|
Re: Final CH2K8 08-09 Rosters (Details Inside)
Been meaning to post this for a while, so I'm not sure if there's been another update since the first version I tested.
Anyway, I ran 3 legacy sims with the DC/Cubs roster to see how the Big 10 compared to actual stats, in preparation for working on the 2009-10 rosters.
I was totally impressed with the results. Just based on the Big 10, these ratings produce very realistic results. Looking mainly at ppg and rpg, the teams' starters were typically within 1-2 ppg and about 1.0 rpg compared to the actual stats.
Just a few random observations (these are not criticisms, rather just some data trends I observed on an admittedly limited data set of 3 simulations):
Purdue's main 3 (Johnson, Hummell & Moore) scored around 3 ppg less than they did in real life, and this was about the biggest statistical variation I saw. The results were very consistent, though. For example, Hummel and Johnson averaged the exact same ppg in separate legacies (10.3 and 10.1 ppg, respectively).
The largest individual variation in scoring I saw was Kevin Coble of NW, whos ppg ranged from about 12 ppg to 18 ppg in the three sims (his actual ppg was 15.5 ppg).
On MSU, Green's simulated stats were closer to Ibok's real-life stats, and Ibok's simulated stats were closer to Green's real-life stats.
Overall, Iowa and Indiana outperformed their real-life counterparts by winning several more games per year in the simulations. Michigan, Wisconsin, and Penn St underperformed as far as team wins go.
Regarding Iowa, Jermaine Davis and Jared Cole had much better simulated stats than in real-life (roughly double ppg and rpg). This may have contributed to Iowa's overall team performance being slightly better.
I did notice that players that didn't play very much (except for 'garbage time') sometimes had better ratings than the players who actually had more minutes played during the actual season. For example, IU's Gamble and Barnett played only 28 minutes in real-life, but their ratings based on this limited playing time had them simming out ahead of players like Moore and Roth, who actually played over 900 minutes combined. I suspect there is a point for bench players where more generic ratings could be applied to off-set elevated stats for those guys who only play in 'garbage-time.' (i.e., the last few minutes of lopsided games).
A few guys seemed to overperform for whatever reason. The following players ppg and/or rpg simmed out to roughly 1.5 to 2 times their actual stats:
Rowley, Nash and Ryan of NW (while Peljusic underperformed)
Kelvin Grady of Michigan (while Novak & Douglass underperformed)
Dallas Lauderdale & PJ Hill of OSU (while Diebler slightly underperfomed).
The only actual mistakes I found involved redshirts, and are listed below (I realize these may have already been fixed in a newer version, but just in case, here they are):
Ben Cronin of Michigan and Billy Oliver of Penn St should be redshirted.
Jonas Suotamo of Penn St didn't play in 2008-09.
Domique Keller of Illinois did play, but has a redshirt on the roster. Based on looking at his ratings, I don't think they were updated to reflect actual 2008-09 stats.
Anyhow, thanks again for all the hard work. This is a very impressive undertaking, and is much appreciated. These results were closer to real-life than I would have though possible. Excellent job! Your system is very good.
Last edited by Seryn; 05-15-2009 at 01:40 PM.
|
Operation Sports
05-15-2009, 01:34 PM
Re: Final CH2K8 08-09 Rosters (Details Inside)
Linear Mode
