College Basketball Player Ratings (Statman, 2011)
Posted: Fri Apr 15, 2011 12:12 am
page 1 of 4
Author Message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Sun Jan 30, 2011 2:31 am Post subject: College Basketball Player Ratings Reply with quote
After a long hiatus, I've decided to dust off my college player ratings/rankings. I did the top 7 conferences a couple nights ago:
http://www.pointguardu.com/f136/statmans-ratings-56243/
I'd love to find an area to start posting my work regularly. If I could find a semi easy way to get the stats - I could update nationally practically daily. I emailed Justin K - wondering if he'd be interested in having my stuff in his cbb-ref blog, since that specific blog has been pretty much dead.
Anyway - I'd love input from those of you interested in checking it out. Also, if you know where I should post (update) my work to get to the most college basketball fans' eyes. For time being - I'll be updating in the above thread. I may even dust off my pro ratings in that thread.
Thanks,
Dan Dickey
danthestatman@gmail.com
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
Chilltown
Joined: 16 Apr 2010
Posts: 15
Location: Boston
PostPosted: Mon Jan 31, 2011 5:02 pm Post subject: Reply with quote
This looks very interesting. I'd love to have you expand a bit on how you account for team defense and strength of schedule in these rankings. Mind sharing some details?
Back to top
View user's profile Send private message Visit poster's website
Crow
Joined: 20 Jan 2009
Posts: 825
PostPosted: Tue Feb 01, 2011 4:00 pm Post subject: Reply with quote
Where would Irving rank is he qualified on minutes?
These are college basketball ratings. I really only follow the NBA so I see these guys as future draft picks. How do or would you translate to pro potential?
Have you done a cut of the data for just when they play the top 64 or 32 teams? If not, would you consider doing so to isolate performance against on average better talent?
Would you weight NCAA tournament performance into a draft rating and moderately or heavily? Would you use prior season data and progress or just the last season?
What would your lottery pick list look like right now? If you compared to Draftexpress' top prospect list or 2011 mock draft are there any notable ranking differences you want to highlight and say more on?
Have you thought about or pursued writing in some fashion at draftexpress again in or beyond its forum?
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 4:18 am Post subject: Reply with quote
Chilltown wrote:
This looks very interesting. I'd love to have you expand a bit on how you account for team defense and strength of schedule in these rankings. Mind sharing some details?
Actually, I use Sagarin's team ratings, where, in essence, SoS and team quality (game results and differentials) are built in. I compile the individual weights for the team, scale them to the team rating. I then make playing time adjustments to individual ratings (scaling toward team rating depending on PT), then readjust all the ratings to exactly match the team rating again.
It'd be like PER, if there was a pre set team PER (based on SoS, results, and point differentials) that the individual PERs got adjusted to (positive and negative weights scaled appropriately).
The premise is that each player is part of the team whole. If a player is considered completely average for that team (looking at the weights and PT), then his rating would be exactly the same as that team's rating. An average player, statistically, for VMI would have pretty impressive looking stat totals (really high pace, weak SoS), while an average player stat totals for a Wisconsin player would look quite pedestrian. The average player for Wisconsin would have a much better rating than the average player for VMI obviously (much higher Sagarin rating), despite a much weaker stat line.
Well, I think I rambled a bit there - hope it helps. I use the same approach when I do NBA ratings (I just use team point ratios for team quality), which I'll probably do near All Star game time in that same thread. Players playing for multiple teams make it harder for me to compile everything quickly. My NBA rankings will probably look fairly similar to Mike G's, since I think there are a little similarities in approach.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 5:20 am Post subject: Reply with quote
Crow wrote:
Where would Irving rank is he qualified on minutes?
I'm at work at the moment - that spreadsheet is at home. Hopefully I'll remember to look it up and maybe compare his per minute rating (ignoring playing time adjustments) to other guys (natioanl PGs, ACC guys). If I'm not too sleepy I'll try to look it up before my kids wake up this morning.
Crow wrote:
These are college basketball ratings. I really only follow the NBA so I see these guys as future draft picks. How do or would you translate to pro potential?
It's not bad. Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season. Believe it or not, I don't have last season's stats - Cousins would have been real high I guess. Maybe Evan Turner was the highest? I'm forgetting. I thought I did an NCAA tourney player rating thing - I'll see when I get home.
Here's a flashback - I found a thread in which I ranked every player in EVERY conference, and gave my all conference teams, in 2008. I did national rankings too. The ratings here are quite similar to what I do at the moment - some slight changes I've made since then (I think I used Pomeroy's team ratings then, now I use Sagarin's - and playing time adjustments are slightly different):
http://www.pointguardu.com/f136/statman ... ndex2.html
I can divide the ratings into skill sets. I'd LOVE to do ratings over the last, say, 12 years, dividing the ratings into individual skillsets (pace, team quality, SoS, etc would all still be part of each skillset) - and see how these skillsets might best translate to future NBA production. If someone can get me the full past seasons' stats - I'd get right on that Wink
Crow wrote:
Have you done a cut of the data for just when they play the top 64 or 32 teams? If not, would you consider doing so to isolate performance against on average better talent?
I haven't done that, and it would be too difficult for me to attempt to do that with the limitations of me gathering the stats.
Quote:
Would you weight NCAA tournament performance into a draft rating and moderately or heavily? Would you use prior season data and progress or just the last season?
I'm not sure really. One would think maybe tourney performance should be weighted more - I'm not in a position to try to do that. I would look at each season seperately I think. I think comparing player ratings at similar ages could be important (Wall's freshman year rating compared to, say, Evan Turner's freshman year) in projecting possible NBA potential.
Quote:
What would your lottery pick list look like right now? If you compared to Draftexpress' top prospect list or 2011 mock draft are there any notable ranking differences you want to highlight and say more on?
I may look into this when I get home. I'm an Arizona alumn - I love Derrick Williams obviously. I'm biased. I also want him to come back as a junior.
Quote:
Have you thought about or pursued writing in some fashion at draftexpress again in or beyond its forum?
I've never written for Draft Express. Jonathan Givony at Draft Express expressed interest a few years back - contacted me about trying to incorporate my ratings into their database. He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe. Givony was intrigued my ratings had pegged him as a top player nationally. Well, their database guy said they couldn't incorporate the ratings into the database - so the whole thing was dropped.
As for writing, I'd love to actually start - but life kinda gets in the way. Working full time, watching my kids when I don't work, working out, kid's sports, spending quality time with the wife, etc. If I got paid to produce, say, 5 articles a week for someone (in essence, I did it for a living) - I have zero doubt I could produce some very interesting stuff. But, I've never even gotten my work together in the same blog. I've never saved past stuff into my own folders (hundreds of hours of past articles were lost in my old Wildcat Sports Report blog). In essence, I've not been organized well enough to show past quality work - and I've not devoted nearly enough time to keep stuff current to stay fresh in people's minds. I'm hoping to SOMEHOW change that, and squeeze in some cool things in that PGU thread as often as I can - until I may find a different, better, place to produce my work.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 10:43 am Post subject: Reply with quote
Crow wrote:
Where would Irving rank is he qualified on minutes.......
What would your lottery pick list look like right now? If you compared to Draftexpress' top prospect list or 2011 mock draft are there any notable ranking differences you want to highlight and say more on?
I'll touch these couple things, now that I'm home. First - Irving's rating when ignoring the playing time adjustment (pretty much a per minute production rating) is an awesome 206. His overall rating is For a freshman PG, that's pretty phenominal. To compare, his teammate Nolan Smith is ranked 9th overall - and his per minute rating is 183. Singler's is 143. Limited data, but very impressive.
If we set the minutes minimum to, say, 25%, Kyrie would be the 8th ranked player in the ACC with an overall rating of 152. He'd been #1 in the ACC (and 3rd or 4th nationally) obviously if he never got hurt and continued his same production.
For comparison, Josh Selby, acclaimed frosh PG with limited minutes for Kansas (due to eligibility) had a per minute rating of 110.
Among all players that played at least 2% of their team's minutes (Irving played 28.8%) - Kyrie ranked 5th in rating per minute. Derrick Williams was #1 at 228, then Jimmer at 225, Kemba at 221, Sullinger at 207, Kyrie at 206, Jordan Hamilton at 204, then others below 200.
As for Draft express - here's where my top guys, and where DE projects them:
Jimmer - #13
Kemba - #7
Derrick Williams - #5
Sullinger - #4
Jordan Hamilton - #17
JaJuan Johnson - #38 (steal?)
Jon Leuer - #54 (steal?)
Klay Thompson - #35 (steal?)
Nolan Smith - #25
Here's DE top guys declaring - and where they ranked (among all guys above 25% team minutes) in my ratings:
#1 Irving - #46 with a 152 (despite limited minutes)
#2 Perry Jones - #166 with a 127 (very possible bust - at least for years)
#8 Terrence Jones - #15 with a 169 (quite nice for a frosh)
#9 Harrison Barnes - #287 with a 116 (not close to ready)
#11 John Henson - #62 with a 146
#14 Singleton - #38 with a 154
#15 Marcus Morris - #12 with a 170
#17 Jeff Taylor - #88 with a 139
#18 Burks - #85 with a 139
#19 Leonard - #10 with a 172
Just looking at ratings, and a little of my feelings (what I've seen) - here's the college guys I'd consider in the lottery kinda in order:
Kyrie Irving
Derrick Williams
Jared Sullinger (body type worries me slightly - future weight issues?)
Kemba Walker
Terrence Jones
Jimmer Fredette
Jordan Hamilton
Those would probably be the only college guys I'd consider in the lottery. I'd probably avoid lottery probables Perry Jones & Harrison Barnes, for fear of bustage.
I'd seriously look at JaJuan Johnson, Klay Thompson, & Jon Leuer as possible late 1st into 2nd round "steals".
