Layups: The Good Old “Fouling When Up By 3” Debate
Posted by Neil Paine on September 25, 2009
At TrueHoop today, Henry gave a nice rundown of the evidence on both sides of the age-old coaching debate abut whether or not to intentionally foul when leading by 3 points in the final seconds of a game, and presented some new research as well, conducted by Mavs advisor Wayne Winston:
In his new book "Mathletics," Wayne Winston -- Indiana University professor and consultant to the Dallas Mavericks -- tackles this issue, and offers some new data. First, he faults two of the main studies on the topic, by Lawhorn and Annis, for failing to account for the many different ways games can unfold. Neither, he says, embraces the idea that a game has an unknowable number of possessions left.Then Winston does something delightfully simple. He asks: Has it worked? He presents, for the first time I'm aware, the evidence:
"A student in my sports and math class, Kevin Klocke, looked at all NBA games from 2005 through 2008 in which a team had the ball with 1-10 seconds left and trailed by three points. The leading team did not foul 260 times and won 91.9% of the games. The leading team did foul 27 times and won 88.9% of the games. This seems to indicate that fouling does not significantly increase a team's chances of winning when they are three points ahead."
But perhaps David Thorpe had the best approach: "I would not have a hard and fast rule," he explained. "I'd have a fluid game to game strategy."
September 25th, 2009 at 3:29 pm
Does he break it down to see if the strategy worked, though? By that I mean, some of those wins come after the team hits a three at the buzzer to send it into overtime, or even tie the game? I think that would be a better indicator of what strategy is more successful.
September 25th, 2009 at 6:00 pm
Doesn't seem like it. That would be the logical next step, although I kind of like their approach of not worrying about how you won or lost at the end (just whether you did or not), too... Because you really have no idea what's going to happen at the end, it's almost better the way they did it, because it replicates the knowledge the coach has at the time of the decision. There are arguments for calculating the % of times the specific strategy worked, though (average points scored by the opponent on the possession when you don't foul vs. average points/poss. when you do foul). I can see reasons why you'd do it both ways.
September 26th, 2009 at 2:07 am
Stats aside, fouling when up by 3 simply doesn't make sense to me. You're taught at an early age that stopping the clock in the final minutes is the easiest way to allow the opposing team back in the game. If the player hits the 1st FT, and misses the 2nd, a player could easily grab an offensive rebound for an easy put-back. If I were a coach, I'd always advise my team to just defend the 3 as best they can.
September 26th, 2009 at 7:34 am
Don't those stats show that neither strategy is superior?
Also, the numbers don't nicely match up with the article by Lawhorn at 82games, where he says there were 32 occurances of the defensive team fouling to maintain a 3pt lead in the last 10 seconds ... of course he might not be looking at the same number of seasons ...
Both articles show that the conventional wisdom in the NBA is not to foul (almost 10-1 preference), of course, this assumes that the 27 instances of the fouls to defend such a lead were equally distributed throughout the last 10s. It may be that in the NBA teams tend to only implement this strategy closer to the end of the game, so the analysis of the historical data will not provide ANY answer to what is a better strategy as it does not evaluate the qn 10s or less properly. One might suggest that given the 10-1 ratio of data, that this is not the best way to get a reliable measure of the 'rare' event either.
September 26th, 2009 at 7:40 am
Further, one might ask if all 27 (or 32) instances of foulling in last 10s were implementations of a strategy to deliberately foul - could it be bad execution of the defence?
And what was the context of the losses, were they games where the fouller gave up an and-1, or 3 free-throws?
Like always, you get tit-bits, and never the whole story, unless you have all the PbP data for the last 3 or 4 seasons yourself ... which I don't :(
September 27th, 2009 at 12:21 pm
I find a contradiction in the claim that coaches must consider a number of details before making a decision, and the fact that 90% of them decide not to foul. That percentage suggests a definite rule designed in advance.
September 29th, 2009 at 5:17 pm
Anybody else remember this game: http://www.basketball-reference.com/boxscores/200303090NYK.html ?
Nationally televised Wiz v. Knicks in Mike's last season. Spree's doing a good job on 40 year old Mike, and then Shandon Anderson checks in and starts talking smack. Michael goes off. Bumps his chin on the court diving for a loose ball and then balls out his whole team because their 40 year old owner is the only player trying. Anyway the game comes down to a 3 point lead with very little time left (I'm thinking it was under 3 secs. Maybe more like 1 sec). Length of the court inbound hits Jordan right at the top of the circle and before he can even begin to turn, Sprewell wraps him up. Both arms trapped just grabs him. Mike hits the first FT. Misses the second hoping for a miracle tip in. Knicks win.
Don't know if that was a Van Gundy call or just Sprewell making a good play, but I'm with the people saying it's very much a matter of circumstance. That was the perfect situation where you could completely prevent any opportunity for the opponent to get a good look and avoid a 3 or 4 point play.