Alex Table Tennis - MyTableTennis.NET Homepage
  New Posts New Posts RSS Feed - New USATT rating calculation
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Forum Home Forum Home > General > General

New USATT rating calculation

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

Joined: 11/15/2008
Status: Offline
Points: 126
Post Options Post Options   Thanks (0) Thanks(0)   Quote DonnOlsen Quote  Post ReplyReply Direct Link To This Post Topic: New USATT rating calculation
    Posted: 06/12/2018 at 11:38am
This is from the results of an Ohio tournament

Pre Tournament Rating:Not Rated
Adjusted To:2184
Post Tournament Rating:2184

WinnerLoserEventScore+/-Head to Head
Wahab, Wale (2184)  Lamp, Jonathan (1922)Open RR3,7,20vs
Dubina, Samson (2446)  Wahab, Wale (2184)Open RR12,-9,6,-7,120vs
Lin, Hong (2496)  Wahab, Wale (2184)Open RR-10,7,4,110


This initial rating seems too low.  Is this algorithmic-derived?
Donn
Back to Top
pgpg View Drop Down
Silver Member
Silver Member
Avatar

Joined: 11/18/2013
Location: United States
Status: Offline
Points: 880
Post Options Post Options   Thanks (0) Thanks(0)   Quote pgpg Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 11:54am
This looks like initial rating assignment of Average(BestWin+WorstLoss) = (1922+2446)/2 = 2184.

After that there were no points gained or lost due to difference in ratings. Sounds like old system behavior.

Edit: I guess I see your point that loss to 2446 player was in 5 games, so one could argue that he's still underrated. More matches will fix that. 


Edited by pgpg - 06/12/2018 at 12:01pm
USATT: ~1800
Dr. N Barricade - AK47 blue - Air UpUpUp (0.6mm)
Back to Top
Vince64 View Drop Down
Gold Member
Gold Member


Joined: 07/13/2013
Location: Renton, WA.
Status: Offline
Points: 1096
Post Options Post Options   Thanks (0) Thanks(0)   Quote Vince64 Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 11:56am
Originally posted by pgpg pgpg wrote:

This looks like initial rating assignment of Average(BestWin+WorstLoss) = (1922+2446)/2 = 2184.

After that there were no points gained or lost due to difference in ratings. Sounds like old system behavior.
What he said. Lol
https://www.facebook.com/SeattlePacificTableTennisClub/timeline
Back to Top
DonnOlsen View Drop Down
Super Member
Super Member
Avatar

Joined: 11/15/2008
Status: Offline
Points: 126
Post Options Post Options   Thanks (0) Thanks(0)   Quote DonnOlsen Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 1:18pm
Thank you for your comments.

Edit: I guess I see your point that loss to 2446 player was in 5 games, so one could argue that he's still underrated. More matches will fix that. 

Yes, this is certainly true, however there are victims ahead, as they will lose many more points than if his rating was considerably higher.  This is why I have argued, forever, that we should skew our initial ratings in a manner that results in higher, not lower, ratings, so that the established player community (with presumably correct ratings) may gain points as the new player loses points down to the player's correct rating, not the opposite.  

I have seen this negative phenomenon many times, resulting in much disgruntlement from those negatively affected.  An example here would be if this player entered U2200 and dominated 2100 players that anticipated that they would be competitive (even, perhaps, win) in the event.  There is the possible compensatory salvation of a well-timed rating adjustment, however this is not reliable.

Thanks again. 
Donn
Back to Top
hunkeelin View Drop Down
Silver Member
Silver Member
Avatar

Joined: 07/22/2013
Location: United States
Status: Offline
Points: 557
Post Options Post Options   Thanks (0) Thanks(0)   Quote hunkeelin Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 1:52pm
Originally posted by DonnOlsen DonnOlsen wrote:

Thank you for your comments.

Edit: I guess I see your point that loss to 2446 player was in 5 games, so one could argue that he's still underrated. More matches will fix that. 

Yes, this is certainly true, however there are victims ahead, as they will lose many more points than if his rating was considerably higher.  This is why I have argued, forever, that we should skew our initial ratings in a manner that results in higher, not lower, ratings, so that the established player community (with presumably correct ratings) may gain points as the new player loses points down to the player's correct rating, not the opposite.  