If I was a team in need of truish PG types (statwise) - Jordan Taylor (#13 in my ratings), Darius Morris (#17), Brad Wanamaker (#19), and Demetri McCamey (#29) would be the guys I'd maybe consider if they went in this draft. They all maybe would be 2nd rounders (or free agents) if they were in this draft.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 307
PostPosted: Wed Feb 02, 2011 11:24 am Post subject: Reply with quote
So, a guy like Fredette really worries me, and I say this as a huge fan of Steph Curry. Curry is a liability right now on defense, and I'm not sure how much he will ever improve. I haven't actually watched "the jimmer" (why did his parents...), so can someone fill me in on this side of the ball?
Do your ratings take into account defense at all?
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
bchaikin
Joined: 27 Jan 2005
Posts: 690
Location: cleveland, ohio
PostPosted: Wed Feb 02, 2011 11:26 am Post subject: Reply with quote
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season.
if you had stephen curry ranked that high, what about other players that went to similarly lower tiered schools like david holston of chicago state or lester hudson of tennessee martin? how high did they rank compared to curry?...
He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe.
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Crow
Joined: 20 Jan 2009
Posts: 825
PostPosted: Wed Feb 02, 2011 1:36 pm Post subject: Reply with quote
Thanks for the replies.
College basketball rankings have a place on their own. College basketball rankings and draft rankings can be different or very different and often are. Stats from play against top 64 or 32 teams is one of the better strategies for improving draft rankings from boxscore stats in my opinion. Draft rankings will also be affected by the physical assessment and the "pro-level" skills assessments including defense. I see you did college stat performance ratings by skill in 2008.
Raw on/off +/- would be of limited value generally. There are few cases of NBA prospects where you'd see or accept less than good or would get wildly excited about a great number.
Adjusted +/- just for play against top 64 or 32 teams (and probably for 2 seasons where available) would seem worthwhile to do to me along with that boxscore data cut, even though the sample size will be small, if one was analyzing and actually advising the draft selections. I know of no one doing it publicly. I would guess that 0-10% of teams do that. I wouldn't weight it too heavily but I'd look for the strong and the weak and factor that information into overall assessments to some degree. It can be done from a public source for Euroleague players now.
I never saw Aleks Maric play but I wondered about him some at his draft opportunity and I still wonder if he might eventually get to the NBA. Does any NBA team scout him now? Is he clearly not then, not now a NBA pro prospect? I assume / trust that the pro scout get it right a high percentage of the time but wonder occasionally.
I assume they got it right on R Hendrix. I thought he might be an off the bench NBA role player.
Bo McCalebb is having a good individual boxscore season in the Euroleague.
None of Maric, Hendrix or McCalebb look special on Euroleague Adjusted +/- with the first two near neutral and McCalebb +1 (his +1 was on defense and that seems worth noting).
Leagues are different, some guys can play their true position / right role in one league but not another, often because of the different level of size and athleticism, or at least not as many minutes and maybe not as good pay.
Roman Sato looks good in europe. Better on Adjusted +/- than individual stats. Same for Ricky Rubio and Llull. Freeland better on individual stats. The number of guys who will come over and do well are very few. But getting one can be a pretty big deal. Would be interesting to compare the relative predictive of individual stats and Adjusted +/- though what kind of role you are thinking about giving them probably should affect how these metric predictions are weighted.
Jamont Gordon's Adjusted +/- is currently one of the weakest tracked over there. Not saying it defines him forever but it is a mark in the file, something to watch and see if it improves. Would be nice to compare to an NCAA Adjusted +/- estimate. On top of what the boxscore and the eyes say.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 6:46 pm Post subject: Reply with quote
bchaikin wrote:
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season.
if you had stephen curry ranked that high, what about other players that went to similarly lower tiered schools like david holston of chicago state or lester hudson of tennessee martin? how high did they rank compared to curry?...
He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe.
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
Oh man, now I HAVE to see if I can find the stats from those past years, and rerun them with my slight changes. Damn my lack of organization (and changing computers!) McCalebb ranked high I remember, I think maybe Hudson too - just not as high as Wood. The others aren't ringing a bell at the moment.
Hopefully I'll get back to you soon on this.....
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
tawtaw
Joined: 25 Jun 2008
Posts: 28
Location: Oregon
PostPosted: Wed Feb 02, 2011 7:25 pm Post subject: Reply with quote
Good stuff Statman. I've been following your posts for years, and appreciate them. I like how you've adjusted for pace, SoS and team. It's a simple and effective way to do it for NCAA stats.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 7:58 pm Post subject: Reply with quote
EvanZ wrote:
So, a guy like Fredette really worries me, and I say this as a huge fan of Steph Curry. Curry is a liability right now on defense, and I'm not sure how much he will ever improve. I haven't actually watched "the jimmer" (why did his parents...), so can someone fill me in on this side of the ball?
Do your ratings take into account defense at all?
Well, maybe more than your more known linear metrics (PER). It obviously includes steals, blocks, personal fouls. It is based on team ratings, better teams play better defense (or at least have better ratios between offensive and defensive efficiencies) - so that's a factor.
My playing time adjustment can really make a difference for high minute guys with seemingly bad production. When I do NBA stats - past Bruce Bowen may be the worst rated player according to PER every season, with my ratings he's at least fairly close to being an average player. Usually low production guys that get nice minutes are good defenders - they do things that don't show up in the boxscore.
Same with the other side of the spectrum - big production, low minute guys will see their overall rating drop a bit, the assumption being there are probably some decent reasons they aren't on the court more.
"Glue" guys in college - high minute, lowish production guys from quality programs - will almost always rate higher than the average D1 guy. They may rate a little lower than a higher production, lower minute guy from the same team - but not nearly as much as PER would.
A good example would be Marcus Ginyard from the '08 UNC team. 28.2 mpg, 6.9 ppg, 4.5 rb, 2.2 ast, 1.1 stl, 1.6 to, 50 TS% - he ended up with a 123 rating. Another player, Reggie Williams of VMI - 35.1 mpg, 27.8 ppg, 9.7 rpg, 3.9 ast, 2.2 st, 3.0 to, 59 TS% - he had a slightly LOWER 123 than Ginyards. That's an extreme case - VMI was super paced, very weak SoS.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 307
PostPosted: Wed Feb 02, 2011 8:23 pm Post subject: Reply with quote
I happen to be a fan of Reggie Williams now that he plays for GSW. He's a very good shooter.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 8:37 pm Post subject: Reply with quote
bchaikin wrote:
if you had stephen curry ranked that high, what about other players that went to similarly lower tiered schools like david holston of chicago state or lester hudson of tennessee martin? how high did they rank compared to curry?...
OK - found the final 2008 ratings. I went ahead and did a basic linear weights per game ranking of all players that played 25% or more of their team minutes (2728 players qualified nationally). Just their stats per game, no other adjustments. I then divided their final rank by my ratings by their linear weights rank - to get a list of guys whose stats look awesome, but my ratings have them rated MUCH lower than their basic stats would suggest. I end the list with a guy you mentioned:
Rnk Name Team Conf M/g P/g R/g A/g S/g B/g T/g TS% RAT LWrnk
351 Reggie Williams VMI BSth 35.1 27.8 9.7 3.9 2.2 0.7 3.0 0.585 123 1
73 Lester Hudson Tennessee-Martin OVC 36.9 25.7 7.8 4.5 2.8 0.7 3.8 0.595 149 2
1332 Chavis Holmes VMI BSth 30.9 18.3 4.3 2.5 2.4 0.4 2.1 0.561 93 50
388 Jonathan Rodriguez Campbell ASun 35.0 20.9 10.1 2.5 1.6 0.6 3.6 0.569 121 20
477 Manny Ubilla Fairleigh Dickinson NEC 38.4 20.8 4.8 4.9 1.5 0.3 3.8 0.593 116 25
230 Ryan Toolson Utah Valley State ind 37.9 23.4 3.2 2.9 0.8 0.1 2.6 0.649 132 14
1160 Sean Baptiste Fairleigh Dickinson NEC 36.8 18.5 6.3 1.4 1.0 0.2 1.7 0.575 97 76
57 Arizona 'AZ' Reid High Point BSth 35.3 23.9 11.0 2.4 1.7 0.5 2.1 0.582 152 4
320 Paul Stoll Texas-Pan American ind 33.1 14.2 2.6 7.2 2.5 0.1 3.5 0.726 125 26
84 Marqus Blakely Vermont AE 34.0 19.0 11.0 2.3 2.0 2.7 2.5 0.555 147 8
1504 Travis Holmes VMI BSth 27.5 15.6 5.9 2.8 2.2 0.6 2.7 0.519 90 146
615 Bryan Smithson NC Asheville BSth 36.2 16.4 4.0 4.1 1.9 0.1 2.4 0.584 112 60
226 Ben Woodside North Dakota State Sum 36.6 20.7 2.7 5.1 1.5 0.1 3.1 0.599 132 23
883 Bruce Price Tennessee State OVC 34.0 17.6 4.7 5.0 2.2 0.1 4.0 0.531 104 92
113 David Holston Chicago State ind 33.3 23.1 3.1 5.1 2.1 0.1 3.9 0.603 143 12
I apologize about how messy that looks. Anyway, Hudson ranked a respectable 73rd nationally - although his stats were arguably the 2nd most impressive looking. Holston ranked 113, although his stats looked good enough to be at least a top 15 player nationally. Three of these guys (the Holmes kids from VMI & Baptiste from FDU) actually were rated as slightly below average D1 players despite their solid looking stats.
