Today’s article is a repost from our friends at Universal Tennis Ratings. What experiences have your junior players had with UTR-based draws vs. PPR-based draws? We’d love to hear from you in the Comments below!
Randomness cuts two ways. The good news is that random selection eliminates bias, so choosing the winning raffle ticket, say, by pulling a random number from a fishbowl ensures a fair result. But randomness also means chance—and chance can be a proxy for luck. As we all know,luck is often grossly unfair.
This matters a lot in tennis, because most tournament matches get set up randomly. Traditionally, tournaments seed only the top one-quarter, or 25 percent, of entrants. The classic method spaces seeds evenly throughout the draw, based on their relative strength: #1 and #2 at opposite ends, #3 and #4 in one half of the draw or the other, and so on. But once the seeds are planted, human intelligence disappears from the process. Random selection takes over and places the remaining 75 percent of the players in their first-round matches. This unseeded 75 percent represents, theoretically, the “bottom” three-quarters of the draw. It is chaos down there.
For example, consider the Stockton Junior Excellence tournament in Stockton, California, held in early January. In the first round, Luke Neal (UTR 9.79), brand-new to the 14-and-unders and hence unseeded, drew a tough opponent, fourth-seeded Hugo Hashimoto (10.23). Hashimoto finished 2015 as the USTA’s top-ranked 12-and-under boy, and dispatched Neal, 8-3. Meanwhile, the unseeded Nick Toman (9.17) drew another unseeded opponent, Jesse Ding (8.05), and won decisively, 8-0, as might be expected from their UTRs, which differ by more than 1.0.
Traditionally, tournaments seed only the top one-quarter, or 25 percent, of entrants… But once the seeds are planted, human intelligence disappears from the process.
In the second round, Toman played the #5 seed, Pavan Murugesh (9.20), and pulled out a three-set win, 3-6, 6-3, 6-3. Toman lost in the next round, but his lucky draw at the outset (and of course, his good play) helped him harvest 270 Northern California points and 60 national points from the event under the USTA system, which awards points based on the most advanced round a player reaches. The flip side was Neal’s misfortune. Since the random draw matched him against the formidable Hashimoto—and the back draw was cancelled due to rain—Neal left Stockton with no USTA singles points at all.
Were this an anomaly, we could dismiss it as just an unfortunate weekend for Neal and a fortuitous one for Toman. But it is not unusual; in fact, it is more normal than exceptional. The standard practice of seeding the top 25 percent of the draw and consigning the other 75 percent to chance produces real inequities on a predictable, consistent basis. “This isn’t a hunch,” says UTR’s founder, Virginia tennis coach Dave Howell. “It’s reality.”
Howell, who specializes in competition analysis for UTR, has analyzed scores of tournaments and thousands of match results to diagnose the systemic problems in tournament draws.
To begin with, tournament directors generally seed their draws according to the entrants’ rankings—whether from the USTA, another national federation, the ITF, or the ATP and WTA. This in itself produces problems, because such rankings correspond very imperfectly with players’ actual tennis prowess.
In almost all cases, rankings derive from a points-per-round (PPR) tally that assigns points based on how many rounds a player survives in a given tournament. There’ll be more points for reaching the third round than the second, more for the semifinals than the quarters, and so on. PPR applies at all levels, from juniors to the pros.
Playing more events means more chances to amass points—so those who’ve entered fewer events may find themselves with unrealistically low rankings, even if they are playing very good tennis. For example, strong ATP newcomers like Noah Rubin (14.87) and Taylor Fritz (15.56) have recently been underrated because they had just turned pro and only played a few ATP matches. Furthermore, playing more events naturally costs more money, so those who can spend more on tennis tournaments are likely to win more PPR and so boast higher rankings.
At the junior level, the USTA system resets a player’s PPR total to zero each time he or she moves up to a higher age category—moving from the 12s to the 14s, for example. Such resetting can plant land mines in a draw. The aforementioned Luke Neal, for example, had garnered plenty of USTA points in the 12-and-unders, but started from scratch again in the 14s. Hence his ranking in the 14s, based on PPR, dropped sharply, so he was unseeded in February at the Johnson Ranch Junior Age Division Singles Championships in Roseville, California.
Had UTRs been used to seed the tournament instead of the USTA’s PPR rankings, Neal would have been the #4 seed there. Instead, in the first round he became one of those land mines for Joseph Teh (9.48). Teh, an older boy with more PPR, was the tournament’s sixth seed; UTR would have seeded him fifth. “Normally, you would never have the #4 and #5 seeds play each other in the first round,” says Neal’s mother, Kim Neal, an attorney in the Bay Area. The two boys, whose UTRs were close, played a competitive match, with Neal winning, 7-5, 6-4. Viewed by rankings or seedings, this looked like an upset, but in UTR terms it was no such thing. In fact, Neal’s UTR was slightly higher than Teh’s.
