[When Rosia and I first talked about this project, I never imagined that it would become as comprehensive and authoritative as this. Amazing job, Rosia. There's no way to put this without sounding self-promoting, so I won't even try. But this is a big, gaudy plume in TW's cap as a blog that takes you places where the mainstream sources simply won't, or cannot - and in a way that has little or nothing to do with me. Many thanks for this brilliant piece of statistical work! Pete]
A few months back, Pete suggested that I might like to look at the efficiency of the Slam performances of the current top group of ATP players. The question could be framed as "How much has each of these guys done with the Slam opportunities he's had so far?".
As well as looking at the current top 25, for good measure I also tossed in Mario Ancic, one of last year’s alternates for Shanghai who has been out for months due to illness and injury, and my local hero Tim Henman, who has just played his last Slam. Marat Safin was in my sample until last week, but even though his ranking has dropped, I've left him in, as one of the small current group of Slam winners. As well, I gathered similar statistics for a number of past Slam champions of the Open era, though they are included with the obvious caution that these represent entire careers, not works-in-progress. But still, I feel that they can provide some kind of benchmark for interpreting the figures.
I kept the actual numbers simple. They have been updated to include the 2007 US Open. The most important ratio for GOAT-hunters is probably the Slams won as a percentage of Slams played in. However, most of the men in our sample haven’t won a Slam, so in addition I’ve looked at the average number of matches won per Slam played in. I also give absolute figures for Slam wins, finals, semifinals, quarterfinals and fourth round appearances – as well as first-round exits. These were used to generate a second table, which I’ve called "Slam Consistency", in which I’ve calculated the percentage of occasions on which a player reached those rounds, or better – i.e. Roger Federer has reached the fourth round or better in just under 71% of the Slams he has played in, while reaching the final or winning it 41.2% of the time. Finally, I’ve looked at the simple win/loss record of each man in the Slams he has played in.
Many of the numbers speak for themselves, but I’ll highlight a few. Rafael Nadal is the only one of the current top 25 never to exit in the first round of a Slam, a distinction he shares (so far) with Bjorn Borg. Richard Gasquet, of a similar age, has played two more Slams than Nadal, and exited in the first round in 47.1% of them. Ivan Ljubicic, so recently in the Top 10, has rarely performed at Slams, and has exited in the first round in 15 of the 32 Slams he has played to date, or 45.5% of the time. I usually have difficulty perceiving Ljubicic as a threat in most Slams, and this record is the reason; in addition he and Juan Monaco share the distinction of being the only two top 20 players with a Slam win/loss record of under 50%, but Ljubicic has had far more time in which to accumulate his record.
Looking at average matches won per Slam, only four of my current sample (highlighted in purple) have won three matches or more per Slam played. This group includes Federer, Nadal and Roddick, excludes former world number ones Safin, Hewitt, Ferrero and Moya – and, perhaps surprisingly, includes David Nalbandian. In addition, the latter's record of reaching the last 16 or better in Slams is exceeded only by Federer and Hewitt. Other than the fact that he hasn't actually won a Slam, his consistency in the Slams that he has played is up there with that of those who have won Slams - and better than some. At the other end of this particular scale, only Ivan Ljubicic, Juan Monaco and Paul-Henri Mathieu have so far managed to win less than one match per Slam played. Novak Djokovic, Sunday's losing US Open finalist, is sixth on my list in terms of average matches won per Slam, at 2.75. His contemporary Andy Murray has some way to go in order to catch him, after the progress he has made in 2007.
I think it’s telling to compare these numbers with those of my list of former Slam champions, all of whom averaged more than three wins per Slam for their entire careers. Bjorn Borg’s insanely good Slam record leaps out here – his average performance was to win enough matches to be in the semifinals each time, and then some. He won over 40% of the Slams that he competed in (a number that Roger Federer, currently at 35.3% could exceed if he were to win, say, five out of the next six Slams, but he’d have to either stop playing, or post a similar average for the rest of his career, to stake a claim to that career record). I don't know whether anyone will ever surpass Borg's record of reaching or winning the final 58.3% of the time; even Federer is "only" at 41.2% on this measure, though that's well ahead of Pete Sampras, on 34.6%.
Of course, Borg’s numbers reflect the fact that he was a prodigy who began winning Slams when still young, and then effectively retired when he was in his mid-20s, before any big decline in Slam performance could be detected. To him, losing Slam finals to John McEnroe seems to have felt enough like a giant decline to call it quits. The picture heading this post is of Borg and McEnroe shaking hands after Borg lost his fourth US Open final, in 1981, having already lost the Wimbledon final to McEnroe in the same year. Immediately afterwards (before the trophy ceremony) he picked up his racquets, stalked out into the New York night, and retired from tennis (aside from a few odd and ends, and his surreal and ill-fated attempted comeback in Monte Carlo in 1991).
