Predicting Slam outcomes mostly calls for observation, judgment, and a lot of guessing. In thinking ahead to the coming Australian Open recently, I did some parlor calculations using my middle-aged personal computer. My goal was to study correlations in Slam outcomes numerically and, in particular, find help in predicting Australia 2002.

By comparing individual-player performances, I determined how well each of the four preceding Slams predicted the results of Australian Open 2001. I then weighted the 2000 Slams to produce the composite prediction that best fit the actual results of Australia 2001. (A description of the method is given in the footnote at the end of this column.^*)

Here is the weighting that best predicted Australia 2001.

2000 U.S. Open, 30.1%
2000 Wimbledon, 15.9%
2000 Roland Garros, 7.0%
2000 Australian Open, 47.0%

I was surprised that Australia 2000 proved a stronger predictor than U.S. 2000, which was much closer in time to the target event.

The final step was to take the above weights and apply them to the subsequent results of the 2001 Slams. The object was to predict player performance in Australia 2002. Here, then, is the outcome--the top eight players in order of probable success at Melbourne Park 2002, according to the calculations.

1. Agassi
2. Hewitt
3. Clement
4. Kafelnikov
5. Safin
6. Sampras
7. Grosjean
8. Henman

Note that Pat Rafter should be in second place but is not shown here, as he is not playing in Australia 2002. Three other members of the world's official top nine failed to attain our short list--Kuerten, Ferrero, and Haas. They are replaced here by Clement, Safin, and Sampras.

I'll now use the results of the calculations in sorting out what might happen on the Rebound Ace courts in Melbourne Park this month.

OUR PRIME EIGHT

Andre Agassi's top place in our ranking in large part reflects his triumph in Australia 2001, his second Australian Open championship in two years and his third in six tries. His relentless heavy-hitting style from the baseline, backed by speed, stamina, and a heavy serve, thrives on slow hard courts and hot temperatures like those often seen at Melbourne Park. Even though his achievements fell off in the second half of 2001, Agassi, now 31, unquestionably stands with the front-runners for 2002. Agassi won the tune-up event at Kooyong last year, and again this year he won his first two matches there. As of this writing, he awaits Sampras in the final.

Our #2, Lleyton Hewitt, finished 2001 as world Number One and is officially seeded first at Melbourne. His blinding speed and his shot-making and serve-returning abilities are probably the best in pro tennis today. It seems likely that his U.S. Open victory in 2001 at age 20 will be the first of several, perhaps many, Slam triumphs for him. His Masters Cup win over Agassi on a slowish indoor court at Sydney in November indicates that he is ready to surpass the American veteran henceforth. Last week he played the Hopman Cup in Perth but, stricken with chicken pox, he withdrew after winning two matches. But he played and won an exhibition match this week.

Two French stars, Clement and Grosjean--#3 and #7 above--have both done well the last two years at Melbourne Park. Last year, the two met in the semis, where Clement won in five sets. Both are speed merchants and back-court shot-makers in the pattern of Hewitt and, like Hewitt, should not be considered clay-court specialists. Both are smallish in stature, in their early 20's. Last year, Clement won three matches at Wimbledon and three more at U.S. Open, but late in the season Grosjean became the stronger of the two, winning the Paris indoors on slowish courts and reaching the final at Masters Cup. Both men started 2002 poorly. Clement lost two matches at Hopman Cup and both lost in the first round in the tune-up this week at Sydney. A third French star, Nicolas Escude, whose heroics brought the Davis Cup to France in 2001, won two matches at Sydney this week but then withdrew with a groin injury.

Kafelnikov and Safin, ranked #4 and #5 here at age 27 and 21, respectively, again lead the Russian contingent. Kafelnikov won the Australian in 1999, was runner-up in 2000, and reached the quarters last year--excellent results though the pattern is downward. Safin's powerful serve and baseline game can be overwhelming--strong enough to triumph at Melbourne Park--if the Russian star is healthy and playing his best. His success in reaching our top eight was remarkable in that he was not right physically throughout 2001. Both stars showed mediocre results in the early 2002 events.

Narrowly behind Safin at our #6 is Pete Sampras, a two-time champion at Melbourne, who also possesses weaponry ample to collect the 2002 crown. Pete reached only the round of sixteen last year at both Melbourne and Wimbledon, but defied skeptics like me in a dazzling run at U.S. Open, where he in turn defeated Rafter, Agassi, and defending champion Safin. Without benefit of a rest day before the final, he lost to Hewitt in straight sets. In early 2002 play, Sampras lost a three-setter to Todd Martin in the first round at Adelaide, then won his first two matches at Kooyong.