Howell’s analysis of tournaments has shown that UTR compellingly predicts the winner of a given match. Consider upsets. Conventionally, when a player beats someone with a significantly higher ranking, that’s considered an “upset,” even if the winner has a higher UTR. In the UTR scheme, however, an upset is a match won by a player whose rating is more than 1.0 below the opponent. These are rare; in general, 3 percent or less of tournament matches qualify as UTR upsets. For example, Howell analyzed the USTA National Championships for Girls 18s (Level 1) tournament held in San Diego in August, 2015, whose draw included 256 players. After examining 190 matches with UTR tools, Howell identified only one bona fide upset.
To reveal even more, Howell uses UTRs to divide the draw in half. He’ll compare performances of the upper or “A” half (the stronger players), and the lower or “B” half (the weaker entrants). Using results from eight important USTA Midwest Level 2 18s and 14s events for boys and girls, he reviewed 178 matches that pitted A versus B players. The B players were victorious in only five of these 178 contests, or 2.8 percent. (Overall, tournament data show that UTR A players will win at least 90 percent of their matches against B opponents.)
In another investigation, Howell looked at 35 main and back draws and found that in 26 of them, no B player beat an A player. In six more draws, only one B player survived a matchup with an A. In three draws, he found three cases of B’s winning twice over A’s. (In two of those three, it was the same B player who won both times.) And there were zero draws in which B’s unhorsed A’s as many as three times.
Given the overwhelming dominance of the A players, how does the random draw for the lower 75 percent affect a tournament’s outcome?
First, it undermines one key reason for seeding. We seed the top players so that the best competitors will play each other late in the competition; we don’t want Djokovic and Federer to face each other in the first round and have one of them disappear from the tournament. Top players have earned their seedings via strong track records, and while they must win to advance, we don’t want them to duke it out with each other until they have notched at least a victory or two over less august contenders.
Seeding does insulate the top players from each other at the outset. But seeding only 25 percent of the draw (say, 16 seeds from a draw of 64) does not protect them from the unseeded-but-strong A players in the 17-32 group. It’s quite possible for one of the best contestants to draw another A athlete and have a “one and done” tournament and a quick ticket to the back draw.
For example, Catalin Mateas (13.44) of Braintree, Massachusetts, who now plays for the Duke varsity, caught a series of tough first-round draws at the Orange Bowl International tournament in Florida. Though unseeded himself, Mateas’s luckless run shows how a strong junior player can repeatedly have a short tournament a thousand miles from home. In 2012 Mateas lost to the #2 seed, Sumit Nagal (14.28) of India, then the following year was defeated by another second seed, Sebastien Tatlot (13.96) of France. In 2014 Mateas completed his trifecta by drawing the #1 seed, Russian Andrey Rublev (14.85), and once again went out in the first round.
On the other hand, perhaps a seeded entrant gets a different kind of “bad draw”—facing a B player at the outset. Here we encounter the second problem with randomizing 75 percent of the draw: dull, one-sided matches. We know going in that A players will win 90 percent of the time versus B opponents; such predictable victories aren’t much fun to watch, nor to play, on either the winning or losing side of the net.
Tournament directors can solve both these problems with an easy, straightforward change: let’s call it the “UTR 50 percent solution.” Simply seed the top 50 percent of the players based on their UTRs. The result will be that first-round matches become even more predictable affairs, as the A seeds beat their B opponents 90 percent of the time. But the rest of the tournament gets far more interesting, competitive, and fair. Starting in the second round, the vast majority of B players will play each other in the back draw in matches they have a decent chance of winning. Meanwhile, the seeded A players will duke it out in the second round and later against opponents who present a true challenge.
The 10 percent of B players who catch fire in the first round and defeat their seeded opponents can congratulate themselves on a real accomplishment: they’ve beaten someone who was supposed to beatthem. That’s a far different feeling from advancing due to a “lucky draw”—i.e., vanquishing a fellow B participant who perhaps couldn’t put up much resistance. At the same time, the seeds who advance can pride themselves on having won their PPR against stiff opposition—those points were earned, not just bestowed by a lucky draw.
As in all sports, luck will always play a part in tennis. But the UTR 50 percent solution can remove much of the luck from the draw sheet, and keep it where it belongs—on the courts!
Comments