How it works

For each starting pitcher, SALCI estimates the distribution of strikeouts they'll record in their next start. It combines four factors:

• Stuff — pitch quality and whiff potential from Statcast.
• Matchup — the opponent lineup's strikeout vulnerability.
• Workload — recent batters faced + projected innings.
• Location — command and called-strike rate.

Those inputs roll into a single SALCI score (0–100) and a grade (S through F). The model also runs a Monte Carlo simulation of the pitcher's likely K distribution, from which we report a floor, an expected count, and a ceiling.

A floor bettakes the Over on a pitcher's strikeout total when a sportsbook offers a line at or below SALCI's projected floor — the K count the model is roughly 70% confident the pitcher reaches. When the book prices a low alternate line at or under that floor, the model reads the Over as having a margin of safety.

A risk gate excludes lines priced more expensively than −350, so the record reflects prices we'd realistically take. Outcomes are settled against the official MLB line score and published in the Track Record.

The ceiling bet (the Under-side mirror of the floor) is paused while we rebuild it. Our own track record flagged it: the Under fired exactly on the weak starters where the sportsbook's number is sharper than the model's, so we were betting against the better line and the ROI went deeply negative.

We'd rather pull a losing play than keep it on the board. We're testing the opposite side of those same spots and will only bring a ceiling bet back once it proves out — measured by closing-line value, not a hot streak. Floor bets and hitter props are unaffected.

Every starter we score is logged in SALCI History with the predicted floor, expected count, book line, and ultimately the actual K count from the game log. The accuracy chart on that page shows hit-rate by grade and by date, so the model is auditable in public.

Day-to-day predictions are stored before first pitch. The accuracy tally updates after the game ends and the K count is confirmed against MLB's official line score.

The Track Record shows how the bets we actually post have performed — this is separate from model accuracy above. It spans floor bets (pitcher K Over) and our recommended pitcher-K and hitter anytime-hit plays. (Ceiling bets are paused — see above.)

Past performance does not predict future results. Variance dominates small samples, the model is wrong some of the time, and hot or cold stretches are expected. We publish the full record so it's auditable — not as a promise of returns.

SALCI is a prediction model. It is not a betting service or financial advice. Sportsbook lines are displayed for context because they're the reference market most readers use to gauge a pitcher's expected K count.

If a sportsbook line sits well below the model's expectation, that's information about the gap between model and market — not a recommendation. The model is wrong sometimes, the market is wrong sometimes, and variance dominates any small sample. Treat the numbers as analysis, not advice.

We do not place wagers and do not benefit from any user's wagering decisions. If you choose to bet, do so legally in your jurisdiction and within your means.

For entertainment and informational purposes only — not financial, investment, or betting advice, and not a guarantee of profit. Must be 21+ (or the legal age in your jurisdiction). If you or someone you know has a gambling problem, call 1-800-GAMBLER.

• MLB Stats API — schedules, lineups, box scores, post-game line scores.
• Baseball Savant (Statcast) — pitch-level metrics for stuff, whiff, and command inputs.
• odds-api.io — sportsbook main lines and game totals (context).
• the-odds-api.com — alt-line ladders for the same context.