Fencing ratings

Discussion in 'Fencing Discussion' started by gillaspy, Apr 1, 2007.

  1. peet

    peet Podium

    Joined:
    Jan 8, 2003
    Messages:
    2,253
    Likes Received:
    241
    They are. As has been reported in some of those threads and elsewhere, the USFA is considering lots of different ideas about tournaments, ratings and the like.

    Change happens slowly though, especially in the USFA.

    -p
     
  2. Sciurus-Rex

    Sciurus-Rex Rookie

    Joined:
    Jun 15, 2005
    Messages:
    1,230
    Likes Received:
    234
    With respect to the fencer in question -- gawd bless 'm for even trying to wrangle the data -- these are bad examples on which to base an opinion. The calculations are constantly being jiggered, values realigned, and the people involved keep trying to justify whether the results *seem* to match their expectations under the current system. ... Yes, it's an interesting experiment, and, if anything, reveals potential problems in DESIGNING such a system. But people here would be doing themselves a disservice to put too much weight behind the experience in order to discount the value of an ELO system itself.
     
    Redblade and Wise-Epeeist like this.
  3. Allen Evans

    Allen Evans Podium

    Joined:
    Oct 25, 2003
    Messages:
    4,772
    Likes Received:
    812
    All politics are local.

    I think that SR's comments points out something that we've overlooked in the VA experiments: that the ELO systems seem sensitive to the sort of mixes of fencers you see at a local event, and may not be applicable at this level of fencing -- potentially a huge weakness. Of course I would suspect that distribution of skills at chess isn't too much different than the distribution of fencing skills: a lot of people at the bottom and only a few at the top, and chess seems fine with an ELO system.

    A
     
  4. swordsen

    swordsen Rookie

    Joined:
    Jul 11, 2000
    Messages:
    1,369
    Likes Received:
    26
    To say as an earlier poster did that people will stop going to events because they won't be able to chase whatever letter they are working on is just plain foolish. Those obsessive folks will go because they have something better, a constantly changing little number that they can glory in when it goes up and agonize over if it goes down. It will result is some folks analyzing who else is going to tournaments very closely to see if it is a good investment for them to go. Which will encourage preregistration. It will probably up the use of askFred a lot.
    In fact, peet should just take one of these algorithms and plug it into Fred and let a system grow and see who people like it. It would be an interesting addition to the program but probably a PITA to do.
     
  5. peet

    peet Podium

    Joined:
    Jan 8, 2003
    Messages:
    2,253
    Likes Received:
    241
    This has already been done some, and I'm still working on more. Last year I ran all FRED's data through the Table Tennis rating system, and it was very interesting. There's no front end to the data, so it can't be shown to the public, but the numbers were interesting.

    It was mostly a proof of concept that one could use the existing data and run it through any number of algorithms, and then compare the outcomes to help the USFA make decisions.


    -p

    Apparently according to the TT system, I'm not quite as good an epeeist as my USFA rating would have you believe....
     
  6. tkexi991

    tkexi991 Rookie

    Joined:
    Oct 28, 2005
    Messages:
    212
    Likes Received:
    18
    While I get frustrated at my difficulties towards getting the next rating (wrong people make the 8, I get stuck in a pool so unbalanced that 4of5 of us make the 8, etc.) the idea that folks on national teams or who have placed in the top 8 or above in international competition are only Us is ridiculous and a lot easier to fix vs. scrapping the entire system. The winner of XYZ International Competition doesn't belong in a NAC fencing Div III Women's Epee. If you place in the top 16 in an International Tournament, you should get a C for that year period.
     
  7. wbowman

    wbowman Made the Cut

    Joined:
    Sep 6, 2004
    Messages:
    570
    Likes Received:
    21
    I remember a numerical system being proposed quite some time ago, and it occurred to me then that there would be a major problem with a major lag in the reporting of results and then dissemination of any changes back out (in time to affect a fencer's next competition). With the increased use of computer-based tournament-management and reporting systems, that would be less of a problem, but there would still be places that didn't get results into the central database in a timely fashion, if at all.
    After seeing all the math involved in the current VA Division discussion (see Cvl's post) and understanding almost none of it, it appears to me that whatever flaws the current system has, it is still relatively transparent. Even the most mathematically-challenged of us can understand how it works and track our results. I'm not sure I really want to be at the mercy of someone who knows more math than I do to tell me whether my rating has changed.
     
  8. eac

    eac Rookie

    Joined:
    Oct 12, 2005
    Messages:
    1,374
    Likes Received:
    117
    peet: Is there any chance of publishing the results of that system, say on a peripheral website just so we can see what it would look like? I'm very interested in those.
     
  9. FoilyDeath

    FoilyDeath Rookie

    Joined:
    Mar 2, 2005
    Messages:
    330
    Likes Received:
    10
    Personnaly, my main problem with the rating system is the flukyness of it all...you got your A, you keep it. You could lose to 900 Cs, you could still have your A. Thats a big advantage of a ranking system like the UK has. Of course, you get the problem of fencers attending a whole bunch of comps getting way higher rankings than people who can't afford too then.
     
  10. qatet

    qatet DE Bracket

    Joined:
    Jan 27, 2003
    Messages:
    722
    Likes Received:
    100
    Care to describe?
     
