This is perhaps one of the most difficult questions to answer in all sports: what does set a champion apart from a good player? The growing availability of match-related data, particularly at higher-level tournaments, allows us to address the question from a new point of view, which we might call a data-driven perspective.
In other words: we can identify a player of the highest level and analyse, in the most scientific way possible, what characteristics his many matches that end in a victory have in common, and what distinguishes them from the few defeats. In part because of his excellent start to the 2021 season, but also due to his continuity at high levels (despite the young age), our first choice was Stefanos Tsitsipas, a 22-year-old Greek who has already been able to defeat Federer, Nadal and Djokovic at least once.
RESUMÉ
Tsitsipas began his career in the ITF Junior circuit at the age of 14 in 2013 and, by the end of 2016, he had won 11 Futures titles: five in singles and six in doubles. He played his first ATP match in the Rotterdam tournament in 2017, losing to Tsonga (who would eventually win the tournament). In the same year, he made his debut as a qualifier in a Grand Slam tournament, losing to Ivo Karlovic in the first round of the French Open. Again in 2017, he failed to qualify for the main draw of the US Open, but in that same period he won a Challenger tournament for the first time in Genoa.
The 2018 season was a turning point, particularly at the Toronto Masters 1000 tournament: Tsitsipas defeated four top tenners (Thiem, Djokovic, Zverev and Anderson), before surrendering to Nadal in the final. That year, he also won his first ATP title on the indoor courts of Stockholm, and he finished the season by winning the Next Gen tournament and rising to the 15th spot in the world rankings. Since then, Tsitsipas has been a constant presence at the highest levels: he has won six more titles (among which the 2019 ATP Finals and this year’s Masters 1000 in Monte Carlo) and has reached the semi-final of the Australian Open twice, in 2019 and in 2021. This year, he’s also played his first Grand Slam final at Roland Garros (defeated by Djokovic in five sets). By now, his place among the very best is consolidated, and in 2021 he is second in the Race to Turin.
STATISTICAL OVERVIEW
Before delving into the analysis, in search for winning and losing patterns of the Greek champion’s game, we need to define a set of matches that will be discussed. We will focus on matches played in singles at the highest level, i.e. the Grand Slam tournaments (Australian Open, Roland Garros, Wimbledon and US Open). Let’s try to frame Tsitsipas’ style of play with some statistics, the average values of which are shown in Figure 1, separating the three playing surfaces. First of all, the number of aces Tsitsipas hits is quite remarkable, especially on grass and on hardcourts (over 10).
In addition, this report confirms, through the number of winners, the vivid image of the brilliant all-court player we’ve come to know and appreciate. On average, Tsitsipas hits over 30 winners per match on each surface: a considerable tally, even if it relates to matches played with the best-of-five format. The number of points played at the net on grass is also notable. It’s noticeable how Tsitsipas significantly changes his style of play, often trying to hit as many volleys as he can on faster turfs (on average, he does it over 30 times per game at Wimbledon).
This second group of statistics sheds light on a few elements: first of all, at the moment, Tsitsipas has obtained most of his Major wins on clay (after all, he was a set away from the French Open title) and hard. As for the grass, the Greek has never managed to go past the round of 16 (2018). At a first glance, these data seems counterintuitive, given his familiarity with the net accompanied by an excellent percentage of success: not only does Tsitsipas volley frequently, but on grass he gets the point in over 70% of these occasions.
Given the effectiveness of his serve (over 75% of points won on first serve and over 65% on the second), the disappointing results he has achieved on grass feel like a bit of a puzzle. It should also be noted that Tsitsipas, due to his young age and due to the pandemic effectively cancelling the 2020 edition, has only played four times at the Championships. Therefore, the unexpected defeat in the first round of the 2019 edition against Thomas Fabbiano undoubtedly weighs on the stats, and the same does this year’s early loss against Frances Tiafoe.
In addition, on a more analytical level, the statistics relating to the percentage of break points saved may perhaps come to our aid and provide us with an explanation. Despite his excellent serve, Tsitsipas saves less than 60% of the break points he faces on grass (as of the 2019 edition), a percentage comparable to what he gets on clay but much lower than what he achieves on hardcourts (on which he saves about 70% of chances). Therefore, so far on the lawns Tsitsipas has perhaps lacked that extra bit of calm under pressure to thrive in clutch situations – this could be due to his ball toss, which (especially before 2021) could represent, in the most critical moments, his Achilles’ heel.