The other side of the spectrum:
Rnk Name Team Conf M/g P/g R/g A/g S/g B/g T/g TS% RAT LWrnk
6 Roy Hibbert Georgetown BE 26.3 13.4 6.4 1.9 0.5 2.2 1.7 0.620 177 147
24 Brian Butch Wisconsin B10 24.6 12.4 6.6 0.8 0.6 0.8 1.4 0.519 162 563
105 Greg Stiemsma Wisconsin B10 11.5 3.5 3.1 0.7 0.4 1.1 0.7 0.579 144 1890
116 Doneal Mack Memphis CUSA 12.5 6.9 1.6 0.5 0.4 0.2 0.5 0.543 142 1757
33 Darnell Jackson Kansas B12 24.3 11.2 6.7 1.1 0.8 0.5 1.3 0.643 158 393
118 Sasha Kaun Kansas B12 17.7 7.1 3.9 0.3 0.4 1.2 0.8 0.599 142 1288
11 Blake Griffin Oklahoma B12 28.4 14.7 9.1 1.8 1.0 0.8 2.3 0.573 170 112
49 Josh Duncan Xavier-Ohio A10 22.4 12.4 4.7 1.3 0.4 0.6 1.6 0.646 154 492
8 Chris Douglas-Roberts Memphis CUSA 28.7 18.1 4.1 1.8 1.2 0.5 2.1 0.611 173 75
143 Derrick Caracter Louisville BE 16.9 8.3 4.5 0.5 0.6 0.9 1.8 0.571 139 1328
41 Danny Green North Carolina ACC 22.3 11.5 4.9 2.0 1.2 1.2 1.9 0.584 156 376
200 Willie Kemp Memphis CUSA 13.8 5.0 1.0 1.5 0.6 0.1 0.5 0.536 134 1775
43 Matt Howard Butler Horz 24.5 12.3 5.5 0.9 0.7 1.1 1.4 0.635 155 374
36 Eric Coleman Northern Iowa MVC 23.8 12.3 8.8 2.0 0.6 1.3 2.2 0.579 157 306
25 Robbie Hummel Purdue B10 28.5 11.4 6.1 2.5 1.3 0.7 1.4 0.624 161 205
9 Mario Chalmers Kansas B12 30.0 12.8 3.1 4.3 2.5 0.6 1.9 0.650 171 72
64 Darrell Arthur Kansas B12 24.7 12.8 6.3 0.8 0.5 1.3 1.9 0.565 151 503
51 David Padgett Louisville BE 23.5 11.2 4.8 2.0 0.4 1.1 1.5 0.665 153 395
271 Durrell Summers Michigan State B10 10.9 4.9 2.4 0.6 0.1 0.2 0.9 0.585 129 2071
2 Kevin Love UCLA P10 29.6 17.5 10.6 1.9 0.7 1.4 2.0 0.636 214 15
95 Robin Lopez Stanford P10 24.5 10.2 5.7 0.6 0.5 2.3 1.8 0.559 145 673
104 Austin Daye Gonzaga WCC 18.5 10.5 4.7 1.0 0.6 1.6 1.7 0.589 144 736
211 Shawn Taggart Memphis CUSA 17.1 5.9 4.2 0.4 0.5 0.9 0.6 0.546 133 1447
115 Aron Baynes Washington State P10 24.0 10.4 6.0 0.3 0.7 0.7 1.5 0.614 142 780
215 Juan Palacios Louisville BE 17.7 6.0 3.8 1.2 0.8 0.2 1.0 0.533 133 1443
5 Aleks Maric Nebraska B12 29.3 15.7 10.2 1.9 1.3 1.7 2.4 0.596 182 33
17 Derrick Rose Memphis CUSA 29.2 14.9 4.5 4.7 1.2 0.4 2.7 0.554 166 104
I posted more guys here because I wanted to get to the #1 pick. Not all major schools, but most. Pretty much all on good teams, fighting for PT and stats with other very good players. Some were backups who probably would have been stars at small schools. Many were eventually drafted or saw the NBA, despite their non overwhelming college stats.
I remember people kinda lambasting me about Derrick Rose only being ranked 17th when I posted this back then. Seriously, what other metric would have had a guy who was short of 15/5/5/56% from a mid major conference in the top 50, let alone 17th? Heck, he was 3rd on his own team in win shares (according to cbb-reference).
Anyway - I hope this gives a better idea on how well the ratings seem to adjust for the small school, big stat guys. Also, how the ratings may tab overlooked guys also (Aleks Maric? Who knew?).
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Last edited by Statman on Wed Feb 02, 2011 9:26 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 8:55 pm Post subject: Reply with quote
tawtaw wrote:
Good stuff Statman. I've been following your posts for years, and appreciate them. I like how you've adjusted for pace, SoS and team. It's a simple and effective way to do it for NCAA stats.
Thank you so much for the kind words. I really have faith in my work, which is why I'm willing to share it with the best stat minds out there, even though it's possible some may be skeptical or critical.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
page 2
Author Message
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Wed Feb 02, 2011 9:28 pm Post subject: Reply with quote
You might consider pulling together a big polished article with stats and a good balance of text in early March and then shop it to various places pre-NCAA tournament. Maybe you chop it into several articles if they prefer. Then maybe try another one pre-draft. If you get a couple articles printed in visible places maybe you can take it up in some fashion next season.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 9:35 pm Post subject: Reply with quote
bchaikin wrote:
He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe.
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
Oh, I saw McCalebb did rank a quite high 21st in 2008, with a 164 rating. His rank by basic linear weights was 10th. I'm not sure where my 2007 stats are - maybe an old computer or hard drive. Anyway - I'm gonna post the top 2008 guys in my PGU thread here in the next 20 minutes or so, since I did find them.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 9:41 pm Post subject: Reply with quote
Crow wrote:
You might consider pulling together a big polished article with stats and a good balance of text in early March and then shop it to various places pre-NCAA tournament. Maybe you chop it into several articles if they prefer. Then maybe try another one pre-draft. If you get a couple articles printed in visible places maybe you can take it up in some fashion next season.
That's the very thing I was thinking also. I'm not sure how to really shop articles around really - I've tried in the past and almost always get no responses. I'm guessing the articles may not even be looked at - it's hard to tell. If anyone knows any specific contacts from known sites that you think may be interested in this type of work - please let me know.
I'm guessing Draft Express may be interested in articles like this from time to time - I think I may contact Jonathan G. again at some point before the tourney.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
bchaikin
Joined: 27 Jan 2005
Posts: 684
Location: cleveland, ohio
PostPosted: Thu Feb 03, 2011 12:10 am Post subject: Reply with quote
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season... Hudson ranked a respectable 73rd nationally - although his stats were arguably the 2nd most impressive looking. Holston ranked 113, although his stats looked good enough to be at least a top 15 player nationally.
here's my question then - curry, hudson, and holston each played their last 2 years in school in 07-08 and 08-09. looking at just the lower conferences, these 3 players these 2 years had the highest scoring rates per 40 minutes (curry 32.6 pts/40min, hudson 29.1 pts/40min, and holston 28.7 pts/40min). nobody else from the lower conferences that played in both seasons was really close...
i really don't see anything specific that separates one of these three players from the other statistically. overall curry shot the best but not by much, hudson was the much better rebounder, holston the much better passer. all had high steal rates, hudson was by far the best shot blocker. none committed fouls at a high or concerning rate...
so my question is in your rating system what specifically separates curry so much from these other two players (a top 5 ranking vs 73th and 113th)? was it something based on their stats? their defense outside of steals, blocks, and defensive rebounding? or perhaps something about the team they played on, like game pace, or team defensive ranking, or level of competition played against?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Thu Feb 03, 2011 1:39 am Post subject: Reply with quote
bchaikin wrote:
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season... Hudson ranked a respectable 73rd nationally - although his stats were arguably the 2nd most impressive looking. Holston ranked 113, although his stats looked good enough to be at least a top 15 player nationally.
here's my question then - curry, hudson, and holston each played their last 2 years in school in 07-08 and 08-09. looking at just the lower conferences, these 3 players these 2 years had the highest scoring rates per 40 minutes (curry 32.6 pts/40min, hudson 29.1 pts/40min, and holston 28.7 pts/40min). nobody else from the lower conferences that played in both seasons was really close...
i really don't see anything specific that separates one of these three players from the other statistically. overall curry shot the best but not by much, hudson was the much better rebounder, holston the much better passer. all had high steal rates, hudson was by far the best shot blocker. none committed fouls at a high or concerning rate...
so my question is in your rating system what specifically separates curry so much from these other two players (a top 5 ranking vs 73th and 113th)? was it something based on their stats? their defense outside of steals, blocks, and defensive rebounding? or perhaps something about the team they played on, like game pace, or team defensive ranking, or level of competition played against?...
Curry vs. Hudson vs. Holston 2008
You have it right really - it's all pretty much a function of team quality in relation to SoS. Davidson was the 12th ranked team by Sagarin.
Tennessee Martin (Hudson) was the 245th ranked team by Sagarin.
Chicago St. (Holston) was the 259th ranked team.
Davidson was 29-7 against the best schedule of the three (a bit better than Chicago St. - MUCH, MUCH better than Tenn-Martin).
Tenn Martin was 17-16 against a very weak schedule.
Chicago St. was 11-17 against an ok schedule (relatively).
Davidson also had a slower pace than Tenn Martin & Chicago State - repressing Curry's stats a little more.
Anyway - Hudson ahead of Holston mainly because a slower pace, and also more PT (91% team minutes to 83%)
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Thu Feb 03, 2011 9:28 am Post subject: Reply with quote
Your system involves a team strength based adjustment to the player ratings . It is at the season level.
Adjusted +/- adjusts player ratings based on play by play game data of players in lineups.
The two are thus very different.
But one could view your system as a hybrid or distant cousin- it is a individual statistic based metric with a team and season level adjustment.
Until Adjusted +/- is available in some form for top NCAA players, there is some room for products such as your product or the product from the fellow who has previously looked at college player performance against the top 65 teams (he almost certainly relied on computer coding to compile the game boxscores).
I don't know exactly how you compile the initial player rating and how fully defense is represented, but I'd hope that your system team adjusts players fairly for both offensive and defensive contributions, perhaps filling the gap on missing shot defense in some fashion so that offense or the counted in the boxscore parts of defense are not over-weighted.
