Wednesday, August 16, 2023

Should computer rated player be able to be disqualified and promoted in USTA League play?

For those that play USTA League, it is pretty well known that new players, and those with expired ratings or a few other scenarios, are required to self-rate to determine the level they will play at when they join a league/team.  Since the self-rating isn't perfect, there is a system in place where a player's results are evaluated to determine if they are out of level, and if so, they can be disqualified from that level and promoted mid-season.  This is known as the three-strike DQ process.

It is also well known that players that have a current computer (C) rating are not subject to this three-strike DQ process, meaning they can play the entire year and improve and be out of level, even way out of level, during the year, and they are "golden" and can't be DQ'd or promoted.

See my FAQ question 16 for more discussion.

What can be interesting to look at is how players get their strikes.  Generally speaking it is when they win very easily over decent to good at-level players, but there are times when score lines that don't look out of whack can end up being strikes.

I was recently made aware of once such situation at NorCal Sectionals where a self-rated player lost a match in straight sets, and got their third strike!  On the surface that may not seem possible, but indeed it is, and in this case it was because the opponent they lost to was rated significantly higher than they were and so the loss was a "good" loss and rated very high.  And yes, my ratings agree with the three strikes and DQ.

So how is it fair for a self-rated player to be disqualified for being out of level if they are playing someone, at the same level, who is expected to and does beat them in straight sets?  How high does this opponent have to be rated?

For starters, the threshold for a player to get a strike isn't just the top of their level.  The USTA gives some leeway for players to improve and be above level before they begin accruing strikes.  The leeway given is greater at lower levels because normal improvement can happen faster at lower levels.  I think the threshold may be too high in the first place, but they are what they are.

For discussion purposes, lets say the threshold in question for this case was 0.2 higher than the top of the level and lets assume the level is 4.0, that puts the strike threshold at 4.20, meaning the self-rated player got their rating above that three times.  For them to have such a rating and be expected to lose in straight-sets to an opponent means that opponent's rating is well north of that, perhaps as high as 4.35-4.40.

This means we have a 4.0 match being played between someone rated around/over 4.2 vs someone around 4.4!  How is a "normal" 4.0 player supposed to compete?

I don't think anyone would be surprised to know that to advance and do well in playoffs, you need to have players rated near the top of the level, and even into the next level.  Players do improve, especially those that get into the competition and are trying to advance and put in hours of practice to get better, and there is something to be said for not penalizing someone that puts in this work.  But at what point is there too much improvement?

With self-rated players the USTA does set a threshold the player cannot improve past, but should C rated players have such a threshold too?

The situation I outline may be an outlier and you don't want to change the entire system for just a few outliers, but I also don't think implementing something to catch the outliers means the other 99% of players will be affected.  In fact, that 99% of players may be affected in a positive way if they don't have to face such out of level players.

Note, it is unfortunate but the problem is exacerbated at times by players willing to manage scores, throw matches, or other shenanigans to get bumped down or ensure they don't get bumped up, so they can be one of these "top of level" players.  A threshold for C rated players to be DQ'd would most likely be targeting these players as they genuinely should be at the higher level.  Might it catch the occasional legit player who just improved "too much"?  Sure, but regardless of how a player gets there, isn't it more fair to the 99% to promote the out of level player and not force at-level player to compete with them to advance?

I realize it is hard to determine where to draw the line, but I think having some threshold for C rated players is a worthy discussion.  It could perhaps be a bit higher than the self-rate strike threshold to give a little more leeway, but a case could be made it should be the same, or that if it is higher the self-rate strike threshold should be a bit lower to start.

What do you think?  Should Computer rated players get a "golden ticket"?  Or should there be some threshold they too can't exceed without being promoted?

26 comments:

  1. Yes, computer rated players should be able to disqualify and be promoted but all matches they play before disqualification and promotion should stand.

    ReplyDelete
  2. I don't think C-rated players should be able to be DQ'd, unless they self-rated dishonestly or are now playing college tennis maybe. The rating system leaves a lot to be desired, but if a player is in the system correctly, they shouldn't have to be promoted midseason.