Considering the length of Jimmy Connors’ Slam career, between 1970 and 1992, his average match-winning performance per Slam is astonishing – better than anyone except Roger Federer's (with only a tiny margin between the two) and Bjorn Borg’s, which stretched only from 1973 to 1981. A few others have won a higher percentage of the Slams that they competed in, and reached a higher proportion of finals, but (like Andre Agassi) Connors was notably consistent in getting through early rounds to at least quarterfinals – he reached semifinals or better in over half the Slams he played in. Given the very high numbers posted by his contemporaries at various times in his career (Borg, McEnroe and Lendl), these numbers look even more impressive - it's not as though he didn't have rivals who would fight tooth and nail over every Slam title.
Using these statistics, the only current player who stands out to me as definitively belonging in the elite group of multi-Slam champions for all time is Roger Federer, perhaps more for his efficiency of the past few years than consistency over his entire career, though he’s certainly being pulled up in these measures too by his recent performances, and on most of these measures, at this point in his career he is ahead of Pete Sampras by a small margin, though he has yet to catch Borg.
Rafael Nadal has started out his career strongly, being more efficient and consistent so far relative to his peers. However, I think it's too soon for these numbers to be definitive. He's topped the performances of all the former world number ones so far (and has more Slam titles and finals to his name than any of them), but what will his career trajectory be from here? He stands out so far as not having reached a semifinal or better in Melbourne or New York - unlike Roddick, Davydenko or Djokovic, the rest of the current top five. Borg, incidentally (to whom I can’t help but compare Nadal), reached his first US Open semifinal at age 19 – and won his first Wimbledon at 20.
When I look at the Slam career of Andy Roddick, especially in the consistency department, he doesn’t really belong among that elite group of Slam Champions, but a tier below. Roddick would very likely have at least two more Slam titles if not for Roger Federer – who has knocked him out in no less than five Slam finals or semifinals. For the record, if Roddick had won all five of those matches, he’d have an 80% Slam win/loss record. However, he still would have recorded less success statistically in reaching the later rounds of Slams. For the latter reason, Hewitt and Safin also don’t belong in the elite group. In Roddick's case, though, it's clear that one of the reasons that his Slam efficiency/consistency record is worse than those of Federer and Nadal is his record at Roland Garros. He's turned up in Paris for seven years in a row, and has only won four matches in total there.
I considered finding similar numbers for one of our herd of potential GOATs, Rod Laver, winner of 11 Slam titles, and his contemporary, Ken Rosewall, but eventually decided that the comparison would be a misleading one. Laver won his first calendar Slam in 1962, turned professional soon thereafter, and therefore wasn’t permitted to play in any of the Slams until the Open era began in 1968 (he said in his autobiography of the first Open era Wimbledon: "Newcombe, Roche, Pilic, Taylor, had never been away. They were amateurs in 1967, pros in 1968. They couldn’t know the elation of being recognized as respectable men again."). This means that in some of his best years (age 24 to 29) there were twenty Slams that he simply couldn’t play in, though he won a significant number of important professional titles during those years. Who can credibly argue that Laver (or Rosewall) couldn’t have stacked up some impressive Slam numbers in those lost years? That’s a large number of missed opportunities. We know that Laver reached the first two Open era Slam finals (he lost to Rosewall in the final at Roland Garros in 1968, and went on to win Wimbledon that year, beating Tony Roche in the final). Laver, of course, then went on to win the only calendar Slam of the Open era to date, in 1969.
Even so, I might have looked up the numbers, for Laver and some other early multi-Slam winners, but there’s something else that troubles me about the direct historical comparison – the inclusion of the Australian Open. This might seem like a quibble, and perhaps it's just the beginning of a slippery slope. Even in the early years, the Australian Open was contested with best-of-five matches (and no tiebreaks), but it seems to me that the field at most of the Masters Series events today, let alone the Slams, is demonstrably much deeper. For some years up until 1981, the Australian Open was played with a draw of only 64 men. Further back, when Rod Laver won the tournament in both 1962 and 1969, the years of his Grand Slams, he played only five rounds. In 1962, when still an amateur, everyone he defeated was from Australia (he beat Roy Emerson in the final, 8-6, 0-6. 6-4, 6-2). In 1969 the field was tougher, and his five rounds included an Italian and a Spaniard (Andres Gimeno, who he defeated in the final 6-3, 6-4, 7-5). There were again three Australians in his path – Roy Emerson, Fred Stolle and Tony Roche. However, Laver said of it "Winning the Australian isn’t that big a thing. It has the weakest field…". He made clear that winning Roland Garros that year was his biggest challenge - more gruelling, with three five-setters to win from the quarters onwards, and a more international field.
