Winning Percentage Calculator

MLB and the NBA do not record ties at all; an MLB game that reaches a tied score in extras simply continues, and an NBA game goes to overtime until one side wins. Their winning percentage formula is the simple wins divided by (wins + losses). The NFL has counted a tie as half a win and half a loss since 1972, so a 10-5-1 record produces (10 + 0.5) / 16 = .656, not 10 / 15 = .667. The NHL ran a similar half-win convention from 1917 through the 2004-05 season, then replaced ties with the overtime/shootout system in 2005-06 and now ranks teams by total points, where a regulation win is 2 points, an overtime or shootout win is 2 points, an overtime or shootout loss is 1 point, and a regulation loss is 0. Most professional soccer leagues (English Premier League, La Liga, Bundesliga, MLS) and most cricket Test/ODI leagues use the 3-1-0 system: 3 points for a win, 1 for a draw, 0 for a loss, then rank by total points and goal difference rather than by a raw winning percentage. The pick of convention in this calculator decides which formula is applied.

Bill James introduced the Pythagorean expectation in his 1980 Baseball Abstract, a self-published annual that grew into the most-cited foundation of modern sabermetrics. James noticed that a team's actual winning percentage tracked closely to runs scored squared divided by (runs scored squared + runs allowed squared), and named it Pythagorean because the formula resembles a^2 / (a^2 + b^2). The original 1980 fit on MLB seasons came in at a typical error of 3 to 4 percent and immediately exposed teams that were over- or underperforming their run differential, which often regressed toward the Pythagorean estimate in the following season. The technique was adopted across pro sports (Daryl Morey adapted it for the NBA at Stats LLC in the 1990s, the Football Outsiders group fitted it for the NFL in the early 2000s) and is the reason a team's run or point differential is now considered a more reliable indicator of true team quality than its won-lost record over short samples.

Because higher exponents fit higher-scoring, lower-variance sports better. Baseball scores in single digits per game with massive game-to-game variance, so an exponent of 2 captures the wide gap between teams that score 5 runs per game and teams that score 4. The NBA scores 100+ points per game with relatively small per-game noise, which means a team that outscores opponents by even 3 points per game wins a high fraction of its games; a 13.91 exponent (Daryl Morey's fit from Basketball-Reference data) reproduces that steepness. The NFL sits in between at 2.37 (Pro-Football-Reference's fit), and soccer's 1.83 reflects very low per-game scoring with large draw probability. The exponents are not chosen theoretically; each one is the value that minimizes prediction error against decades of historical results in that league. Using the wrong exponent (NBA exponent on a baseball season, say) would produce expected winning percentages near 1.000 or 0.000 for any team with even a small run differential.

It was a 1972 rule change. From 1932 to 1971 the NFL excluded tied games entirely from the winning percentage calculation, so a 10-2-2 team had the same .833 winning percentage as a 10-2 team. That created standings quirks: in 1968, Baltimore finished 13-1 and Cleveland 10-3-1, but the Browns' tie was ignored entirely and Cleveland's percentage was 10/13 = .769 versus Baltimore's 13/14 = .929. The half-win, half-loss convention adopted in 1972 includes ties in the denominator so they cannot be ignored, and matches how most sports historians and statisticians had been computing the percentage informally for decades. Ties are now rare in the NFL (the 10-minute regular-season overtime period adopted in 2017 produces fewer than one tie per season on average), but the convention still applies in tiebreakers for playoff seeding. Cup competitions follow the opposite logic: a tie that goes to penalties is recorded as a win for the team that advanced, because the practical outcome is identical to a win.

A team is mathematically eliminated from reaching a target winning percentage when even winning every remaining game cannot get them there. The math is simple: total games at season's end will be (current games + remaining games), and to finish at target WP T, the team needs at least ceil(T * total_games) wins. If their current wins plus remaining games is below that threshold, they are eliminated. For example, a baseball team at 60-80 with 22 games left and a target of .500 needs 81 wins from 162 games. They have 22 wins possible from here, plus 60 already, for a maximum of 82 wins, so they are not yet eliminated, but it requires a 21-1 finish. Drop to 60-82 with 20 left and the max becomes 80, which is below the 81 threshold, so they are out. This calculator runs that exact check under whichever convention you select (ties affect the denominator in NFL, cup, and soccer-points modes), so it works for any sport.