Thursday, June 09, 2005

Going, Going, Gonzaga

It's been real slow around here, which is somewhat of a shame since the season moves too fast and the off-season provides a time to catch up on things I missed in frantic weeks of February and March. I am working on things behind the scenes (that is, in the rare times I am sitting at my computer), so posting will be light for the next couple of months. But I'll tackle one of those off-season topics in this post.

Of the eight teams that have made the tournament each of the last seven years, Gonzaga has the most unusual postseason record.

Year    W-L   Seed  Season ended
1999   25-6    10   Elite Eight
2000   24-8    10   Sweet 16
2001   24-6    12   Sweet 16
2002   29-3     6   First Round
2003   23-8     9   Second Round
2004   26-3     2   Second Round
2005   26-4     3   Second Round

Gonzaga's post-season performance has been correlated to their seed, but negatively so. Their consecutive sweet 16 runs came as a relative unknown with a double digit seed. In the post-Cinderella era, since 2002, their best performance was in 2003 as a 9 seed, when they took top-seeded Arizona to double overtime in the second round.

When I judge a team's ability from past seasons, I put more weight on the tournament seed rather than the tournament performance. The seed is based on 30 games, and usually the committee does a good job seeding teams. However, with Gonzaga the seed gives you an impression that they have improved significantly over the years, but this doesn't seem to be true.

The 'Zags (who calls them the Bulldogs anymore?) have been living off the reputation established in '99-'01 for the last couple of years. So much so that well-respected basketball people have picked them to go a long way in their brackets (in a least one case, all the way). Gonzaga's reputation was exceeded by their regular-season performance during the sweet 16 years. They snuck up on us. But their reputation has since caught, and passed their play on the court. Let's demonstrate that with a little more data.

Yr   W-L  Seed  Season ended   OE  DE  Pythag Actual Diff
99  25-6   10   Elite Eight   112  93   .882   .806  +.076
00  24-8   10   Sweet 16      112  94   .868   .750  +.118
01  24-6   12   Sweet 16      122 102   .878   .800  +.078
02  29-3    6   First Round   113  93   .890   .906  -.016
03  23-8    9   Second Round  111  99   .773   .742  +.031
04  26-3    2   Second Round  117  95   .902   .897  +.005
05  26-4    3   Second Round  115 101   .789   .867  -.078

I've added the offensive and defensive ratings for each season and the pythagorean winning percentage. Then I've added the actual winning percentage and how that differs from the pythagorean. Only the OE/DE is for the full season, everything else is pre-tournament. You can't really compare the pythagorean winning percentage from one year to the next because of scheduling differences, but it does make sense to compare it to their actual record from the same season. What I want to compare is the difference from year to year.

The sweet 16 teams were underrated according to their record. Then came a three-year run where their real record was pretty close to what Pythagoras said it should have been. Finally, last season Gonzaga prevailed in most of their close games - they were 7-2 in games decided by five points or less before the post-season - so their record was a little bloated. A well-timed win against Oklahoma State - arguably the most impressive non-conference road win that any team had - went a long way to fuel the perception that Gonzaga was better than they were.

The other columns in the chart tell the story - Gonzaga's defense slipped last season. Unlike in 2001 when their defense was similarly porous, there was no increase in offense to compensate. Anyone following the possession stats knew their defense was a weakness. It was one of the things that grabbed me when I first posted the possession numbers in December. (And now for a little commercial on the possession stats. I know some people wonder: what can the possession stats do for me? I mean, we knew UNC had a great offense - they led the nation in points per game. We knew Wake Forest had a poor defense. They were ninth in the ACC in points allowed per game. So why use the possession stats? One example is this: Gonzaga allowed five points per game fewer than Wake Forest, but on a possession basis their defense was actually worse. However, based on the points per game, Wake Forest's weakness got a lot more play in the media.)

There's been another transformation with Gonzaga over the years. Here is the year-by-year difference in made threes between Gonzaga and their opponents. Also listed are the Gonzaga players that made at least ten three-pointers during the season.

       3PM 
Year  Diff    Shooters 
1999   116    Frahm (93), Santangelo (64), Hall (55),
              Floyd(24), Calvary (21), Leasure (14),
              Dench (12) 
2000    52    Frahm (90), Santangelo (67), Floyd (33),
              Dench (26), Calvary (22)  
2001     9    Dickau (71), Stepp (64), Bankhead (37),
              Calvary (29), Forbes (25) 
2002     8    Dickau (117), Stepp (47), Bankhead (35),
              Violette (11) 
2003    -9    Stepp (98), Bankhead (56), Skinner (51),
              Brooks (12), Reisman (12) 
2004   -35    Stepp (79), Bankhead (41), Raivio (23),
              Skinner (18), Morrison (17) 
2005  -105    Raivio (77), Morrison (37)