The Australian Open draw was still of only 96 players from 1981 through to 1988, the year that it moved to Melbourne Park from Kooyong. In that year, it acquired a 128 draw like the other Slams, and changed its surface from grass to Rebound Ace. So for much of its history, the Australian Open hasn’t been statistically directly comparable to the other Slams.
While we're discussing "easier" Slams, some of you will be aware that the first Open-era US Open in 1968 had a draw of 96 men (128 from 1969). In the Open era there have been four years during which the US Open has flirted with best-of-three matches for the men – all rounds before the semifinals in 1975, and before the fourth round in each of the following three years. Roland Garros tried best-of-three as well, before the round of 32, from 1973-75. That was after having a draw of only 64 for just one year, in 1972, before reverting back to the original 128. But even Roland Garros had a smaller draw towards the end of the amateur era – Laver played six rounds to win it in 1962, though seven in 1969. In any event, there's certainly an argument that says that Roland Garros had less prestige before Borg began winning it in successive years in the mid-1970s, while also being the reigning Wimbledon champion, and that he helped the crowds (and players) return. Only Wimbledon has remained free of any shorter draws or reductive experiments in the period we’re discussing. And in fact, only Wimbledon and Roland Garros haven’t changed their surfaces (if we exclude slowing down the Wimbledon grass), given that the US Open was played on grass until 1975, when it switched to Har-Tru clay before settling on Deco-Turf in 1978.
Returning to the Australian Open for a moment, I’d note that even during the 1970s and 1980s, many of the big names in the sport didn’t make it a priority to compete there. Air travel was a lot more difficult back then, and it wasn’t considered as prestigious as the other Slams. Thus, the strength of the field in those years remained weaker. Jimmy Connors played there exactly twice, winning in 1974; Bjorn Borg played it only once, in 1973. He thought about playing it again later in his career, but the incentive would have been a US Open win following wins at Wimbledon and Roland Garros, to give him a calendar Slam (this was in the era when the Australian Open was played in December). This never happened for Borg, so he never made the trip again. John McEnroe played five Australian Opens between 1977 and 1992, and missed ten. Three of those he played were after 1988 (he managed to get himself defaulted in the fourth round in 1990, but that’s another story). Even Andre Agassi, for his own reasons, missed eight straight Australian Opens before going there and winning the whole thing in 1995, and three times afterwards (no doubt it’s easier when you have your own plane). But looking at current players, Lleyton Hewitt, Andy Roddick and Roger Federer have never missed playing there so far. Rafael Nadal missed it in 2006 because of rehabilitation after an injury, not a lack of desire to compete there.
Back in the seventies and eighties, though, some of the top guys were effectively reducing their own Slam chances to three per year. We could say it’s their own fault – but this was a different era, and they couldn't have known how thoroughly the general view and actual difficulty and prestige of the Australian Open would change in the future. Whether that would have changed their behaviour, it's hard to say. After all, Bjorn Borg did quite a few things for money (and other stubborn reasons) that I couldn't imagine today's top pros doing - but then, they are less in need of money. Can anyone seriously imagine Rafael Nadal missing Roland Garros one year, like Borg did in 1977 when he would have been a virtual lock for the title, to play World Team Tennis? And then there was the 1982 Roland Garros that might have been Borg's last Slam, which he didn't play because, under new rules governing minimum numbers of tournaments to be played, he refused to go through qualifying, as he was the defending champion. (And, OK, by that stage he had all but announced he was quitting tennis anyway).
So, today’s professionals tend to turn up for all four Slams – in former years, many of them didn’t. They basically make four Slam attempts each year. More chances taken can of course mean greater absolute numbers of Slams stacked up. Sampras, incidentally, mostly played in Australia, and gained two of his Slam titles there. Nothing wrong in that. But I'd still contend that it means something different now than it did in earlier years - the further back, the less easy to compare.
I often find the GOAT debate somewhat pointless, and, because of the immutable differences between eras, generally prefer the idea that each of the greats was or is the greatest in his own time. Perhaps we could institute an Order of the GOATs of the Round Table? That said, looking at the Open era, Roger Federer is currently making very efficient use of his time at the Slams (four per year, as opposed to three, for some former greats, or three plus the old Australian Open, for others), and his eventual career numbers, on the measures here, look on track to exceed those of Pete Sampras, with whom any comparison is entirely fair, as the playing field hasn't changed much. We should bear in mind that Sampras played 18 more Slams than Federer has so far, and if he starts winning less efficiently at some point, these percentages can go down.
In conclusion: these efficiency and consistency stats are one extra way to frame the inevitable GOAT discussion, perhaps, in addition to absolute numbers of Slam wins, and right now, it appears that they are stacking up in Roger Federer’s favour.
-- Rosangel
