SquashTalk> Columns> The Spin > To Fix the World Rankings [last update was 12-jul-05 ]


Pro Squash Rankings:

... Time to Fix the Algorithms ...

The Pro Rankings Belie Current Form © 2005 SquashTalk.com
by Ron Beck
(Ron Beck is the editor of SquashTalk and spent five years on the WPSA pro tour)

Since founding SquashTalk in 1999, I have been following the pro rankings closely, from month to month. Both the men's pro (PSA) and women's pro (WISPA) rankings, which are essentially the world rankings, since WSF does not issue rankings, use a similar rankings methodology. The ranking systems are as a whole pretty good, but have several serious flaws, which come to the fore more frequently than one might expect.

'It's time for the squash community to put pressure on the PSA and WISPA to bring their ranking systems up-to-date!   [Read also: WISPA Rankings - A Perspective, just provided by WISPA to SquashTalk]

Probably the most egregious problem I have observed with the world rankings happened during 2000 and 2001. Leilani Joyce, won the British Open in 1999 but then played an extremely sparse tournament schedule, especially in 2001 (she played 9 events outside of New Zealand in 2000 and only six total events in 2001). During that time, she lost to Cassie Jackman (the player she succeeded at #1) both times she played her in 2000 and 01 and lost to Sarah Fitz-Gerald (the player who succeeded her at #1) four of the five times they played in 2001.

Leilani Joyce achieved world #1 on November 1, 2000 and stayed #1 through September 30 2001. During that time she played a total of only FOUR EVENTS! Of those four, she only won one of those events, the 2001 Hong Kong Open, an excellent tournament for her in which she defeated both Fitz-Gerald and Owens. But in the other three she lost to Carol Owens once and Sarah Fitz-Gerald twice.

During that same 11 month time period, Sarah Fitz-Gerald played in 10 events, WINNING EIGHT of them!

A cynical view would say that Joyce limited her tournament appearances in 2001 to maximize the time she would remain at #1, knowing that Fitz-Gerald was the consensus best women's player during that entire period. Of course, Fitz-Gerald, gracious at all times, expressed surprise at her own rise to #1 in October 2001, but really, she was clearly the player to beat in every event she entered in 2000, 01 and 02, winning 26 of the 30 ranking events she played in during those three years.

It's not as clear cut a case this time around on the WISPA circuit, but Vanessa Atkinson, the current world #2, has been in fairly dominant form, ever since losing to Rachael Grinham in the finals of the 2004 British Open. It's a clear case where up to her surprise loss in the Dutch Open last month, she has been very dominant over a five month span, the best player out there, in effect the #1, but the inertia of the rankings computer kept her at #2. In this case - there was no ducking competition by Rachael Grinham, she just was having trouble contending with the on-fire Atkinson. And Atkinson and Holland deserved to have the #1 spot, at least for a month or two.

The situation actually got so counter-intuitive that WISPA itself issued a press release in mid April, after Vanessa's win in the Texas Open, indicating that Vanessa had achieved a mid-month "informal" ascension to #1.

The problem and solution here are clear - a player's results reflect the past 12 months play, and are equally weighted whether the result happened 12 months ago or last week. So a player who was dominant 10, 11, and 12 months ago gets identical credit to the player who is dominant for the past three months. The solution is clear - change the ranking algorithm to give more weight to newer results. One simple way to achieve this would be to weight the results by 3 month quarter: Current quarter worth 4x, last quarter 3x, 2nd quarter 2x, and oldest quarter 1x.

Here's the second problem that needs to be addressed: The system disincents the top players from entering smaller events. Jonathon Power and Sarah Fitz-Gerald are both players who have taken the "high road" here. In other words, they have entered events on occasion in which their rankings would surely suffer even if they won events! This played a role in the 2001 situation with Fitz-Gerald. She had made a decision to play her way into top championship form by entering every event in sight. As a result, there were several occasions when she won an event and by winning the event lowered her ranking points!

This also happened with Jonathon Power and Peter Nicol at the beginning of 2005, where to the delight of North American fans they entered a few events with second tier prize money. As a consequence, by losing before the finals, they negatively impacted their rankings.

You may ask, why does this happen? It is due to the current "star" system which places a weighting factor into the rankings computer based on the prize money bracket of an event.

Ask the PSA and WISPA administrators (I have), and their line of reasoning is that the lesser prize money events should be set up to give the lower ranked players a chance to play in ranking events and accumulate ranking points, by discouraging top players from going to those minor events.

But wait. Let's look at this from the fans and sponsors point of view. When an event gets organized in a place like Salt Lake City or Ottawa or Mexico City or Copenhagen, the organizers and fans are looking to see the best squash possible. They want to see a Vanessa Atkinson or a Jonathon Power or a James Willstrop. They NEED to see those players, in order to build up excitement and commitment within the local community to stage a larger (prize money) event the following year. If the French have an event in Paris, they would like, in fact need, to see Thierry Lincou appear.

