THE PICK
SF San Francisco GiantsAWAY · 1471 ELO
@ Seattle MarinersHOME · 1502 ELO SEA
42.0% WIN PROBABILITY 58.0%
The model makes Seattle Mariners a 58.0% favorite — home advantage included.
One game is one game: a 58.0% edge still loses 42.0 times in 100.
Prediction history
6 UPDATES Every time the model moved this pick — logged in full. It updates until 15 minutes before gametime, then locks.
58.0% Seattle Mariners ▲ +0.2
Ratings shifted on the latest results run.
57.8% Seattle Mariners ▼ -0.3
Ratings shifted on the latest results run.
58.1% Seattle Mariners ▲ +0.8
Ratings shifted on the latest results run.
57.3% Seattle Mariners ▼ -0.9
Ratings shifted on the latest results run.
58.2% Seattle Mariners ▼ -0.3
Ratings shifted on the latest results run.
58.5% Seattle Mariners OPENING LINE
First model run when the matchup posted. Elo gap of 31 points at open.
The factors
1 IN THE NUMBER Everything the model sees for this game. Only backtest-validated signals move the probability — the rest is context, shown anyway.
PROBABLE STARTERS
Landen Roupp FIP 3.28 · ERA 4.27 · 97 IP
Bryce Miller FIP 3.12 · ERA 2.18 · 57.2 IP
GAME-TIME CONDITIONS · RETRACTABLE ROOF · NIGHT GAME
68°F DEW 54°F HUM 61% WIND 5 MPH CLOUD 0% 1015 hPa
SCHEDULE & TRAVEL
SF REST 4D SEA REST 4D SF TRAVELED 678 MI INTERLEAGUE
IN THE NUMBER
PITCH +1 ELO
The numbers
AWAY
San Francisco Giants ELO RATING1471
RANK#24
RECORD41-55
LAST 105-5
STREAKW2
HOME
Seattle Mariners ELO RATING1502
RANK#16
RECORD48-49
LAST 104-6
STREAKW1
Ratings come from every MLB result this season — updated daily, weighted by
margin of victory, with Seattle Mariners getting the standard home bump. An Elo difference
of 31 points plus home advantage at T-Mobile Park produces
the 58.0% call. Full methodology →
What could break the pick
San Francisco Giants come in hot on a 2-game win streak — recent form the season-long rating hasn't fully priced in.
The edge is real but modest; single-game variance eats edges like this regularly. Starting pitchers, lineup cards, and bullpen usage move real outcomes in ways a rating system can’t see.