Yesterday's picks: how the model did as MLB chalk held and the Knicks favorite fell

Verdexed graded two sports on June 8: a full Monday MLB card and the lone basketball game on the board, Game 3 of the NBA Finals. The honest one line read is that the day pulled in two directions. The baseball favorites the model leaned on mostly did their job, while its single marquee call, a modest home favorite in the Finals, did not. There was no NHL to grade, with the Stanley Cup Final idle between Game 3 on June 6 and Game 4 on June 9.

That split is worth stating plainly before any numbers. A solid night on a stack of baseball chalk does not erase a high profile miss on the one game most people were watching, and a model that grades in public has to own both at once.

The scorecard

On the baseball side, the model worked a full Monday slate, which on its terms is a long list of win probability calls rather than a few headline spots. The favorites it leaned on came through more often than not. The Seattle Mariners, the AL West leaders and the better side against Baltimore, won 6-3. The Philadelphia Phillies handled Toronto 5-2 behind a dominant start. The Houston Astros edged the Los Angeles Angels 5-4 in a tight one. None of those were upsets, and that is the point: when a probability model leans on the stronger team and the stronger team wins, the chalk is doing exactly what it is supposed to do. The baseball portion of the day graded on the right side of even.

The basketball portion is where the miss lives. Verdexed's game model had New York as a modest home favorite for Game 3 at Madison Square Garden, roughly a 2.5 point favorite by the market, while explicitly flagging the game as closer to a true toss-up than the public read of a dominant 2-0 leader. The Knicks lost outright, 115-111. The straight favorite did not deliver. The one piece of context that keeps this from being a clean blown call is that the model had already downgraded its confidence on that side and pointed to value on the Spurs as a live home dog, which is exactly how the night played out. The favorite still lost, and the scorecard records that as a miss, but the model's own caution on the spot was warranted rather than ignored.

On calibration the takeaway is qualitative rather than numeric, since Verdexed does not surface a single day Brier figure publicly. A day where the baseball chalk holds and a flagged coin flip lands on the underdog side is close to what a well calibrated book expects, not a result that should move a longer run record.

Best call / worst miss

The sharpest spot was on the diamond. The Mariners were the stronger side against Baltimore and won 6-3, powered by a Josh Naylor grand slam in a five run fifth, with Randy Arozarena adding three hits as Seattle ran its recent stretch to 10 wins in 13 games. That is a clean example of a favorite the model would have leaned on cashing without drama. The Phillies were close behind, beating Toronto 5-2 as Cristopher Sanchez struck out 10 over seven innings and Adolis Garcia provided a two run homer.

The worst miss is the one everyone saw. The Knicks, the model's modest home favorite and a team up 2-0 in the Finals, lost Game 3 at MSG by four, 115-111. Victor Wembanyama imposed himself for a full night with 32 points, eight rebounds and six assists, Stephon Castle added 23, and San Antonio got the road response it needed to cut the series to 2-1. Jalen Brunson and OG Anunoby both scored well for New York, but the home favorite could not close. For a model that priced the Knicks as the better side, the favorite losing on its own floor is the day's clear blemish, even with the toss-up flag attached.

What it means

One Monday is noise, not a pattern, and the right frame is the longer run rather than a single box score. The baseball card going better than even on its favorites is ordinary, and the Knicks dropping a game they were a small favorite to win is well within the variance any Finals game carries. The series itself still tilts toward New York, which leads 2-1 with the model's longer run read on the matchup unchanged by one result.

The sober verdict: a decent night on baseball chalk, undercut by a high visibility miss on the lone basketball game, with the model's pre game caution on that exact spot keeping the call honest. Nothing here argues for reading more into June 8 than that. The model's value, if it has one, shows up across hundreds of these calls, not across one Monday where the favorites split between two sports.

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.