Currently for Fortnite Tracker, I feel that the “average placement” (for tournaments) feature is flawed and doesn’t say much about a player’s skill. This statistic would be a lot more useful if the geometric mean was used instead of the arithmetic mean where the geometric mean is (x1 * x2 * … xN) ^ 1/N and the arithmetic mean is (x1 + x2 + … xN) / N. Basically I feel that the geometric mean is a lot more fair due to the nonlinear nature of tournament leaderboards.
The arithmetic mean treats the difference between 1st and 100th as the same as the difference between 9,900th and 10,000th even though there is a much, much higher jump for the first one. The geometric mean accounts for this and I feel is much more fair and rewards high placements accurately.
Example:
A player places 1st and 1000th in 2 events. AM: about 500. GM: about 31
A player places 500th twice in 2 events. AM: 500. GM: 500.