It's sometimes said that serve-and-volley players reach their peak at relatively late age, and Pat Rafter is usually cited as an example. Tim Henman, our #8, has finished in the world's top twelve for the last four years. But except at Wimbledon where his net skills are well suited, he has never passed the fourth round in a Slam. Henman, now age 27, achieved a fine 10-3 record at the non-clay Slams last year, and began the new year lustrously by winning the tournament at Adelaide last week, defeating Rusedski and Philippoussis. The possibility of a career break-through for the stylish Britisher in 2002 remains open.

THE NEAR-PRIMES

The chances are good that one of our elite eight, above, will be the tournament winner. But as we descend in our list, we find several other stars who are plausible dark-horse candidates.

At #9 in our ranking is American Todd Martin, the player who last year at Melbourne defeated Pete Sampras and eventually reached the quarters. A finalist in 1994, Martin has the serving, serve-returning, baseline, and net skills that would seem perfect for winning at Melbourne. He began 2002 by defeating Sampras at Adelaide. Martin could reach the late rounds, but at age 31 his chances for a first-time Slam triumph seem faint.

I prefer the overall chances of Roger Federer, our #11, who is clearly a rising player at age 20. He almost single-handedly carried the Swiss team over the U.S. in Cup play last year, and subsequently ended Sampras's long undefeated run at Wimbledon. Federer later won three matches at U.S. Open before bowing to Agassi, thus completing a 9-3 record in the non-clay Slams. Federer's pattern of upward jumps in the annual rankings seems likely to continue, and a good performance at Melbourne Park seems likely. Disturbingly, he began the new year with losses to Hewitt and Robredo at Hopman Cup, but he won his first four matches, reversing the loss to Robredo and defeating Roddick, to reach the final at Sydney, now in progress.

Robredo, Roddick, and Youzhny are strong teen-agers who require our attention. At #16 in our calculated ranking is Moscow-born Mikhail Youzhny, 19, who last year achieved a combined 7-3 record in the three non-clay Slams including a victory over American Andy Roddick at Wimbledon. Youzhny finished 2001 unimpressively, however, and began the 2002 season with an early loss at Qatar and another, to Ivanisevic in three sets, at Auckland. Meanwhile Roddick, our #18, recorded a 6-2 won-loss record in the Wimbledon and U.S. main draws last year, though his results slightly declined thereafter. Roddick's rocket serves will lose a little of their effectiveness at Melbourne, but the young American's strong baseline game should stand him well. Finally, Spain's Tommy Robredo, who ranks #37 in our tally, at 19 had good wins over Todd Martin and Ferrero at U.S. Open before bowing to Roddick. He began the new year by leading the Spanish team to victory at Hopman Cup.

Inexperienced players have won the Australian sometimes in recent years, but none were as inexperienced as Youzhny, Roddick, or Robredo. (Agassi won in his first appearance there in 1995 at age 24, Sampras won in his fourth try at age 22, and Courier won in his third try at age 21.)

The Rebound Ace courts are sometimes said to be equalizers, helping clay-court baseliners compete effectively against the big servers and net artists. Spanish star Carlos Moya, #10 here, was a Melbourne finalist in 1997, Garros champion in 1998, and world's Number One briefly in early 1999, Moya reached the quarters at Melbourne Park last year, losing to Grosjean. Also raised on clay and showing considerable past success on hard courts is Brazilian Gustavo Kuerten, our #15 and a three-time champion at Roland Garros. Kuerten is a sweeping striker of the ball capable of ending points from either back or front court, who ought to relish the true and relatively slow bounce at Melbourne Park. He has never passed the second round there in five tries, however. Kuerten's downhill finish to year 2001, including three losses in round-robin play in Masters Cup, bodes poorly. Another past finalist at Melbourne and former world Number One is Marcelo Rios, whose brilliance of early 1998 has never quite returned. I watched him early in his comeback last summer in Washington, when he seemed the old Marcelo except slightly lacking in his former shot-making consistency. Rios scored well at Sydney this week, winning twice and finally losing to Federer in a close three-setter.

Big servers Rusedski and Ivanisevic, our #12 and #14, respectively, are always threats on hard courts. Of the two, Rusedski probably has the better net skills, Ivanisevic the stronger ground game. They met this week in Auckland, where Rusedski defeated Ivanisevic in three sets. Ivanisevic had won all nine of their previous meetings. Clouding the Croatian's chances are recurring questions about his shoulder. Another likely contender is Mark Philippoussis, who missed all four Slams in 2001 and thus does not appear in our ranking. The tall Australian, at 6-4 and age 25, slimmer by 20 pounds, likes to hammer the ball full power. When playing his best he is overwhelming, especially on hard courts. Two years ago, his record in the non-clay Slams was an impressive 8-3. He returned to the wars from knee surgery in late 2001 and began the new year with a good start at Adelaide, defeating Todd Martin in the quarters before losing a split-set final to Henman.

