Yesterday's picks: how the model did as its firmest side cashed in Milwaukee and its Rays lean got buried at Fenway

Verdexed's model had a winning Sunday on a full Major League Baseball board, and the honest headline is that the sides it trusted most did the heavy lifting. Its single firmest lean, the Milwaukee Brewers at home, cashed in the most dramatic way available, and two of its other shared leans landed alongside it. The one clear blemish came at Fenway Park, where a slight lean on Tampa Bay ran into a buzzsaw. Netted out, July 19 graded as a solid day for the higher-confidence tier, with one loud miss that deserves to be named rather than buried.

The scorecard

The model ran a full Sunday slate. Its exact daily graded hit-rate was not published on the public accuracy surface at writing, so the day is described here by the sides the model actually shared and the confirmed final scores, not a fabricated percentage.

Of those named sides, the picture was clean. The firmest number on the board, the Brewers as a mid-60s home favorite, held. Its modest lean on the Seattle Mariners held. And the direction it shared in the night's marquee coin flip, the Los Angeles Dodgers at the New York Yankees, held as well. The one side that missed was its slight lean on the Rays at Boston. A fifth read, Texas at Atlanta, was flagged as provisional in the morning writeup because the Rangers had not confirmed a starter, so it is left out of this tally rather than counted after the fact.

That leaves the high-confidence tier doing exactly what a chalk-leaning model wants from it. The day's firmest side cashed, and the coin flips broke toward the model more often than not. A single Sunday tells you very little about calibration, but a clean top of the board is a good data point rather than a bad one. Anchored to Verdexed's published MLB model accuracy, which sits around 68 percent on the season, nothing about July 19 moved the running record materially in either direction.

Best call / worst miss

The best call was the firmest one. Milwaukee, the model's top side, trailed the Marlins 1-0 into the bottom of the ninth after Eury Perez carved the Brewers up for six innings on a single hit. Then William Contreras turned a 1-2 cutter from closer Pete Fairbanks into a 399-foot, three-run walk-off homer, and the Brewers won 3-1 to complete a sweep. The model does not, and cannot, predict a ninth-inning walk-off. What it graded correctly was the side: Milwaukee to win. It got there the hard way, and it still got there.

Two more landed cleanly. The Mariners handled the Giants 6-3 behind ten Logan Gilbert strikeouts, matching the model's lean. And in the marquee, Yoshinobu Yamamoto threw a complete-game four-hitter as the Dodgers beat the Yankees 8-2 in the opener of a doubleheader. The model framed that last one as a near coin flip rather than a conviction side, so it earns partial credit at best, but the direction it shared was the right one.

The worst miss was at Fenway. The model leaned slightly to Tampa Bay on the road in a two-ace matchup, and Boston won 6-1. Sonny Gray, 11-1 on the year, was the reason, and the game was never close. There is no spin available here. The model liked the wrong side, and it lost by five. On a board this size a slight lean falling is unremarkable, but it was the cleanest loss of the day, and it belongs in the ledger next to the wins.

What it means

One winning Sunday is noise, not a trend, and the model's design is built to look exactly like this across a long summer. The firmest sides carry the day more often than not, the coin flips split, and the occasional lean gets buried. Sunday fit that shape. The encouraging read is that the top of the board held, which is where a calibrated model most needs to be right. The sober read is that it held on a ninth-inning swing the model had no way to foresee, a reminder that being on the right side and being lucky can look identical in a final box score. Anchored to a running MLB accuracy near 68 percent, this was a good day inside a good season, and no single slate should be read as more than that.

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.