← the lab

Monday, October 19

Two picks, and the reads behind them.

Posted before the games, graded after. Every public pick since Sep 25, 2026, win or loss.

FanaticsFlat $50 stakePick 1
Freddy Peralta Over 13.5 Outs
Cleveland Guardians @ Tampa Bay Rays8:10 PM ET
−120Fanatics
last 5 5 of 5 · 100%last 10 10 of 10 · 100%last 20 15 of 20 · 75%season 29 of 34 · 85%

Next best −130 at theScore BetSame 13.5 line

What is the pick for Guardians at Rays?

Freddy Peralta over 13.5 Outs at −120 with Fanatics.

Why over 13.5?

Freddy Peralta has gone over 13.5 outs in 85% of his 34 games this season, and in 10 of his last 10.

Full game log and prices for Freddy Peralta →

DraftKingsFlat $50 stakePick 2
Chase DeLauter Over 0.5 RBIs
Cleveland Guardians @ Tampa Bay Rays8:10 PM ET
+255DraftKings
last 5 4 of 5 · 80%last 10 6 of 10 · 60%last 20 9 of 20 · 45%season 49 of 143 · 34%

Next best +240 at theScore BetSame 0.5 line

What is the pick for Guardians at Rays?

Chase DeLauter over 0.5 RBIs at +255 with DraftKings.

Why over 0.5?

Chase DeLauter has gone over 0.5 RBIs in 34% of his 143 games this season, and in 6 of his last 10.

He faces Freddy Peralta.

Full game log and prices for Chase DeLauter →

What has to be true before the model posts a play

  • A real history behind it. At least 20 games logged, so a hot week cannot masquerade as a trend.
  • Something you can actually bet. Between -250 and +700, because a 40-to-1 shot with a great number on it is still a 40-to-1 shot.

Prices are the ones on the board when the card was posted. They move. Check yours before you bet.

Trend leans

Every line on the slate where recent form runs furthest from the price, graded. These are not picks and they are not part of the card's published record. A lean, not a probability.

True of every pick on this card

  • These are historical-rate reads. No park, weather, or lineup-spot adjustment is applied yet.
  • Retail prices carry roughly 7% vig on props, so every line here is priced with the book's cut in it. Shopping the price across books is what narrows it.
  • This card is forward-tested in public and every settled result is on the results page. The sample is still early, so treat it as research rather than a signal.

built Oct 19, 10:04 AM · selection and wording are generated from the data, not written by hand