Friday, February 4, 2011

What Peter Sullivan Means...

In case you missed it this past weekend, Brown forward Peter Sullivan missed his first career game - an overtime loss at Penn.

Sullivan's solid three-point shooting and knack for getting to the line as a freshman allowed him to force his way into the starting lineup as a rookie - a spot he has maintained until injuring his shoulder against Princeton. Now the Bears must face the prospect of going forward without their captain, as Sullivan's younger brother Matt, a sophomore, is expected to take his starting role.

How much of an impact will this have on Brown? Ask a variety of people and you'll get pretty much the same answer, one of either "He's irreplaceable, they're screwed" or "They might go winless in the Ivies." You'd be hard-pressed to find anyone that would say that it might not have a significant impact whatsoever, even though that might indeed be the case.


There is often a myth or legacy that comes along with certain players' past, which can supersede present realities. As a rookie, with possession-eating seniors Mark McAndrew and Damon Huffman leading the way, Sullivan could be choosier about his shot selection and could afford to take the ball to the basket only when the situation was advantageous.

His usage rate in 2008 was 16.7 percent and his shot percentage was 17.0 percent (with five players on the floor, obviously 20 percent is average) and the more selective approach allowed him to record the seventh best eFG% and fourth best TS% in the nation.

Since McAndrew and Huffman graduated, however, Sullivan's performance has hovered around the national average. In 2009, he and Matt Mullery had to make up for the lost possessions, and Sullivan's volume shooting approach knocked his eFG% down over 18 points to 47.5 percent. After a brief eFG uptick last year, when Sullivan slightly stepped down the percentage of shots he took, his eFG% hit rock bottom at 43 percent this season.

He has managed to buoy his offensive rating (now 99) by getting to the line at an impressive rate (27th nationally). That offensive rating, though, is slightly below the national average - though slightly above for a player using as high a percent of possessions as Sullivan (26.4 percent).

Sullivan's quality is only one aspect of the equation, however. It's a difference that we're after, and a difference implies a measurement against something else.


In order to model Sullivan's departure correctly, you can't just ignore his possessions and measure the strength of the rest of the team. Rather, you have to take his possessions and distribute them to the players that will have to use them in his absence, and then use their efficiency rates to judge how those possessions would have been used.

As mentioned at the start of the piece, Sullivan's primary replacement will be his brother Matt, also a 6'5 swingman. Matt has put together an outstanding year in a reserve role, after struggling as a rookie starter last year. His efficiency rating sits at a healthy 114, and he has posted a 63.6 eFG% from the field while taking 20 percent of the shots while on the floor.

For the analysis, I doubled Matt's minutes (roughly 240 extra) which accounts for about half of Peter's lost minutes. Since Brown doesn't go much deeper at the true swingman spot, I split the rest of Peter's minutes among the two senior guards, Garrett Leffelman and Adrian Williams, as well as giving some minutes to reserve forwards Andrew McCarthy and Dockery Walker.

With Peter using the most possessions on the team, as well as an above average percentage of shots, I had to allocate more than a proportional amount of possessions and shots to the players to whom I gave his minutes. Thus, I lowered the shooting percentages of each slightly and raised their turnover rates to account for the marginally worse shot selection and greater offensive stress.

The end result took a Brown team which had raw efficiency numbers of 99 offensively and 102 defensively down to 95 offensively and 102 defensively. At 69 possessions per game, Peter's absence would be expected to make the Bears roughly three points worse per game.


While it seems easy to conceptualize three points a game (Brown would have won at Penn!!!), it's actually trickier than just tacking three points onto the Bears' score. The three points are merely an average - some nights he might be worth six points, others zero, etc.

Thus, the concept works better when applied to expected win totals over an extended number of games. Brown has 10 Ivy games remaining in the 2010-2011 season, and with Peter Sullivan included in the numbers, it was expected to win 4.3 of those 10 (with six home games left and a visit to Dartmouth, it's easy to see how a 4-6 record could be achieved down the stretch).

Re-running the simulations with the Bears' new offensive and defensive ratings, Brown is expected to win 3.4 of its final 10 games, meaning that Sullivan was worth roughly 0.9 wins per 10 games. That's hardly the disaster that most folks in Providence would have expected.

While taking a pro forma approach to a team's possession structure in the absence of a key player is hardly an exact science, the good news is it takes very large movements to affect win totals. That is, while one can quibble with the adjustments, it would take rather extreme ones to lower Brown's expected win total to less than two.

In other words, while it is always sad to see a senior potentially finish his career in the doctor's office rather than on the floor, the effect that losing Sullivan should have on this Bears team in the wins and losses column should be relatively minor.

No comments:

Post a Comment