Yesterday's picks: the model's chalk got routed, the Phillies bailed it out

Thursday was a losing day for Verdexed's model. Baseball was the only league still on the board, the published card named six games, rain wiped one out before first pitch, and of the five sides that were actually graded the model went 2-for-5. Worse than the bare record is where the misses landed: of the four games the model read as a favorite, only one held. This was a day the chalk got routed, and the honest way to start is to say so plainly.

The scorecard

Five named sides reached a final score, and the model finished 2-for-5 on them. The damage was concentrated in the part of the board that is supposed to carry a recap: the favorites. Verdexed leaned the Phillies, Blue Jays, Yankees and Tigers as the priced-up sides on the slate. Three of those four lost. The lone favorite to survive was Philadelphia, the day's single highest-confidence side near a 60 percent read, and even that one backed in rather than cruised. The only other win came on a near coin flip, a slight Cubs lean in a game the model itself filed close to even.

That is a poor return on the confidence tier, and it is the number that matters most here. A model earns its keep when the sides it likes best hold up; on Thursday they went 1-for-4. Verdexed's accuracy and edges pages at verdexed.com/mlb were not cleanly retrievable at the time of writing, so the exact graded accuracy percentage and Brier figure for the day cannot be confirmed and are not invented here. But the named-side ledger is unambiguous, and it reads as a down day for the favorites.

Best call / worst miss

The best call was also the only favorite that held. The model's headline side was the Phillies behind Cristopher Sanchez in Washington, and Philadelphia won 10-5, so the pick landed. The candor it deserves is that it landed sideways. Sanchez, the pitching edge the entire lean was built on, got tagged for four runs in the first inning, and the Phillies trailed 5-0 before Bryce Harper and Derek Hill homered in the ninth to rally past the Nationals. Right side, wrong reason. A win is a win, but a clean pitching-matchup read that survives only because the offense erupted late is a reminder that the model's premise and the model's result are not the same thing.

The worst miss was the Yankees. Verdexed read New York in the low-to-mid 60s at Fenway behind Cam Schlittler, and the Yankees lost 6-3 to a last-place Red Sox club. The score flatters Boston less than it should: New York committed a season-high four errors, and all six Red Sox runs were unearned. The American League's best record (48-31) fell to its worst (32-46) on a sloppy defensive night, exactly the kind of variance a probability model cannot price into a starter matchup. Right behind it, the Blue Jays were a clear home favorite near -155 and lost 6-5 as Kevin Gausman was rocked for six runs, and the Tigers, a slight favorite, lost a 2-1 pitchers' duel in which Houston rookie Tatsuya Imai, the arm the model doubted on a 6.15 ERA, struck out 10. Three favorites, three different ways to lose.

What it means

One Thursday does not redraw the model. The favorites went 1-for-4, but the sample is four games, and three of the four were decided by a single run or by a defensive meltdown that no projection accounts for. The Yankees did not lose because the model misread the matchup; they lost because they kicked the ball around. The Phillies did not win because the projection nailed it; they won despite their ace getting hit. That is noise wearing the clothes of a pattern.

The sober read is to anchor to the longer run rather than to one card. Verdexed's own framing is that high-confidence MLB sides tend to land somewhere around three in five across a month, with cold stretches mixed in, and a 1-for-4 day on the favorites sits inside the range a sound model absorbs without alarm. The month-to-date figure could not be confirmed from the accuracy page on this pass, so it is left qualitative here rather than stated as a number. What Thursday earns is an honest entry in the ledger, not a verdict: the model had a bad day with its chalk, the one favorite that held needed a ninth-inning rescue, and the board moves on to Friday.

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.