The most frustrating thing about GOAT debates, aside from the absence of civility, is the lack of definition. It’s not clear what we’re discussing. It should be about the MAP — Most Accomplished Player. It needs to be data-driven with agreed-upon criteria.
First let’s go over what shouldn’t be considered, not just for the MAP but in any discussion about whether Player X was better than Player Y. There should be no style points, no aesthetics, no impressions, no feelings. I’m talking to you, Federer fans. If we want to have a debate over who’s the most aesthetically pleasing great player ever, Federer would probably win easily. (Shout-out to Suzanne Lenglen on the women’s side, with honorable mention to Evonne Goolagong.)
Likewise, you can’t take away points for lack of style or for technical weaknesses. I know people who discredit Steffi Graf because she spent a career running around her backhand. But the way a player plays is not an end; it’s a means to an end, with the end being the results. If Graf can win 22 majors avoiding her backhand, so be it. I have a friend who thinks Serena Williams can’t be considered the greatest-ever women’s player because he says her footwork is often poor. Even if that’s true, it’s ridiculous. It’s the results that matter. If one wants to argue that she could have won 30 majors by now if her footwork were better, fine, but that’s a different debate.
There are also no points for personality, or being a credit to the game. Federer doesn’t get a bonus for being a globally admired tennis ambassador. I mean, he does, but not in a MAP debate. John McEnroe and Jimmy Connors don’t get penalized for boorish behavior, and Ivan Lendl doesn’t get docked points because some people found him to be dour. I have found that when fans debate the merits of those 1970s-‘80s stars, it turns into a personality discussion. People were understandably drawn to or away from McEnroe and Connors. You’re allowed to have your favorites, but that doesn’t change the data if we are judging who was more accomplished as a tennis player.
So what should the criteria be? I haven’t tried to devise a points system, though I would love to hear suggestions. What I’ll do here is offer broad guidelines that should be accounted for in data-driven MAP criteria. First, we need to allow for different kinds of accomplishments. I see three broad categories: Dominance, consistency and longevity.
I would weigh dominance the most, followed by consistency and then longevity. But I haven’t thought through how much weight I would give each. One challenge here is that consistency can overlap with both dominance and longevity. Are Federer’s 10 straight Slam finals from 2005-2007 (which is pretty insane) an achievement of dominance or of consistency? Consistent dominance? Dominant consistency?
There’s a reason why a calendar year Grand Slam hasn’t been accomplished since 1969 on the men’s side and 1988 for the women. It’s really, really hard. That dominance accomplishment adds a lot of points on the men’s side for Rod Laver, who did it twice. You can argue that he wasn’t playing against all the best players the first time, in 1962. But he was when he won the calendar year Grand Slam seven years later, after the start of Open tennis. I would rank Novak Djokovic’s four straight Slams very high in the dominance category. But I would give even more points to Federer’s run of winning three Slams in three of the four years of 2004-07. Here’s an often-forgotten stat that’s mind-blowing: Eighteen Slams were played from 2003 Wimbledon through the 2007 U.S. Open, and Federer won 12 of them. That’s Djokovic’s entire Slam haul, and a sustained period of dominance that no one has approached before or since. I would also give a lot of points to Pete Sampras for his six straight years as year-end No. 1. That’s dominance. (And a stat I think is often overlooked because of the recency effect. Poor Pete. He retired in 2002 considered the greatest ever, and within 15 years it was widely believed that at least two, and possibly three, players had surpassed him.)
There should be some points awarded for domination of one event, like Nadal at the French (10) and, to a lesser extent, Federer at Wimbledon (8 plus three runners-up). But there should also be points for domination of multiple majors. I remember thinking after Federer won the U.S. Open in 2008 that the same guy won both the U.S. Open and Wimbledon five consecutive years, and that he could just quit and be considered the greatest ever right then. (This is for another blog entry or debate, but I think Federer weakened his legacy with the public in the post-2009 period, because a lot of people have forgotten how dominant he was before 2008. He is remembered by many younger fans as not being the best player in the world from 2010 to 2016.)
But I’m digressing. On to longevity. In this discussion, longevity means being great — not good — for a long time. I wouldn’t have given Federer a lot of credit for longevity before his three Slams starting in 2017. That’s because while he was very good from 2013-16, he wasn’t great. Almost great, but not great. His last three Slam titles have given him a huge boost in the longevity category.
