skip to main content

2006 Rewind: The Stats, and Rationale, of Phil’s Final-Hole Decisions

September 15, 2020 Mamaroneck, N.Y. By Lucius Riccio
Some ill-timed decisions on the 72nd hole cost Phil Mickelson a chance to win the 2006 U.S. Open at Winged Foot. (John Mummert/USGA)

It’s no secret that the U.S. Open is the most difficult, the most coveted of golf’s major championships. The course setup demands the most from golfers. It is not only the most challenging, it is the most mentally exhausting. With that stress and pressure, how good are the best golfers at making key, rational decisions as they face each shot?

When it comes to making a key decision in a U.S. Open, I’m afraid that the one that first comes to my mind as we approach this year’s U.S. Open is Phil Mickelson’s decision to use driver off the tee on the 72nd hole of the 2006 U.S. Open at Winged Foot. I say “afraid” because I was one of the millions of fans rooting for Phil to win and one of the most disappointed when he didn’t.

Let’s consider Mickelson’s decision to use driver on the 18th tee the last time the U.S. Open was contested at this year’s host venue. The general consensus was that Phil was “crazy” to use driver. Johnny Miller in the TV booth was practically yelling at him to put the driver away. But was his decision to use driver irrational?

Mickelson claimed he needed the distance provided by a driver to help get to the green of the 450-yard hole in two (No. 18 is listed at 469 yards for this year’s championship). It’s an argument that eventual champion Geoff Ogilvy used to justify his use of a driver off the tee, but that Jim Furyk rejected when he chose 4-wood. Both Ogilvy and Furyk were able to reach the green area in two shots.

Was Phil’s choice of driver based on a clear understanding of his personal trade-off of distance vs. accuracy between a driver and a 4-wood and an evaluation of those consequences? Did the possibility of a double bogey enter his mind? Also, did the fact that he had hit only two fairways all day, regardless of whether he hit a wood or an iron, play into his decision?

Mickelson was denied two previous major championships by players (Payne Stewart in the 1999 U.S. Open at Pinehurst and David Toms in the 2001 PGA at Atlanta Athletic Club) who chose more conservative layup shots and went on to get up and down to make championship-winning pars on the 72nd hole. Should those outcomes have influenced Phil’s strategy when he was in that same situation?

Taking all factors into account, was Phil being rational to choose driver?

Phil had three choices on the 18th tee. He could use driver (he had decided to carry only his “cut” driver so he had to play a cut), use his 4-wood, or hit a 3-iron. Of those three, the questions are: which gave him the best chance at making par for an outright win? Which would guarantee at least a bogey and no worse than a tie? Which would eliminate the chance of a double bogey and a loss?

What would a “Moneyball”-style data analytics review say of the choice? That day he hit only two fairways but parred more than half of the par 4s. He averaged par on the 18th over the three previous days – a birdie, a par, and a bogey. But is that relevant to Sunday? There likely is no perfect historical data to come up with those probabilities.

In cases with little or no data, management scientists often turn to a method called the “Delphi Technique.” This method aggregates the collective wisdom of a panel of experts. There is enough research to show that a panel of informed, independent, and representative thinkers provides, in the long run, the best decisions more often than any other structure, more often than any one expert and more often than an uninformed group.

I queried a number of experts using a modified Delphi technique to get a consensus of what those probabilities would look like. The half-dozen people on the panel all had significant golf credentials (one had played in a U.S. Open and other USGA events), were students of the game, all had technical degrees, all but one were present at Winged Foot, and together they represented a range of opinions on how Phil should have played the hole. By aggregating their responses, these are the estimates of the probabilities of the various outcomes, given the club used:

Score/Percentage For Particular Clubs Used Off Tee On 72nd Hole Of 2006 U.S. Open


Driver 53 Percent 33 Percent 14 Percent 4.61
4-Wood 52 Percent 38 Percent 10 Percent 4.59
3-Iron 38 Percent 53 Percent 9 Percent 4.71

Interestingly, looking at these computations, the driver was not automatically the worse choice. The expected or average score was estimated to be essentially the same. The 4-wood and 3-iron choices reduced his chances of losing it outright, but not by much. The difference between 14 percent and 9 percent may be too small to be significant. The most obvious conclusion was that a 3-iron was likely not a good choice. As it turned out none of the contenders used a 3-iron, supporting that conclusion and giving credence to the model.

As such from this analysis, it could not be said with certainty that Phil was irrational for using driver. The bad result was more likely due to poor execution during the shot rather than poor thinking before the shot. With that swing, he was probably going to hit that corporate tent with whatever club he used.

Now, given that he hit driver and it ended up in a terrible spot, was it rational to attempt to reach the green for his second shot? He had just pulled off a similar shot on the hole before, so it was doable. But was it probable? Analytically, the rationality of this shot was more difficult to figure out, since there was little historical data of play from that part of the course. But it appears that he was not thinking rationally by not considering his position, the alternatives, and their respective risks and uncertainties.

The rational, justifiable thing to do was probably not to try the heroic play, but get back to safe harbor and reduce his risks of an all-out failure. Was it possible to play safe? Did he have any other choices? Did he even consider that there was another option? All the data from my research of hundreds of golfers indicates that there is a reason it is called the fairway, and getting back there is always the smart move. (Read: Payne Stewart and David Toms, in 1999 and 2001, respectively.)

Phil’s “rational” choice on the tee which resulted in a bad outcome, followed by an “irrational” choice for his second which resulted in an even worse result, is different than Jean van de Velde’s disaster in the 1999 British Open at Carnoustie, which has to be considered the “Edsel” of all golf decision-making. Interestingly, using rational analysis, van de Velde irrationally used driver off the tee, but the result was “good” (luckily) in that it was in a fairway (albeit the wrong one) with a clear shot to the green. Van de Velde’s second shot I propose was “rational,” but it turned out badly, far worse than anyone could have predicted and about as unlucky as could be, because his 4-iron cleared the water and hit the grandstand in the only possible place that could cause the ball to bounce backward across the burn.

Phil was not, and is not, an “idiot,” as he said after the round. His thinking might not have been completely rational, but that is not the same thing. His free-wheeling approach to the game is what makes him great and is also what occasionally holds him back. Yet even at age 50, he continues to entertain legions of fans who watch and wonder what he will do next.

Lucius Riccio is a lecturer at Columbia University who has been a USGA committee member since 1979. He is the author of “Golf’s Pace of Play Bible” and a statistical advisor to Golf Digest.