THE MOST SIGNIFICANT PATTERNS AND KEY ELEMENTS OF TSITSIPAS’ GAME
So far, we have focused on Tsitsipas’ game by examining one aspect at a time: let’s now try, with the help of technology, to consider several aspects at the same time in order to develop a multivariate analysis. In particular, we will try to understand which of the various statistics discussed above (that represent our input variables) are key, and how they impact victory and defeat in a match (these two represent our output variable).
For greater clarity, we will ensure that the classification algorithm used automatically returns, based on the variables available, a model consisting of a set of rules representing the statistically most significant patterns that lead the Greek to victory or defeat. Below, we illustrate the three most important rules so calculated:
1 – “If Tsitsipas hits at least 1.3 more winners per set than the opponent and commits fewer than 6 double faults, then he is the winner of the match”. This pattern is very strong: in the 17 Grand Slam matches in which this combination has taken place, Tsitsipas has always been the winner. We conclude that when the Greek manages to be proactive (more winners than his opponent) and remains focused (limited number of double faults), he wins the match.
2 – “If Tsitsipas hits more winners than the opponent per set (or the difference in favor of the opponent isn’t greater than 0.5 per set), if he plays fewer than 20 points at the net, and if the match does not last more than 55 games, then he wins the match”. In this case, the pattern occurred in 22 matches and, in all 22 cases, Tsitsipas is the winner. We can perhaps interpret it this way: if the Greek doesn’t let his opponent dictate too much (the difference in terms of winners can be unfavourable, but contained), doesn’t let the match drag on for too long (no more than 55 games) and does not spend too many energies rushing forward, he then wins the match.
3 – “If Tsitsipas commits an average of more than 1.4 unforced errors than those of his opponent per set, if he does not hit at least 3.75 more winners per set, and if the average number of hits per rally is less than 4.6, then the Greek is defeated.” This pattern occurs in fewer matches: only seven. In this case too, however, the prediction is extremely precise: in all seven cases, Tsitsipas lost. It seems natural to infer that, if in a match that is played with short exchanges Tsitsipas proves less concentrated and does not have a very significant contribution from his own winning shots, then the Greek is forced to concede.
Let’s now summarise which are the decisive elements of Tsitsipas’s game, that is, the most decisive with respect to the outcome of the match. To obtain this result, we will evaluate which elements of his game (for example, winning strokes or unforced errors) appear in patterns such as the three mentioned above, allowing to predict the outcome of the match with great precision. The more a game feature appears as a relevant condition within these patterns, the more we can define it as a key element of the Greek champion’s game. We will therefore be able, on the basis of the data, to draw up a feature ranking of the various aspects of his game, distinguishing those that, alone or in combination with others, prove to be key.
As can be seen in Figure 3, which shows the feature ranking, the average difference between the winners of Tsitsipas and those of the opponent is the most relevant element to his wins, and therefore earns the value 1 (maximum) in the feature ranking. In second position, we find the average difference of double faults and, in third position, the number of unforced errors. Of course, in these cases (double faults, unforced errors) lower values correspond to a greater probability of winning, so the correlation is associated with a negative coefficient (inverse correlation). Similarly, and with almost the same weight, we observe how an excess of net points, as already mentioned, tends to decrease the chances of Tsitsipas, thus constituting another example of inverse correlation.
In fifth position is to be found Tsitsipas’ number of double faults, in this case measured independently of those of his opponent. The fact that two out of five of the most relevant features are associated with double faults, an element that generally seems marginal in a match’s economy, can perhaps be traced back to the difficulties that Tsitsipas encounters (less and less as time goes by) with his ball toss, and consequently with the serve, in moments of great tension and fatigue.
The relevance of double faults, as well as that of unforced errors, also reminds us how, in addition to his undoubted technical qualities as a proactive player (aces, winners), the Greek needs to stay focused to highlight his excellent skills as a fighter, raising his level of play just when the number of games grows or when the going gets tough. This is a feature that he shares with the great champions, perhaps to a greater extent than with his other remarkable qualities.
Article by Damiano Verda; translated by Alessandro Valentini; edited by Tommaso Villa