25 February 2013

NCAA Week 4 plus bubble thoughts

1) Florida 10-4
2) Indiana 9-3

3) Louisville 10-5
4) Gonzaga 10-2
5) Duke 12-3
6) Michigan 9-4

7) Kansas 12-3-1
8) Wisconsin 9-8
9) Syracuse 12-5
10) Pitt 9-6-1
11) Ohio St 6-7
12) Michigan St 8-6
13) Miami 10-3-1

14) Georgetown 10-3-1
15) Arizona 9-4
16) Oklahoma St 7-5-1

17) Minnesota 9-8
18) Creighton 8-6-1
19) Virginia 5-2-6
20) VCU 9-6
21) Colorado St 8-5-1
22) San Diego St 6-7
23) Marquette 7-6-1

24) Middle Tennessee St 1-3-1
25) Missouri 7-8

Ranked Teams
30) New Mexico 14-4
29) St Louis 8-4-1
36) Memphis 6-3
51) Butler 8-5-1
37) Notre Dame 7-6
45) Oregon 7-5-1
27) St Marys 6-3-2
49) Akron 4-3-1
75) Louisiana Tech 4-2-1

I have no idea how Louisiana Tech got a bunch of votes. They've played almost no meaningful games (especially in the last couple of weeks). They're likely to win their conference but aren't the best team in that conference, which is Denver at #50, so it's unclear if they'd merit any bubble conversation. I rather doubt it.

Here's the list of bubble teams identified by the bracket matrix as teams not getting a solid percentage of inclusions or as low at large teams. Maybe 3 or 4 of these are thus far looking like legit tournament teams. 7 of them will get in anyway. 11 won't.
19) Virginia. 6 non top 100 losses is undoubtedly their problem here. Right now they're in but they cannot afford any more weird losses.
26) Kentucky 6-8, presumably this is because of the injury to Noel. They did get a big win at home against Missouri this week, so they could remain in as well. With Noel they were a no-brainer for inclusion and at a higher seed.
27) St Marys 6-3-2. Not sure what this is all about. They're ranked after all. The win over Creighton during bracketbusters this weekend really helped here. Swept by Gonzaga, but Gonzaga's really good this year.
39) Ole Miss 4-6-1. I'm not sure that they actually should merit inclusion, but they're ranked highly enough to do so. A sweep of Tennessee and a win over Missouri (at home) are basically their only resume fodder. They're also the only decent team Middle Tennessee St has beaten. Loss to South Carolina hurts here.
40) Iowa 4-7-3. Very fringe votes for inclusion. They're really high on the MOV/tempo metrics and have 3 wins against the top 25 (Wisconsin, Minnesota, and Iowa St, all at home), and lost by only 4 to Indiana (also at home). Their losses @ Virginia Tech, @ Purdue, and @ Nebraska all hurt a lot here and other than Iowa St, they've only got a win against N. Iowa non-conference of any note. I think they should be in, but they probably will not be without a deep Big Ten tournament run.
43) Baylor 5-9-2. Possibly safe, with a number of decent wins (Kentucky, Oklahoma St, BYU, Lehigh, and St John's). Oklahoma St and early win over Kentucky are the only resume fodder and they're losing a lot of games lately (6 of the last 8, with the only wins coming over in-conference cannon fodder). I could see them slide out.
50) Denver. 3-8. Probably has to win the WAC to get in but did get a nice road win during bracket busters this weekend.
52) Stanford. 5-10-2 Probably shouldn't get in and doesn't look like they will.
55) California. 5-9. Probably shouldn't get in either but looks like they might. Win @ Arizona and sweep of Oregon seems to be the keys here. They've only got a win over Denver to point to out of conference, where they lost to Harvard and Creighton, were blown out by Wisconsin, and lost by 1 to UNLV.
56) Villanova 6-9-1. Wins over Louisville, Syracuse, and Marquette are most likely why they would get in. Swept by Providence, loss to Columbia and only one win over the other Philly teams (St Joe's) would be arguments why not. I'd be fine with them in instead of Stanford or maybe Baylor though.
57) Maryland 3-5-3. Win over Duke is pretty much the only reason they're even in the conversation. NC State and Stony Brook are the only other good wins, with a close loss to Kentucky also. I doubt they get in without a couple of ACC tournament wins (over the Carolina teams).
58) Southern Miss 1-6-1. Shouldn't get in. Seems like a safe bet to miss out.
62) Tennessee. 7-9-1. Win over Wichita St is about it, and then the blowout over Kentucky put them on the radar. Unlikely to make it. They do have a lot of wins over quality teams, but most of them are modest wins rather than impressive stock.
64) Boise St 6-6-2. Win over Creighton on the road helped. Lost by 4 to Michigan St also. Otherwise, win over UNLV and a sweep of Wyoming is about it. I could see them getting in, but they've got a lot more buzz surrounding them than I suspect is warranted.
67) Temple 7-6-2. Wins over Syracuse and St Louis. Lost to Duquesne. Otherwise, wins over LaSalle and Villanova is about it. I guess I could see them getting in as well, but it's not a strong case.

