Part I of Brian Sommer’s three-part series on Moneyball, golf’s numbers gane, and what statistics truly show
Data Can Describe Greatness. That Does Not Mean It Explains It
Baseball had statistics long before Bill James. That is worth remembering, because the popular version of Moneyball makes the revolution appear almost embarrassingly simple: old baseball men trusted their eyes; young baseball men trusted their spreadsheets; and eventually arithmetic defeated intuition.
It is a wonderful movie. It is also a much tidier story than reality. Baseball was already a game dominated by statistics. Batting averages, runs batted in, home runs, earned run average, pitcher wins, stolen bases, errors. Scouts carried stopwatches. Newspapers printed box scores every morning. General managers had files stuffed with statistics. James – the Kansas born father of sabermetrics – did not discover numbers. What he challenged was the authority baseball had given certain numbers and the explanations attached to them.
Were RBI really a clean measure of individual offensive ability, or did they partly depend on how often teammates happened to be standing on base? Did a pitcher’s win total tell us how well he pitched, or did it also reflect run support, defense, bullpen performance, and circumstance? Was batting average really the best representation of offensive value? Was a stolen base valuable simply because everyone applauded when it worked, or should we also count the cost of getting thrown out?
James’s great contribution was not “more data.” It was a question: What does the evidence actually allow us to say?
That mighr be the Moneyball lesson golf needs most to learn. Golf has never possessed more data, but I am not convinced we have become equally good at knowing what the data mean.
Golf Has Its Moneyball Toys
Launch monitors can tell us club speed, ball speed, path, face angle, dynamic loft, spin axis, attack angle, carry, peak height, landing angle, and impact efficiency. Force plates can measure pressure patterns beneath a golfer’s feet. 3D systems can map pelvis rotation, thorax rotation, side bend, flexion, extension, sequencing, and angular velocities. High-speed cameras can freeze the golf swing at practically any point we desire. Putting systems can measure launch, skid, roll, face rotation, impact location, and path.
In a world of golf, we can measure golfera is looking at. We can measure how long they look at it. We can measure heart rate. We can record routines. We can catalogue thoughts. We can compare a 15-handicapper to a PGA Tour database and identify nearly every measurable way in which the two differ.
Wonderful. I use data. I find much of it fascinating. But there is a sentence I think golf needs to repeat more frequently: Precision of measurement is not certainty of explanation. The launch monitor mightt tell me precisely what the club did. That does not mean it has told me why the golfer did it. A force plate might describe precisely how pressure moved. That does not establish that the measured pressure movement caused the successful swing. A photograph might show precisely where an elite player positioned the clubface through impact. That does not mean wehave discovered the source of that player’s greatness. We might simply have become exceptionally good at describing greatness after it has already occurred. That is not the same thing as explaining it.
Scottie Scheffler and the Photograph of Greatness
Consider Scottie Scheffler. He wins by eight shots. Naturally, everyone wants to know why. The Golf Channel recently showed a still image of Scheffler through impact and pointed to the extraordinary stability of the clubface. What wasn’t discussed, however, was how little the face and loft appeared to change through the impact interval.
It is an interesting observation. It might be entirely correct. A stable presentation of the face and predictable loft through impact would quite reasonably contribute to predictable ball flight. But then comes the sentence that interests me: “That’s why he’s so good.” Is it? Or is that what being extraordinarily good happens to look like when we photograph Scheffler?
Those are different claims.
Suppose we measure Scheffler perfectly. We know the exact face angle. The exact dynamic loft. The exact pressure trace. The exact sequencing. The exact movement of his pelvis. The exact path of his hands. The exact behavior of the shaft. We now possess an extraordinary description of what Scottie Scheffler does when he hits a golf ball. We still have not necessarily answered the deeper question:
Why is Scottie Scheffler capable of doing it so exceptionally, and so often? That is the problem. We photograph greatness, measure everything in the photograph, and then announce that we have explained greatness. We haven’t. We have described it.
What About Everyone Else Who Can Do It?
There is an even larger problem with the explanation. Scottie Scheffler is hardly the only golfer in the world who can control a clubface. The PGA Tour is populated by absurdly talented players. They hit the center of the face. They control trajectory. They control loft. They know their distances. They have routines. They work with outstanding swing coaches. They employ trainers, sports psychologists, physiotherapists, nutritionists, and statisticians. They use launch monitors, force plates, and high-speed video. They practice relentlessly. They travel with teams devoted to extracting every possible fraction of performance. Many of them can produce impact conditions that would astonish almost every golfer who has ever played the game.
Yet they are not all Scheffler. That matters – because if the characteristic we identify in the exceptional performer is also present in many performers who are not exceptional to the same degree, then that characteristic cannot, by itself, explain the difference we are actually interested in. This is one of the more simple tests of a performance claim – and one of the least frequently applied.
