Alex Table Tennis - MyTableTennis.NET Homepage
  New Posts New Posts RSS Feed - Interesting USATT Rating result
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Interesting USATT Rating result

 Post Reply Post Reply
Author
DonnOlsen View Drop Down
Silver Member
Silver Member
Avatar

Joined: 11/15/2008
Location: Maryland, USA
Status: Offline
Points: 619
Post Options Post Options   Thanks (0) Thanks(0)   Quote DonnOlsen Quote  Post ReplyReply Direct Link To This Post Topic: Interesting USATT Rating result
    Posted: 11/26/2019 at 4:08pm
At the Westchester November:  Andy Hou

Pre Tournament   663
Adjusted to        1597
Post Tournament 1991

Points gained:  1328

After he was adjusted, he had five 50 point wins, one 45 point win, one 35 point win, one 30 point win, one 16 point win, and one 13 point win.

In his seven match loses, he lost 0 points.

I would say the USATT adjustment mechanism failed here.

Thanks.
Advances in a field are best achieved by constructs that fully reflect the phenomena of interest and are rooted in a theory that specifies their determinants, mediating processes, and its effects.
Back to Top
mjamja View Drop Down
Platinum Member
Platinum Member


Joined: 05/30/2009
Status: Offline
Points: 2337
Post Options Post Options   Thanks (0) Thanks(0)   Quote mjamja Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 4:47pm
Any idea of how long between the 663 tournament and the Westchester 1991 tournament?

I have mentioned before that for players with initial ratings below 1000 the adjustment calc should be like an unrated player instead of averaging the old rating in with the "best win + worst loss /2"  calc.

Maybe this should also be done for any player who gains more than 400 points in the initial pass.  That kind of gain indicates the initial rating is out of date and should not be used in the adjustment calc.

Mark
Back to Top
Lightspin View Drop Down
Super Member
Super Member


Joined: 07/11/2018
Location: USA
Status: Offline
Points: 126
Post Options Post Options   Thanks (0) Thanks(0)   Quote Lightspin Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 4:50pm
There is just too much randomness to the readjustment process.  At the US Open in 2015 (I think) A Japanese woman who made it very far to the open was given an initial rating of 1400 or so and ended up being 1800.  I had to point this out to the USATT and eventually they gave her another rating which was still low based on her wins.  The only reason I noticed is because a coach from our club lost 50 points to this Japanese "1800" player while the coach was one of the top 5 rated women in the country.  I wondered how the heck this happened as the coach routinely would destroy 2400-2500 rated players. 

There needs to be some standard, routine process for readjustment.  Not some, well gee we will readjust this guy here because we feel like it and not readjust this person because we feel like it.  Either have a standard way or just do away with the readjustment process already. 
Back to Top
ttssbba View Drop Down
Member
Member


Joined: 09/26/2018
Location: san jose
Status: Offline
Points: 20
Post Options Post Options   Thanks (0) Thanks(0)   Quote ttssbba Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 5:16pm
if you trace back, this initial rating also seems too low:

Back to Top
Vince64 View Drop Down
Gold Member
Gold Member


Joined: 07/13/2013
Location: Kent, WA
Status: Offline
Points: 1378
Post Options Post Options   Thanks (0) Thanks(0)   Quote Vince64 Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 5:41pm
Originally posted by ttssbba ttssbba wrote:

if you trace back, this initial rating also seems too low:

There was definitely a error there. His initial rating should have been 1145 after the tournament not the 663. Best win 1165 + worst loss 1127 divided by 2 is 1146 and he would have lost 1 point to the 1383 player
https://www.facebook.com/SeattlePacificTableTennisClub/timeline
Back to Top
pongfugrasshopper View Drop Down
Platinum Member
Platinum Member
Avatar

Joined: 03/22/2015
Location: USA
Status: Offline
Points: 2035
Post Options Post Options   Thanks (0) Thanks(0)   Quote pongfugrasshopper Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 5:57pm
Originally posted by Vince64 Vince64 wrote:

Originally posted by ttssbba ttssbba wrote:

if you trace back, this initial rating also seems too low:

There was definitely a error there. His initial rating should have been 1145 after the tournament not the 663. Best win 1165 + worst loss 1127 divided by 2 is 1146 and he would have lost 1 point to the 1383 player
Argh!!! How in the world are we still having these issues years after Simply Compete took over ratings? 
Back to Top
ttssbba View Drop Down
Member
Member


Joined: 09/26/2018
Location: san jose
Status: Offline
Points: 20
Post Options Post Options   Thanks (0) Thanks(0)   Quote ttssbba Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 6:13pm
is there a similar problem in this example?