As for consistency, here I’m thinking steadiness in the rankings, such as X years or consecutive years in the top 2 or 3, or X number of consecutive years winning Slams or reaching Slam finals.
Now we have to address which criteria to use. Broadly, I would look at 1) Slams; 2) Rankings; 3) Significant non-Slams (such as World Tour Finals and Masters Series events); 4) Other titles. I’m not sure how I would weigh each, but maybe something like 45 percent Slams, 25 percent rankings, 20 percent significant non-Slams and 10 percent other tournaments.
But each of those categories needs to have its own internal system awarding a certain number of points. For example, you shouldn’t get X number of points for winning a Slam and 0 points for all other results at a Slam. The Slam conversation can be largely about number of Slam titles won, but it can’t be totally about that. You have to award some points for a final, and a small number for a semifinal. Otherwise, you’re giving no value to someone winning a Grand Slam semifinal, which is a very important match. If you asked players if they would want to win a Slam semi even if they knew they would lose in the final, they would of course say yes. You’d rather win a Slam semi than lose it. So if we do an internal Slam points system, maybe we give 100 points for winning a Slam, 50 for a final, and say 20 or so for a semi. I think I would cut it off at the semis. Yes, it’s arbitrary, but you have to draw the line somewhere. You could convince me to give some points for a Slam quarter, but it would have to be a small number, like 5. It wouldn’t be enough to make a difference.
One problem with majors is that the Australian Open wasn’t played by most of the top players for a pretty long time, basically from the early 1970s to around the mid-1980s. Connors and Borg, in particular, almost surely would have added to their Slam totals if they had played it more often. Borg played it once, as a 17-year-old. Connors played twice, winning it once. Furthermore, there wasn’t one specific year when it started to be a first-rate tournament again, adding to the difficulty of judging the Australian. Many of top Europeans returned a bit before the top Americans.
After the Slams I would look, at least since the 1980s, at the rankings. (For the men, the computer rankings have been around since 1973, but they were suspect in the early years, when there was a huge premium of quantity over quality.) And again, within the rankings, there needs to be an internal system. I would give the most points for being year-end No. 1. I’d give some for being year-end No. 2 and No. 3. But I’d cut it off somewhere around No. 3 if the discussion is about the most accomplished players. Even if you want to stick to the GOAT nomenclature, keep in mind that the debate is about greatness, not goodness. Total weeks at number 1, or in the top 2 or 3, should also be given robust consideration. That would fall under the consistency category.
Then I would consider important non-Slams, such as the World Tour Finals and Masters events. There has to be a separate discussion about how much to weigh the year-end championships. I have mixed feelings about this. It’s a big deal, and you have the best players there. But it’s a round-robin format, played indoors in the winter, and it’s all a bit weird. Plus it’s a format that only goes back to the 1970s, so it’s unfair to use it when comparing a player of today against, say, Laver or Rosewall or anyone before them. I would weigh it similarly to how the ATP ranking points system does now — fewer points than a Slam but considerably more than a Masters series.
Another problem in this category is that the Masters events only date back to 1990, and they have been taken more seriously by the current generation than by previous ones. And even today the top players occasionally skip some of them to be fit for majors that follow shortly thereafter. Some have started skipping Cincinnati to be ready for the U.S. Open. So this category is fraught with complications. Furthermore, not all Masters events have the same value to most players. Indian Wells and Miami are viewed as more important than Shanghai. Monte Carlo isn’t mandatory.
Then I would factor in other tournament titles and finals. Sure, these are the lifeblood of the tour and they matter, but it’s not where greatness is established. So I would give them some, but not much, consideration. Connors has the most tournament titles, and no one thinks he’s the most accomplished player of all time.
I would not consider doubles, since I view this conversation as being about the most accomplished singles players. Keep doubles separate. Besides, no great player since McEnroe has spent much energy on doubles.
I also wouldn’t use head-to-head in a MAP debate. I will explain why in a future post, but let’s leave it at that for now. Suggestions, comments and civil criticism are welcome.
One thought on “Changing the GOAT Debate to a Data-Driven MAP Discussion”