How to Win a GOAT Argument: A Statistical Framework for Tennis History

The standard advice for ending a GOAT argument is two words: count the Slams. Whoever has the most majors wins. It is clean, it fits in a tweet, and it is the reason most tennis history debates die in the same place they started — with one person reciting a number and the other person walking away unconvinced.

We want to take that advice seriously, because it is not stupid. It is roughly right. The trouble is that "roughly right" is exactly the zone where confident people lose arguments to slightly more careful people. This piece is about where Slam-counting holds up, where it quietly breaks, and how to build a position that survives contact with a skeptic. Not a verdict. A framework you can defend.

Why "just count the majors" works better than its critics admit

Start with the steelman, because the number deserves it. Grand Slam titles correlate with almost everything else we care about. A player who wins 20-plus majors has, by necessity, beaten elite fields over two weeks, on the sport's biggest stages, repeatedly, across many years. You cannot fluke your way to that total. Variance gets squeezed out over a long enough sample.

So the major count is a useful proxy. It compresses thousands of matches into one integer, and that integer tracks real greatness most of the time. When someone leads on Slams and weeks at No. 1 and head-to-head and Masters-level titles, the argument is essentially over, and people who keep arguing are usually arguing about something other than tennis.

The problem is that the headline cases — Novak Djokovic, Rafael Nadal, Roger Federer, and a step back, the field around Pete Sampras and Björn Borg — do not line up that cleanly. The proxies disagree. And when your proxies disagree, the number you happen to recite first stops being evidence and starts being a preference dressed as evidence.

The denominator problem nobody puts in their argument

Here is the first crack, and it is the one that wins debates.

A Slam count is a numerator. It tells you how many majors a player won. It says nothing about how many they entered, how long they were healthy enough to compete, or how often they showed up at all. Borg retired at 26. He played his last major in 1981 and finished with 11. Had he played another five seasons in his prime, the integer everyone recites would be different, and the integer is the entire argument.

This is why "Slams won" is a weaker statistic than Slams won per major entered, or per prime season, or per final reached. A player who reaches 30 finals and wins 22 is a different animal from one who reaches 24 finals and wins 22, even though the raw trophies are close. The first player got more chances and converted fewer of them. The second was nearly automatic once they arrived.

We are not saying conversion rate is the "true" stat. We are saying that the moment you pick Slams won over Slams converted, you have made a value judgment — you have decided that accumulation matters more than efficiency — and you almost certainly made it without noticing. That is the thing to notice. Most GOAT arguments are two people defending different hidden value judgments while both believing they are reciting facts.

How a "dominance" number actually gets built

Fans throw around words like "dominant" as if they were measured. Sometimes they are. It is worth walking through how an analyst would actually construct a dominance figure, in the order it happens, because seeing the steps shows you where the choices hide.

Step one: pick the unit. Matches? Sets? Tournaments? Weeks ranked No. 1? Each unit answers a different question. Match win percentage rewards consistency. Weeks at No. 1 reward sustained peak. They are not interchangeable, and a player can dominate one while looking ordinary on another.

Step two: pick the window. Career totals favor longevity. Best-five-consecutive-years favors peak. Djokovic's 2011, 2015, and 2021 seasons look superhuman in a peak window and merely excellent in a flat career average that includes injury years and the slow start everyone has. Choose the window and you have half-chosen the winner.

Step three: adjust for the field, or decline to. This is where it gets hard, and we will come back to it. You can leave raw numbers alone, or you can try to weight wins by opponent quality. Both are defensible. Neither is neutral.

Step four: aggregate. You sum or average across the window. The instant you do this, you have assumed that a match in a first round and a match in a final count the same, unless you weight them — and weighting is another buried choice.

By the time you have a single "dominance score," you have made at least four decisions, each of which a reasonable person could have made differently. The number on the screen looks objective. The path to it was a series of preferences. This is not a flaw in analytics. It is what analytics is. The honest move is to state your choices out loud so the other person can argue with the choices instead of the conclusion.

Era adjustment: the thing everyone demands and nobody can deliver cleanly

The most common counter to any Slam count is "different era." It is also the most abused word in tennis history, because people invoke it as a trump card and then refuse to specify what they mean.

So let us be precise about what era adjustment can and cannot do.