84) Arizona St 5-6-2. Win over UCLA and a sweep of Colorado is about it here. Getting swept by Washington does not help. Loss to Utah and DePaul also doesn't help. They also got smacked by Arkansas, Creighton and Arizona. They're a fringe candidate, but they shouldn't even be a candidate.
86) St John's 5-7-4. Close wins over Cincinnati and Notre Dame, and Detroit is about it. Losses to San Francisco and UNC Asheville do not help. Again, a fringe candidate, but shouldn't be even involved. Blow out losses to every good team in their conference (Louisville, Syracuse,Pitt, Georgetown twice), plus Baylor should suggest this isn't a very competitive team.
NR) Charlotte. 5-6-2 I don't think a team I'm not even tracking should be included as even a fringe possible bubble team. Something tells me there's a huge flaw in the RPI system that they are. Win at Butler and at Davidson is about it here, plus a win against LaSalle. Destroyed by Miami, blown out by St Louis, lost at home to VCU, swept by Temple. Otherwise a lot of close games against mediocre teams. If I had to guess, they're behind George Mason in the #120 range.

18 February 2013

NCAA Week 3

1) Florida 9-3
2) Indiana 8-3
3) Louisville 11-5

4) Michigan 8-4
5) Duke 14-3
6) Gonzaga 10-2
7) Syracuse 11-4
8) Pittsburgh 9-5-1
9) Kansas 10-3-1

10) Miami 11-3
11) Wisconsin 10-8
12) Michigan St 8-4

13) Ohio St 4-7
14) Minnesota 8-7-1

15) Oklahoma St 7-4-1
16) Arizona 9-4
17) VCU 8-4-1
18) Colorado St 8-3-1
19) Creighton 8-5-1
20) Georgetown 10-3-1

21) Cincinnati 8-7
22) Kentucky 5-8 (about to start dropping though)
23) Missouri 6-7
24) Marquette 8-6
25) San Diego St 5-7

Ranked teams out of my top 25
49) Butler 7-4-1
36) New Mexico 13-4
42) Oregon 6-4-1
51) Notre Dame 6-6

I'm not sure what the NBA commentators for the ASG last night were talking about with "about 20 teams that could make the final four". I see maybe 10-12 good contenders, with even several of those as unlikely and draw dependent (because they're playing too slow in the cases of Pitt and Wisconsin, to a lesser extent Florida and Miami, same deal). There's a pretty hard gap between the top 10-12 teams and the rest, and even in there, a sizeable gap between the top 2-3 teams and the rest. There's a few teams beneath that gap that could make a run (Arizona?, VCU?) but not many. Last year Louisville was the only "surprise" final four team and even then they were 15-8-1 coming in, which was quite respectable and were one of the top ranked defenses, (this year they're the #1 ranked, but they're scoring at a much better clip than last year).

17 February 2013

HS draft for the NBA

Since this topic comes up occasionally, I thought I should quantify it somewhat to see if the data supports the conclusion that I reached quite some time ago that the one and done rule the NBA imposed was arbitrary and dumb.

The list of HS players drafted into the NBA isn't very long, so data-driven conclusions are unlikely to reach very far unfortunately. But they are still useful to examine to see if there are any hints. I would suggest that we are likely being imprisoned by media narratives surrounding parts of the limited data rather than the whole from this analysis.

First off.
Starting just with the 1995-6 classes, we can see right away two Hall of Famers, both clearly the class of their draft class. Kevin Garnett is leaps and bounds above any other player drafted in 1995. The next best player in that draft is Rasheed Wallace (drafted right ahead of him at 4th), but he's way behind in value.