One limitation of your system is that it assumes the outside the boxscore impacts of players are consistent with their boxscore contributions. In the NBA that is often not the case, especially with shot defense impact.
In addition to college basketball ratings based on your method or stats from play against top teams or Adjusted +/-, it might be quite interesting to see EZPM for college basketball.
Last edited by Crow on Thu Feb 03, 2011 12:17 pm; edited 1 time in total
Back to top
View user's profile Send private message
EvanZ
Joined: 22 Nov 2010
Posts: 264
PostPosted: Thu Feb 03, 2011 9:32 am Post subject: Reply with quote
Crow wrote:
In addition to college basketball ratings based on your method or stats from play against top teams or Adjusted +/-, it might be quite interesting to see EZPM for college basketball.
On my long, long-term todo list. I started looking into scraping ESPN for college PBP. It looks feasible.
This would be one of those "grand challenges" I referred to in the other thread. Laughing
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Thu Feb 03, 2011 12:14 pm Post subject: Reply with quote
Good to hear it is on the list, even though it is a long one.
How much of a college team strength based adjustment should go to the players and how much to the coach? It is will vary and be hard to settle. You could perhaps track player transition to the pros by college / coach, compare college and pro performances and try to apply analysis of the similarities and differences back onto the college rating system in a targeted fashion but there will be small samples and variation which will make it guesswork.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Thu Feb 03, 2011 6:39 pm Post subject: Reply with quote
Crow wrote:
I don't know exactly how you compile the initial player rating and how fully defense is represented, but I'd hope that your system team adjusts players fairly for both offensive and defensive contributions, perhaps filling the gap on missing shot defense in some fashion so that offense or the counted in the boxscore parts of defense are not over-weighted.
I can tell you, my playing time adjustment was something I came up with because of concern that defense just isn't represented well in linear weights metrics. Linear weights metrics (PER the famous one) are great, because pretty much anyone can do them if they tried, and the general public can see the correlation between the boxscore stats and their rating. However, we ALL know that obviously isn't the FULL picture of the player and his true impact to the team - it's just our best attempt at trying to get there.
Anyway - the playing time adjustment I make I believe does bridge the gap (albeit, probably just a small amount) a bit toward adjusted +/- - without having access to A+/- data and the programming power to run the regressions at such a large level. I mean, what are the chances a coach plays a lowish production guy, on a good team, big minutes if he ISN'T a good defender (considering there are 4 & 5 star kids who can fill it up sitting for him)? I mean - those "glue" guys HAVE to be defenders to see the court - because they sure as hell aren't scoring the points, getting all the boards, etc. Same with the high production - low minute guys. It's hard to believe a coach is gonna bench a kid all the time who statistically is impressive if he also plays good defense. That kid is probably benched because he gets lost on D, maybe fouls at bad times, maybe doesn't give full effort, or any other myriad of issues.
Quote:
One limitation of your system is that it assumes the outside the boxscore impacts of players are consistent with their boxscore contributions. In the NBA that is often not the case, especially with shot defense impact
I'm not sure where you get this exactly. Yes, defensive stats are in the weight (Pts, OR, DR, Ast, Stl, Blk, TO, FGmiss, FTmiss, and PF - in essence, all the known box score stats). However, that in no ways means that's the only attempt at representing none stat producing contributions.
But, yes, I do use % of team minutes as an adjustment of overall rating. That impact, however, is very different for every player - even players on the same team playing the same minutes. Say, a HIGH producing kid (say a 200 rating per minute before PT adjustment) playing 20 minutes a game will have a VERY different adjustment to his rating than a low production kid (say a 80 rating per minute) playing 20 minutes a game on that same team. That first kid will see his rating drop drastically (ALOT if the overall team is BAD - the assumption being his production outside the boxscore stats must be horrible not to be seeing more pt. Not nearly as much if the team is great - the assumption being that some of the reason he sits is that there are other very good players needing court time). The 2nd kid will see his rating maybe come up if his production is worse than team average (it'll come up quite a bit if he plays on a great team). His rating would actually come down a little if he plays on a team bad enough to have an overall rating lower than 80.
Anyway, the ONLY way the final adjustment would be the same for two players is if they play the EXACT percentage of team minutes for teams with the EXACT same rating (or the same team obviously). Those two players would then end up with the same per minute rating and final rating. Of course - that doesn't happen - there's always at least slight differences in PT or rating - almost always both.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Last edited by Statman on Thu Feb 03, 2011 6:55 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Thu Feb 03, 2011 6:44 pm Post subject: Reply with quote
EvanZ wrote:
Crow wrote:
In addition to college basketball ratings based on your method or stats from play against top teams or Adjusted +/-, it might be quite interesting to see EZPM for college basketball.
On my long, long-term todo list. I started looking into scraping ESPN for college PBP. It looks feasible.
This would be one of those "grand challenges" I referred to in the other thread. Laughing
I was so happy when I saw Statsheet had +/- data last season. Not so happy when I realized all the games weren't necessarily tallied, or tallied correctly.
I completely support you in your quest. Moral support of course... Wink
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Sat Feb 05, 2011 4:24 pm Post subject: Reply with quote
bchaikin wrote:
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
OK - I just recrunched the 2007 season (I now use Sagarin and a few minor tweaks than I did back then). I posted it in in my current ratings thread (see my sig). Anyway, Durant ended up #1, but not by that much. As for Dashaun Wood, he ranked #11 with a 174 rating.
As for the 3 other guys you mention - they amazingly were tightly bunched. Schilb was 71st with a 148, Wilson 78th with a 146, McCalebb 80th with a 145. McCalebb was definitely on the worst team of the group, and shot the worst, hurting his final rating despite more eye popping stats. Like I said earlier - he ended up ranked #21 overall in 2008 - so he appeared to improve.
Wright ranked high because his team was very slow paced, so he accounted for a bit more of his team's production than Schilb or Wilson. So, big minutes and more production on a better team. Production was about the same as McCalebb (around 33% - Durant was around 31% for comparison), but he played more than Bo AND was the leader of a much better team. His big minutes made the PT adjustment not hurt him as much as it did Schilb.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Sat Feb 05, 2011 5:04 pm Post subject: Reply with quote
I don't have advice from experience on shopping bb articles. There are a number of folks here that do, of course, if they were willing to help publicly or privately.
Minutes given by a coach may be a generally good proxy for overall talent and productivity though there are going to be exceptions either way that an advanced method would help get at.
Your team strength adjustment is something of a substitute for performance against just top teams or Adjusted +/-. But in another way it also uses where a player gets his scholarship as a proxy for talent level. Probably a pretty good one in general but again there will be exceptions and guys who change their relative ranking from high school to college or from first college to second college for reasons other than winning or participation in a more prominent program.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 593
Location: Where the wind comes sweeping down the plains
PostPosted: Sat Feb 05, 2011 7:13 pm Post subject: Reply with quote
Statman wrote:
I can tell you, my playing time adjustment was something I came up with because of concern that defense just isn't represented well in linear weights metrics. Linear weights metrics (PER the famous one) are great, because pretty much anyone can do them if they tried, and the general public can see the correlation between the boxscore stats and their rating. However, we ALL know that obviously isn't the FULL picture of the player and his true impact to the team - it's just our best attempt at trying to get there.
Anyway - the playing time adjustment I make I believe does bridge the gap (albeit, probably just a small amount) a bit toward adjusted +/- - without having access to A+/- data and the programming power to run the regressions at such a large level. I mean, what are the chances a coach plays a lowish production guy, on a good team, big minutes if he ISN'T a good defender (considering there are 4 & 5 star kids who can fill it up sitting for him)? I mean - those "glue" guys HAVE to be defenders to see the court - because they sure as hell aren't scoring the points, getting all the boards, etc. Same with the high production - low minute guys. It's hard to believe a coach is gonna bench a kid all the time who statistically is impressive if he also plays good defense. That kid is probably benched because he gets lost on D, maybe fouls at bad times, maybe doesn't give full effort, or any other myriad of issues.
This is absolutely correct. In my overall ASPM, MPG is worth about 3 points/100 possessions (in other words, a really high minutes guy is worth 3 points/100 possessions more than a really low minutes guy, given all other rate stats are the same. However, in the defensive ASPM regression, the MPG coefficient was negligible--most of the effect was in the offensive ASPM. I was rather puzzled by this, though I think the fact is that most "bench sitters" in the NBA are good defensive players--rather like the replacement level for baseball defense being set at league average.
_________________
GodismyJudgeOK.com/DStats
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Tue Feb 08, 2011 11:50 am Post subject: Reply with quote
I just did complete national rankings for this season, and learned that there are a couple stud PGs that I hadn't really been privy to yet - Norris Cole of CLeveland State and Charles Jenkins of Hofstra - who appear to be as good as any PGs in the nation this side of Jimmer and Kemba.
Anyway - I posted the top 99 players through Sunday's games (2-6-11)
http://www.pointguardu.com/f136/statman ... post342235
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Last edited by Statman on Tue Feb 08, 2011 7:32 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 593
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Feb 08, 2011 12:46 pm Post subject: Reply with quote
Have you ever calculated aging curves for your ratings, Statman? That would be interesting, in terms of calculating who would be the best player if they were all the same age (useful for drafting purposes, perhaps).
_________________
GodismyJudgeOK.com/DStats
Author Message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Sun Jan 30, 2011 2:31 am Post subject: College Basketball Player Ratings Reply with quote
After a long hiatus, I've decided to dust off my college player ratings/rankings. I did the top 7 conferences a couple nights ago:
http://www.pointguardu.com/f136/statmans-ratings-56243/
I'd love to find an area to start posting my work regularly. If I could find a semi easy way to get the stats - I could update nationally practically daily. I emailed Justin K - wondering if he'd be interested in having my stuff in his cbb-ref blog, since that specific blog has been pretty much dead.
