June 18, 2014


Yesterday was Day Two of Week One of our ten weeks of camps at MDTTC. As usual, it's always surprising watching beginners on day one who seem to have no concept or potential to do a proper shot suddenly start making the shots properly on day two. There are four in the camp who are basically beginners. 

Sometimes it's a coach can get caught off guard by what he doesn't know about the players. After a day and a half of coaching, I asked three of the beginners (ages 6, 8, and 10) if they thought they were ready to play some games. All three looked rather confused. After some questioning, I realized the obvious - they didn't know how to play a game. None knew how you score a point or how many points a game was up to. So I stopped everything and went over the rules with them. I made a game out of it, seeing who could correctly answer (or guess) the answer to basic rules. When I first asked how many points a game was up to, they argued whether games were to ten or twenty. One guessed that each player served six times in a row, and the other two immediately agreed with this. So I made a mental note to myself to remember to go over the basic rules sooner next time. 

Ratings Quirks

Here are two interesting quirks of the USATT rating system. 

1) Can You Predict the Odds in a Match from their Ratings?

The USATT rating system is set up so that players exchange points based on the difference in their ratings, and whether it's an upset or not. (Here's an explanation of the system.) For example, if the two players are rated 100 points apart (technically from 88 to 112 points), if the higher-rated player wins, he gains 4 points, the lower player loses 4. If it's an upset, then the lower-rated player gets 20 points while the higher-rated player loses 20. Based on this, the higher-rated player should win 20 times for every four times he loses, or 5/6 of the time. There have been discussions of this on online forums for years.

On the one hand, it's fairly obvious that there are more upsets among lower-rated players than higher ones. For example, a player rated 1100 is more likely to upset a player rated 1200 than a player rated 2100 is going to upset a player rated 2200. This is because the ratings are far more volatile at the lower levels - players are far more likely to make sudden dramatic improvements at those levels than at the advanced levels. So there are far more upsets at the lower levels than at the higher levels.

On the other hand, the ratings should spread out to match this probability, so at all levels the probability of a hundred-point upset should be the same at all levels. But this seems to contradict the previous statement.

So who's right? Surprisingly, both are basically right. The part left out is that many of these under-rated players at the lower levels get adjusted upward. The USATT system goes through four steps in processing the ratings. The first step only finds players who need to be adjusted, and then adjusts them upwards. So mathematically, the system is making predictions for these players with their adjusted ratings. So when an 1100 player beats a 1200 player, but is adjusted up to 1500, as far as the rating system is concerned, a 1500 player just beat a 1200 player. In other words, it doesn't see an 1100 player beating a 1200 player; it sees a 1500 player beating a 1200 player. 

Among 1100 players who beat 1200 players but are not adjusted, I would guess that the percentage of their wins against 1200 players would be about the same as a 2100 player (who is not adjusted upwards) against a 2200 player. And so if you look only at the adjusted ratings of players (for those who were adjusted) and use the USATT rating system to predict matches, it'll be pretty consistent at all levels, i.e. the chances of an 1100 player beating a 1200 player would be about the same as a 2100 player beating a 2200 player - about 1/6.

2) Can a Player Gain Ratings Points by Losing a Match?

Yes, but it's extremely rare. The only case I know of where this happened was at the Southern Open in 1996, though I'm sure there are others. Xu Huazhang, a former Chinese national team member who had recently moved to the U.S., had been going to school and so had a rating of only 2572. He went undefeated in the tournament, winning both the Open and Elite Singles. He was adjusted upwards, and came out 2777. But when we looked at the results, we realized a strange thing. The USATT adjustment factor - a mathematical formula which used to be public but is no longer listed in the explanation of the system - normally adjusts a player upward partially based on their best win and worst loss. But since Xu had no losses, a different formula is used. We discovered that if Xu had lost one of his two matches against Jim Butler, he would have achieved a rating over 2800! So by winning both matches, he lost at least 23 rating points. When I discovered this I reported it to the USATT Ratings Committee, but nothing was done. And since the adjustment formula is no longer online, I can't show why this happened.

Facebook Friends

Are you on Facebook? Are you an avid table tennis person? Then send me a friend request! Here's my Facebook page. One advantage - as soon as I finish a blog each morning (Mon-Fri), I post a note on Facebook about it, including a content listing.

Why the New Poly Balls Can Be Heavier

Here's the ITTF Technical Leaflet on this - see the parts in green, which are new. The key part is right at the end, where it says that instead of the standard range for the weight of the ball from 2.67 to 2.77 grams, with the new balls it can be from 2.65 to 2.82 grams until Jan. 1, 2016. I don't have any accurate way of measuring the Nittaku Poly ball I blogged about on Monday, but I'm guessing it pushed the upper limit here, while most balls were, if anything, probably on the lower range. (Anyone have figures on that?)

100-Day Countdown to Change in the ITTF's Presidency

Former USATT President Sheri Pittman Cioroslan is doing an article every day during the last 100 days of Adham Sharara's ITTF presidency. Previous ones are linked from the USATT News page, as well as in my past blogs. Twenty-six down, 74 to go!

  • Day 75: Until the 1950s, Table Tennis Was Considered a “Jewish Sport”

Youngest Player at the Japan Open

Here's a picture of Japan's Harimoto Tomokazu, showing all the media attention he's getting.

Top Ten Shots at the Korean Open

Here's the video (5:16). 

Ping Pong Beer

Here it is - straight from Belgium!

World Ping-Pong Federation

Here's the cartoon!

Send us
your own coaching news!