Yesterday's picks: how the model did (August 5)

Wednesday's board was full and this desk graded nothing on it. The August 5 column published no stated sides, no edges and no confidence tiers, because per-game model probabilities were not retrievable at publication. A day with no picks produces no record, and the honest version of that is a blank line, not a backfilled one.

The scorecard

Zero stated directions, zero correct, zero incorrect. That is the entire entry for August 5.

Two things belong apart here. The first is the model, which logs and scores its own predictions whether or not a column quotes them. The second is this column, which can only grade what it published. On Wednesday the second was empty and the first could not be reached: no slate-wide accuracy figure, no Brier value and no confidence-tier breakdown for August 5 was retrievable from Verdexed's public accuracy surface at the time of writing. That is a fourth consecutive morning without a retrievable slate figure, and it is the more serious of the two gaps.

What the column did publish were structural notes with sourcing flagged on every line: which clubs the market had priced as home favorites, which starters were reported rather than confirmed, and one line calling Angels at Orioles the closest thing on the board to a straightforward mismatch on records alone. Those are observations about the market and the schedule. They are not model picks, and nothing below should be read as a hit rate.

The confirmed results on the six games the column named:

**Cubs 7, Dodgers 6.** Chicago completed the sweep at Wrigley. **Cardinals 3, Yankees 1.** St. Louis took the series in the Bronx after a 68-minute rain delay. **Mets 6, Guardians 5** in 10 innings. **Orioles 5, Angels 2.** **Blue Jays 5, Astros 4** in 10 innings. **Rays 4, Rockies 0**, a shutout at Coors Field behind Nick Martinez.

Measured against the column's three directional notes rather than any model output, one held and two did not. Baltimore, the named mismatch, won. Houston and Cleveland, the two named priced home favorites, both lost, and both lost in extra innings.

Best call / worst miss

There was no call, so there was no best one, and that is the finding rather than a way around it.

The nearest thing to a call that landed was Baltimore. Tyler O'Neill and Coby Mayo homered back to back in the fourth, Trevor Rogers carried a two-hitter into the seventh, and the Orioles beat the Angels 5-2 at Camden Yards. Baltimore entered the series at 55-58 and hosted a Los Angeles club at 43-70, which is the gap the column described. It is also the softest kind of correct read, since it required nothing beyond setting two records side by side.

The nearest thing to a miss was the chalk the column pointed at. Toronto beat Houston 5-4 on Ernie Clement's two-run homer off Bennett Sousa in the tenth. Cleveland lost 6-5 in ten when Bo Bichette singled home Francisco Lindor off Matt Festa, after Brayan Rocchio fumbled a bases-loaded grounder in the ninth to let the tying run score. The column had also cited a trend crediting the Guardians with wins in each of their last six meetings with New York. That broke as well. Every piece of directional flavor in Wednesday's column pointed at home favorites, and the home favorites lost twice in extra innings.

The largest game on the board was the one the column said the least about. Pete Crow-Armstrong homered twice and drove in four as the Cubs finished the sweep 7-6. Shohei Ohtani also went deep twice, and Max Muncy hit a three-run homer in a five-run eighth that got Los Angeles within one. It was the Dodgers' sixth straight loss, their longest skid since July 2025. A Dodgers game had headlined the card on each of the three previous mornings. On the one morning the desk stayed silent, it was the loudest game of the day.

What it means

A blank day changes nothing in the running record and should not be dressed up as either discipline or failure. The stated-side ledger stands where it did on Tuesday night. Per this desk's own prior recaps, five graded cards read 2-0, 1-1, 2-2, 3-2 and 4-2, or 12-7. Nineteen decisions carry essentially no information, and adding a zero to the front of that does not change the picture in either direction.

The real signal from Wednesday is not about the model at all. It is about what this column can see. Four consecutive mornings have passed without a retrievable slate-wide accuracy or Brier figure, and Wednesday was the first day on which even the named games carried no numbers. Declining to publish a probability that cannot be sourced is the correct behavior, and the picks desk made the right call in printing nothing. But a transparency column that cannot reach the model's scored output is not, on that day, measuring anything. That limitation should be weighted more heavily by a reader than any single day's win or loss would be.

There is no way to know from here whether the model would have been right on August 5, and no guess about that counterfactual belongs in a record.

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.