Anyway - I'd love input from those of you interested in checking it out. Also, if you know where I should post (update) my work to get to the most college basketball fans' eyes. For time being - I'll be updating in the above thread. I may even dust off my pro ratings in that thread.
Thanks,
Dan Dickey
danthestatman@gmail.com
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
Chilltown
Joined: 16 Apr 2010
Posts: 15
Location: Boston
PostPosted: Mon Jan 31, 2011 5:02 pm Post subject: Reply with quote
This looks very interesting. I'd love to have you expand a bit on how you account for team defense and strength of schedule in these rankings. Mind sharing some details?
Back to top
View user's profile Send private message Visit poster's website
Crow
Joined: 20 Jan 2009
Posts: 825
PostPosted: Tue Feb 01, 2011 4:00 pm Post subject: Reply with quote
Where would Irving rank is he qualified on minutes?
These are college basketball ratings. I really only follow the NBA so I see these guys as future draft picks. How do or would you translate to pro potential?
Have you done a cut of the data for just when they play the top 64 or 32 teams? If not, would you consider doing so to isolate performance against on average better talent?
Would you weight NCAA tournament performance into a draft rating and moderately or heavily? Would you use prior season data and progress or just the last season?
What would your lottery pick list look like right now? If you compared to Draftexpress' top prospect list or 2011 mock draft are there any notable ranking differences you want to highlight and say more on?
Have you thought about or pursued writing in some fashion at draftexpress again in or beyond its forum?
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 4:18 am Post subject: Reply with quote
Chilltown wrote:
This looks very interesting. I'd love to have you expand a bit on how you account for team defense and strength of schedule in these rankings. Mind sharing some details?
Actually, I use Sagarin's team ratings, where, in essence, SoS and team quality (game results and differentials) are built in. I compile the individual weights for the team, scale them to the team rating. I then make playing time adjustments to individual ratings (scaling toward team rating depending on PT), then readjust all the ratings to exactly match the team rating again.
It'd be like PER, if there was a pre set team PER (based on SoS, results, and point differentials) that the individual PERs got adjusted to (positive and negative weights scaled appropriately).
The premise is that each player is part of the team whole. If a player is considered completely average for that team (looking at the weights and PT), then his rating would be exactly the same as that team's rating. An average player, statistically, for VMI would have pretty impressive looking stat totals (really high pace, weak SoS), while an average player stat totals for a Wisconsin player would look quite pedestrian. The average player for Wisconsin would have a much better rating than the average player for VMI obviously (much higher Sagarin rating), despite a much weaker stat line.
Well, I think I rambled a bit there - hope it helps. I use the same approach when I do NBA ratings (I just use team point ratios for team quality), which I'll probably do near All Star game time in that same thread. Players playing for multiple teams make it harder for me to compile everything quickly. My NBA rankings will probably look fairly similar to Mike G's, since I think there are a little similarities in approach.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 5:20 am Post subject: Reply with quote
Crow wrote:
Where would Irving rank is he qualified on minutes?
I'm at work at the moment - that spreadsheet is at home. Hopefully I'll remember to look it up and maybe compare his per minute rating (ignoring playing time adjustments) to other guys (natioanl PGs, ACC guys). If I'm not too sleepy I'll try to look it up before my kids wake up this morning.
Crow wrote:
These are college basketball ratings. I really only follow the NBA so I see these guys as future draft picks. How do or would you translate to pro potential?
It's not bad. Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season. Believe it or not, I don't have last season's stats - Cousins would have been real high I guess. Maybe Evan Turner was the highest? I'm forgetting. I thought I did an NCAA tourney player rating thing - I'll see when I get home.
Here's a flashback - I found a thread in which I ranked every player in EVERY conference, and gave my all conference teams, in 2008. I did national rankings too. The ratings here are quite similar to what I do at the moment - some slight changes I've made since then (I think I used Pomeroy's team ratings then, now I use Sagarin's - and playing time adjustments are slightly different):
http://www.pointguardu.com/f136/statman ... ndex2.html
I can divide the ratings into skill sets. I'd LOVE to do ratings over the last, say, 12 years, dividing the ratings into individual skillsets (pace, team quality, SoS, etc would all still be part of each skillset) - and see how these skillsets might best translate to future NBA production. If someone can get me the full past seasons' stats - I'd get right on that Wink
Crow wrote:
Have you done a cut of the data for just when they play the top 64 or 32 teams? If not, would you consider doing so to isolate performance against on average better talent?
I haven't done that, and it would be too difficult for me to attempt to do that with the limitations of me gathering the stats.
Quote:
Would you weight NCAA tournament performance into a draft rating and moderately or heavily? Would you use prior season data and progress or just the last season?
I'm not sure really. One would think maybe tourney performance should be weighted more - I'm not in a position to try to do that. I would look at each season seperately I think. I think comparing player ratings at similar ages could be important (Wall's freshman year rating compared to, say, Evan Turner's freshman year) in projecting possible NBA potential.
Quote:
What would your lottery pick list look like right now? If you compared to Draftexpress' top prospect list or 2011 mock draft are there any notable ranking differences you want to highlight and say more on?
I may look into this when I get home. I'm an Arizona alumn - I love Derrick Williams obviously. I'm biased. I also want him to come back as a junior.
Quote:
Have you thought about or pursued writing in some fashion at draftexpress again in or beyond its forum?
I've never written for Draft Express. Jonathan Givony at Draft Express expressed interest a few years back - contacted me about trying to incorporate my ratings into their database. He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe. Givony was intrigued my ratings had pegged him as a top player nationally. Well, their database guy said they couldn't incorporate the ratings into the database - so the whole thing was dropped.
As for writing, I'd love to actually start - but life kinda gets in the way. Working full time, watching my kids when I don't work, working out, kid's sports, spending quality time with the wife, etc. If I got paid to produce, say, 5 articles a week for someone (in essence, I did it for a living) - I have zero doubt I could produce some very interesting stuff. But, I've never even gotten my work together in the same blog. I've never saved past stuff into my own folders (hundreds of hours of past articles were lost in my old Wildcat Sports Report blog). In essence, I've not been organized well enough to show past quality work - and I've not devoted nearly enough time to keep stuff current to stay fresh in people's minds. I'm hoping to SOMEHOW change that, and squeeze in some cool things in that PGU thread as often as I can - until I may find a different, better, place to produce my work.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 10:43 am Post subject: Reply with quote
Crow wrote:
Where would Irving rank is he qualified on minutes.......
What would your lottery pick list look like right now? If you compared to Draftexpress' top prospect list or 2011 mock draft are there any notable ranking differences you want to highlight and say more on?
I'll touch these couple things, now that I'm home. First - Irving's rating when ignoring the playing time adjustment (pretty much a per minute production rating) is an awesome 206. His overall rating is For a freshman PG, that's pretty phenominal. To compare, his teammate Nolan Smith is ranked 9th overall - and his per minute rating is 183. Singler's is 143. Limited data, but very impressive.
If we set the minutes minimum to, say, 25%, Kyrie would be the 8th ranked player in the ACC with an overall rating of 152. He'd been #1 in the ACC (and 3rd or 4th nationally) obviously if he never got hurt and continued his same production.
For comparison, Josh Selby, acclaimed frosh PG with limited minutes for Kansas (due to eligibility) had a per minute rating of 110.
Among all players that played at least 2% of their team's minutes (Irving played 28.8%) - Kyrie ranked 5th in rating per minute. Derrick Williams was #1 at 228, then Jimmer at 225, Kemba at 221, Sullinger at 207, Kyrie at 206, Jordan Hamilton at 204, then others below 200.
As for Draft express - here's where my top guys, and where DE projects them:
Jimmer - #13
Kemba - #7
Derrick Williams - #5
Sullinger - #4
Jordan Hamilton - #17
JaJuan Johnson - #38 (steal?)
Jon Leuer - #54 (steal?)
Klay Thompson - #35 (steal?)
Nolan Smith - #25
Here's DE top guys declaring - and where they ranked (among all guys above 25% team minutes) in my ratings:
#1 Irving - #46 with a 152 (despite limited minutes)
#2 Perry Jones - #166 with a 127 (very possible bust - at least for years)
#8 Terrence Jones - #15 with a 169 (quite nice for a frosh)
#9 Harrison Barnes - #287 with a 116 (not close to ready)
#11 John Henson - #62 with a 146
#14 Singleton - #38 with a 154
#15 Marcus Morris - #12 with a 170
#17 Jeff Taylor - #88 with a 139
#18 Burks - #85 with a 139
#19 Leonard - #10 with a 172
Just looking at ratings, and a little of my feelings (what I've seen) - here's the college guys I'd consider in the lottery kinda in order:
Kyrie Irving
Derrick Williams
Jared Sullinger (body type worries me slightly - future weight issues?)
Kemba Walker
Terrence Jones
Jimmer Fredette
Jordan Hamilton
Those would probably be the only college guys I'd consider in the lottery. I'd probably avoid lottery probables Perry Jones & Harrison Barnes, for fear of bustage.
I'd seriously look at JaJuan Johnson, Klay Thompson, & Jon Leuer as possible late 1st into 2nd round "steals".
If I was a team in need of truish PG types (statwise) - Jordan Taylor (#13 in my ratings), Darius Morris (#17), Brad Wanamaker (#19), and Demetri McCamey (#29) would be the guys I'd maybe consider if they went in this draft. They all maybe would be 2nd rounders (or free agents) if they were in this draft.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 307
PostPosted: Wed Feb 02, 2011 11:24 am Post subject: Reply with quote
So, a guy like Fredette really worries me, and I say this as a huge fan of Steph Curry. Curry is a liability right now on defense, and I'm not sure how much he will ever improve. I haven't actually watched "the jimmer" (why did his parents...), so can someone fill me in on this side of the ball?
Do your ratings take into account defense at all?
