Of course if we're getting into this much precision, we maybe should be calculating each player's age to the day, then weighing their precise age on the date of each game by the number of minutes played in that game.Cross posted as this may be the better place to discuss.
Then we'd need to account for the fact that this season isn't complete yet, so my proposed method would bias the 24-25 age younger than the 23-24 age just because we haven't completed the season -- the individual player ages in future games are higher than the player ages of games already played.
But that's my engineering/actuarial background talking. The ratio between the amount of effort needed and how interesting the output would be is way too high, so lets just stick with your analysis instead.