The following year we get Kobe Bryant picked 13th, again the best player of that class, but it's a much better draft. He's followed by Steve Nash, Ray Allen, Ben Wallace (who went undrafted), and Marcus Camby. Iverson and Stojakovic are also better than Rasheed was, for comparison of how deep this class is. Jermaine O'Neal was also in this draft at 17th and falls in a class behind these players but still had a respectable career (if injury riddled and complete with an infamous fight). He is roughly as productive and valuable as Stephon Marbury was (a one-done shoot first point guard who went 4th).

1997 Tracy McGrady, picked 9th, is the 3rd best player in a draft following Tim Duncan (just ahead of KG's class as a top tier Hall of Famer), and Chauncey Billups (a pretty clear HoF candidate too). It drops way off after those three (Antonio Daniels?).

1998 Finally a bust? Al Harrington as a HS pick is rather mediocre, but he was picked 25th. And despite being a rather average to below average player is still playing (because he can shoot reasonably well). Rashard Lewis picked 32nd is a capable NBA player and 5th best player in this draft (roughly as good as Antawn Jamison who went 4th). Dirk Nowitzki is an international player of note picked at 10th is the best player in this draft, followed by Paul Pierce and Vince Carter. Brad Miller goes undrafted and comes in at 6th. Korleone Young is drafted 40th and does pretty much nothing. Second round picks rarely do though. Famously this is the Olowakandi draft (he was terrible and picked 1st) and also famously Nowitzki was traded for Robert "Tractor" Traylor, who went 6th and has a short unimpressive career while Dirk has a surefire HoF career.

1999. Here some real worry probably started to creep in.  Jonathan Bender goes 5th, mostly on the strength of comparisons to KG. While he never really puts up any numbers anyway, he also has a very brief injury riddled career. Leon Smith is a total basket case of a human being, is drafted at 29th and doesn't amount to much either. More than likely, his career, what there was of it, IS the case against drafting HS players. But I have a hard time seeing much generalisation to be made of his circumstances to any high school athlete. Manu Ginobili goes 57th and is probably the 2nd best player in this draft (behind Shawn Marion).

2000 A historically bad draft class all around. Michael Redd is the only good player drafted (43rd), while Mike Miller (5th) and Kenyon Martin (1st) are the only other reasonably productive players picked and a pair of others (Turkoglu and Jamal Crawford) are decent. And that is all. Darius Miles goes 3rd, the highest a HS player has gone so far, probably because nobody else seems very good anyway, and doesn't amount to very much (but isn't nearly as bad as Marcus Fizer, picked after him at 4th). DeShawn Stevenson goes 23rd, and appears to have a niche as a bearded defensive player (and is still in the league today) but few other appreciable skills.

2001 Odd class all around. Kwame Brown goes 1st. He is a below average player, but sticks around in the league still today and isn't nearly as terrible as Olowakandi was. Pau Gasol is clearly better than anyone else from this class (picked 3rd), followed by Tony Parker (a young French point guard at the time, picked 28th), and Tyson Chandler (picked 2nd and another very good HS player). Despite the flops of Brown and Eddy Curry (a HS player picked 4th, a high volume scoring big man who can't do anything else at all, ends up with a heart condition to boot), this is actually a very deep draft with many highly skilled players scattered in it (about 11-12). Diop, while styled as an international player, played HS ball in the US and is picked 8th. Doesn't amount to much.

2002 Amare Stoudemire becomes the first HS player to win a rookie of the Year award after being picked 9th. He remains the best player from this draft, although Yao Ming was too injury prone to actually claim this title. Stoudemire's a problem child player for much of his career though and a terrible post defender.

2003 LeBron. Enough said there. Kendrick Perkins goes 27th and isn't all that productive but has a mean scowl and excellent post defense skills. Travis Outlaw goes 23rd and somehow ends up with a "Travis Outlaw's contract" style career, which is to say, he gets dramatically overpaid and isn't a very good player. Darko Milicic goes 2nd and isn't all that productive either. We could consider him either international or HS. Ndubi Edi and James Lang are also HS players drafted in this class. Neither amounts to much of anything. Both go late (26th and 48th).

2004 Dwight Howard. Again, enough said. A ton of HS players are picked this year in the first round. Al Jefferson goes 15th and is the 3rd best player in this class (behind Iguodala). Josh Smith goes 17th and while something of a coaching problem (for putting up ill-advised jump shots, among other reasons), is actually quite good (probably 6th best in class behind Luol Deng). JR Smith is an even bigger coaching problem at 18th, but again, at least a skilled scorer. Dorell Wright goes 19th and takes a while to get some PT, but turns out okay (better than JR, but with less minutes). This class is marred basically by the Bassy Telfair pick at 13th, who had a documentary film produced about his HS to NBA leap and the knee injuries to Shaun Livingston (picked 4th). (don't click on that if you're squeamish, his leg literally bends sideways suffices to say). And then the pick of Robert Swift at 12th, which amounts to a total dud, but certainly not any bigger than Araujo at 8th.