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
bchaikin
Joined: 27 Jan 2005
Posts: 690
Location: cleveland, ohio
PostPosted: Wed Feb 02, 2011 11:26 am Post subject: Reply with quote
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season.
if you had stephen curry ranked that high, what about other players that went to similarly lower tiered schools like david holston of chicago state or lester hudson of tennessee martin? how high did they rank compared to curry?...
He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe.
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Crow
Joined: 20 Jan 2009
Posts: 825
PostPosted: Wed Feb 02, 2011 1:36 pm Post subject: Reply with quote
Thanks for the replies.
College basketball rankings have a place on their own. College basketball rankings and draft rankings can be different or very different and often are. Stats from play against top 64 or 32 teams is one of the better strategies for improving draft rankings from boxscore stats in my opinion. Draft rankings will also be affected by the physical assessment and the "pro-level" skills assessments including defense. I see you did college stat performance ratings by skill in 2008.
Raw on/off +/- would be of limited value generally. There are few cases of NBA prospects where you'd see or accept less than good or would get wildly excited about a great number.
Adjusted +/- just for play against top 64 or 32 teams (and probably for 2 seasons where available) would seem worthwhile to do to me along with that boxscore data cut, even though the sample size will be small, if one was analyzing and actually advising the draft selections. I know of no one doing it publicly. I would guess that 0-10% of teams do that. I wouldn't weight it too heavily but I'd look for the strong and the weak and factor that information into overall assessments to some degree. It can be done from a public source for Euroleague players now.
I never saw Aleks Maric play but I wondered about him some at his draft opportunity and I still wonder if he might eventually get to the NBA. Does any NBA team scout him now? Is he clearly not then, not now a NBA pro prospect? I assume / trust that the pro scout get it right a high percentage of the time but wonder occasionally.
I assume they got it right on R Hendrix. I thought he might be an off the bench NBA role player.
Bo McCalebb is having a good individual boxscore season in the Euroleague.
None of Maric, Hendrix or McCalebb look special on Euroleague Adjusted +/- with the first two near neutral and McCalebb +1 (his +1 was on defense and that seems worth noting).
Leagues are different, some guys can play their true position / right role in one league but not another, often because of the different level of size and athleticism, or at least not as many minutes and maybe not as good pay.
Roman Sato looks good in europe. Better on Adjusted +/- than individual stats. Same for Ricky Rubio and Llull. Freeland better on individual stats. The number of guys who will come over and do well are very few. But getting one can be a pretty big deal. Would be interesting to compare the relative predictive of individual stats and Adjusted +/- though what kind of role you are thinking about giving them probably should affect how these metric predictions are weighted.
Jamont Gordon's Adjusted +/- is currently one of the weakest tracked over there. Not saying it defines him forever but it is a mark in the file, something to watch and see if it improves. Would be nice to compare to an NCAA Adjusted +/- estimate. On top of what the boxscore and the eyes say.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 6:46 pm Post subject: Reply with quote
bchaikin wrote:
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season.
if you had stephen curry ranked that high, what about other players that went to similarly lower tiered schools like david holston of chicago state or lester hudson of tennessee martin? how high did they rank compared to curry?...
He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe.
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
Oh man, now I HAVE to see if I can find the stats from those past years, and rerun them with my slight changes. Damn my lack of organization (and changing computers!) McCalebb ranked high I remember, I think maybe Hudson too - just not as high as Wood. The others aren't ringing a bell at the moment.
Hopefully I'll get back to you soon on this.....
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
tawtaw
Joined: 25 Jun 2008
Posts: 28
Location: Oregon
PostPosted: Wed Feb 02, 2011 7:25 pm Post subject: Reply with quote
Good stuff Statman. I've been following your posts for years, and appreciate them. I like how you've adjusted for pace, SoS and team. It's a simple and effective way to do it for NCAA stats.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 7:58 pm Post subject: Reply with quote
EvanZ wrote:
So, a guy like Fredette really worries me, and I say this as a huge fan of Steph Curry. Curry is a liability right now on defense, and I'm not sure how much he will ever improve. I haven't actually watched "the jimmer" (why did his parents...), so can someone fill me in on this side of the ball?
Do your ratings take into account defense at all?
Well, maybe more than your more known linear metrics (PER). It obviously includes steals, blocks, personal fouls. It is based on team ratings, better teams play better defense (or at least have better ratios between offensive and defensive efficiencies) - so that's a factor.
My playing time adjustment can really make a difference for high minute guys with seemingly bad production. When I do NBA stats - past Bruce Bowen may be the worst rated player according to PER every season, with my ratings he's at least fairly close to being an average player. Usually low production guys that get nice minutes are good defenders - they do things that don't show up in the boxscore.
Same with the other side of the spectrum - big production, low minute guys will see their overall rating drop a bit, the assumption being there are probably some decent reasons they aren't on the court more.
"Glue" guys in college - high minute, lowish production guys from quality programs - will almost always rate higher than the average D1 guy. They may rate a little lower than a higher production, lower minute guy from the same team - but not nearly as much as PER would.
A good example would be Marcus Ginyard from the '08 UNC team. 28.2 mpg, 6.9 ppg, 4.5 rb, 2.2 ast, 1.1 stl, 1.6 to, 50 TS% - he ended up with a 123 rating. Another player, Reggie Williams of VMI - 35.1 mpg, 27.8 ppg, 9.7 rpg, 3.9 ast, 2.2 st, 3.0 to, 59 TS% - he had a slightly LOWER 123 than Ginyards. That's an extreme case - VMI was super paced, very weak SoS.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Back to top
View user's profile Send private message Send e-mail
EvanZ
Joined: 22 Nov 2010
Posts: 307
PostPosted: Wed Feb 02, 2011 8:23 pm Post subject: Reply with quote
I happen to be a fan of Reggie Williams now that he plays for GSW. He's a very good shooter.
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 8:37 pm Post subject: Reply with quote
bchaikin wrote:
if you had stephen curry ranked that high, what about other players that went to similarly lower tiered schools like david holston of chicago state or lester hudson of tennessee martin? how high did they rank compared to curry?...
OK - found the final 2008 ratings. I went ahead and did a basic linear weights per game ranking of all players that played 25% or more of their team minutes (2728 players qualified nationally). Just their stats per game, no other adjustments. I then divided their final rank by my ratings by their linear weights rank - to get a list of guys whose stats look awesome, but my ratings have them rated MUCH lower than their basic stats would suggest. I end the list with a guy you mentioned:
Rnk Name Team Conf M/g P/g R/g A/g S/g B/g T/g TS% RAT LWrnk
351 Reggie Williams VMI BSth 35.1 27.8 9.7 3.9 2.2 0.7 3.0 0.585 123 1
73 Lester Hudson Tennessee-Martin OVC 36.9 25.7 7.8 4.5 2.8 0.7 3.8 0.595 149 2
1332 Chavis Holmes VMI BSth 30.9 18.3 4.3 2.5 2.4 0.4 2.1 0.561 93 50
388 Jonathan Rodriguez Campbell ASun 35.0 20.9 10.1 2.5 1.6 0.6 3.6 0.569 121 20
477 Manny Ubilla Fairleigh Dickinson NEC 38.4 20.8 4.8 4.9 1.5 0.3 3.8 0.593 116 25
230 Ryan Toolson Utah Valley State ind 37.9 23.4 3.2 2.9 0.8 0.1 2.6 0.649 132 14
1160 Sean Baptiste Fairleigh Dickinson NEC 36.8 18.5 6.3 1.4 1.0 0.2 1.7 0.575 97 76
57 Arizona 'AZ' Reid High Point BSth 35.3 23.9 11.0 2.4 1.7 0.5 2.1 0.582 152 4
320 Paul Stoll Texas-Pan American ind 33.1 14.2 2.6 7.2 2.5 0.1 3.5 0.726 125 26
84 Marqus Blakely Vermont AE 34.0 19.0 11.0 2.3 2.0 2.7 2.5 0.555 147 8
1504 Travis Holmes VMI BSth 27.5 15.6 5.9 2.8 2.2 0.6 2.7 0.519 90 146
615 Bryan Smithson NC Asheville BSth 36.2 16.4 4.0 4.1 1.9 0.1 2.4 0.584 112 60
226 Ben Woodside North Dakota State Sum 36.6 20.7 2.7 5.1 1.5 0.1 3.1 0.599 132 23
883 Bruce Price Tennessee State OVC 34.0 17.6 4.7 5.0 2.2 0.1 4.0 0.531 104 92
113 David Holston Chicago State ind 33.3 23.1 3.1 5.1 2.1 0.1 3.9 0.603 143 12
I apologize about how messy that looks. Anyway, Hudson ranked a respectable 73rd nationally - although his stats were arguably the 2nd most impressive looking. Holston ranked 113, although his stats looked good enough to be at least a top 15 player nationally. Three of these guys (the Holmes kids from VMI & Baptiste from FDU) actually were rated as slightly below average D1 players despite their solid looking stats.
