Model: each starter's first-inning hold rate and the opposing offense's first-inning scoreless rate, combined against the league baseline, nudged by first-inning strikeout rate, then multiplied into a game probability. Hold rates are pulled toward the league average because a 20 to 25 start sample overstates how good or bad a pitcher really is. Fair price is that probability expressed as American odds, shown so it can be compared against a book rather than read on its own. Our own testing found this model does not beat the market, so treat it as research, not a recommendation. Streaks and recent form are shown because people look for them, not because they predict the next game. 21+.