Trimmed Mean
Drops your best and worst results, then averages what is left — the Olympic judging approach.
avg(results after removing best and worst)
Best at
A balanced reading that neither a lucky win nor an unlucky collapse can swing.
Where it misleads
Needs at least three matches to have anything left after trimming, and it discards your genuine peak along with your fluke.
How it works
Trimmed Mean removes the best and worst result(s) before averaging. It's like the Olympic judging approach: throw out the highest and lowest scores.
How it works: Drop the top and bottom placement(s) from your results, then calculate the mean of what remains. With finishes [1, 2, 3, 3, 9], dropping the best and worst gives mean(2, 3, 3) = 2.67.
When to use it: When you want a balanced view that isn't skewed by either a lucky win or an unlucky catastrophe. Works with both placements and raw scores.
Watch out for: Requires at least 3 matches to work (need something left after trimming). The number of results to trim is configurable.