The new ITTF World Ranking: big changes, bigger challenges

A new World Ranking regulation has been recently proposed to start its effect from 1st January 2018. Some substantial changes will modify the way the World Ranking is calculated, which will obviously have a big impact on the international level.

In this article, we will first try to understand the current World Ranking, paying attention to its most important rules and the way they affect the players’ position. Then, we will then see the changes to be introduced by the new system, focusing on their impact on the ranking. Finally, we will analyze the prediction ratio of both rankings in order to compare them.

About the World Ranking concept

People love rankings. We do really love them. Not only in sports, but almost anything that can be measured has the so-called “best of the World” title.

Many of the rankings are based on subjective opinions. As an example, the well-known IMDB top films ranking has “The Shawshank Redemption” as the best film of the history, based on 1.8 million votes from their webpage users. On the other hand, most of the sports have a World Ranking which tries to be as much realistic as possible, usually based on the players’ merits.

Table tennis is not an exception, and its World Ranking provides an ordered list with all the players that fulfill certain requirements. Depending on gender and age, there are the next ranking lists:

  • Men
  • Women
  • Men U21
  • Women U21
  • Men U18
  • Women U18
  • Men U15
  • Women U15

All of them, including historic data starting in 2001, can be found at the ITTF official website.

Apart from allowing to compare and predict the performance of the players, the World Ranking is used for the next purposes:

  • Seeding in open tournaments, like in the WTTC where there are 16 seeds.
  • Qualification round configuration, like in ITTF World Tour events, where groups are configured based on the players’ ranking.
  • Selection in tournaments with limited entries, like in ITTF World Tour events.
  • Invitation for restricted Top-like events, like in the ITTF World Cup.
  • Special fines, like when a Top-16 players cancels its entry for an ITTF World Tour event after the deadline was reached.
  • Maximum number of player entries per country, like in the WTTC where having a Top-100 player grants one more entry.

2017 World Ranking system

ITTF World Ranking example
ITTF World Ranking example

As of March 2017, Women World Ranking has 596 players, being Ding Ning the first ranked with 3474 points and Lekeasha Johnson the last ranked with 398 points. How did they end with that amount of points? How much does “1 point” represent? Let’s take a look at the regulation.

Basic principles

Two basic rules determine how the points are obtained:

  • For each match, a certain number of rating points is on stake, depending on the ranking points difference between the players. The winner gets those points, while the loser has half of them subtracted. A multiplying factor (1x, 1.5x or 2x) is applied, depending on the importance of the tournament.
  • Bonus points are awarded on certain tournaments to players reaching remarkable positions. The amount of points depends on the importance of the tournament, having five ranges. These points are valid only for 12 months, and are not considered for the rating points calculations.

However, you might be wondering about something. How do players start to be ranked? Do the players start from 0 points? They don’t.

First ranking

There is a policy in place to determine the first ranking points for a new player. First of all, the player must have 4 wins against already ranked players to get ranked.

For example, Anna Dmitrieva, a 16 years old player from Rusia, had her first ranking entry in March 2017 Women World Ranking, after winning its 4th match against a ranked player in the 2017 Swedish & Cadet Open. Her starting  points are 1480, which rank her as World’s #340 out of 596 ranked players.

Those 1480 points are based on the highest ranked player she defeated, with some subtractions (10 fixed points and some more for all the lost matches she had). Therefore, a new player could even start as World’s #2 if he/she managed to defeat the first ranked player and win the tournament they take part of.

Full regulation

In case you are interested in knowing more details, the complete regulation regarding all the mentioned variables (amount of points on stake, multiplying factors, bonus points, tournaments considered for the ranking, etc..) can be read in these documents:

Stats and insights

March 2017 World Ranking has 794 men and 596 women ranked players. However, there are many players who did not have activity during the last four months. Consequently, they are removed from the current listing. In this chapter, all the presented graphs will consider them as well.

First, we will see some rough numbers to have an idea of the amount of data we are working with:

numbers

How do all these players distribute around the globe? Where are the hot spots of the table tennis world?

heat_map_men_with_legend

 

Men’s hottest spot is located around center-Europe. France and Japan are areas with many top-level players, too.

heat_map_women_with_legend

 

Women’s map does not differ much from Men’s. Center-Europe is still the hottest area. However, France is not as relevant as Japan, while South and North Korea emerge.

If we only focus on the top 200 players, we can discover the list of top nations:

countries

At this moment, Japan is at the top of both Men and Women rankings, having 24 and 33 players, respectively, in the Top 200. China and South Korea are not far from them in both categories, having 15 men and 22 women. While France, Germany and Sweden excel in Men’s (15, 14 and 11 players),  Russia and Chinese Taipei (12 and 10 players) are the next relevant nations in Women’s category.

Regarding the points distribution, the amount ranges from its minimum allowed, 200, to the #1 current score, with more than 3000 points. The next histograms depict this fact:

histogram_men

histogram_women

Comparing Men and Women histograms, one aspect is easily noticed: Men’s ranking has more players in the lower ranges. Women have them distributed in a more equal way around center ranges. In fact, we have the next indicators:

  • Men’s points mean is 1202.95, with a standard deviation of 604.26 points.
  • Women’s points mean is 1567.61, with a standard deviation of 587.07 points.

