It is mid-May. Team A is 22-14 and sitting in first place. Team B is 16-20 and struggling. They play each other, and surprisingly, the model highlights Team B as the value play.
How can a sub-.500 team be the right pick against a division leader?
The answer often lies in run differential.
The problem with W-L records
A win is a win in the standings, whether a team wins 10-0 or 3-2. But for evaluating how good a team actually is, those two games are completely different.
In a small sample size—like the first 40 games of an MLB season—win-loss records are highly susceptible to sequence luck.
- A team might go 8-2 in one-run games, driven by a few lucky bounces or perfectly timed hits.
- Another team might lose five games by a single run, while winning their games by blowout margins.
Over a 162-game season, luck in one-run games tends to regress toward the mean. The team winning all the close games will start losing some; the team losing the close ones will start winning some.
Enter Pythagorean Expectation
In the 1980s, Bill James introduced a formula called Pythagorean Expectation. It uses a team's runs scored and runs allowed to calculate what their winning percentage should be.
The basic formula is:
`(Runs Scored^2) / (Runs Scored^2 + Runs Allowed^2)`
While modern analytics have tweaked the exponent slightly, the core truth remains: Run differential predicts future wins better than past wins do.
If Team A is 22-14 but has been outscored by their opponents over the season (-5 run differential), their record is a mirage. They have been winning close games and getting blown out in their losses. The math says they are actually a slightly below-average team.
If Team B is 16-20 but has outscored opponents by +25 runs, they are suffering from bad sequence luck. They are hitting well and pitching well overall, but the timing of those runs has cost them wins.
How the market reacts
The betting public largely reacts to W-L records and recent streaks. If "lucky" Team A plays "unlucky" Team B, the public money will heavily favor Team A.
This inflates the price on Team A, creating value on Team B.
Our model doesn't care about the standings. It evaluates run creation, run prevention, and underlying metrics to determine a team's true baseline strength. When a team's actual performance diverges wildly from their Pythagorean expectation, it is one of the best opportunities to find edge in the moneyline market.
The takeaway
Next time you look at a matchup and wonder why a struggling team is priced competitively against a winner, check the run differential.
In baseball, the truth isn't always in the win column. It's in the runs.
For informational use only. Past results don't guarantee future performance. Bet responsibly.