Bonn1997 wrote:The number of seasons has no bearing on the reliability of each player's percentage. The number of observations for each player does. And that number is not even one-twentieth of what you'd need for each player. You could have 44 shots spread over five thousand seasons and it still wouldn't make the FG% any more reliable.

Not only that, but I'd wager that a players abilities change over time so that you couldn't expect the same performance from year to year. The difference wouldn't be explained only by some "expected" variability. You would have to introduce a new element to account for players declining ability after it reaches some theoretical peak. For example, I doubt that Melo will still be shooting as well in his 3,475th season as he might in his 10th season. Just a thought.

BasketballJones wrote:

Bonn1997 wrote:The number of seasons has no bearing on the reliability of each player's percentage. The number of observations for each player does. And that number is not even one-twentieth of what you'd need for each player. You could have 44 shots spread over five thousand seasons and it still wouldn't make the FG% any more reliable.

Not only that, but I'd wager that a players abilities change over time so that you couldn't expect the same performance from year to year. The difference wouldn't be explained only by some "expected" variability. You would have to introduce a new element to account for players declining ability after it reaches some theoretical peak. For example, I doubt that Melo will still be shooting as well in his 3,475th season as he might in his 10th season. Just a thought.