Back to Top
mjamja View Drop Down
Platinum Member
Platinum Member


Joined: 05/30/2009
Status: Offline
Points: 2337
Post Options Post Options   Thanks (0) Thanks(0)   Quote mjamja Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 6:38pm
Originally posted by Vince64 Vince64 wrote:

Originally posted by ttssbba ttssbba wrote:

if you trace back, this initial rating also seems too low:

There was definitely a error there. His initial rating should have been 1145 after the tournament not the 663. Best win 1165 + worst loss 1127 divided by 2 is 1146 and he would have lost 1 point to the 1383 player

I think one or both of the 1100 players was not rated.  Therefore they could not be used to figure his initial rating.  Looks like he had no wins against rated players for the first pass so that "secret formula" was used to get I initial rating.

When several unrated players play each other some odd numbers can come out initially.

Mark
Back to Top
Lightspin View Drop Down
Super Member
Super Member


Joined: 07/11/2018
Location: USA
Status: Offline
Points: 126
Post Options Post Options   Thanks (0) Thanks(0)   Quote Lightspin Quote  Post ReplyReply Direct Link To This Post Posted: 11/26/2019 at 10:24pm
Originally posted by ttssbba ttssbba wrote:

is there a similar problem in this example?



What a joke.  Yes Houston we have a problem.  I guess he is in the "lets not readjust him for whatever reason" club.  Going undefeated and beating a 2566 player in the finals of the open? Maybe the rating chairman, or whomever is in charge, doesn't like the guy....

But then again, he was readjusted in the next tournament:

https://usatt.simplycompete.com/t/tr/13969?u=167100&uai=167100


Edited by Lightspin - 11/26/2019 at 10:26pm
Back to Top
Ttunderthesun21 View Drop Down
Super Member
Super Member


Joined: 03/21/2019
Location: USA
Status: Offline
Points: 119
Post Options Post Options   Thanks (0) Thanks(0)   Quote Ttunderthesun21 Quote  Post ReplyReply Direct Link To This Post Posted: 11/27/2019 at 8:39am
Yeah I noticed the same thing the other day, I was wondering how you would go about getting this fixed. I'm assuming it's a mistake, so you just contact USATT and ask them to look at it again?
Back to Top
Ttplayer91 View Drop Down
Beginner
Beginner


Joined: 03/21/2019
Location: Va
Status: Offline
Points: 6
Post Options Post Options   Thanks (0) Thanks(0)   Quote Ttplayer91 Quote  Post ReplyReply Direct Link To This Post Posted: 11/27/2019 at 8:46am
Originally posted by Vince64 Vince64 wrote:

Originally posted by ttssbba ttssbba wrote:

if you trace back, this initial rating also seems too low:

There was definitely a error there. His initial rating should have been 1145 after the tournament not the 663. Best win 1165 + worst loss 1127 divided by 2 is 1146 and he would have lost 1 point to the 1383 player

Although inaccurate, I think this may be "normal", his best win was on a initially rated 0, not 1127. When you only beat unrated players, you end with an abnormally low rating ( although it never appears to be initial rating of half your worst loss, I haven't figured out the exact formula).

I have observed this several other times, common in lower rated events or smaller tournaments with lots of new players.

