Explanation
The Pythagorean Rating formula originates from baseball statistician Bill James, who crafted a formula that settles the discrepancy between how many games a team should win vs. how many they actually won by using a team’s margins of victory over a season. Specifically, it looks like:
For instance, the 2023 Baltimore Orioles finished 101–61, the best record in the American League, but their margins were not quite as great as their record. They scored 807 runs and allowed 678, which works out to a Pythagorean record of just 94–68, seven full wins below their actual mark. This overperformance was immediately realized in the playoffs, where they were first-round exits.
The brilliance of this framework is that it can be applied to any sport, so long as the exponent is tuned to minimize the MSE. For instance, basketball uses a team’s point margins and has an exponent tuned to 13.91. Hockey uses a team’s goal margins and has an exponent tuned to 2.15. In this school of thought, I personally tuned Bill James’ formula to VCT by using round-differentials.
The below Pyth% is a mathematically-proven way of seeing the true strength level of a team relative to the year — the true rate at which they should win maps.
- In keeping with Bill James’ framework, only domestic events are used to calculate Pyth%. This is because in domestic splits, a team’s schedule is balanced to play all opponents, or at least an even distribution of opponents by strength. Adding internationals would skew the teams’ Pyth%.
For example, at LOCK//IN, NAVI played Krü (2023 Pyth% of 37.6%), TS (2023 Pyth% of 47.7%), Lev (2023 Pyth% of 47.3%), and eventually Fnatic (who were the best team of 2023). Meanwhile, a team like Sentinels just played Fnatic (again, the best team of 2023), where they got stomped 6–13 and 7–13. The two teams clearly got different luck when it came to their LOCK//IN draw. If internationals were included in Pyth%, NAVI’s 2023 value would be unfairly skewed upwards, and Sentinels’ 2023 value would be unfairly skewed downwards.
- If you’re interested in seeing international Pyth%, those are calculated separately and located within the all-time category in the “Internationals” filter.
| # | Team | W-L | Pyth% | Actual Win% | Luck | RW | RL |
|---|