Offered here are odds for winning Australia 2002. Several young and rising players (Hewitt, Roddick, Federer, and Youzhny) are shown more favorably here than in our above ranking.

Hewitt, 3-1
Agassi, 7-1
Sampras, Henman, each 15-1
Safin, Federer, Philippoussis, each 20-1
Clement, Kafelnikov, Roddick, Grosjean, each 30-1
Martin, Kuerten, Youzhny, Enqvist, Moya, each 45-1
all others, 50-1 or longer

THE DRAW AND THE PREDICTIONS

Here are the eight sections of the draw and my predictions. Within each section players are listed in the order of their official seed. Also shown is the weighted composite score of each player obtained in my prediction exercise. I generally followed these scores in making the choices.

--Hewitt (4.6),Clement (4.3), Lapentti (1.2), Schalken (1.2), A. Martin (0.8), Rios (0.8). A healthy Hewitt should survive this treacherous road. The choice is Hewitt.

--Haas (1.7), Federer (2.9), Moya (2.9), Malisse (1.3), T. Martin (3.0). Our computer analysis shows Federer, Moya, and Martin close. I choose the youngest of the three, Federer.

--Agassi (7.3), Roddick (2.1), Gambill (0.6), Pavel (1.0), Ljubicic (0.6). The official seeds agree with our analysis. Agassi and Roddick should meet in round four. Agassi.

--Sampras (3.7), Safin (3.7), Santoro (1.4), Escude (1.6), Youzhny (2.2). A safe path for Pete until round four against Safin or Youzhny. Sampras.

--Grosjean (3.6), Ivanisevic (2.4), Portas (0.9), Novak (1.1), Kiefer (1.5), Arthurs (1.2), Hrbaty (2.5). Grosjean's slow start into 2002 belies his strong number here. Looking for the magic of last year's Wimbledon, Ivanisevic.

--Kafelnikov (4.2), Corretja (1.0), Arazi (1.1), Robredo (1.3), Mirnyi (1.5). Unseeded Mirnyi has the firepower to unseat the favored Russian star, but probably not on these slowish courts. Kafelnikov.

--Henman (3.3), Canas (1.6), Enqvist (1.1), Rusedski (2.7), Philippoussis (no score). An unfortunate draw for three premier candidates. Rusedski and Philippoussis must meet in round two, the winner to face Henman. In the biggest story of the first week, Philippoussis.

--Kuerten (2.4), Johansson (2.2), El Aynaoui (0.5), Vinciguerra (1.8). Recalling the hard-hitting Swede of two years ago in Canada, the choice over a recently out-of-form Kuerten is Johansson.

Having survived the early rounds, Hewitt should find a way over Federer. Likewise, Agassi should win over Sampras and Kafelnikov over Ivanisevic. Philippoussis, now approaching his promise of four years ago, should comfortably prevail over Johansson. In the semis, Agassi should outlast Hewitt, while Philippoussis should outpound Kafelnikov. Then in the final, Agassi for the fourth time should claim this title--the Slam that ought to be his favorite tournament.

THE WOMEN SUPERSTARS

The withdrawal of Lindsay Davenport because of knee trouble leaves three clear favorites among the women. Almost surely at least one of them will reach the final. All three--Venus Williams, Jennifer Capriati, and Serena Williams--are large and strong, able to dominate weaker opponents with endless heavy artillery off the serve and from both sides. Venus and Serena have superior court mobility, enabling them to neutralize the power of any opponent into the corners. The defending champion, Capriati, has excellent mobility and power but may be temporarily weakened by minor hip injuries, one at Hong Kong last week, another at Sydney this week. Meanwhile Serena's chances abruptly came under question when an ankle injury forced her to withdraw from her semi in Sydney. Venus has won four Slams, Capriati two, and Serena one.

The second rung consists of three-time champion Martina Hingis (1997-1999), four-time champion Monica Seles (1991-1993, 1996), Belgian teen-agers Clijsters and Henin, both of whom penetrated the world's top eight for 2001, and French star Amelie Mauresmo, who reached the finals at Melbourne in 1999 at age 19.

Recent player performance strongly suggests that Venus Williams will win the tournament. She captured U.S. Open late last summer without losing a set, indeed without losing more than four games in any set. Her victims included Serena, Capriati, Clijsters, and Testud. She did not compete again until January, when she won the Australian tune-up at Gold Coast, defeating Henin in the final.

But for her latest injury, Serena would have been Venus's most dangerous rival. Serena was runner-up at U.S. Open, showing victories over Hingis, Davenport, and Henin. She then won the year-end indoor event in Munich, which brought together most of the world's top sixteen. Meanwhile Capriati's results slipped during the fall, though she began 2002 by winning the exhibition event at Hong Kong, defeating Dementieva in a three-set final. But she lost to an aggressively playing Alexandra Stevenson this week at Sydney upon incurring the aforementioned hip injury.