I have seen this negative phenomenon many times, resulting in much disgruntlement from those negatively affected.  An example here would be if this player entered U2200 and dominated 2100 players that anticipated that they would be competitive (even, perhaps, win) in the event.  There is the possible compensatory salvation of a well-timed rating adjustment, however this is not reliable.

Thanks again. 

If a player win 2 or more opponent above a certain rating, his initial rating will get adjusted. So don't worry about innocent opponent losing points. 
FH: h3 2.2 39
t05: 2.1
viscaria: 93g
Back to Top
DonnOlsen View Drop Down
Super Member
Super Member
Avatar

Joined: 11/15/2008
Status: Offline
Points: 126
Post Options Post Options   Thanks (0) Thanks(0)   Quote DonnOlsen Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 2:14pm
Thanks for the comment.

If a player win 2 or more opponent above a certain rating, his initial rating will get adjusted. So don't worry about innocent opponent losing points. 

Yes, this is true.  Another negative consequence pertains to the event's prize money.  I know many players over 2000 rated are very strategic about events they enter due to their interest in winning the prize money and realizing a return on their money invested.  In some scenarios, although the loss is not rating points, it is in the opportunity loss of money not won.

Thanks...
Donn
Back to Top
mhnh007 View Drop Down
Platinum Member
Platinum Member
Avatar

Joined: 11/17/2009
Status: Offline
Points: 2655
Post Options Post Options   Thanks (0) Thanks(0)   Quote mhnh007 Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 2:36pm
Don't you get a 'T' (temp) beside the rating until you play with enough players?  I think in Canada, you do not get dink for point if loosing to a player with a T-rating.  I am not sure if you get points for beating a player with a T-rating though.  Can some one chime in please...

EDIT:  You get point winning against T-rating player.


Edited by mhnh007 - 06/12/2018 at 4:13pm
Back to Top
pgpg View Drop Down
Silver Member
Silver Member
Avatar

Joined: 11/18/2013
Location: United States
Status: Offline
Points: 880
Post Options Post Options   Thanks (0) Thanks(0)   Quote pgpg Quote  Post ReplyReply Direct Link To This Post Posted: 06/12/2018 at 3:47pm
Originally posted by mhnh007 mhnh007 wrote:

Don't you get a 'T' (temp) beside the rating until you play with enough players?  I think in Canada, you do not get dink for point if loosing to a player with a T-rating.  I am not sure if you get points for beating a player with a T-rating though.  Can some one chime in please...

Never ran into this in USATT, this 'T-rating' label must be a Canadian thing... Wink

It does make sense, though. RatingsCentral deals with it by quoting uncertainty range for your rating, which is a more scientific way of doing this.  

As far as future potential hurt feelings (likely mitigated by adjustments as other pointed out) and prize money implications for this specific player - this situation is likely considered to be an edge case and an effort to fix it by USATT (e.g. by taking game scores into account) is probably too high compared to the scale of problem. 
USATT: ~1800
Dr. N Barricade - AK47 blue - Air UpUpUp (0.6mm)
Back to Top
BH-Man View Drop Down
Premier Member
Premier Member
Avatar

Joined: 02/05/2007
Location: United States
Status: Offline
Points: 4359
Post Options Post Options   Thanks (0) Thanks(0)   Quote BH-Man Quote  Post ReplyReply Direct Link To This Post Posted: 06/13/2018 at 3:02am
Originally posted by DonnOlsen DonnOlsen wrote:

Thank you for your comments.

Edit: I guess I see your point that loss to 2446 player was in 5 games, so one could argue that he's still underrated. More matches will fix that. 

Yes, this is certainly true, however there are victims ahead, as they will lose many more points than if his rating was considerably higher.  This is why I have argued, forever, that we should skew our initial ratings in a manner that results in higher, not lower, ratings, so that the established player community (with presumably correct ratings) may gain points as the new player loses points down to the player's correct rating, not the opposite.  

I have seen this negative phenomenon many times, resulting in much disgruntlement from those negatively affected.  An example here would be if this player entered U2200 and dominated 2100 players that anticipated that they would be competitive (even, perhaps, win) in the event.  There is the possible compensatory salvation of a well-timed rating adjustment, however this is not reliable.

Thanks again. 
 