Nevertheless, in both categories we can see that the top ranges have fewer players than the rest. The next graphs zoom in the top positions of the ranking to appreciate the points difference:

top25_points

Although both graphs seem to have a similar shape, there is an important difference: the Y-axis scale. The points difference between #1 and #25 in the Men’s ranking is 996, while in Women’s ranking it is 677. As a consequence, progressing to the top positions of the ranking is much harder for Men than for Women.

2018 World Ranking system

The new World Ranking regulation for the next season has already been presented. According to the official website, “changes might occur during the test period”, which finishes on May 31st. The intention of its action is, according to the ITTF:

One of our main goals is to massively increase the prize money at the ITTF, to also ensure that the top table tennis players are “stars” on the international scene and at the same time to improve the quality of events.

To achieve this, we need to prove to our partners (media and sponsors) and our fans that our events are the best and most important. For them to be the most important, we need best players playing at the events. Our tools need to support that.

The upcoming changes have a huge impact on the current World Ranking, as its basic principles are affected, which dramatically shakes the order of the ranking list. Let’s take a look at the new regulation, highlighting the differences with the current one.

Simpler than ever

The new regulation is much simpler than the current one. It is ruled by the next principles:

  • Players are rewarded with ranking points depending on the round they end their participation in the tournaments. There are fixed tables with the amounts, separated by tournament kind. Points are also given for each won match in qualification stages.
  • Different ranking lists are published for senior, under 21, junior and cadet, whose points are rewarded separately.
  • Only the best eight results are considered for each player’s ranking. This means that, from all the rewarded points the player has, the eight highest ones are chosen, including only one continental event.
  • Rewarded points are valid for certain time, expiring in:
    • 48 months for Olympic Games, with a 25% decreasing rate every 12 months.
    • 24 months for World Team and Singles Championships, with a 50% decreasing rate after 12 months.
    • 12 months for any other event.

That is all. The provisional full regulation as of March 2017, including all the points tables, can be found in this document.

The differences

  • The new ranking is not adjustive anymore, but accumulative. Although bonus points already existed before, they are the only kind of points that the ranking now considers.
  • There are different ranking scores for each age category. Nowadays, each player has the same points on each list. This means that current non-senior lists are only a filter over the senior list.
    • Fan Zhendong is the current U21 top ranked player, while with the new regulation he would not be ranked at all, as he did not take part in U21 competitions
  • The new regulation establishes a virtual maximum amount of points to be achieved, as only eight results are considered. In the current ranking, there is not any maximum amount to be achieved.
    • The maximum amount is achieved when winning at the Olympic Games (3000 p.), World Championships (3000 p.), World Cup (2550 p.), Grand Finals (2550 p.), and four ITTF Platinum Events (4 x 2100 p). It sums up to 19500 points.
    • As of March 2017, Ma Long and Ding Ning would be the top ranked players with 17100 and 14655 points.
  • The minimum amount of points that a ranked player can have in the new ranking drops from 200 to 6, which is the minimum reward for a tournament result.

What is this mess?

The new World Ranking for 2018 has an official webpage with a test version. We would like to thank Zoltan Bencsik, ITTF Rankings Manager, for his assistance and effort accepting our feedback regarding some mistakes that occur in it.

For the rest of this article, we are considering an improved version of the current ranking, which fixes some issues which are still present on the referenced website. Therefore, the information in this article might differ from the one you can find online at the moment of reading this.

The following illustration shows how the top 16 players of the ranking would be affected by the new regulation. The ranking to be used is the published on March 2016, applying the current and the new regulation:

mess_men_with_arrows

Only the top 3 players remain at the same position in both rankings. Wong Chun Ting is the one having the biggest benefit, as he would be seed number four. Vladimir Samsonov has a good improvement too, ascending four positions.

Three players lose their top-16 priviliged position: Fang Bo, Timo Boll and Yan An. The three of them had a much better ranking due to a few but successful participations. As the new ranking encourages players to compete more frequently, it strongly affects them.

Let’s see now the Women’s ranking differences:

mess_women_with_arrows

In this case, only two players keep their position: Ding Ning and Zhu Yuling. The most damaged player, which is not really, is Li Xiaoxia, as she announced her retirement in January. Apart from her, Wu Yang and Doo Hoi Kem would leave the top-16, losing 30 and  46 positions respectively.

The players ascending the most are Suh Hyowon, Tie Yana and Yang Haeun, which achieve a top-16 seed, along with Jeon Jihee and Hu Melek, which account for the total of five differences.

Old Ranking or New Ranking?

Players usually keep an eye on their ranking position, as it strongly influences their life. As an example, playing the qualification stage of an ITTF World Tour event makes the player start competing two days before than a seeded player.

Consequently, having a realistic and trustworthy World Ranking is an important matter. How can we measure the confidence level of a ranking? Let’s first take a look at some ranking theory.

