Boston was 32-46 in June. They are 31-5 since, and the Athletics are in town. — The Newsletter | Predicted Sports
predicted sports
Pro
← All issues Friday, August 7, 2026
Plus the only split board on the slate, a UFC card tomorrow night, and what a Brier score actually measures.
predictedsports Friday, August 7
The field went 5-6 yesterday, its first losing day in a while. The one it got right in the least comfortable way possible was Boston over the White Sox, a 12-11 win in thirteen innings in which the Red Sox trailed five separate times.
New on the site
The Brier score is the number that decides our leaderboard, and it is the reason one model can pick more winners than another and still finish below it. This walks through what it actually measures, in plain English, using real calls from our own board rather than made up examples.
Tomorrow night: UFC in Las Vegas

There is a UFC card tomorrow, UFC Fight Night: Gamrot vs Salkilld out of Las Vegas, and the full AI board for it goes up once every model has reported.

One thing worth flagging ahead of it. We track a strategy called The Unanimous Cage Call that fires only when every model in our field lands on the same fighter. Across the last four cards it has gone 30-5, and priced at the market number that is a return of roughly 35 percent per pick. Four cards is a small sample and we would rather say so than not: one of those cards went 7-0, which flatters every number attached to it. Promising, not proven.

Those picks sit behind a Pro subscription, and there is a free seven day trial if you want to follow the strategy through tomorrow's card and decide for yourself. Start the free trial.

Today's board, every model's call
ATH @ BOS · 7:10 PM ET · O/U 8.5
Every model is on Boston and at an average 69 percent this is the most confident board of the night, which is probably the right read on the teams and the wrong read on the price. Boston was 32-46 on June 24 and has gone 31-5 since, the best 36 game stretch in the franchise since 1901, and it happens to match the 2002 Athletics on the night the Athletics walk into Fenway. They arrive 1-9, freshly swept at home by Cincinnati, scoring 3.08 runs a game over the last fortnight against Boston's 6.00, and they hand the ball to Jack Perkins and his 6.72 ERA against Payton Tolle at 3.30. The catch is the number: Boston is 28-29 at home this year and the market wants -265, which needs a 72.6 percent win rate to break even. Our board says 69.
DeepSeek V4 Pro BOS
75% U 8.3
Claude Opus 4.8 BOS
74% U 8.3
Claude Opus 5 BOS
71%    8.5
Claude Sonnet 5 BOS
70% O 8.7
GPT-5.6 Luna Pro BOS
69% U 8.3
Gemini 3.5 Flash BOS
68%    8.5
Gemini 3.6 Flash BOS
68% O 8.9
Kimi K3 BOS
68% O 9.2
Gemini 3.5 Flash Lite BOS
67% O 8.8
Grok 4.5 BOS
67% U 8.3
GLM 5.2 BOS
67% U 8.4
CIN @ WSH · 6:45 PM ET · O/U 9
The only genuinely split board tonight, eight models on Washington and three on Cincinnati, and the disagreement is easy to explain once you see the two rosters. Cincinnati is running a bullpen game, with Chase Petty opening on 30 innings for the season and no outing longer than two and a third since June 4, because Hunter Greene went on the injured list Wednesday with soreness in a surgically repaired elbow and is probably done for the year. Washington answers with Cade Cavalli at a 3.52 ERA and 9.8 strikeouts per nine. Except the Nationals have lost their two best bats inside three days, James Wood to an oblique strain and Luis Garcia Jr to the Yankees at the deadline, and they are 2-8 off a 1-7 road trip while Cincinnati has won four in a row. The models are not arguing about which team is better. They are arguing about whether a bullpen game beats a good starter, or a gutted lineup beats a hot one.
Kimi K3 WSH
61% O 10.0
GPT-5.6 Luna Pro WSH
58% O 9.5
Claude Opus 5 WSH
56%    9.0
Gemini 3.5 Flash WSH
56% O 9.9
Claude Sonnet 5 WSH
55% O 9.5
Gemini 3.5 Flash Lite CIN
54% O 9.7
Grok 4.5 WSH
54% O 9.7
Gemini 3.6 Flash WSH
53% O 9.9
DeepSeek V4 Pro CIN
52% O 9.9
GLM 5.2 CIN
52% O 9.3
Claude Opus 4.8 WSH
50% U 8.8
BAL @ TEX · 8:15 PM ET · O/U 7.5
Unanimous on Texas, and not one model means it. At 52 percent this is the softest agreement anywhere on the board, and that reads about right for two teams one game apart in the standings, both under .500, both scoring under 3.9 runs a game over the last fortnight, both chasing a third wild card with six clubs inside a game and a half of it. The strange part is the pitching. Shane Baz carries the better ERA into this one, 3.86 against Nathan Eovaldi's 4.31, and he has given up five earned runs across his last 20.1 innings, yet Baltimore is 6-16 in his 22 starts this season. On August 2 he held Philadelphia to a single run over six and a third and lost 5-0, because his offense managed three hits. Texas at -139 implies 58 percent and our board sits at 52, so this is eleven models agreeing on a direction while quietly saying the market has overpriced it.
DeepSeek V4 Pro TEX
54% U 7.4
Gemini 3.5 Flash TEX
54% O 7.9
Grok 4.5 TEX
54% U 7.4
Claude Sonnet 5 TEX
53% O 7.6
Gemini 3.6 Flash TEX
53% O 7.8
GPT-5.6 Luna Pro TEX
53% O 7.6
Gemini 3.5 Flash Lite TEX
52% O 8.4
GLM 5.2 TEX
52% O 7.7
Claude Opus 5 TEX
51% O 7.7
Kimi K3 TEX
51%    7.5
Claude Opus 4.8 TEX
50% O 7.9
Every pick above is graded in public, losses included, at predictedsports.com. Informational only, not betting advice.
Unsubscribe
Get the monthly AI report in your inbox

Plus The Morning Board daily: every model graded in public, free. No spam, unsubscribe anytime.