    Was the match you were talking about singles or dubbs? I assume singles. If an S-rated player is getting a strike at 4.0 via losing a singles match in straight sets, then they shouldn't get a strike no matter what that match is rated for them. If the USTA is allowing a C-rated player rated 4.40 to play 4.0 then don't penalize their opps. That makes no sense. Also, how is it even possible for a 4.0 C, even rated 4.00 to improve so much in a year that their rating is around 4.40? That seems unbelievable.

    Separate question: if a 4.0 S-rated player strikes out during a year and is promoted to 4.5 midseason for example, but let's say has played 10 rated 4.0 matches total, could this player still obtain a 4.0 C rating at the end of the year hypothetically? In other words, will the USTA still calculate his rating like everyone else?

    ReplyDelete
    Replies
    1. Yes, a 4.0C improved so much that they are around 4.4 now. It is rare, but double bumps do happen, e.g. a 4.0 getting to at least 4.51 and being bumped up to 5.0 at year-end. Double bumps are more common at lower levels where improvement is easier, and it almost always involves playing up too.

      Delete
    2. To your other question, no, players that are DQ'd/promoted mid-year cannot end the year at a lower rating than they are promoted to.

      Delete
    3. Oh, right. I've seen double bumps before even for C players. Ray Ohara is the latest I've seen. He's an interesting case. Played some in 2019(2020 season), then not again until 2022. Got a 3.5C for 2019 and/or 2020(which would've been 2021 year-end ratings. So, like many beginners to USTA, maybe just getting back into tennis and struggling more at the start. But, he was clearly playing at a solid 4.0 level for most of 2022. And I guess enough to be 4.5 now. And he's winning at 4.5 sectionals this year now. These types of cases need to remedied somehow. If nothing else, it substantially hurts the ratings of his 3.5 opps when he's a 3.5 with a rating of at least 4.01.

      That's interesting DQ'd players can't be lower than their promoted rating. I guess it depends how they were promoted. If grievanced and upheld, makes sense. But, if they are promoted via striking out, then all of their matches should still be calculated at year-end like everyone else, and if it comes out to be the rating they were before promotion, then they should be able to be that lower rating for the following year. If it was a C player, that C player would retain his/her rating.

      Delete
    4. I believe the idea is if someone is so high as to be DQ'd/promoted, there is no legitimate way they should be allowed down at the end of the year. Some players if DQ'd would tank matches to try to get back down, so just remove that incentive/option.

      Ray, I think the USTA just missed on. Clearly he isn't a 3.5 or 4.0, and my ratings had him as a 4.0 after his first year, the USTA "missed" having him a 3.5. I don't mind when I miss like this and my rating is more accurate/appropriate. But yeah, with the missed year and COVID double year, ample time for a double bump to happen.

      Delete
    5. I see. If it's a promotion via striking out, I think that player should still have all his matches calculated like anyone else though, but just my opinion. If that player is so concerned with staying at his previous rating before striking out, he would've been tanking or managing matches before striking out as well but I see your point. How do they calculate a player's rating then if he strikes out if the calculations say he should be at his previous rating at year-end?

      Delete
    6. I think they just put a floor on their year-end rating and they can't be lower than that.

      Delete
  3. I think the problem would be better solved by not allowing appeals down to a lower level. The vast majority of the people in our area (midwest) that are consistently killing people, then managing wins seem to be people that appeal down. To me, that fixes 99% of this issue.

    ReplyDelete
    Replies
    1. Valid point. I see limited value in allowing auto-appeals down.

      Delete
  4. The rating system is terrible. We recently played a male player at mixed 7.0 sectionals that C rated as 3.5. He was previously bumped up to 4.0 and had a 12:1 win ratio. He appealed down to 3.5 and it was granted. If there is no DQ for C rated players how do we catch crap like this with a system that leave a lot to be desired.

    ReplyDelete
    Replies
    1. If they appealed down, they should be subject to strikes as an 'A' now.

      Delete
  5. I'm actually OK with the current system for C-rated players, even if it's not perfect. There are more pressing fixes I wish USTA would implement, specifically:

    1) Don't allow automatic appeals down. If you want to appeal, fill out a form and have a committee review (just like medical appeals).

    2) Every USTA match should count towards ratings, no matter the league/season. Fine, have a separate Mixed rating if you want, but must play the level of whatever the highest rating is at year-end.

    3) The 3-strike threshold is MUCH too high. For example, as a 3.5S player, the strike threshold is 3.8. You must get 3 strikes at 3.8 and above to be DQ'd automatically. Lower the threshold to something more reasonable to have the system work as intended.