Martina Hingis hurt her ankle early last October and exited the tennis wars. She returned at Sydney this week, winning three matches and, as of this writing, now awaiting the final. Meanwhile Kim Clijsters, 18, won the October event in Luxembourg, lost narrowly to Davenport at Munich (7-6 in the third set), and with Henin swept all her matches in leading Belgium to victory in Fed Cup. This week, she defeated Henin in straight sets in a quarter-final meeting at Sydney but then lost to Hingis. Clijsters is a large and strong player at 5-8 1Ž2 and 151 pounds, and now, foremost among the current teen-aged crop, seems ready to join the topmost superstars.

Here are the odds, wildly influenced by the current injury report, as I see them.

Venus Williams, 3-2
Capriati, 3-1
Serena Williams, 8-1
Hingis, Clijsters, each 15-1
Henin, 35-1
Seles, 50-1
Mauresmo, Testud, Dementieva, each 75-1
all others, 100-1 or longer

THE DRAW AND PREDICTIONS

Here are the eight sections of the main draw and my predictions. To an unprecedented degree, the highest-seeded player should prevail in every section. Allowing myself to choose only four of the top eight, as is customary here, requires some difficult choices.

--Capriati, Tulyaganova, Grande, Panova. Capriati.
--Mauresmo, Shaughnessy, Montolio, Tanasugarn. Shaughnessy.
--Clijsters, Sanchez-Vicario, Sugiyama, Likhovtseva. Clijsters.
--Henin, Dementieva, Serna, Suarez. Dementieva.
--S. Williams, Farina Elia, Nagyova, Torrens Valero. Only because of Serena's late injury, the choice is Farina Elia.
--Hingis, Coetzer, Schett, Kremer. Hingis.
--Seles, Testud, Bedanova, Schiavone. A player showing a rising pattern of success, Schiavone.
--V. Williams, Maleeva, Raymond, Hantuchova. Venus Williams.

To reach the semis then--I choose Shaughnessy, Clijsters, Hingis, and Venus. To reach the final--Clijsters and Venus. Finally, the only player making the picking easy--Venus Williams--will capture her first Australian.

THE NATIONS

On the men's side at Melbourne last year, the U.S. contingent, paced by tournament champion Agassi, led all nations in singles matches won. France was second and Australia third. But Australian strength in doubles gave the Australian men the highest total in singles and doubles combined. Meanwhile the American women dominated, led by champions Capriati in singles, the Williams sisters in women's doubles, and Morariu in mixed. The commanding U.S. lead on the women's side gave the U.S. first place overall.

With American Andy Roddick now on the scene and with Australian Rafter now absent, the U.S. is likely to score highest on the men's side in 2001 and is certain to do so among the women.

FINAL NOTE

As the first Slam of the new tennis year, the Australian Open has large implication for what lies ahead. Last year's winners, Agassi and Capriati, dominated the first half of 2001 and maintained their leads in the year's standings until near the end. Meanwhile it will be interesting to watch for progress among the younger stars already close to the top--Roddick, Federer, Youzhny, and, on the women's side, Clijsters, Henin, and Dementieva.

For fellow tv watchers amid the Northern hemisphere winter, please enjoy two weeks of middle-of-the-night tennis from Melbourne.

--Ray Bowers

^* Footnote on the Calculations. A numerical score is first obtained for each player's performance in each tournament. A player's score consists of his matches won, refined for sets won and lost, tiebreaks, and subsequent-round performance of last opponent.

Correlations between tournaments are measured by comparing individual numerical scores. To do this, player scores in two events--in our case Australia 2001 and, in turn, each of the predictor tournaments--are aligned in two columns on a spreadsheet. The spreadsheet software (Excel) then readily calculates the straight-line relations of the two sets of data--i.e., slope and intercept, along with the correlation coefficient, which measures the tightness of the relation. The correlation coefficients or slopes can provide our preliminary weights.

A larger spreadsheet is then created for multiplying the weights by each player's numerical scores in each predictor tournament, thus obtaining a weighted composite performance. The spreadsheet should be designed so that the weights can be altered throughout the spreadsheet by single keystrokes, thus permitting refinement of weights by systematic tinkering to attain maximum correlation coefficient between the composite predictor scores and the actual results in Australia 2001. Where a player does not compete in one of the predictor tournaments, a percentage of his weighted average score in the other events is awarded. I calculated that an absentee-score percentage of 15.3% gave highest correlation to target-event actual results. This value along with the four tournament weightings listed early in this column produce a correlation coefficient, 0.58832086, that is higher than any coefficient produced by altering the absentee-score percentage or any pair of tournament weights by 0.1%.

I hope that Tennis Server readers will provide expert comments on my method.

--Ray Bowers