TLS Rating vs Tennisrecords

Discussion in 'Adult League & Tournament Talk' started by Thatshot, Aug 23, 2017.

  1. OnTheLine

    OnTheLine Professional

    Joined:
    Mar 8, 2017
    Messages:
    1,057
    I think you are right and that is part of the problem. If you have a C rating, it seems to be VERY sticky.

    If you play only at level, even if you crush it at level. A 3.0 on my 3.5 team had a stellar 2016, played 53 matches, lost only 8 matches total including wins at districts and sectionals ... did not get bumped, but had only played at level. This year, she is only playing "up" and I am certain she will get the 3.5C rating.

    It makes no sense that if you want to go up to the next level that you seemingly have to play up, thus diluting the entire rating system in the process.
    To get more favorable pairings ... captain. I think that was a good move on my part. I have a favorite partner, we play like soulmates, it is absolute poetry, otherwise I play with people who are weakest on the team and we have fun and figure out how to win or come close.

    On TLS, you need at least 3 matches per league for it to calculate anything, and it seems to scrape data weeks before it does an update so a lot of matches just aren't there.
     
  2. MisterP

    MisterP Professional

    Joined:
    Jul 13, 2012
    Messages:
    1,214
    I think our level of play is also kind of sticky. We all have a range that we play within, and that fluctuates a lot depending on many things, but who we play has something to do with it. When I play a high level 4.0, the mental and physical demands on me are higher and I have to raise my game. Conversely, if I were to play a low 3.5, the demands on me are lower, which will eventually affect my performance. I may bagel them in the first set, and possibly even lose the second set because it's difficult to stay focused when the level drops that much.
     
  3. OnTheLine

    OnTheLine Professional

    Joined:
    Mar 8, 2017
    Messages:
    1,057
    Quite insightful, and I wish it didn't apply to me, that I played at my best no matter the opponent. Alas, I play up to my opponent and conversely can play down to my opponent at times.
     
  4. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    It is true that having a good record at-level does not guarantee a bump up. But it certainly does happen too.

    When it doesn't happen, it is often because the player happens to rack up the good record playing against low to mid at-level players. This is sometimes because playing up is common or even rampant in their league and they end up playing lower level players, or they play on a strong team and always play court 1 and opponents know they are strong and stack playing weaker players against them. In this case, they are expected to win and "just" winning may not improve their rating.

    The reports I do will show this clearly as I show both the average opponent NTRP level and average opponent dynamic rating. So for example, if a 3.0 plays only at level, if they played only at-level players, their average opponent NTRP level would be 3.0 and if they played on average an average 3.0, you'd expect their average opponent dynamic rating to be 2.75. And if this player plays up at 3.5, you'd expect those to be higher. These stats will often explain why someone with a good record doesn't have a higher rating.

    And of course, for the same reason, playing up does not guarantee a bump up. I've seen many cases where someone plays up, but does so against others playing up. This is no different than playing those same players at-level.
     
  5. Startzel

    Startzel Hall of Fame

    Joined:
    Jan 21, 2015
    Messages:
    2,366
    I like the fact you feel the need to justify your sandbagging on here lol.
     
    J011yroger likes this.
  6. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    This is a huge part of it too. If you play to the level of your opponent when you play at-level, that effectively dilutes the "good" you've done from playing up and doing well. To some this may appear to be sandbagging by the player, but it is just human nature and difficult to maintain one's best tennis for an entire match against weaker players.

    And the ratings tend to be sticky because a lot of players are pretty stagnant in their playing level. And the ranges are large enough that players can improve and still be correctly the same level. Remember, 80-85% of players stay the same level at the end of each year. Of course, one could argue this is the effect of a system that is too stagnant and not a fair representation ...
     
  7. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    This is a huge part of it too. If you play to the level of your opponent when you play at-level, that effectively dilutes the "good" you've done from playing up and doing well. To some this may appear to be sandbagging by the player, but it is just human nature and difficult to maintain one's best tennis for an entire match against weaker players.

    And the ratings tend to be sticky because a lot of players are pretty stagnant in their playing level. And the ranges are large enough that players can improve and still be correctly the same level. Remember, 80-85% of players stay the same level at the end of each year. Of course, one could argue this is the effect of a system that is too stagnant and not a fair representation ...
     