Yup,

Over time, an influx of better players who get initially rated too low has a real deflationary drag overall on the ratings. What was a 2000 level player in mid 2000s is now a 1800-1900 player depending on where they play. We see this at the top end too. Some of it is we simply have developed players to higher skills and some of it is they are coming into USA, then also blooming. If the bloom is not rapid fire, then it is a dragging effect on ratings overall.

This is a natural equilibrium given a situation where there are the same overall points, yet more of the same number have gotten better and there are no more points to give out. Over time, the same rating number simply becomes a stronger level.

It is real tough to keep the rating points consistent with level over time. One could argue subjective judgement of officials to correct this, but that could get out of hand too.

Ratings points for ranking purposes, even if over time have deflated significantly, still do the job for ranking and seeding players. Even if the same numbers do not mean the same (they got stronger), relative to everyone, the ranking is pretty much still fair.

Yet, the hardcore BACKBONE supporting USA table tennis DEMANDS "Justice" with the ratings points and unfortunately
Korea Foreign Table Tennis Club
Search for us on Facebook: koreaforeignttc
Back to Top
BH-Man View Drop Down
Premier Member
Premier Member
Avatar

Joined: 02/05/2007
Location: United States
Status: Offline
Points: 4359
Post Options Post Options   Thanks (0) Thanks(0)   Quote BH-Man Quote  Post ReplyReply Direct Link To This Post Posted: 06/13/2018 at 3:08am
pgpg, Were you forming back in the day when we debated the daylights outta this on about.com forum?

Your "butthurt" theory is a true one and would have added good fuel to the debate back then.

On another note of the "Strategic Professional Sandbaggers" Donn makes good note of, East Coast has a few, just look at who is always taking a picture with the Westchester Owner and Head Managing Coach every month like clockwork...

At the top end of every stacked division awaits players who are several levels better players than the rest of the field. To reach the finals, you have to play at a level 2 or more higher than your rated level. To reach the finals month after month in this situation, you have to be 2-3 levels better consistently. Over time, a rating would kick in and such a player would be closer rated to his/her truer playing level. A loss in group stage is an excellent tool to keep the rating in check, while collecting the champion's prize check.
Korea Foreign Table Tennis Club
Search for us on Facebook: koreaforeignttc
Back to Top
pgpg View Drop Down
Silver Member
Silver Member
Avatar

Joined: 11/18/2013
Location: United States
Status: Offline
Points: 880
Post Options Post Options   Thanks (0) Thanks(0)   Quote pgpg Quote  Post ReplyReply Direct Link To This Post Posted: 06/13/2018 at 7:57am
Originally posted by BH-Man BH-Man wrote:

pgpg, Were you forming back in the day when we debated the daylights outta this on about.com forum?

Your "butthurt" theory is a true one and would have added good fuel to the debate back then.

On another note of the "Strategic Professional Sandbaggers" Donn makes good note of, East Coast has a few, just look at who is always taking a picture with the Westchester Owner and Head Managing Coach every month like clockwork...

At the top end of every stacked division awaits players who are several levels better players than the rest of the field. To reach the finals, you have to play at a level 2 or more higher than your rated level. To reach the finals month after month in this situation, you have to be 2-3 levels better consistently. Over time, a rating would kick in and such a player would be closer rated to his/her truer playing level. A loss in group stage is an excellent tool to keep the rating in check, while collecting the champion's prize check.

I did catch the tail of about.com forum, and, yes, it was 'lively'... 

I actually don't have a well-developed 'butthurt' theory, mostly just an observation that  it's really hard to get an accurate rating estimate from just 3 matches, so some inaccuracies are unavoidable in the short term.

In general, if people cared less about the ratings (and perhaps about winning as well) at an amateur level, TT world would be in better place. I'm beginning to enjoy non-sanctioned tournaments more, may be because of that.

As far as your sandbagger example - I'm following Westchester results quite closely, and lately I don't see the same people winning U2350 and similar events, so I guess better competition takes care of that. 
USATT: ~1800
Dr. N Barricade - AK47 blue - Air UpUpUp (0.6mm)
Back to Top
 Post Reply Post Reply
  Share Topic   

Forum Jump Forum Permissions View Drop Down



This page was generated in 0.141 seconds.
Mark all posts as read :: Delete cookies set by this forum

Cookies and JavaScript must be enabled on your web browser in order to use this forum


Copyright © 2003-2013 MyTableTennis.NET - All Rights Reserved. Disclaimer