The other side of the spectrum:
Rnk Name Team Conf M/g P/g R/g A/g S/g B/g T/g TS% RAT LWrnk
6 Roy Hibbert Georgetown BE 26.3 13.4 6.4 1.9 0.5 2.2 1.7 0.620 177 147
24 Brian Butch Wisconsin B10 24.6 12.4 6.6 0.8 0.6 0.8 1.4 0.519 162 563
105 Greg Stiemsma Wisconsin B10 11.5 3.5 3.1 0.7 0.4 1.1 0.7 0.579 144 1890
116 Doneal Mack Memphis CUSA 12.5 6.9 1.6 0.5 0.4 0.2 0.5 0.543 142 1757
33 Darnell Jackson Kansas B12 24.3 11.2 6.7 1.1 0.8 0.5 1.3 0.643 158 393
118 Sasha Kaun Kansas B12 17.7 7.1 3.9 0.3 0.4 1.2 0.8 0.599 142 1288
11 Blake Griffin Oklahoma B12 28.4 14.7 9.1 1.8 1.0 0.8 2.3 0.573 170 112
49 Josh Duncan Xavier-Ohio A10 22.4 12.4 4.7 1.3 0.4 0.6 1.6 0.646 154 492
8 Chris Douglas-Roberts Memphis CUSA 28.7 18.1 4.1 1.8 1.2 0.5 2.1 0.611 173 75
143 Derrick Caracter Louisville BE 16.9 8.3 4.5 0.5 0.6 0.9 1.8 0.571 139 1328
41 Danny Green North Carolina ACC 22.3 11.5 4.9 2.0 1.2 1.2 1.9 0.584 156 376
200 Willie Kemp Memphis CUSA 13.8 5.0 1.0 1.5 0.6 0.1 0.5 0.536 134 1775
43 Matt Howard Butler Horz 24.5 12.3 5.5 0.9 0.7 1.1 1.4 0.635 155 374
36 Eric Coleman Northern Iowa MVC 23.8 12.3 8.8 2.0 0.6 1.3 2.2 0.579 157 306
25 Robbie Hummel Purdue B10 28.5 11.4 6.1 2.5 1.3 0.7 1.4 0.624 161 205
9 Mario Chalmers Kansas B12 30.0 12.8 3.1 4.3 2.5 0.6 1.9 0.650 171 72
64 Darrell Arthur Kansas B12 24.7 12.8 6.3 0.8 0.5 1.3 1.9 0.565 151 503
51 David Padgett Louisville BE 23.5 11.2 4.8 2.0 0.4 1.1 1.5 0.665 153 395
271 Durrell Summers Michigan State B10 10.9 4.9 2.4 0.6 0.1 0.2 0.9 0.585 129 2071
2 Kevin Love UCLA P10 29.6 17.5 10.6 1.9 0.7 1.4 2.0 0.636 214 15
95 Robin Lopez Stanford P10 24.5 10.2 5.7 0.6 0.5 2.3 1.8 0.559 145 673
104 Austin Daye Gonzaga WCC 18.5 10.5 4.7 1.0 0.6 1.6 1.7 0.589 144 736
211 Shawn Taggart Memphis CUSA 17.1 5.9 4.2 0.4 0.5 0.9 0.6 0.546 133 1447
115 Aron Baynes Washington State P10 24.0 10.4 6.0 0.3 0.7 0.7 1.5 0.614 142 780
215 Juan Palacios Louisville BE 17.7 6.0 3.8 1.2 0.8 0.2 1.0 0.533 133 1443
5 Aleks Maric Nebraska B12 29.3 15.7 10.2 1.9 1.3 1.7 2.4 0.596 182 33
17 Derrick Rose Memphis CUSA 29.2 14.9 4.5 4.7 1.2 0.4 2.7 0.554 166 104
I posted more guys here because I wanted to get to the #1 pick. Not all major schools, but most. Pretty much all on good teams, fighting for PT and stats with other very good players. Some were backups who probably would have been stars at small schools. Many were eventually drafted or saw the NBA, despite their non overwhelming college stats.
I remember people kinda lambasting me about Derrick Rose only being ranked 17th when I posted this back then. Seriously, what other metric would have had a guy who was short of 15/5/5/56% from a mid major conference in the top 50, let alone 17th? Heck, he was 3rd on his own team in win shares (according to cbb-reference).
Anyway - I hope this gives a better idea on how well the ratings seem to adjust for the small school, big stat guys. Also, how the ratings may tab overlooked guys also (Aleks Maric? Who knew?).
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
Last edited by Statman on Wed Feb 02, 2011 9:26 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 242
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 8:55 pm Post subject: Reply with quote
tawtaw wrote:
Good stuff Statman. I've been following your posts for years, and appreciate them. I like how you've adjusted for pace, SoS and team. It's a simple and effective way to do it for NCAA stats.
Thank you so much for the kind words. I really have faith in my work, which is why I'm willing to share it with the best stat minds out there, even though it's possible some may be skeptical or critical.
_________________
Dan
My current national college player rankings (and other stuff):
http://www.pointguardu.com/f136/statman ... post355594
page 2
Author Message
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Wed Feb 02, 2011 9:28 pm Post subject: Reply with quote
You might consider pulling together a big polished article with stats and a good balance of text in early March and then shop it to various places pre-NCAA tournament. Maybe you chop it into several articles if they prefer. Then maybe try another one pre-draft. If you get a couple articles printed in visible places maybe you can take it up in some fashion next season.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 9:35 pm Post subject: Reply with quote
bchaikin wrote:
He was very intrigued by the fact my ratings had Dashaun Wood (out of Wright State) rated real highly nationally in 2007. People knew very little about the kid - but he blew up somewhere in Europe.
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
Oh, I saw McCalebb did rank a quite high 21st in 2008, with a 164 rating. His rank by basic linear weights was 10th. I'm not sure where my 2007 stats are - maybe an old computer or hard drive. Anyway - I'm gonna post the top 2008 guys in my PGU thread here in the next 20 minutes or so, since I did find them.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Wed Feb 02, 2011 9:41 pm Post subject: Reply with quote
Crow wrote:
You might consider pulling together a big polished article with stats and a good balance of text in early March and then shop it to various places pre-NCAA tournament. Maybe you chop it into several articles if they prefer. Then maybe try another one pre-draft. If you get a couple articles printed in visible places maybe you can take it up in some fashion next season.
That's the very thing I was thinking also. I'm not sure how to really shop articles around really - I've tried in the past and almost always get no responses. I'm guessing the articles may not even be looked at - it's hard to tell. If anyone knows any specific contacts from known sites that you think may be interested in this type of work - please let me know.
I'm guessing Draft Express may be interested in articles like this from time to time - I think I may contact Jonathan G. again at some point before the tourney.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
bchaikin
Joined: 27 Jan 2005
Posts: 684
Location: cleveland, ohio
PostPosted: Thu Feb 03, 2011 12:10 am Post subject: Reply with quote
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season... Hudson ranked a respectable 73rd nationally - although his stats were arguably the 2nd most impressive looking. Holston ranked 113, although his stats looked good enough to be at least a top 15 player nationally.
here's my question then - curry, hudson, and holston each played their last 2 years in school in 07-08 and 08-09. looking at just the lower conferences, these 3 players these 2 years had the highest scoring rates per 40 minutes (curry 32.6 pts/40min, hudson 29.1 pts/40min, and holston 28.7 pts/40min). nobody else from the lower conferences that played in both seasons was really close...
i really don't see anything specific that separates one of these three players from the other statistically. overall curry shot the best but not by much, hudson was the much better rebounder, holston the much better passer. all had high steal rates, hudson was by far the best shot blocker. none committed fouls at a high or concerning rate...
so my question is in your rating system what specifically separates curry so much from these other two players (a top 5 ranking vs 73th and 113th)? was it something based on their stats? their defense outside of steals, blocks, and defensive rebounding? or perhaps something about the team they played on, like game pace, or team defensive ranking, or level of competition played against?...
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Thu Feb 03, 2011 1:39 am Post subject: Reply with quote
bchaikin wrote:
Durant was easily the highest rated player 4 years ago. Beasley was the next season - Love was a close 2nd the that season, followed by Curry and Hansbrough. I think either Griffin or Curry were tops the next season... Hudson ranked a respectable 73rd nationally - although his stats were arguably the 2nd most impressive looking. Holston ranked 113, although his stats looked good enough to be at least a top 15 player nationally.
here's my question then - curry, hudson, and holston each played their last 2 years in school in 07-08 and 08-09. looking at just the lower conferences, these 3 players these 2 years had the highest scoring rates per 40 minutes (curry 32.6 pts/40min, hudson 29.1 pts/40min, and holston 28.7 pts/40min). nobody else from the lower conferences that played in both seasons was really close...
i really don't see anything specific that separates one of these three players from the other statistically. overall curry shot the best but not by much, hudson was the much better rebounder, holston the much better passer. all had high steal rates, hudson was by far the best shot blocker. none committed fouls at a high or concerning rate...
so my question is in your rating system what specifically separates curry so much from these other two players (a top 5 ranking vs 73th and 113th)? was it something based on their stats? their defense outside of steals, blocks, and defensive rebounding? or perhaps something about the team they played on, like game pace, or team defensive ranking, or level of competition played against?...
Curry vs. Hudson vs. Holston 2008
You have it right really - it's all pretty much a function of team quality in relation to SoS. Davidson was the 12th ranked team by Sagarin.
Tennessee Martin (Hudson) was the 245th ranked team by Sagarin.
Chicago St. (Holston) was the 259th ranked team.
Davidson was 29-7 against the best schedule of the three (a bit better than Chicago St. - MUCH, MUCH better than Tenn-Martin).
Tenn Martin was 17-16 against a very weak schedule.
Chicago St. was 11-17 against an ok schedule (relatively).
Davidson also had a slower pace than Tenn Martin & Chicago State - repressing Curry's stats a little more.
Anyway - Hudson ahead of Holston mainly because a slower pace, and also more PT (91% team minutes to 83%)
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Thu Feb 03, 2011 9:28 am Post subject: Reply with quote
Your system involves a team strength based adjustment to the player ratings . It is at the season level.
Adjusted +/- adjusts player ratings based on play by play game data of players in lineups.
The two are thus very different.
But one could view your system as a hybrid or distant cousin- it is a individual statistic based metric with a team and season level adjustment.
Until Adjusted +/- is available in some form for top NCAA players, there is some room for products such as your product or the product from the fellow who has previously looked at college player performance against the top 65 teams (he almost certainly relied on computer coding to compile the game boxscores).
I don't know exactly how you compile the initial player rating and how fully defense is represented, but I'd hope that your system team adjusts players fairly for both offensive and defensive contributions, perhaps filling the gap on missing shot defense in some fashion so that offense or the counted in the boxscore parts of defense are not over-weighted.