2005 Andrew Bynum goes 10th, aside from injuries, he's probably the 4th best player in this class (just ahead of undrafted Jose Calderon) but well behind Deron Williams and way behind Chris Paul. Monte Ellis goes 40th and becomes a famous ballhog but otherwise is a solid player (especially for a second round pick). Louis Williams goes 45th and ends up as a very good instant offense shooter as well. Martell Webster goes 6th and ends up just below average (behind Francisco Garcia who went 23rd). Andrey Blatche is a headcase picked at 49th but having a productive season this year with the Nets. Amir Johnson goes 56th, gets limited minutes with the Pistons, but plays very productively in them (followed also with somewhat less limited minutes with Raptors and still more productive play). Gerald Green goes 18th and wins a dunk contest, but not much else is done. CJ Miles goes 34th and is also largely unproductive.

I don't count very many "busts" here such that the problem was NBA teams needed to protect themselves from themselves. There's a good volume of late-first round and second round picks (picks which rarely turn out to be suggestive of productive NBA careers). This suggests that many NBA teams already recognized the dangers of HS evaluations and adjusted accordingly. The famous busts of Kwame Brown (thanks be to Michael Jordan), and Telfair's documentary, and a few problems of immaturity (Miles, Stoudemire, the Smiths, Blatche), appear to be the cause of concerns here. There are also numerous cases where these players had to play a year or two of pro ball to begin to become productive however. Garnett or Kobe don't start out as a perennial All-Star for example and McGrady doesn't win a couple scoring titles right away either. Howard or LeBron are both very good immediately but neither is rookie of the year, and neither is in the discussion for top 5 best players in the league for at least a couple of seasons. So it is possible an extra year of seasoning may help these players mature better or polish their games. That argument is not invalidated by saying the age restriction is dumb. It also takes most college players a year or two to begin to adjust to the skill level jumps, and usually year 3 is when any player begins to really flourish.

What it suggests is the NBA prefers to have someone else pay to develop the players, pushing risks off on someone else (in this case, the players themselves and colleges who recruit them for one-and-done runs). For the most part, college freshmen who are drafted are a relatively safer bet than HS players, in the defence of this practice. But college freshmen are rarely drafted outside of the lottery or first rounds. HS players were. In addition, statistically valid issues can be raised with international player scouting and that of 4 year seniors who are drafted, both groups that fail on a very high rate relative to HS players and underclassmen who are drafted. There are prominent examples of both groups who succeed, who are held up as exemplars (particularly Tim Duncan or Steve Nash types as 4th year college players) but we rarely hear about 4th year seniors who are terrible professionals who are drafted. This may be because such players are rarely lottery picks in the first place (Jimmer Fredette so far is in this category, Terrence Williams from a few years back, Acie Law and Al Thornton in the same draft, and then Luke Jackson, as but a few examples). But if such a comparison is fair and honest, we should admit that many of the HS players who have failed and busted or been mediocre are of similar draft status to that of international or college players and have similar expectations leveled upon them. Whatever the valid reasons may be behind such a policy, it doesn't appear that "protecting" owners from their supposed stupidity is one of them.

HS great picks would be
KG
Kobe
LeBron
Howard
McGrady
Chandler
Stoudemire
Jefferson
Lewis (mostly because he was second round)
Josh Smith

Value picks
Bynum
Wright
JR Smith
Jermaine O'Neal
Williams
Ellis
Johnson
Perkins
Blatche
Harrington

HS Busts would be
Brown (#1 picks aren't supposed to be below average)
Swift
Telfair
Darius Miles
Bender
Webster
Livingston
Eddy Curry

Lost picks would be
Young
Lang
Ebi
Leon Smith
Outlaw
Green
CJ Miles
Stevenson

This looks like a fairly normal distribution, but it has some serious top heavy options.

11 February 2013

Week 2

Not much changed, despite the apparent chaos of upsets. The bigger sort seemed to occur among the middle, bubble quality teams with a few (Illinois, Virginia), playing up, and a few (BYU, Maryland), playing down.