What we can measure has changed. Racquet technology shifted from wood to graphite to modern poly strings, and that shift altered the physics of the rally. Court speeds have converged — the gap between a fast grass court and a slow clay court is narrower now than it was in the 1990s, which is part of why a single player can win on every surface in a way that was rarer before. Sports science has extended prime years; the top men now compete seriously past 35, which simply was not normal in Borg's day. These are documented, directional changes. We can describe them.

What we cannot do is run the experiment. We cannot put 1980 Borg on a 2015 court with 2015 strings against a 2015 field and read the score. Every "he'd dominate today" or "he'd never survive today" claim is a counterfactual with no data behind it. The honest version is that surface and equipment changes make raw cross-era comparison genuinely uncertain, and the uncertainty does not resolve in favor of whoever feels strongest about it.

There is one era variable we can at least gesture at with numbers: field depth. A reasonable proxy is how often the same handful of players reach the late rounds. When three men split most of the majors for fifteen years, you can read that two ways — either the era was historically deep and three giants rose above it, or it was top-heavy and thin underneath. Both stories fit the same data, which is exactly why "the era was weak" and "the era was the strongest ever" are argued with identical conviction by people looking at the same bracket.

The mechanism matters more than the slogan. If your opponent says "weak era," the productive response is not "no it wasn't." It is "weak by what measure — variance in finalists, ranking points behind No. 1, depth of the round of 16." Force the measure. The argument either becomes specific or it evaporates, and both outcomes are wins.

The three value judgments hiding in every ranking

Underneath the data, three forks in the road decide most GOAT positions before any number is consulted. Naming them is most of the battle.

Peak versus longevity

Is the greatest player the one who hit the highest ceiling, or the one who stayed near the ceiling longest? A player with a blinding three-year peak and an otherwise good career is a different proposition from a player who was top-three for fifteen years and rarely transcendent. There is no equation that resolves this. It is a taste. Borg-versus-the-modern-totals arguments are almost always this fork in disguise — a short, incandescent career against a long, accumulating one.

One surface versus all surfaces

Nadal's clay record is, statistically, the most lopsided dominance of any player on any surface in the Open era — the win rate at Roland Garros sits in a range no one else approaches anywhere. Does specialized supremacy outrank balanced excellence across all four majors? If you weight versatility, you rank one way. If you weight peak ceiling on any single surface, you rank another. Same player, two defensible placements, depending entirely on a value you chose before you opened the spreadsheet.

Trophies versus the eye

Some greatness shows up in the ledger and some shows up only on the court — shot quality, the way a player bent matches without always winning the tournament. The ledger is more reliable and less romantic. The eye is more honest about what you actually watched and less defensible in an argument. Most people use both and pretend they used only the first.

You do not have to resolve these forks. You have to declare which side you are on. "I rank for longevity and versatility, so my list looks like this" is an unbeatable opening, because your opponent can no longer catch you in a contradiction. They can only disagree with your values, and disagreeing with stated values is not the same as refuting an argument. That asymmetry is the whole game.

A field guide for stress-testing any GOAT claim

Here is the practical part. When someone hands you a ranking — or when you are building your own — run it through these checks before you commit. The goal is not to find the right answer. It is to find out whether a claim is load-bearing or just loud.

Five checks. A claim that survives all five is a real position. A claim that fails three of them is a feeling wearing a jersey.

The most useful single move in the table is the second one. Almost every loud GOAT take rests on a raw total and falls apart the moment you ask "out of how many tries, and over how many healthy years." Not because the total is wrong, but because the person reciting it has never once divided it by anything.

A second-order tip: when you find yourself certain, check which of the three value forks you are standing on, and ask whether you would still be certain if you stood on the other one. If your conclusion survives swapping peak for longevity, it is strong. If it collapses, you do not have a tennis argument — you have a preference, which is fine, as long as you say so.¹

A more honest version of the advice

So return to "just count the Slams." It is not wrong. It is incomplete in a specific, fixable way. The honest rewrite goes like this: count the Slams, then divide by something — entries, finals, prime seasons — and then say out loud which you value more, the height of the peak or the length of the climb. Do that, and you have not ended the debate. You have made it a real one.

That is the part the slogan skips. The number was never the argument. The number was the start of the argument, and the people who win these things are the ones who know what their own number actually assumes. Tennis history rewards the player who showed up over and over; the debate about it rewards the fan who knows the difference between a fact and a value, and is willing to name which one they are holding.

The honest rule, the one you can use tonight: never recite a total without naming what you divided it by and which kind of greatness you decided to count.