The old WPSA Tour (the predecessor North American pro tour in the eighties and early nineties) had this problem licked. Their ranking system added both a weighting factor AND a scaling factor. So no matter how minor the event, if the world #1 wanted to come, he (she) was welcome, and if a winner would improve their ranking points. The weighting factor would simply mean that the IMPACT on the ranking would be much smaller in a smaller event but the scaling factor would mean that it would always be a POSITIVE impact.

And of course there would be a natural correction - there are only so many events that one can conceivably play. So a Power or Nicol or Lincou would limit their presence in smaller events, to avoid wearing down physically. But if they felt the interest or commitment to compete in a particular event, they would be properly incented to go.

So the simple fix for the disincentive problem is to introduce a scaling factor. The way this would work would be a simple process of dividing the ranking total point by a sum of the weighting factors for all the event's entered, essentially "normalizing" the ranking results, and making sure that even the world #1 player would not be penalized by winning a small event.

What many squash fans who follow pro squash may not know is that the top players meet each other on many more occasions than appear in any statistical summaries, rankings, or historical record.

These meetings happen in the National Leagues in the UK, Germany, Holland and France. The National Leagues are lucrative opportunies for squash pros to make a living and hone their games. For all but the top few world ranked pros, more of their income is earned in the national leagues than in tournament matches. Pros who are based in Europe, such as Joe Kneipp, David Palmer, Thierry Lincou, and countless others, often are engaged in more than one league meeting per week.

Logic says that pros playing each other in such competitions (they are not exhibitions - the league teams are playing for league positions and championships) where they are earning money and where the results are recorded and count in league standings, should have their results count for something in their relative rankings.

Most probably the weightings and scalings (see above) should ensure the impact of a league match is less than a significant PSA or WISPA event. But common sense tells me that if a David Palmer beats a Thierry Lincou five straight times in a national league match, that should count for something in the bigger picture.

I suggest that the PSA or WISPA experiment with the introduction of league results with a minor weighting factor. For instance, maybe give it the value of a first or second round win in a one star event.

This could also address the rampant problem of well-paid pros who put forth desultory efforts in a league match because of the training cycle they happen to be in or the proximity of a tournament appearance they are focused on.

It would also cause media, such as SquashTalk, to provide more close coverage of these leagues - more people would know about the leagues, follow them - and there would be more value for the sponsors.

The current situation of Stewart Boswell, former top-five player (between #4 and #6 from Nov. 01 to Sept 03), who has returned to the tour from back injury and must work his way back up the rankings ladder from zero (he has gone from 252 to 62 in five months), by qualifying in low level events, does not make a lot of sense. The same situation was presented to WISPA with Sarah Fitz-Gerald's comeback from knee injury in 2000 and 2001 and Cassie Jackman's subsequent comeback from back surgery. This is a more difficult problem without a clear, simple answer. Maybe it can be addressed through a series of specially controlled exemptions, as was partially employed in Fitz-Gerald''s case.

This injury comeback rankings "penalty" not only puts the recovering player at a big disadvantage, it impacts the entire game. It imbalances the draws. In Fitz-Gerald's situation, it created several cases where Fitz-Gerald, clearly the best player in the events, but unseeded, met and defeated the first or second seeds in the first or second rounds of the events. This unbalances the events, and diminishes spectator interest.

In Boswell's case, because the PSA pool is much deeper, it literally puts him in a position of having to play steadily in the most minor events before he can even get into the QUALIFIER for a major event. A Boswell, who two years ago was a seeded player in most of the major events, now may not even be able to play in the qualifier in Hong Kong or the US Open this year, despite having won easily, from a qualifying position (i.e. playing two extra matches than the seeded players), five straight events this summer. [And by the way, why was he playing in these five "minor" events, where his competition decidely outmatched? Because with a ranking of 252, or in 80s, 90s or 100s, there are a very limited number of events where he can even get entry into the qualifier.]

The above is all presented as simply food for thought for the squash community and PSA and WISPA Boards. The basic point is that rankings can be changed and improved. They are simply math formulas that can be adjusted or modified. They should NOT be viewed as a black box in the hands of a Horizon or any other supplier of the ranking calculator.

As squash seeks to improve its visibility, fan interest and stature, it would do well to examine closely how rankings are calculated. My suggestions are only one person's view and one way to go. But the evident problems with the rankings, as evidenced by the examples I described, really beg to be addressed.

Fitz-Gerald and Atkinson have both been at #2 at junctures when logic would place them at #1.