

This is the daily accounting, no spin and no cherry-picking. A favorite-leaning model can survive a coin-flip day; what it cannot control is defense, and Tuesday's marquee loss came on gloves, not on a blown projection.

The scorecard

The model went into Tuesday leaning several home favorites, with Los Angeles priced as the cleanest side on the card. Justin Wrobleski entered at 10-2 with an ERA in the high-2.00s and drew a Colorado club carrying the worst record in the National League and a starter, Michael Lorenzen, sitting near a 7.00 ERA. That is close to the largest single-game edge a probability model produces in July, and it is exactly the pick that busted.

Elsewhere the leans mostly held. Philadelphia, behind ace Zack Wheeler, was a comfortable model favorite at Cincinnati and delivered a 4-1 win. Toronto rebounded from a Monday blowout to beat San Francisco 9-3, and Milwaukee edged St. Louis 4-3 in the opener of a doubleheader. Verdexed's exact graded accuracy figure for the day was not cleanly retrievable at publication, so this recap describes the slate qualitatively rather than attaching a hit-rate the model did not confirm. The shape is a split leaning mixed: multiple favorites cashed, but the one the model was surest about is the one that got away, which is the outcome that carries the most weight on any given day.

Best call / worst miss

The best call that landed was Philadelphia, and it doubled as a bounce-back story. A day after the Phillies were routed 15-1 at Kansas City, the model still trusted Wheeler on the road at Cincinnati, and the trust paid. Trea Turner drove in the game's first run, Kyle Schwarber followed with a three-run swing for his 31st home run of the season, and the Phillies built a 3-0 lead before adding an insurance run in the eighth. Eugenio Suarez's solo shot was all Cincinnati managed. Final: Phillies 4, Reds 1. When the model backs an ace the morning after that same team got blown out, and the ace answers, that is the pick worth naming first among the winners.

The worst miss is Los Angeles, and there is no softening it. The Dodgers led 3-2 into the bottom of the eighth with the model's biggest edge of the slate on the line, and the inning came apart on defense rather than pitching. A potential inning-ending double play turned into a run-scoring error by shortstop Miguel Rojas, a Jake McCarthy bunt and a second-baseman throwing error by Alex Freeland into the camera well plated two more, and Colorado stole three runs on two errors to win 4-3. Wrobleski had done his job with seven innings of one-run ball, and Ohtani's 300th career homer went to waste. The side the model priced as its safest lost the game it should have closed, and it lost it the most frustrating way a favorite can.

What it means

One upset built on two errors is not a crack in the model, but it is a real losing result at the top of the board and it should be logged as one. The Dodgers pick was correct on process: a 10-2 starter at home against the worst team in the league is the kind of edge a probability model is designed to price, and Wrobleski pitched exactly the game the projection asked for. Defense is the variable no forecast fully captures, and two eighth-inning miscues are the sort of tail outcome that occasionally erases a clean edge. That is the tax a favorite-leaning system pays; most nights the chalk holds quietly, and once in a while the safest pick unravels on a play the model never modeled.

The read cuts both ways, though. The same slate that cost the model its top lean also rewarded the Wheeler call and several other favorites, and a mixed Tuesday does not move a season-long picture in which Verdexed's MLB model has graded in the high-60s percent range on win accuracy. This is one day inside a much larger sample, the Dodgers loss sits squarely in the variance a public record is built to absorb, and nothing about a two-error eighth, in either direction, should be read as a trend. The record is the record, graded in public, and it moves one honest day at a time. Tuesday was a split that stung at the top.

Responsible play

**This is a record of past model output, not betting advice.** Verdexed grades its model in public for transparency and entertainment. Past results do not predict future outcomes, nothing here is a recommendation to place a wager, and no pick is ever a guarantee, even a strong model loses regularly.

**You must be 21+** (or the legal age in your jurisdiction) and located where sports wagering is legal to bet. Sports betting is not available everywhere; know and follow the laws where you are. Never bet more than you can afford to lose, and never chase losses.

If gambling is affecting you or someone you know, help is available. In the US, call or text **1-800-GAMBLER** (1-800-426-2537), or visit ncpgambling.org. Please play responsibly.