Dan Lewis at Armchairgm.com posted a story after the NCAA Championship bemoaning the fact that the game was much closer than it should have been because Georgia Tech gained an advantage by frequently fouling UConn during the final minutes. More specifically, because UConn is a notoriously poor free throw shooting team, Georgia Tech gains an advantage by breaking the rules ('intentionally' fouling) and forcing UConn to convert their free throws. I disagree with the premise of Dan's argument and I'll explain below. First, here's the post from Armchairgm.com:
"If you saw tonight's UConn v. G'Tech game, you undoubtedly heard the commentators suggesting that Tech start fouling Huskies as the game came closer and closer to a close. To wit, at one point, the color man (Billy Packer?) noted that Jim Calhoun was in a tough position, as such a strategy would all but require that Coach C pull Taliek Brown from the game. Here you have the Huskies starting 2-guard being taken out of the game not because he's failing in his role, but because the Yellow Jackets are violating the rules. One really cannot give a good explanation as to why such rule breaking should result in a net loss for the victimized team."After initially reading the post, I had a couple of comments. First, it wasn't clear to me how Georgia Tech committing fouls ('rule breaking') results in a net loss for UConn. If Taliek Brown is a below average free throw shooter compared to his teammates, it makes sense to replace him with someone who is a better free throw shooter. While you lose something in the way of ball handling, you gain something in the way of free throw conversion percentage.
Dan argues that Brown is taken out 'not because he's failing in his role, but because the Yellow Jackets are violating the rules.' Well, Brown is failing in his role -- if he could shoot free throws he wouldn't be taken out. Is Manny Ramirez being punished unfairly from an offensive perspective (he's one of the best hitters in baseball) because in the late innings he's replaced by someone who is better defensively? While the Red Sox lose something offensively, strategically they think it's to their advantage to strengthen their defense. This is exactly what Calhoun is doing by replacing Brown.
Second, it's important to remember that ball handling is secondary in situations where the losing team must foul. From the perspective of the offense, the most important pass is the first one (the inbounds), because after that all a player needs to do is hold the ball until they are fouled -- it's very rare that a player has to beat five men down the court.
Dan argues that substituting a better free throw shooter for Brown isn't a net gain for UConn. Here's his take:
"How can this be a net gain for UConn? If it were, why would Tech go for the foul? The problem with that analysis is that it takes almost no ability whatsoever to commit a foul. With Brown in the game, GT's optimal strategy is to put in guys who are good at stealing the ball, and if they can't do that quickly (that is, on the in bounds pass), foul. If Calhoun takes Brown out for a lesser ballhandler who, admitedly, is a better free throw shooter, Georgia Tech's optimal strategy is . . . exactly the same. It has to be gain for GT."What Dan overlooks is that when trailing teams foul in the waning minutes of a game they do so primarily to stop the clock and secondarily in the hopes that the fouled team misses the free throws. If the fouled team converts their free throws it's not a net gain for the trailing team. They foul in the last few minutes because they have to. Their primary goal is to stop the clock and then make up the points their next time down the court.
Think of it this way. Let's say that the semifinal game between UConn and Duke is close with Duke leading and UConn being forced to foul in the last minutes of the game. Is it a net gain for Duke by having J.J. Redick (95% ftpct) handle the ball in the waning minutes of a close game instead of Chris Duhon (72% ftpct)? Conceivably Duhon is a much better ball handler than Redick, but I'm guessing any coach asked will take Redick every time. Especially when you consider that for UConn to gain an advantage from using the fouling strategy, they have to come down and convert three's.
But there's a wrinkle. J.J. Redick is one of the best free throw shooters in the country and we can estimate the probability that he makes both of his free throws. Redick shot 95.3% from the line so the probability that he makes both free throws is 0.953*0.953 = 0.91 or 91%. Now, given that after Redick converts his free throws, the trailing team has to come down and convert a three to make up for the two made free throws and gain ground. UConn ranked fifth in the country in 3-point shooting, and the probability that they convert the shot is 40.2% (which is just their 3-pt shooting percentage). We can estimate the expected value ('expected value' is what will happen on average) of Redick making his free throws when compared to the expected value of UConn making their three's. Redick's expected value is 91%*2 = 1.82 points for every two free throws taken. UConn's expected value is 40%*3 = 1.21 points for every 3-pointer taken. It's very clear that fouling Redick is a net loss for UConn because on average they will be giving up more points than they will score (1.82 points for Duke compared to 1.21 points for UConn).
Now if instead, Duhon gets the inbounds pass and is fouled, it actually is in UConn's best interest to foul him. Because he shoots 72% from the line, the probability that he makes both shots is 0.52 (or 52%) and Duhon's expected value is 52%*2 = 1.04 points for every two free throws taken.
Of course someone may argue that it might be better for UConn to trade 2-pointers for free throws instead of taking all 3-pointers (because 3's are tougher to convert). Well if UConn fouls Redick we know his expected value is 1.82 points per 2 foul shots taken. UConn shoots 48% from the field, which means that their expected value is 0.48*2 = 0.96 points per shot taken -- it's still a net gain for Duke.
So the real story here is that the observant coach should shuffle his lineup in such a way that the probability of any of his players making two free throws is greater than the probability that the trailing team converts either a 3- or 2-pointer the next time down the court. And that's exactly why Taliek Brown shouldn't be in the game down the stretch -- not because he's unimportant to UConn's success, but because he can't shoot free throws.