Someone says: Great players do X. Fine. I want to know: How many less-great players also do X? And: How many great players do not? Those two questions can become extremely inconvenient. Great players have routines. So do thousands of mediocre golfers. Great players visualize. So do players who miss cuts. Great players are confident. So are golfers who shoot 92. Great players work hard. So do countless players who never become great. Great players use coaches. So do players who never reach the Tour. Great players produce certain biomechanical patterns. Perhaps. How many outstanding players organize their movement differently? The existence of X among successful performers makes X interesting. It does not automatically make X the source of their success.
The Survivorship Problem
This is where golf continually falls into survivorship bias. We study the winners. We interview champions. We photograph elite performers. We catalogue what they do. Then we look for commonalities. Perhaps 10 great players use highly structured routines. We conclude that routines contribute to greatness. But we have looked at only one side of the equation.
How many players ranked 300th in the world also have meticulous routines? How many golfers who never made it out of developmental tours had the same? How many unsuccessful players visualize, journal, meditate, breathe, train, lift weights, practice deliberately, work with psychologists, and follow every fashionable performance system available? Unless we examine that comparison group, we have not identified what separates greatness from everything else. We have identified something greatness happens to be accompanied by. That is very different. And this distinction applies far beyond golf.
Successful entrepreneurs wake up at 5 a.m. Therefore, waking at 5 a.m. creates success. Successful executives journal. Therefore,journaling produces success. Great athletes visualize. Therefore, visualization creates greatness. The question almost nobody asks is: How many people wake up at 5 a.m., journal, visualize, work relentlessly and remain entirely ordinary?
Their existence matters. They are part of the evidence, too.
Ted Williams Was Still Ted Williams
Baseball offers an even cleaner example. Consider Ted Williams.\, generally regarded as the best hitter in baseball history. Study him endlessly. His physical gifts. His strength. His eyesight. His strike-zone discipline. His bat speed. His knowledge of pitchers. His obsessive study of hitting. His mechanics. His preparation. His ability to recognize pitches. His approach to different locations. You can spend a lifetime studying Williams and learn an extraordinary amount about hitting. You should.
But baseball has had other magnificent hitters. Hall of Famers. Players who were bigger. Players who were stronger. Players who studied hitting obsessively. Players with tremendous eyesight. Players with exquisite strike-zone judgment. Players with magnificent mechanics. Players who practiced relentlessly. They still aren’t Ted Williams. That does not mean there was some mystical substance called “Tedness” floating around inside him.
It means our measurements and explanations should retain some humility. We may understand many contributors to greatness without possessing a complete causal account of greatness. There is nothing intellectually embarrassing about saying that. In fact, the embarrassment begins when we refuse to. We can say: “This mattered.” “This appears associated with his skill.” “This might have contributed.” “This helps us understand part of what he did.” All reasonable. But: “This is why Ted Williams was Ted Williams.” That is a much larger claim. And it deserves a much larger burden of proof.
The Part Moneyball Conveniently Simplified
The Moneyball story focuses heavily on the Oakland A’s finding undervalued players through data. Scott Hatteberg. Chad Bradford. On-base percentage. Finding value where traditional baseball scouts failed to see it.
All true. And fascinating. But the Oakland Athletics in that Moneyball story also had some extraordinarily conventional talent. Barry Zito. Tim Hudson. Mark Mulder. Miguel Tejada. Zito won the Cy Young Award. Tejada won the MVP. The rotation was one of the best in baseball. Those men were not statistical tricks. They were terrific baseball players. That does not undermine the Moneyball thesis. It improves it. Moneyball analytics helped Oakland understand value differently. They helped identify inefficiencies. They helped a financially constrained organization allocate resources intelligently. But the spreadsheet did not create Barry Zito’s curveball. The Moneyball numbers could help describe his value. They could identify how much he contributed. They could compare him with other pitchers.
Zito still had to throw the baseball. That distinction is enormously important. Moneyball ball did not prove that numbers create talent. It demonstrated that numbers could help us recognize talent and value more accurately. Recognition is not creation. Description is not source. Prediction is not causation. Golf should tattoo thosedistinctions somewhere near the launch monitor.
Bill James Was Really Challenging Certainty
The most important thing about James was not that he distrusted scouts. It was that he distrusted unearned certainty. Baseball people had spent their lives in the game. They had watched tens of thousands of innings. They had evaluated players for decades. Their experience deserved respect. But James refused to grant them the privilege they most wanted: exemption from evidence.
That strikes me as enormously relevant to coaching. “I’ve coached for 30 years.” Fine. “I’ve seen this hundreds of times.” Good. “My players improve.” Interesting. “My clients tell me it works.” Worth hearing. None of those statements automatically establishes: Therefore, my explanation of why they improved is correct. This is where experience so often inflates. Observation becomes explanation. Explanation becomes causation. Causation becomes prescription. Prescription becomes necessity. And somehow each step happens without anyone noticing that a new claim has been introduced.