I will start posting top 100 records. Any extra record element is non-top 100 losses. For most good teams, this would be a non-factor, but it becomes an issue with many bubble teams, if for no other reason than to explain absurdly high RPIs. I consider games against non-top 100 teams to be irrelevant record padding (college football is much worse on this front with very few games outside the top 30-40 teams). Top 50 records aren't much better as a metric unless a team has a bunch of 50-100 wins but nothing in the 1-50 range (California did this last year, and lost in the first round).

1) Florida 9-3
2) Indiana 8-3
These two may take a lot, a major injury for example, to drop down to the next tier.

3) Michigan 9-3
4) Louisville 9-5
5) Duke 13-2
6) Syracuse 9-3
7) Pittsburgh 8-4-1
I could see Michigan moving up from this tier, and Miami or Gonzaga moving up.

8) Miami 12-3
9) Gonzaga 9-2
10) Ohio St. 4-6
11) Kansas 11-3-1
12) Minnesota 8-6-1
13) Wisconsin 8-7
14) Michigan St 8-4
15) Arizona 8-3
Pretty stable group here I think. Minnesota is the highest team I'm charting that isn't ranked in the AP (they're 26th, they were ranked last week though).

16) Kentucky 4-6
17) Creighton 8-4-1
18) Colorado St 6-3-1
19) Oklahoma St 8-4-1
20) VCU 7-4-1
21) Cincinnati 6-6
22) St Marys 2-3-1
23) Baylor 6-6-2
24) Georgetown 8-3-1
25) San Diego St 5-5
Creighton dropped way too much for their loss this week in the polls.

Honorable Mentions (ranked teams)
27) Virginia 6-1-5 (5!)
28) Marquette 5-4-1
36) New Mexico 12-4
40) Butler 8-4
41) Memphis 4-342) Notre Dame 6-5
47) Oregon 5-4-1




06 February 2013

Early NCAA Ranks

I use a composite of Pomeroy, Sagarin Predictor, and LRMC to rank teams. It comes out fairly close to ESPN's BPI, but there's some twists and turns along the way. It tends, combined with some other factors, to put me in the top 1-2 percentile on ESPN's bracket challenges year after year.

I will look at things like away/neutral records or high/low pace teams once the field is set to adjust picks but I don't generally care until then. A very good team will win either way anyway. Teams that are closely or reverse ranked on this list when matched up tend to be excellent upset picks early.

For once the polls appear to be correct near the top at least. (Louisville is a little underrated and Butler is way overrated). I made the mistake of actually looking at the horrid RPI ranking system, so there's some notes on that.

1) Florida
2) Indiana

3) Michigan
4) Duke
5) Louisville

6) Kansas
7) Syracuse
8) Pittsburgh (huge gap on RPI rank)
9) Ohio State
10) Gonzaga

11) Minnesota
12) Miami (somehow 2nd in RPI?)
13) Arizona (somehow 3rd?)
14) Creighton
15) Wisconsin (perennially underrated, probably because they're boring and slow. Pitt is slower this year though)

16) Kentucky (44th?)
17) Michigan State
18) Colorado State
19) Cincinnati
20) Oklahoma St
21) VCU (49th?)

22) Georgetown
23) San Diego St
24) Ole Miss
25) Baylor

Some other teams of note
26) St Marys (somehow 60th in RPI)
27) UNLV
29) Belmont (little conference high pace team, fun upset watch pick)
34) BYU (on Lunardi's bubble, safely in here)
36) New Mexico (5th! in RPI? 5th!)
37) Oregon (ranked)
38) Missouri (ranked)
42) Middle Tennessee (little conference likely snubbed team if they don't win out, also fun upset watch)
45) Butler (ranked, 14th RPI?). I have them in but they're on my bubble.

Southern is also having the best season for a SWAC team since I've been keeping track of the probable and set fields. That doesn't mean they'll win any games, but they're inside the top 200, which is astonishing for that league. I have them at roughly 148th. I think a number of times a worse team has won their conference tournaments, bumping them up to 300th range, but even then, the better team was only in the 220s.

Memphis, UConn, UCLA, and Virginia would be the last four in at large.

Right now here's my list of bubble teams that wouldn't get in:
Maryland
Illinois
Oklahoma (23rd?)
Arkansas
LaSalle (27th?)
Alabama
Temple
Villanova

Lunardi also has either in or on the bubble listed these teams:
Louisiana Tech
Arizona State
Indiana State
Washington
St John's
Only two of those (La Tech and Indiana St) even made my list to keep tabs on and neither would get in.