    ReplyDelete
    Replies
    1. I generally agree with each of your points.

      Delete
    2. The rating system is obviously far from perfect. But, while the USTA allows players playing too low, they promote players up too often, too. I think if a player is barely over a threshold like 0.02-0.03, that is so minimal and I'd be ok with allowing them to appeal down if they so chose to.

      Agree there should be a separate mixed rating even if that player has a C rating. But, don't agree they'd have to play the higher of the 2 ratings. They just have 2 different ratings, depending on what they are playing. Also, separate singles/dubbs ratings would make sense, too.

      Is the threshold for strikes really 0.3? That seems way too high. Didn't think it'd be that high. I think strikes are misleading often though, but not sure how to improve on what the computer is spitting out.

      Delete
    3. Players can already try the auto-appeal which will grant them an appeal if they are close, that is basically what you are describing.

      Depending on the level, yes, the threshold can be 0.3 or higher (lower levels). I agree it is way too high.

      Delete
    4. I see players who are A rated sometimes that really shouldn't be able to appeal down. But, I see many more players who are rated too high, so I think allowing a little leeway for appeal downs is acceptable.

      I didn't realize certain levels have a higher threshold for strikes. How does anyone know how much though other than those in charge? USTA doesn't publish these thresholds.

      Delete
  6. Have you ever seen a district not enforce DQs?

    ReplyDelete
    Replies
    1. It is hard to say for certain since the USTA isn't transparent with ratings, but I have seen situations where I would have expected a DQ and one didn't happen.

      Delete
  7. The issue with this is that I dont think that giving strikes to Computer Rated Players promotes League engagement. Since the more matches you play, the higher chance you have of one of those generating a strike. This is the biggest issue I see with the Strike system in general.

    ReplyDelete
    Replies
    1. Strikes are based on one's dynamic rating, not individual matches, so an A, or C if they were subject to strikes, has to have sustained above level play to even get a single strike.

      Delete
  8. I've been on both sides of this, and I think the core of the issue is that NTRP has too wide of bands. If there wasn't such a huge gap between the top and bottom of the levels, people would have less incentive and less ability to manage ratings. I was a 1 year varsity player in college, gave the game up for 26 years, and came back 3 years ago as a 3.5S who literally could not hit a topspin backhand.

    I worked hard on that and other parts of the game, when the computer ran in 2021 it put me up to 4.0, after a 6-4 record in 3.5 adult. At that point, my choice was either to be one of the worst 4.0s in the district and essentially not be competitive in matches (and not be asked to play by captains), or mash the appeal button and be one of the better 3.5es. I spent a year with A rating, and kept working and was clearly ready to move up when it happened last year. But, nobody who played me at the end of 2022 when I was still a 3.5 was particularly happy about it.

    If instead I had been promoted into what was essentially a 3.51 - 3.75 band, I probably would have been like "yeah, ok, this is manageable" and just stayed at assigned level.

    There are lots of reasons why a move to finer-grained bands would be difficult of have other ramifications, but it would take away a lot of incentive around ratings management and promote more competitive matches.

    ReplyDelete
    Replies
    1. Smaller bands would make for more competitive play, but yeah, there are challenges with that, namely getting critical mass for leagues in less populated areas.

      Delete
    2. Smaller bands is ideal, but tougher in smaller populated places and it's too much for nationals. Nationals already spans 6-7 weekends as is, so I don't think this is feasible unless there's no advancement, which is the the main point of playing USTA or a big priority of it for most.

      But yea, it's always tricky when someone is at the top of their level and/or at the bottom of their level, for different reasons.

      Delete
  9. I think C-rated players should be able to be bumped up during the season if they exceed a threshold, but it should require more strikes and there shouldn’t be any of the associated penalties for disqualification. The situation I don’t like is when a self-rated player joins late in the season, plays just enough to get a C rating for the next year, and proceeds to become a prolific player and go something like 31-2.

    ReplyDelete
  10. I beat a 3.45: 13-8, and a 2.9: 12-1 at a recent level 5 tournament. Supposedly I am a 2.9. I'd love to get bumped up. My learning curve was a quick one, as I played in HS and I am super fit. playing 3.0/3.5 isn't fun.

    ReplyDelete