Archives For Basketball Strategy

Who Starts?

Kurt —  November 28, 2007

It’s been the question in the comments recently — start Bynum or bring him off the bench?

He’s the best offensive center the Lakers have by far, and plays good defense for the most part (opposing centers are shooting just 37.5% and have a PER of 13.5). On the other hand, he averages 5.4 fouls per 40 minutes and starting him means risking having to sit him with foul trouble at key points. Plus, without him the second unit that was such a force for the team early on is dramatically weakened (and the current starting unit with him is not tearing it up).

However, as Rob L. brought up, the question here is really bigger than just Bynum:

Should a coach go with the best starting five athletes, or find the best complementary starting five? Is the answer to this question absolute? Or does it change from team to team? Shouldn’t a coach be able to take the best five athletes and meld them into a cohesive starting five?

I’m going to email this to a couple basketball people and see what they think, and post the responses as updates. What are your thoughts?

UPDATE #1: From Tom Ziller of Sactown Royalty and AOL Fanhouse:

I think you absolutely have to go with the best complementary starting five. Look at San Antonio, with Michael Finley starting over Manu Ginobili. Ginobili is one of the best five two-guards in the league (and Finley might be in the bottom quartile among rotation-level SGs), but he fits the team better coming off the bench and not necessarily playing all of his minutes with Tony Parker and Tim Duncan. Manu played 62% of his minutes with Parker on the floor last year. In that other 38%, the Spurs had a net per-48 +/- of +11.5, which is comparable with the Manu + Parker +/- of +13.3. Finley’s +/- without Parker: +1.89 points per 48 (versus +6.9 with Parker). It’s not just the starting lineup you have to worry about when constructing a rotation — those eight minutes which straddle the first quarter break count too. And this says nothing of the benefits of subbing a roleplaying rebounder (Chuck Hayes?) for a scoring power forward (Luis Scola) in the starting lineup, even though Scola’s a better player. A basketball starting lineup isn’t like a baseball batting order; the pieces need to fit. Starting your five best players usually doesn’t do that.

UPDATE #2: Next up are some great points from Henry of the legen-(wait for it)-dary True Hoop:

Oh man, I am SQUARELY against the “best five players” approach, and so is Gregg Popovich!

I could write a novel about why. Here are some highlights:
• Who starts is not about who deserves honor. It’s a strategy.

• Pretty much no matter what, you always need at least one big man and at least one real deal point guard. On a lot of teams, that rule alone would defeat the “best five” argument.

• As Daryl Morey explained in a Houston Press interview recently, thanks to substitution patterns, you can think of an NBA game as a series of mini-games. Every time there’s a different lineup, a new mini-game begins, and it lasts until the next substitution. As coach, you have to have a plan to win more than your fair share of those mini games. You have to be able to put effective combinations on the floor at all times. So your beginning game strategy must not crap upon our mid-game strategy. Which is why there is a long history of great players — Manu Ginobili, Adrian Dantley, Vinnie Johnson, Bill Walton in Boston — coming off the bench.

• In trying to win a team game, you have to find combinations of players that work. For instance, Shane Battier is not all that great compared to a lot of players, but when he’s on the floor, statistics show the team is good. 82games has examples of player combinations that are highly effective, like last year Kyle Lowry and Mike Miller were, per 48 minutes, the best combination in the NBA. If the sample size is big enough, that kind of stuff is all you need to know. Guys who beat the other team consistently are your best lineup, whether you consider them your best players or not. So, as coach, I think you need to seek lineups that demonstrate they can perform at a high level together. Not lineups that look like you think lineups are supposed to look, feature the highly paid players, etc. If you want to win, play the guys who are in the habit of scoring more points than the other team.

UPDATE #3: Mike from Knickerblogger says the Knicks are struggling with some of the same issues:

In New York, this is the exact problem the Knicks have. For argument sakes, let’s assume that the starting five are the best five players on the team (Marbury, Crawford, Richardson, Zach, and Curry). While there are some that might disagree with this statement, there are enough Knick fans that would play Zach and Curry over Lee, and just as many that would play Marbury/Crawford over Nate Robinson. (And I’m sure Quentin Richardson’s mom still thinks her son should be starting over Balkman). In any case Isiah has put these guys on the court (when possible) at the start of each game. The problem is these players don’t complement each other in the least, and at least 4 of the 5 starters are poor defenders who need the ball to be effective. The Knicks would be better served to play some of their guys that can contribute without the ball and can play defense (Balkman, Lee, Jeffries, Robinson, etc.)

Basketball isn’t baseball where for the most part it doesn’t matter how you fill in your lineup card, as long as you’re not putting your pitcher up after Barry Bonds. Baseball matchups are mostly mano-a-mano events. Think about it, how many times do basketball players acknowledge a good pass that led to an easy basket? Dozens in each game. Now when was the last time anyone hit a homerun and credited the batter behind him for setting up the pitcher? Ummm never? In basketball the parts on the court have to fit together. A prime exampltes are the Spurs who bring Manu off the bench.