  8. ChaelAZ

    ChaelAZ Hall of Fame

    Joined:
    Oct 23, 2014
    Messages:
    2,099
    Location:
    Arizona
    And what if...if....if I don't get my 4.0 according to the USTA? So my playing 3.5 as a 3.5 is still sandbagging? I defer to your extensive expertise of identifying sandbaggers.

    We should play sometime.
    I think it would be fun.
     
    FedLIKEnot likes this.
  9. Startzel

    Startzel Hall of Fame

    Joined:
    Jan 21, 2015
    Messages:
    2,366
    This is the problem. Even if it’s artificial there needs to more more movement in the rankings to prevent the same folks from winning every year.
     
  10. MisterP

    MisterP Professional

    Joined:
    Jul 13, 2012
    Messages:
    1,214
    I am curious how often TR is updating. Seems like we had a flurry of updates a week or two ago, and now nothing.
     
  11. ChaelAZ

    ChaelAZ Hall of Fame

    Joined:
    Oct 23, 2014
    Messages:
    2,099
    Location:
    Arizona
    Hmmm...have over 10 matches to date this season. Hopefully I will get on the radar sometime.
     
  12. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    I've often thought moving the thresholds up or down 0.25 once in awhile just to change things up might be a good idea. This wouldn't just be bumping players up, but redefining what a 3.0, 3.5, 4.0, etc. is.

    The problem is how people would react and try to manipulate things, but it would make a whole new set of players the "top" of the level for a year.
     
  13. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    Like the great bump of 2010?

    J
     
  14. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    Perhaps it had the same effect, but I think that was an increase in rating to redistribute things. But yeah, probably 6 of one and half a dozen of the other.
     
  15. NumbersGuy

    NumbersGuy Rookie

    Joined:
    Jun 5, 2012
    Messages:
    246
    The trouble would be that a great many teams would then need to reconstitute themselves (as opposed to the few highly-successful ones who currently have to each year), lots of friends who had been playing together won't be able to, and the squawking would be too loud for USTA's tastes. But an interesting idea.
     
  16. OnTheLine

    OnTheLine Professional

    Joined:
    Mar 8, 2017
    Messages:
    1,057
    Beating computers is not so easy while playing at level.

    Formula to move up without playing up?
    Seems either play singles and win by good margins
    OR
    Play with the lowest ranked player on your roster on the highest court where you are "supposed" to lose big, and beat the computer formula by enough
     
  17. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    A bunch of lifers would suddenly be king of the hill, and a bunch of high and mighty folk would be playing third dubs.

    J
     
  18. Moveforwardalways

    Moveforwardalways Hall of Fame

    Joined:
    Mar 7, 2015
    Messages:
    2,729
    If you can't beat players at your level by enough in singles, then you have no business moving up. The vast majority of people who get bumped up do so without playing up. They just get good enough to beat players at their level by enough to move the computer rating.

    For instance, if you are playing against a player whose dynamic rating is 3.3 and you cannot play well enough to get a 6-2 6-2 victory and earn a rating of 3.6 (or whatever) for that match, you should not be moving up. True 4.0 players have no trouble beating 3.3 rated players handily.
     
    schmke likes this.
  19. MisterP

    MisterP Professional

    Joined:
    Jul 13, 2012
    Messages:
    1,214
    That logic completely ignores variation within your game as well as how play styles match up. I've beaten 3 players with 3.7, or 3.8 ratings in singles in the last month in straight sets and during that time won a razor thin match against a guy with a 3.4 (he was a serve/volleyer which I don't see much). Levels go up and down with the day, and some players have more variation than others.

    Now, if 4.0 players are regularly losing matches to 3.4 and below, that's maybe an issue.
     
    Last edited: Nov 14, 2017
    5sets likes this.
  20. downs_chris

    downs_chris Professional

    Joined:
    Jun 14, 2009
    Messages:
    1,142
    Location:
    Northern VA
    ratings are out when? december 1st?
     
  21. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    That's what I was told by our league coordinator.