Back when Gonzaga emerged they were primarily a jump shooting team - and good at shutting down opposing jump shooters. But their original strengths have steadily eroded, with a tremendous drop-off in '05. Blake Stepp and Kyle Bankhead left a perimeter void that was unable to be filled by Derek Raivio singlehandedly. They got almost no perimeter production from the two spot. In fairness, Texas Tech transfer Nathan Doudney was expected to pick up the slack at that position in '05, but got injured in November. Doudney's presence would have mitigated the decrease in threes, but it wouldn't have changed the trend - as Gonzaga has transitioned from Cinderella to perennial top 20 team, they are becoming increasingly dependent on its bigs scoring in the paint. Their FTM/FGA number ranked second in the country, after years of being a fairly pedestrian free throw shooting team. Well, with the exception of the 2001 edition.

And how incredible was that 2001 squad? I don't have access to every team's stats from then, but compare the '01 'Zags to the '05 season's leaders. Their 122 offensive efficiency compared to 2005 leader Illinois' 117. Their 57.9% effective field goal shooting is near Utah State's 58.2%. Their .360 FTM/FGA beats Eastern Michigan's .350. They were a well-balanced team with future NBA draft picks Dan Dickau (61.5 eFG%) and Blake Stepp (53.1) in the backcourt, and Casey Calvary (58.4), along with the quietly efficient Zach Gourde (63.0), and Mark Spink (58.0) up front. In fact, all ten players that averaged double-figure minutes shot better than 52% in eFG terms. With that kind of offensive production, you can go a long way with a poor defense. Sure, their schedule was weak, and they didn't post numbers like that when they played Arizona, New Mexico, Florida, or Michigan State. But they did post a 125 OE in their first round game against fifth-seeded Virginia.

So back to the point here...one of the things I'll be interested in early in the season is where the Gonzaga D goes without Ronny Turiaf. By the numbers, it certainly appears as if the perimeter players are what's holding the defense back. Last season, in addition to giving up more threes than they had in the previous six seasons, Gonzaga's opponents committed turnovers on only 17.8% of their possessions, the lowest in the last seven seasons. If Gonzaga is to return to the sweet 16, their perimeter defense must improve, but is there any reason to think it will?

Friday, June 03, 2005

What Goes On Here

Back in the day - which was like three years ago - I read a column by Rob Neyer. Rob is a baseball columnist for ESPN.com, and despite that I was not much of a baseball fan, I was addicted to his work. In March of 2002, he wrote what I believe to be his only column about college basketball. The subject was the Kansas Jayhawks’ poor first round showing in the NCAA tourney, an 11-point victory over 16th-seeded Holy Cross that felt closer than the final margin indicated.

I am a big fan of Rob’s because his sensible approach to baseball analysis is refreshing. The way he supports assertions not by talking to a couple of “baseball people,” but by using cold hard facts made me a religious reader of his before ESPN.com charged folks to see his work. Rob’s piece on KU was in this same style, pointing out that based on history, the close game more than likely signaled an early exit for the top-seeded Jayhawks, rather than a wake-up call. That column made me wonder why there wasn’t someone writing stuff like that regularly for college hoops.

You see, college basketball provides a forum for any opinion to be supported. You can claim that because Player X leads his team in scoring, that his teammates must be worse off without his presence. You can think expanding the lane would make the game better. And you can believe that Team X has a bad defense because they allow a lot of points. You can believe all of these things, but around here, I like to test such opinions with the mountains of data that this game with its 300+ teams provide us.

So that’s what propelled me to start a blog in the fall of 2002. My site, kenpom.com, didn’t get much traffic back then, and I made sure to make the link to the blog obscure. Which was why I got about 50 hits a week at the peak of 2003 season. That was good because the content was truly pathetic. The problem was I put the cart before the horse. While I wanted to approach college hoops from a fresh angle, I didn’t have the resources or mindset to do so. Tragically, much like Neyer’s Kansas column, there is no trace of the premier season of my blog to be found on the ‘net.

However, you can look at the archives on this site and see that it was a still a struggle for me for most of the 2004 season. Around tourney time in 2004, I figured out a way to compute the efficiency statistics that are featured on this site. That allowed me to do some objective analysis with revealing data that was previously unavailable to me. For instance, it allowed me to see that Saint Joe’s was legit. Reading Dean Oliver’s Basketball on Paper the following summer sparked a whole lot of new thoughts, and this space has progressed from there.

From the beginning the mission of the blog has been simple - provide something useful. That first season I didn’t accomplish much, but the blog has evolved to provide the occasional nugget that hasn’t been presented anywhere else. Statistical data on rules experiments has been one favorite of mine (with more to come in 2006). More than anything, I want to provide a unique examination of the game. If it means turning conventional wisdom on its ear, all the better.