Optimal Slips

What these are, precisely. A leg qualifies on two counts and no opinion: the line has gone one way often enough across every settled event we hold at that exact number, and this batter is not an exception to it, measured on his own record and shrunk toward the line's. No projection, no tier, no model probability enters the calculation. The percentage on each leg is the smaller of those two counts, so it understates rather than flatters.

The model probability is still withheld, and that has not changed. This board's own graded record is 52.8% on 1,547 decided picks since 2026-07-28, measured from the first slate projected entirely from repaired inputs — below the ~54% break-even of the -119 leg the grader prices, and below the 57.7% a two-pick entry has to clear, on a sample too short to conclude much either way. So the model still publishes no tier, no edge and no stake anywhere on this site. These slips do not use it. They are a different claim, made from a different number, and they are graded separately.

Read the stake, not the payout. Each slip shows its win probability, its expected value, and the fraction of a bankroll the arithmetic actually justifies — capped at 3%, and often far less. A large multiple with a small stake is the honest shape of a thin edge, not a recommendation to press it. These are unvalidated forward: the method was tested on history, this board's live record starts now, and every leg here is logged and graded like any other pick.