    J
     
  22. ncgator

    ncgator Rookie

    Joined:
    Oct 14, 2007
    Messages:
    271
    Location:
    Concord, NC
    Whatever happened to Monday after Thanksgiving for publishing year-end ratings?
     
  23. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    You act like they actually give a crap.

    J
     
  24. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    Yes, playing styles can cause results to vary, but this is in both directions, and even with regular ups and downs, play enough matches and the highs/lows even out and the system arrives at a rating that is about right for you on average. If you can't play well enough on average to get bumped up, you won't be bumped up and like @Moveforwardalways says if you can't beat mid/upper at-level players regularly and/or somewhat easily at times, you haven't demonstrated you should be bumped up.

    Now, one could make the argument that your rating should not reflect your "average" performance but instead, like golf, reflect something more like your average good day. That would likely result in more players being bumped up, at least those that have highs and lows and are just steady eddies, but the system as we have it today doesn't work that way.
     
  25. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    Doesn't the current system weight your most recent matches and therefore not count your average or best performance over the season but rather your most recent few?

    J
     
  26. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    Yes, that is correct, but it is still an "average" and unless you are improving and playing better at the end of the year than the beginning (or playing worse at the end than the beginning), the highs/lows still cancel each other out. My point was that the variance @MisterP brought up typically both helps and hurts a player's rating and generally cancels itself out given enough matches. But if the variance goes in streaks, then yes, how they did lately matters more.
     
  27. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    Think about me, I played 16 post season matches so my rating for the regular season 2018 isn't based on regular season 2017.

    J
     
  28. MisterP

    MisterP Professional

    Joined:
    Jul 13, 2012
    Messages:
    1,214
    If we're going to change anything I think there should be separate ratings for singles and doubles matches. I know 4.5 doubles players I could take to the woodshed at singles. Doesn't mean I'm better than they are.
     
    nytennisaddict likes this.
  29. Jim A

    Jim A Professional

    Joined:
    Dec 29, 2008
    Messages:
    1,120
    The league season doesn’t officially end until after this weekend, correct? Last year it was 2 weeks earlier so 12/1 seems right. Gives them 2 work weeks to get everything finalized. We are losing 2 this year, would be shocked if any else gets bumped.


    Sent from my iPhone using Tapatalk
     
  30. Shankadelic

    Shankadelic New User

    Joined:
    Sep 27, 2017
    Messages:
    86
    If USTA followed your recommendation and one is computer rated as a 4.0 in singles and 4.5 in doubles. Are you saying the person could then play singles for a 4.0 team and then singles or doubles for a 4.5 team?
     
  31. kevrol

    kevrol Professional

    Joined:
    Jan 14, 2014
    Messages:
    1,004
    The first problem with this idea is there are a ton of players who don't play enough doubles or singles to get an accurate rating. In Southern only the spring season counts. Most of these seasons are only 8 - 10 matches and folks probably only play 2/3 of those matches at most. So you're talking about trying to get a statistically valid sample size from that. You can't further cut that number in half and have any confidence in the data.
     
  32. BeyondTheTape

    BeyondTheTape Rookie

    Joined:
    Aug 15, 2013
    Messages:
    217
    Just saw this. The Web version of USTA league page has a Countdown in a red border. "Countdown to NTRP Published Ratings" Interesting.

    That date is Dec 1.

    Let the anxiety begin! :eek:
     
  33. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    Yeah, I'd been told 12/1 was the date so that is no surprise, it is interesting they have a timer to the second going and that time is midnight ET on the 1st. Can you say unintentional denial of service attack when everyone starts hitting refresh at that time?
     
    OnTheLine likes this.
  34. kevrol

    kevrol Professional

    Joined:
    Jan 14, 2014
    Messages:
    1,004
    Pretty surprised they are painting themselves into a corner on the release date this year.
     
    schmke likes this.
  35. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    Maybe they need that to put pressure on whomever is doing it to make sure it gets done! :eek:
     
  36. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    Odds of it actually being done on time?

    J
     
  37. OnTheLine

    OnTheLine Professional

    Joined:
    Mar 8, 2017
    Messages:
    1,057
    Weren't they earlier than expected last year?
     