Ranking theory

Ordering a limited amount of entities based on observations does not have any perfect method. That is, there is no statistical/mathematical theory or algorithm that guarantees that we obtain a perfect ranking. In fact, there is not such perfect concept. Thus, it is always up to the organizer to determine which ranking kind to use.

The publication “Survey of the major world sports rating systems(Journal of Quantitative Analysis in Sports, Vol. 7, 2011), by R.T. Stefani, presents a list of all the different ranking systems used by internationally recognized sports federations. The analysis over 159 sports provides these facts:

  • 60 sports have no rating system.
  • 2 sports have a subjective system.
  • 84 sports have an accumulative system, where points never decrease except when ageing.
  • 13 sports have an adjustive system, where ranking is adjusted based on the prediction of the expected result. ELO or Probit are examples of this kind of ranking system.

As developed in the paper “Predicting Outcomes” (Statistics in Sport, Arnold Press, 1998),  by the same author, the chances of guessing a match result using a proper ranking increase 17% the random chance. That is, the winner would be correctly selected 67% (50% + 17%) of the times when draws are not allowed, and 50% (33% + 17%) when draws are allowed.

Using some real data, the first paper exposes the percentage of times a higher-seeded player won against a lower-seeded one, revealing:

  • Accumulative rankings:
    • In 2010 FIBA World Championship of basketball, the ranking accuracy  over 80 games was 69%. That is, 2% above the prediction.
    • In Grand Slam tennis tournaments from 2006 to 2010, the ranking accuracy was 78% over 1032 games. That is, 11% above the prediction.
  • Adjustive rankings:
    • In 2010 FIFA World Cup of football, during the group phase where draws are allowed, the accuracy was 48% over 46 matches. That is, 2% below the prediction.
    • In 2010 FIFA World Cup of football, during the knockout path where draws are not allowed, the accuracy was 88%. That is, 21% above the prediction.

Predicting with the current ranking

Let’s now evaluate the prediction capabilities of the current ranking system. In order to do so, we will consider all the singles matches from the twelve 2016 ITTF World Tour events plus the 2016 ITTF World Tour Grand Finals.

Some numbers from the data sample:

  • 13 events
  • 5779 singles matches
  • 963 players

We will apply the next rules:

  • A prediction will be considered correct when the top ranked player wins the match.
  • A prediction will be considered incorrect when top ranked player loses the match.
  • The ranking to be considered for each match is the one that was valid for seeding purposes of the tournament the match belongs to. Matches with unranked players will be discarded.
  • As draws in table tennis are not allowed, all matches where both players had the same amount of ranking points will not be considered. The result prediction would be a draw in that case, which would make the ranking useless and the prediction always a failure.

The analysis reports the next stats:

prediction_2016

This means that 75% of the times, in 3 of each 4 played matches,  the top seeded player won. There is no difference between Men and Women matches, as both categories have the same guessing percentages.

Contrasting the current ITTF World Ranking with the ranking theory exposed before, it is predicting better than the expected 67%  guessing ratio. This fact makes us conclude that the current ranking is strongly reliable when predicting results.

Predicting with the new ranking

As proved before, the current World Ranking provides a fairly good list of ordered players, having a prediction ratio that exceeds the expectations. How does the new ranking do?

In order to compare the current and the new ranking, we must do it using the same set of matches. As the test version of the new ranking is only available for 2017, we cannot use the previous data sample that we used to validate the current ranking. This comparison will be done using the following 2017 competitions:  ITTF World Tour Events (Hungary, India, Qatar, Belarus) and 2017 Europe Top 16. It includes:

  • 5 events
  • 1451 singles matches
  • 620 players

The rules are the same used in the previous section. This is the prediction comparison between the old and the new ranking:

comparison

 

Please note the difference in the amount of considered matches, which is caused by players which are not ranked in the new ranking system and were in the previous. Most of them are non-senior players, which become unranked as they did not have remarkable result in senior competitions.

First thing to consider is the current ranking prediction ratio with the new data sample. The percentage of correctly guessed matches is 75.61%, which exactly matches the broader prediction made in the previous section. Now, we are sure that the data sample is representative enough.

Consequently, we can state that the new ranking has a lower correct prediction ratio than the current ranking system: 66.94%. However, two important things must be noted:

  • According to the presented ranking theory, the usual ranking prediction ratio is 67%, which is exactly the ratio achieved by the new ranking. Therefore, it is good enough for its purpose.
  • The new ranking tries to encourage players to compete more often, which is beneficial for both the players (which could achieve higher prizes) and the spectators (which will have the chance to enjoy more espectacular tournaments).

Big challenges

“Big changes, bigger challenges”. That’s the title of the article and, definitely, a true statement. The new ranking system introduces many significant changes. However, the resulting differences does not impact much the prediction ratio.

Players like Timo Boll or Wu Yang are now targetted by the ITTF to play more often. The ranking will then make justice for all of them. Will it help table tennis become a more attractive sport for sponsors and media? That is where the challenge is.

4 comments

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s