One limitation of your system is that it assumes the outside the boxscore impacts of players are consistent with their boxscore contributions. In the NBA that is often not the case, especially with shot defense impact.
In addition to college basketball ratings based on your method or stats from play against top teams or Adjusted +/-, it might be quite interesting to see EZPM for college basketball.
Last edited by Crow on Thu Feb 03, 2011 12:17 pm; edited 1 time in total
Back to top
View user's profile Send private message
EvanZ
Joined: 22 Nov 2010
Posts: 264
PostPosted: Thu Feb 03, 2011 9:32 am Post subject: Reply with quote
Crow wrote:
In addition to college basketball ratings based on your method or stats from play against top teams or Adjusted +/-, it might be quite interesting to see EZPM for college basketball.
On my long, long-term todo list. I started looking into scraping ESPN for college PBP. It looks feasible.
This would be one of those "grand challenges" I referred to in the other thread. Laughing
_________________
http://www.thecity2.com
http://www.ibb.gatech.edu/evan-zamir
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Thu Feb 03, 2011 12:14 pm Post subject: Reply with quote
Good to hear it is on the list, even though it is a long one.
How much of a college team strength based adjustment should go to the players and how much to the coach? It is will vary and be hard to settle. You could perhaps track player transition to the pros by college / coach, compare college and pro performances and try to apply analysis of the similarities and differences back onto the college rating system in a targeted fashion but there will be small samples and variation which will make it guesswork.
Back to top
View user's profile Send private message
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Thu Feb 03, 2011 6:39 pm Post subject: Reply with quote
Crow wrote:
I don't know exactly how you compile the initial player rating and how fully defense is represented, but I'd hope that your system team adjusts players fairly for both offensive and defensive contributions, perhaps filling the gap on missing shot defense in some fashion so that offense or the counted in the boxscore parts of defense are not over-weighted.
I can tell you, my playing time adjustment was something I came up with because of concern that defense just isn't represented well in linear weights metrics. Linear weights metrics (PER the famous one) are great, because pretty much anyone can do them if they tried, and the general public can see the correlation between the boxscore stats and their rating. However, we ALL know that obviously isn't the FULL picture of the player and his true impact to the team - it's just our best attempt at trying to get there.
Anyway - the playing time adjustment I make I believe does bridge the gap (albeit, probably just a small amount) a bit toward adjusted +/- - without having access to A+/- data and the programming power to run the regressions at such a large level. I mean, what are the chances a coach plays a lowish production guy, on a good team, big minutes if he ISN'T a good defender (considering there are 4 & 5 star kids who can fill it up sitting for him)? I mean - those "glue" guys HAVE to be defenders to see the court - because they sure as hell aren't scoring the points, getting all the boards, etc. Same with the high production - low minute guys. It's hard to believe a coach is gonna bench a kid all the time who statistically is impressive if he also plays good defense. That kid is probably benched because he gets lost on D, maybe fouls at bad times, maybe doesn't give full effort, or any other myriad of issues.
Quote:
One limitation of your system is that it assumes the outside the boxscore impacts of players are consistent with their boxscore contributions. In the NBA that is often not the case, especially with shot defense impact
I'm not sure where you get this exactly. Yes, defensive stats are in the weight (Pts, OR, DR, Ast, Stl, Blk, TO, FGmiss, FTmiss, and PF - in essence, all the known box score stats). However, that in no ways means that's the only attempt at representing none stat producing contributions.
But, yes, I do use % of team minutes as an adjustment of overall rating. That impact, however, is very different for every player - even players on the same team playing the same minutes. Say, a HIGH producing kid (say a 200 rating per minute before PT adjustment) playing 20 minutes a game will have a VERY different adjustment to his rating than a low production kid (say a 80 rating per minute) playing 20 minutes a game on that same team. That first kid will see his rating drop drastically (ALOT if the overall team is BAD - the assumption being his production outside the boxscore stats must be horrible not to be seeing more pt. Not nearly as much if the team is great - the assumption being that some of the reason he sits is that there are other very good players needing court time). The 2nd kid will see his rating maybe come up if his production is worse than team average (it'll come up quite a bit if he plays on a great team). His rating would actually come down a little if he plays on a team bad enough to have an overall rating lower than 80.
Anyway, the ONLY way the final adjustment would be the same for two players is if they play the EXACT percentage of team minutes for teams with the EXACT same rating (or the same team obviously). Those two players would then end up with the same per minute rating and final rating. Of course - that doesn't happen - there's always at least slight differences in PT or rating - almost always both.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Last edited by Statman on Thu Feb 03, 2011 6:55 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Thu Feb 03, 2011 6:44 pm Post subject: Reply with quote
EvanZ wrote:
Crow wrote:
In addition to college basketball ratings based on your method or stats from play against top teams or Adjusted +/-, it might be quite interesting to see EZPM for college basketball.
On my long, long-term todo list. I started looking into scraping ESPN for college PBP. It looks feasible.
This would be one of those "grand challenges" I referred to in the other thread. Laughing
I was so happy when I saw Statsheet had +/- data last season. Not so happy when I realized all the games weren't necessarily tallied, or tallied correctly.
I completely support you in your quest. Moral support of course... Wink
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Sat Feb 05, 2011 4:24 pm Post subject: Reply with quote
bchaikin wrote:
again, if you had dashaun wood rated highly, was it higher than say players like blake schilb of loyola, jamar wilson of albany, or bo mccalebb of new orleans?...
OK - I just recrunched the 2007 season (I now use Sagarin and a few minor tweaks than I did back then). I posted it in in my current ratings thread (see my sig). Anyway, Durant ended up #1, but not by that much. As for Dashaun Wood, he ranked #11 with a 174 rating.
As for the 3 other guys you mention - they amazingly were tightly bunched. Schilb was 71st with a 148, Wilson 78th with a 146, McCalebb 80th with a 145. McCalebb was definitely on the worst team of the group, and shot the worst, hurting his final rating despite more eye popping stats. Like I said earlier - he ended up ranked #21 overall in 2008 - so he appeared to improve.
Wright ranked high because his team was very slow paced, so he accounted for a bit more of his team's production than Schilb or Wilson. So, big minutes and more production on a better team. Production was about the same as McCalebb (around 33% - Durant was around 31% for comparison), but he played more than Bo AND was the leader of a much better team. His big minutes made the PT adjustment not hurt him as much as it did Schilb.
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Back to top
View user's profile Send private message Send e-mail
Crow
Joined: 20 Jan 2009
Posts: 804
PostPosted: Sat Feb 05, 2011 5:04 pm Post subject: Reply with quote
I don't have advice from experience on shopping bb articles. There are a number of folks here that do, of course, if they were willing to help publicly or privately.
Minutes given by a coach may be a generally good proxy for overall talent and productivity though there are going to be exceptions either way that an advanced method would help get at.
Your team strength adjustment is something of a substitute for performance against just top teams or Adjusted +/-. But in another way it also uses where a player gets his scholarship as a proxy for talent level. Probably a pretty good one in general but again there will be exceptions and guys who change their relative ranking from high school to college or from first college to second college for reasons other than winning or participation in a more prominent program.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 593
Location: Where the wind comes sweeping down the plains
PostPosted: Sat Feb 05, 2011 7:13 pm Post subject: Reply with quote
Statman wrote:
I can tell you, my playing time adjustment was something I came up with because of concern that defense just isn't represented well in linear weights metrics. Linear weights metrics (PER the famous one) are great, because pretty much anyone can do them if they tried, and the general public can see the correlation between the boxscore stats and their rating. However, we ALL know that obviously isn't the FULL picture of the player and his true impact to the team - it's just our best attempt at trying to get there.
Anyway - the playing time adjustment I make I believe does bridge the gap (albeit, probably just a small amount) a bit toward adjusted +/- - without having access to A+/- data and the programming power to run the regressions at such a large level. I mean, what are the chances a coach plays a lowish production guy, on a good team, big minutes if he ISN'T a good defender (considering there are 4 & 5 star kids who can fill it up sitting for him)? I mean - those "glue" guys HAVE to be defenders to see the court - because they sure as hell aren't scoring the points, getting all the boards, etc. Same with the high production - low minute guys. It's hard to believe a coach is gonna bench a kid all the time who statistically is impressive if he also plays good defense. That kid is probably benched because he gets lost on D, maybe fouls at bad times, maybe doesn't give full effort, or any other myriad of issues.
This is absolutely correct. In my overall ASPM, MPG is worth about 3 points/100 possessions (in other words, a really high minutes guy is worth 3 points/100 possessions more than a really low minutes guy, given all other rate stats are the same. However, in the defensive ASPM regression, the MPG coefficient was negligible--most of the effect was in the offensive ASPM. I was rather puzzled by this, though I think the fact is that most "bench sitters" in the NBA are good defensive players--rather like the replacement level for baseball defense being set at league average.
_________________
GodismyJudgeOK.com/DStats
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Statman
Joined: 20 Feb 2005
Posts: 239
Location: Arlington, Texas
PostPosted: Tue Feb 08, 2011 11:50 am Post subject: Reply with quote
I just did complete national rankings for this season, and learned that there are a couple stud PGs that I hadn't really been privy to yet - Norris Cole of CLeveland State and Charles Jenkins of Hofstra - who appear to be as good as any PGs in the nation this side of Jimmer and Kemba.
Anyway - I posted the top 99 players through Sunday's games (2-6-11)
http://www.pointguardu.com/f136/statman ... post342235
_________________
Dan
My current ratings (go to the last page for the most recent stuff):
http://www.pointguardu.com/f136/statmans-ratings-56243/
Last edited by Statman on Tue Feb 08, 2011 7:32 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 593
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Feb 08, 2011 12:46 pm Post subject: Reply with quote
Have you ever calculated aging curves for your ratings, Statman? That would be interesting, in terms of calculating who would be the best player if they were all the same age (useful for drafting purposes, perhaps).
_________________
GodismyJudgeOK.com/DStats