It's always fun to find a good new joke. You can run out and get a good laugh out of your friends when they hear it for the first time. But if you tell that joke over and over again to the same people, it won't be as funny the second and third time around. Once the novelty wears off, a typical joke loses most of its luster.
Over at Truehoop Henry Abbott has been holding an extended discussion with former Mavericks stat guru Professor Wayne Winston on the use of adjusted plus/minus as a metric for measuring basketball performance. Now plus/minus is basically a metric that attempts to show how well the team does when a particular player is on the floor. If Joakim Noah plays 28 minus, and the team scored 75 points but gave up 70 in that time, Jo's basic +/- is +5. Adjusted plus minus is a similar score that is adjusted to correct for the other players on the floor: how this takes place is beyond my math skills (my last math class was "Magic of Numbers" as a college freshman in the year 2000). Now my own personal opinion is that basketball is probably the hardest of the major sports to effectively quantify, and that if the ultimate goal of statistical analysis in basketball is a mathematical model that accurately expresses player value and can be used to predict future performance, the stat work being done in the public domain is probably 10% there. But Winston in his discussion raised an interesting theory, that having 4 shooters on the floor at all times is one key to a successful offense. This led him to defend the utility of Tim Thomas and see Tyrus Thomas and Noah together on the floor as a combination which crippled their offense against Boston in the playoffs.
Now I don't have any particular quarrel with his theories, and I certainly don't have the statistical prowess to prove or disprove anything using numbers. But the case of Tim Thomas raised an issue, for me. I'd love to hear from Winston or any other basketball stathead talk about how they conceptualize time spent on the court. Are all minutes created equal, or is the amount of time a player spends on the floor in a particular lineup a facet of their performance that gets overlooked or flattened in the production of the plus/minus metric? Is comparing a starting unit that sees 12-14 minutes on the court together with a lineup that only sees 4 to 5 on a per minute basis a useful metric of analysis?
Take a look at Tim Thomas's on court data from 82games.com. His most prevalent lineup, Hinrich-Gordon-Salmons-Tim Thomas-Miller, a case, in fact, of 5 shooters on the floor, put up +34 points in only 46 minutes and won 73% of the time. Taken as a single game, it's a blowout. But this lineup occurred 11 times, meaning the average time on the court was only about 4 minutes. This lineup did great, but was probably only on the floor for short bursts.
My contention is short term lineups are intrinsically different from, say, the starting lineups that open a game and log more minutes together, because the coach is using them differently, and opposing defenses are not spending as much time practicing to stop and attack them as they do the higher minute lineups. A coach may go to a guy like Tim Thomas to "accelerate" the game on offense: the shooter comes in, the offense changes, and the defense has to react to that change. By the time they've handled the "acceleration" in the game and adjusted their own offense to exploit whatever weaknesses the unconventional lineups presents, the original team may have already switched back to a more conventional set. In a sense, short minutes lineups are like "jokes" that are really funny the first time around. Could some of their effectiveness come from the fact that they are only used in short bursts? If used more, would they reach a point where their efficacy would plummet as defenses would come to expect and plan for their implementation as they would with a starting lineup, like jokes you have already heard more than once?
One such "joke" offense Bulls fans were familiar with was the small ball lineup of the Scott Skiles era. Skiles liked to shake up a game by throwing Duhon, Gordon, and Kirk out there; Duhon-Kirk-Gordon-Deng-Wallace from 06-07 was +25 in 102 minutes with a win% of 61. Those aren't bad numbers, but anecdotally one remembers a debate over whether Skiles was too apt to go to small ball. Such tactics also can become less useful in a playoff situation: as teams come to know each other, they see the "joke" lineups multiple times and are quicker to react to the situation. The "acceleration" is dampened, and the game regresses to fundamentals, one of the reasons the best team usually wins in a seven game basketball series and the effect of being a team on a hot streak, like in baseball, seems far less important in the NBA postseason.