  11. BySword

    BySword Rookie

    Joined:
    Mar 18, 2006
    Messages:
    361
    Likes Received:
    6
    I believe they rank every fencers in the nation and you get points and advance on the list as you attend tournaments. That's how they seed their tournament I believe.

    so it looks like even a rating system like UK's would have the problem of people being underrated. and even a highly rated fencer can choke one day or the other, I meant no rating system is perfect, why cant we just stick with current not so perfect one we have and focus more on fencing rather than the mere letter and two digits of numbers?
     
  12. oiuyt

    oiuyt Podium

    Joined:
    Apr 26, 2000
    Messages:
    10,240
    Likes Received:
    974
  13. oiuyt

    oiuyt Podium

    Joined:
    Apr 26, 2000
    Messages:
    10,240
    Likes Received:
    974
    The BFA system suffers from under-ranked fencers if there's significant time off (injury, pregnancy, other real-world considerations). There's a limit of 6 competitions (over a 12-month span), so merely hitting up every competition doesn't massively skew things.

    Non-statistical numeric systems are also a part of the current conversation about what the USFA might want to move towards in the future as well as the various statistical systems.

    -B
     
  14. qatet

    qatet DE Bracket

    Joined:
    Jan 27, 2003
    Messages:
    722
    Likes Received:
    100
    Obviously, one of the biggest problems with things like Jeff Snider's fantastic mock-ups (based on George Masin's proposal) is that they are applying a different measure to points than fencers currently use in their own competitions. The system encourages a fencer to win by as large a margin as possible - in ALL bouts. Currently, in DEs one might intentionally give up touches, perhaps gaining information or warming up an action for a later bout. This system punishes fencer for that approach (as refs might like to!).

    A problem in and of itself? No. Not really. But it encourages an approach to DEs that is very different from what one will encounter elsewhere - particularly at the international level, which is a set of results that we do and should care about.

    Now, of course, there are plenty of other things that competitors must get used to when they reach the international level, but it seems like it would make sense to keep something so basic the same from one level to another. Kinda like how the ref cadre is constantly told that the cards that our international fencers get for things like covering with their masks or hitting their weapon on the strip are our responsibility.

    It also means that the pressure on refs and fencers will increase a huge amount. There were bouts in Jeff's equations where a fencer won 5-1, and still had their rating go down as a result. (A nice combination with the current foil debounce problems!) Right now, people are willing to put up with developmental refs at the local level because they know that it really doesn't matter. When suddenly every touch impacts your national standing, imagine how much more tense the ref/fencer relationship will be. Currently, one hauls out the solid refs for things like qualifiers or SYCs - things that impact national events - but there's just not the requisite body of refs to have nationally-rated refs for every U event that every little club holds.

    Perhaps part of the solution is to hold regional events - like the SYCs - and base ratings off of those, rather than making every event everywhere a Serious Important Tournament. But that really seems kinda like NACs. Maybe every division has to have one each year. But that benefits folks in areas with geographically small divisions and sucks for people in the middle of the country. Maybe there's a limit on how many results you can report. But it still turns into a matter of good standings for those who can afford to hunt them down.

    People are so quick to jump on the bandwagon for the numerical rankings - I was, too, when I first heard the idea - but I think that this system raises a huge number of other problems.
     
    CvilleFencer likes this.
  15. peet

    peet Podium

    Joined:
    Jan 8, 2003
    Messages:
    2,253
    Likes Received:
    241
    Well, there's nothing to publish really; it's just a big fat database. Like I say, it was really mostly a proof of concept on the method of simulation. I'm not so sure the table tennis system is really right for fencing anyway.

    When there's something to publish that will be interesting/decipherable to people, I'd be happy to.

    -p
     
  16. downunder

    downunder Podium

    Joined:
    Jul 17, 2003
    Messages:
    3,843
    Likes Received:
    460
    You guys do tend to get very hung up about this rating thing.

    It is a basic method to seed a competition. That is it. I think it does a good job.

    In the real world, you get good poules and bad poules. This helps to even it out in a country with so many fencers (in so many areas) that a 12 month rolling ranking for all fencers like the UK has would be impossible.
     
  17. mrbiggs

    mrbiggs Podium

    Joined:
    Mar 25, 2004
    Messages:
    7,837
    Likes Received:
    477
    It does a good job for now, but in a year or two it won't, for men's epee at least.
     
  18. downunder

    downunder Podium

    Joined:
    Jul 17, 2003
    Messages:
    3,843
    Likes Received:
    460
    You'd be right if you weren't so wrong. It only counts 6 competitions, and only 3 of those can be internationals etc etc.

    The major opens (with the most points) are well attended by most of the serious fencers in the country, and once you get below that level the ranking is just for seeding which will have little impact except at the smallest of competitions (which carry little points so have minimal effect).
     
  19. Sciurus-Rex

    Sciurus-Rex Rookie

    Joined:
    Jun 15, 2005
    Messages:
    1,230
    Likes Received:
    234
    Rumor has it that an alternative ratings shift is being considered: The system will remain the same as it is now, except that the letter range will begin with P and run through T. ... A fencers will be identified as Ps instead, Bs will be Qs, Cs will be Rs, Ds will be Ss, Es will be Ts, and Unrated fencers will still be Us.

    I guess we'll see how invested people are in the ratings themselves versus the effectiveness of a seeding method. There's something vaguely less sexy about claiming you're a Q-fencer working on your P, rather than a B pursuing your A.
     
  20. qatet

    qatet DE Bracket

    Joined:
    Jan 27, 2003
    Messages:
    722
    Likes Received:
    100
    I may be missing something - how exactly is the British system different from our NACs?
     

Share This Page