View Single Post
Old 04-29-2019, 02:58 PM   #26
Bill Veeck
Minors (Rookie Ball)
 
Join Date: Aug 2014
Posts: 30
Quote:
Originally Posted by Syd Thrift View Post
Recalc is, well, how you want the game to recalculate a player's ratings from season to season. If you do 1-year recalc then Roger Maris will have a big bump in his HR rate in 1961...hits 70 about as often as he hits 50, which looks wildly implausible given that nobody had hit 60+ since 1927; it's much more likely that he got "good dice rolls" and overachieved in '61 a bit) but it's less volatile.
I always use 1-year and that does lead to some "exaggerated" results. I always thought that anything beyond 1 year was somewhat unrealistic, but you cite an excellent example in Maris hitting 61 in 61. He obviously had "Good dice rolls" that season in real life.

Using 3 year recalc he'd be expected to hit 36 (39 actual) in 1960, 47 (61) in 1961 and 46 (33) in 1962. (the numbers are adjusted for PA in the 3 years). HR/PAs would be 7.8%, 10.3% and 5.6% for 1-yr recalc vs. 7.3%, 7.9% and 7.8% for 3-yr recalc.

The point being that if you believe, as I do that there is a significant amount of "good dice rolls" factored into real life, by using 1-yr you are adding another layer so he could easily hit as few as 52 or many as 70 in 1961 using 1-yr while he's usually going to be between 40 and 54 using 3-yr calc.

If he hits 70, you probably say to yourself, "that was unrealistic." However for 1-yr recalc it certainly is to be expected as is 52. After thinking about this for a bit, it still gets back to what you are looking to design.

The only thing I am sure of is that it's not unrealistic to get individual player results - or for that matter team wins - that fluctuate more than most of us would think.
Bill Veeck is offline   Reply With Quote