Mar 29 2021 04:24 AM
Mar 29 2021 04:42 AM - edited Mar 29 2021 04:44 AM
@kendzis2908 Nothing wrong with it. The average of the individual percentages would only be equal to the overall percentage if ALL of the games played are the same. For instance, if all 7 teams played 11 matches, thus 77 matches in total, then the total 28 wins reflect 36.36% winning ratio. Exactly the same as the average of the individual winning percentages.
Mar 29 2021 04:50 AM
You haven't done anything wrong - the average of sums is in most situations not equal to the sum of averages. That would only be the case if the number of played games was the same for all players.
Here is a simplified example with just two players.
First, a situation in which both have played 4 games. One player lost all 4, the other won all 4:
Here, the numbers correspond because the number of played games is the same for both players.
But now, a situation in which one player has played only 2 games (and lost them), while the other has played 6 games (and won them).
Since the second player has payed more games, the overall average of 75% is skewed towards the second player. It is a weighted average,
The average 50% of the two averages 0% and 100% is not a weighted average - both players contribute the same to it, even though the second player has played three times as many games as the first one.