Quote:
Originally Posted by Brad K
What I'd really like is the guys who were good after age 30 and especially those who had their best years after age 30 to have a shot at that.
I think I'm going to try your method or some variation of it and accept that it probably kills those 3 good years and out guys from having 6-7 decent years on occasion.
So do I understand you're running recalc every three years? If so, what time of year do you do it? Or do you randomize when during the year it happens?
That might provide some interesting situations... make recalc a possibility but a low probability starting sometime during the second year since the last one and giving in increasing probability as time goes on until there's deadline and its done.
|
3-year recalc is not what you seem to think it is. Let's say the player in question is coming up on his 1985 season stats from RL. 3-year recalc would use his 1984, 1985, and 1986 weighted evenly (33.33% each year) to get ratings. In 1986, it would be 1985, 1986, and 1987 weighted evenly. 3-year recalc double weighted for the current year would take 25% from 1984, 50% from 1985, and 25% from 1986 for a player's 1985 season. 5-year recalc would be 1983, 1984, 1985, 1986, and 1987 weighted evenly (20% each year) for a player's 1985 season. 5-year recalc double weighted for the current year would take one-sixth from 1983, one-sixth from 1984, one-third from 1985, one-sixth from 1986, and one-sixth from 1987 and derive ratings accordingly. It basically smooths out the rough edges of variability that come with 1-year recalc. 5-year recalc is also very useful for dealing with the players who were interrupted for military service, or other long periods of time away from the game.