Yesterday's picks: how the model did (August 12)

There is no graded line for August 12. Verdexed's column for Wednesday's 15-game MLB board stated no sides, claimed no edges and assigned no confidence tiers, because the model's per-game probabilities were not retrievable when it published. The honest entry is zero picks, zero correct, zero incorrect, and that makes eight consecutive slates with nothing for this desk to grade. Baseball was the only sport in play. The NFL sat between the Hall of Fame Game and the start of preseason Week 1, the NBA and NHL are in their offseasons, so MLB was the entire graded universe and it graded to nothing.

The scorecard

Zero stated directions, zero correct, zero incorrect for August 12.

The stated-side ledger does not move. It stands at 12-7 across 19 decisions, where it has sat since August 4.

No trailing accuracy figure and no Brier value are claimed here, because none was retrievable. Verdexed's public accuracy surface could not be reached at publication, and no per-game or per-tier number for August 12 was available anywhere else this desk could source. That is now well past a week without a quotable calibration figure. It is an absence and nothing more, neither a quiet good stretch nor a quiet bad one.

What can be counted is the market, which is a separate instrument and earns the model no credit. The column quoted a firm two-way price in three games. Tampa Bay, near -180, won. Minnesota, near -110 in what the column itself called a near pick-em, won. Philadelphia, near -164 and the shortest number on the board, lost badly. That is 2-1 for quoted chalk, and the one that lost was the one the market liked most.

Best call / worst miss

There is no best call and no worst miss in the graded sense, because there was no call. What can be examined is where the published framing pointed, and Wednesday split it cleanly.

The framing that landed was the headline. The column named Phillies at Cardinals the board's widest gap between price and matchup, noting that the market was charging near -164 for Zack Wheeler and a 2.69 ERA while at least one public model, numberFire's, scored St. Louis at 50.5 percent. The market was wrong and the skepticism was right. **Cardinals 7, Phillies 1** at Busch Stadium, where four consecutive two-out hits turned into a four-run fifth against Wheeler, his second loss in as many starts. A side implied by that framing would have cashed at plus money. It was written down before first pitch and it can be checked afterward, which is the minimum standard. It still was not a stated pick, so it reaches nothing.

The framing that aged worst was Rockies at Diamondbacks, described as the day's clearest class mismatch by record on the strength of Merrill Kelly against Ryan Feltner. **Rockies 6, Diamondbacks 4** at Chase Field. Brett Sullivan's three-run homer off the right-field foul pole built a 4-1 second-inning lead, Zac Veen added a two-run shot into the center field pool, and Kelly finished with six earned runs in five innings. Reading a season record as a matchup edge is a shortcut, and the shortcut failed.

Elsewhere on the board, the structural note that held was Tampa Bay. The column flagged Drew Rasmussen against Jack Perkins as the largest paper mismatch, and **Rays 8, Athletics 4** followed at Sutter Health Park, Liam Hicks capping a four-run fourth with his 16th homer and Rasmussen winning his fifth straight start as Tampa Bay completed a 9-0 West Coast trip. Baltimore at Minnesota, flagged as a near pick-em, went the distance: **Twins 7, Orioles 5**, Alan Roden's go-ahead double doing the damage. **Cubs 8, Nationals 6** at Nationals Park, and **Yankees 10, Mariners 5** in the Bronx, where Trent Grisham homered twice and doubled in the go-ahead run in the seventh.

What it means

A full slate of results with no picks attached says nothing about the model, and the 12-7 ledger behind it says very little. Nineteen decisions is a sample a coin flip could produce in either direction, and treating it as a track record asks more of the number than it can carry.

Quoted favorites going 2-1 is not a finding either. Prices in that range describe teams that win somewhere around 55 to 64 percent of the time before the book's margin comes out, so two of three holding on a Wednesday in August is ordinary variance.

The finding is the gap, and it is getting old. Eight straight slates have passed without a graded decision, and the accuracy and Brier figures this column exists to report have been out of reach for more than a week. Wednesday at least beat a blank page in one narrow sense, in that the reasoning was published in advance and went one for one on its two loudest structural claims. But reasoning is not a pick and cannot be scored. An accountability desk is not damaged by losing days. It is damaged by long runs where nothing gets graded, because silence reads as success to anyone who is not counting. So the count gets published again, and the count is eight slates at zero graded decisions.

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.