  38. J011yroger

    J011yroger G.O.A.T.

    Joined:
    Feb 2, 2007
    Messages:
    15,052
    Location:
    Long Island, NY
    That doesn't sound like the USTA. Maybe you are thinking of something else.

    J
     
    FedLIKEnot likes this.
  39. kevrol

    kevrol Professional

    Joined:
    Jan 14, 2014
    Messages:
    1,004
    No they were much later than expected this last year IIRC.
     
  40. OnTheLine

    OnTheLine Professional

    Joined:
    Mar 8, 2017
    Messages:
    1,057
    I think they said 12.1 last year and they came out 11.30
     
  41. BeyondTheTape

    BeyondTheTape Rookie

    Joined:
    Aug 15, 2013
    Messages:
    217
    Someone had a post that showed the actual day the ratings were published over the last 5 years. maybe @schmke ?

    Generally published that Monday after Turkey day. Last year as @kevrol mentioned was later than normal, that Wednesday.

    historically 12.1 seems to be a general target date.
     
  42. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    Yeah, I did write something up, probably here somewhere, but it is more or less 12/1 most years plus or minus a day. Now sometimes, the "expectation" of early may have been present but I don't know that earlier is ever the official target.
     
  43. nytennisaddict

    nytennisaddict Legend

    Joined:
    Aug 20, 2007
    Messages:
    6,191
    Location:
    Westchester (lower), NY
    I agree with @MisterP that there should be separate singles/doubles ratings (just like the atp, wta, utr, tencap, etc...)
    but also agree, that there is no easy practical way to implement that in a team environment without revamping how teams are allowed to be formed.

    i'm a big fan of the way platform tennis teams are organized.
    there are no ratings. you can recruit whomever you like.
    there are flights rather than ntrp levels.
    if your team wins it's flight, they move up a flight.
    if your team is last in it's flight, they move down a flight.

    teams that want to be competitive, will naturally select the best players than can find (good players gravitate to the best teams)
    teams that want to be social, can be. ie. if joe4.0 gets bumped up to 4.5 but really wants to play with his buddies at 4.0, he still can
    want to play with your daughter, wife, etc... in a men's league.. np
    team full of ex-d1 women's players with no women's 5.0 league to play in.. np
    no more crap about sandbagging... just bad team captaining/recruiting... or you're just a bad team, but eventually youll go to the flight you belong in.
     
    Moveforwardalways likes this.
  44. rod99

    rod99 Professional

    Joined:
    Sep 25, 2005
    Messages:
    1,269
    in the past the ratings have been released gradually over the couple of hours.
     
  45. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
  46. Moveforwardalways

    Moveforwardalways Hall of Fame

    Joined:
    Mar 7, 2015
    Messages:
    2,729
    So, basically ALTA in Atlanta.
     
    nytennisaddict likes this.
  47. OnTheLine

    OnTheLine Professional

    Joined:
    Mar 8, 2017
    Messages:
    1,057
    So TLS has not updated in my region since 10/17 and only on matches through 9/23. Here is my guess, they will not do a full update for all regions before the actual new ratings come out .... total joke.
     
  48. nytennisaddict

    nytennisaddict Legend

    Joined:
    Aug 20, 2007
    Messages:
    6,191
    Location:
    Westchester (lower), NY
    never played ALTA, but if it's that way, i can see why ALTA > USTA in that area.
    sounds like a fun area to live, if you're a tennis addict :p
     
  49. brettatk

    brettatk Rookie

    Joined:
    Jan 19, 2015
    Messages:
    361
    Location:
    Metro Atlanta
    I play both ALTA and USTA. Been playing ALTA much longer though. But there is still a lot of sandbagging going on in ALTA. People are always going to find a way to cheat if they really want to.
     
  50. schmke

    schmke Hall of Fame

    Joined:
    Jan 20, 2010
    Messages:
    2,445
    I'm curious, how does sandbagging occur in a format like ALTA uses? Is it through recruiting new players that should be on a team at a higher level? Players moving from one team to another at a lower level than they should be at? Or do teams tank matches to avoid being promoted or even to get demoted? And what if any provisions does ALTA have to try to prevent any of this?
     

Share This Page