Friday, February 6, 2009

Final Word on Experience

Today’s post concludes my discussion on experience. I’ve previously argued that the most experienced teams are often not significantly more efficient than the year before.

I’ve also pointed out that year–to–year improvements in efficiency are usually because of new players or coaches, not because of the current players getting significantly better. That’s not to say that players don’t improve, but these improvements are a lot less predictable than we might think.

The problem of course is that we don’t know which new coaches or new players will be able to make a difference ahead of time. Certainly McDonald’s All–Americans are good bets to have an impact. But MAA’s don’t have an impact every year. And it is often hard to tell whether a veteran coach or an up–and–coming coach will make a difference.

Given a choice between a number of high variance teams with new parts and a few low variance experienced teams, most preseason publications will pick the low variance experienced teams, even though we know every year that someone will surprise us.

If experience is over-rated, what about youth?

Today I wanted to look at whether the biggest impact from experience might come from freshman teams becoming sophomore teams, not from junior teams becoming senior teams.

According to, the 15 least experienced BCS teams in 2007–08 were Texas, Kansas St., USC, Florida, Syracuse, Purdue, Virginia Tech, St. John's, Oregon St., Wake Forest, Northwestern, Arizona St., Michigan, Villanova, and Boston College.

We probably need to throw out several of these teams because they all lost a significant player to the NBA. See Texas (DJ Augustin), Kansas St. (Michael Beasley), USC (OJ Mayo), Florida (Marreese Speights), and Syracuse (Donte Greene).

I believe I'm stealing this line from Rush The Court, but can we all admit we are surprised to see Speights on this list? Look at his NBA numbers.

And indeed the five teams that lost key players to the NBA have all performed worse offensively than last year.

OEff – Deff – Team Year
107.3 – 87.5 – Texas 2009
123.8 – 91.8 – Texas 2008

109.4 – 90.9 – Kansas St. 2009
117.1 – 90.2 – Kansas St. 2008

106.2 – 87.7 – USC 2009
111.9 – 88.7 – USC 2008

115.1 – 95.8 – Florida 2009
117.7 – 96.0 – Florida 2008

113.6 – 93.4 – Syracuse 2009
114.3 – 94.3 – Syracuse 2008

Another interesting case is Purdue whose defense has improved while the offense has dropped off. I’m not ready to assign this to the loss of Scott Martin, but I think the injuries to Robbie Hummel are taking a toll. Hummel is still an incredibly efficient scorer, but his offensive rebounding and his free throw rate have fallen dramatically which may be due to his more limited mobility. Also, while E’Twaun Moore has taken on a larger percentage of the Purdue shots, his 3 point percentage has fallen from 43% to 35%. The net result is a team that is slightly worse offensively and missing an expected second year bump in production.

108.4 – 84.5 – Purdue 2009
112.2 – 88.5 – Purdue 2008

St. John’s and Virginia Tech are also surprisingly disappointing. While both have improved offensively as their very young players have matured, both teams have also suffered rather important drops in defense. It is very hard to get a read on St. John’s which also lost Anthony Mason Jr. to injury early this year, but Virginia Tech hasn’t quite found the dominant defense in ACC play it had at the end of 2008. Prior to the season ending loss to Ole Miss, Virginia Tech held its final 7 opponents under 100 in efficiency last year, and that included UNC.

110.4 – 96.2 – Virginia Tech 2009
107.9 – 86.7 – Virginia Tech 2008

101.5 – 96.4 – St. John's 2009
98.0 – 93.2 – St. John's 2008

The rest of the teams show clearer gains:

108.6 – 85.5 – Wake Forest 2009
108.6 – 94.1 – Wake Forest 2008

108.9 – 95.1 – Northwestern 2009
104.7 – 104.5 – Northwestern 2008

117.1 – 91.7 – Arizona St. 2009
111.5 – 93.5 – Arizona St. 2008

111.7 – 87.7 – Villanova 2009
110.2 – 91.7 – Villanova 2008

112.0 – 98.0 – Boston College 2009
109.3 – 98.3 – Boston College 2008

109.3 – 96.8 – Michigan 2009
103.3 – 96.3 – Michigan 2008

It should be noted that Wake Forest and Northwestern both added size this year in the freshman class which might explain some of the defensive improvement, but for the most part, the above teams have gotten better by utilizing their existing players more effectively.

In terms of the overall Pomeroy rank, Michigan has raised from 112 to 73, Boston College from 89 to 63, Nova from 43 to 17, Arizona St. from 48 to 14, Northwestern from 158 to 62, Wake Forest from 68 to 18, St. John’s from 123 to 111, and Purdue from 23 to 13. Of the teams without an NBA player defection, only Virginia Tech fell from 31 to 60.

Thus in my view, the younger teams gained more consistently from experience than the older teams. That’s good news for Ole Miss, Indiana, Arkansas, Vanderbilt, St. John’s, Kansas, Colorado, Cincinnati, Ohio St., Iowa, Oregon, Mississippi St., Florida, Virginia, and Georgetown.

Unless these teams lose a player to early entry (See Greg Monroe, Sylven Landesberg, Cole Aldrich, Deonta Vaughn, Nick Calathes) or suffer significant injuries, they should all expect some improvement in efficiency next season.

The Curse of Identification

I can’t end this discussion without a look at Oregon St.

103.4 – 101.8 – Oregon St. 2009
94.4 – 100.8 – Oregon St. 2008

There’s no question Craig Robinson has Oregon St. playing better than they were last season, but how much of that is better coaching, and how much of that was to be expected as players learned from last year?

Realistically, we can’t separately identify the two effects, but I’m still going to give a lot of credit to the coach. This clearly wasn’t a naturally better team when they stepped on the court in November. In fact, the first four games were the worst of the season for Oregon St. The in-season improvement suggests that Craig Robinson is having an impact in my humble opinion.

The Curse of Sample Size

Finally, I want to emphasize that I’m using a very small sample in all these arguments. In fact, because of the wackiness of finding the oldest and youngest teams, Villanova and Texas showed up in both arguments. They were among the youngest last year and the oldest BCS teams this year.

In fact, all I’ve really shown you is that Baylor, Notre Dame, and North Carolina are a bit disappointing this year. But based on my observations over time, there are always a few disappointing older teams each year.

If anyone wants to do a comprehensive study or look at returning minutes instead of the class of the returning players, I could certainly change my opinion. But until then, I stick to the following:
-Youth makes mistakes
-Talent and coaching wins games
-And only some players get better over time