Late last night, the guys over at the Harvard Sports Analysis Collective posted a study they put together on one of college basketball's most controversial strategies: whether or not to foul when up three late in the game.
The results they found are quite interesting. Essentially, the study says that there is no discernible difference in winning percentage for teams that foul up three and for those that decide to play defense. In fact, teams that fouled had a lower winning percentage than teams that opted to rely on their defense during the 2009-2010 season. (I strongly urge you to read the study, especially those of you that are mathematically inclined.)
These days, the general consensus -- correct or not -- has seemingly swayed towards fouling. Don't give your opponent a chance to tie the game with one shot. Force them to get an offensive rebound off of a free throw and score on a putback. The study debunks that line of thinking.
Like any study involving sport, the decision is not as cut-and-dry as numbers are. There are plenty of factors that go into a decision like this. Does the other team have any three-point shooters? How disciplined is your team defensively? Do they understand that they can't foul a player in the act of shooting a three? Who is home and who is away? How good are their free throw shooters? Do you have confidence in your team being able to make two free throws should the situation arise?
The biggest question, however, is how much time if left in the game.
And unfortunately, the HSAC didn't factor in for time left on the clock in their study. The only mention of time left is in one of the comments on the post, and it says that the author believes that most of occur with less than 20 seconds left.
The way I see it, this skews their results.
Let me explain.
Fouling when you are up three is only a valid course of action in a very specific situation. The ideal is when there is somewhere between three and eight seconds left in the game when your opponent has possession, down three points.
If there is much more than eight ticks left, then ratcheting down for one defensive possession is probably the best option for every team. Let's say you are the coach of the team up three with 12 seconds, and you decide to foul. (When a coach makes the decision to foul when up three, more often than not it happens within a second or two of the ball being inbounded to avoid the possibility of the player with the ball being able to draw a foul in the act of shooting a three.) If the other team hits both free throws, they are going to have time to foul you and force you to make two free throws. If your guys hit both, then you are right back into the same situation. Miss one, and your opponent has the ball down just one or two points. If you foul and the other team hits the first and misses the second (or if they miss both/the front end of a one-and-one), then there will be enough time for an offensive possession should they grab an offensive rebound. The best option is to play tough defense and hope that you force a miss.
With fewer than three seconds left, the odds of your opponent getting a good look at a three -- especially when they have to go the length of the court -- is minimal. In this situation, the best strategy usually is to play a soft full court defense, force your opponent to catch the ball in the back court without a head of steam, and hope that they miss a very low-percentage three, usually of the 50-foot runner variety.
That means that the only time where it makes sense for a team to foul is when they are up three in that four-to-eight second range. When there is enough time on the clock that your opponent can get a decent shot at a game-tying three, but not enough time for multiple ensuing possessions. (In other words, there is only enough time left for the opponent to try and get an offensive rebound off of an intentionally missed free throw.)
Now, based on the study and what the authors mentioned in the comments section, they had no way of determining how much time was left on the clock when the possession began. (The time when the possession begins is the only time that matters. You don't intentionally foul at the five second mark after playing 15 seconds of defense.) Without determining this variable, the authors are including a number of data points where intentionally fouling is not only sub-optimal, it isn't even a consideration.
I don't know the lingo for statistics, but they are analyzing the wrong stuff.
That said, I don't want to do anything to discredit this study. It is very interesting work, and certainly is a great piece of information when, undoubtedly, this situation arises during the upcoming season. And I hope that the guys over at HSAC can figure out a way to factor in time, because I would love to see the results. (Hint, hint!!)
If anything, what you should take out of this piece is that regardless of strategy, your team shouldn't lose a game up three on a final possession.