On the flip side, if an unrated player loses all his matches and all were to rated players, they get an initial rating of 1 point less than their worst loss (in cases I have seen). This means someone that's never played ping pong before can enter an open event with many high rated players, lose all their matches without scoring a point and end up with a very high rating. If all you care about is rating, this must be the easiest way to get to 2000 LOLLOL
Back to Top
hunkeelin View Drop Down
Silver Member
Silver Member
Avatar

Joined: 07/22/2013
Location: United States
Status: Offline
Points: 845
Post Options Post Options   Thanks (0) Thanks(0)   Quote hunkeelin Quote  Post ReplyReply Direct Link To This Post Posted: 11/27/2019 at 3:19pm
the rating system is just fine. You need 5~6 tournaments to get to your true rating level and that's common knowledge. 
USATT: Current 2139 as of 11/2019
Back to Top
Lightspin View Drop Down
Super Member
Super Member


Joined: 07/11/2018
Location: USA
Status: Offline
Points: 126
Post Options Post Options   Thanks (0) Thanks(0)   Quote Lightspin Quote  Post ReplyReply Direct Link To This Post Posted: 11/27/2019 at 4:41pm
When there adjustments are applied in a whimsical and arbitrary way, I think there is a problem.  There are also many people from Puerto Rico, Japan and Canada who only play here once in a blue moon and cause havoc to the rating system.   In the first post, the guy should have been adjusted much higher.  The people who lost 50 points and 45 points to this 663/1597/1991 player probably aren't too happy today.  Can you imagine being 1850 and losing 50 points to someone who ended up rated 1991? Isn't that weird?
Back to Top
hunkeelin View Drop Down
Silver Member
Silver Member
Avatar

Joined: 07/22/2013
Location: United States
Status: Offline
Points: 845
Post Options Post Options   Thanks (0) Thanks(0)   Quote hunkeelin Quote  Post ReplyReply Direct Link To This Post Posted: 11/27/2019 at 5:04pm
No, not really i won't be mad. It's part of the game, to stay at your current rating you have to be able to keep underrated players from stealing rating from you and be able to challenge higher rated players. It has always been like that. On my second most recent tournament where I achieve 2050, I had to keep few 1900s from beating me. Right after that tournament 2~3 become 2050. 

USATT: Current 2139 as of 11/2019
Back to Top
DonnOlsen View Drop Down
Silver Member
Silver Member
Avatar

Joined: 11/15/2008
Location: Maryland, USA
Status: Offline
Points: 619
Post Options Post Options   Thanks (0) Thanks(0)   Quote DonnOlsen Quote  Post ReplyReply Direct Link To This Post Posted: 11/27/2019 at 6:21pm
Originally posted by Lightspin Lightspin wrote:

When there adjustments are applied in a whimsical and arbitrary way, I think there is a problem.  There are also many people from Puerto Rico, Japan and Canada who only play here once in a blue moon and cause havoc to the rating system.   In the first post, the guy should have been adjusted much higher.  The people who lost 50 points and 45 points to this 663/1597/1991 player probably aren't too happy today.  Can you imagine being 1850 and losing 50 points to someone who ended up rated 1991? Isn't that weird?

Thank you for your comment.  

From my perspective, you make the most salient point: the problem is not with the personal rating outcome of the 663/1597/1991 player, it is with the ramifications upon the player's much-higher-rated losing opponents.  

Now the system has a considerable number of (effectively) underrated players of which there are at least two possible consequences: 1) They may now be eligible for rated events they would not otherwise be eligible for (given their previously demonstrated standard of play), thus promoting a distortion in the field of competitors for those future events, and 2) Upon those underrated players playing successfully at their actual standard of play, they will take an excessive number of rating points from those currently at their correct rating level due to the underrated players' inappropriate rating level, the product of the 663/1597/1991 player experience.

I agree with the contributors here who advocate for an improvement in the adjustment methods in the USATT rating system.

Thanks.  


Edited by DonnOlsen - 11/27/2019 at 6:22pm
Advances in a field are best achieved by constructs that fully reflect the phenomena of interest and are rooted in a theory that specifies their determinants, mediating processes, and its effects.
Back to Top
 Post Reply Post Reply
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 12.01
Copyright ©2001-2018 Web Wiz Ltd.

This page was generated in 0.113 seconds.

Become a Fan on Facebook Follow us on Twitter Web Wiz News
About MyTableTennis.NET | Forum Help | Disclaimer

MyTableTennis.NET is the trading name of Alex Table Tennis Ltd.

Copyright ©2003-2019 Alex Table Tennis Ltd. All rights reserved.