Previous Rating System

Discussion in 'Fencing Discussion' started by leftweet, Oct 27, 2010.

  1. leftweet

    leftweet Rookie

    Joined:
    Dec 17, 2009
    Messages:
    56
    Likes Received:
    1
    Does anyone know what type of rating system (if any) was used before the alphabetical system (A-B-C-D-E-U)? Or when the current system took effect? I've looked through past threads and haven't found anything on the subject.
     
  2. KD5MDK

    KD5MDK Moderator

    Joined:
    Feb 15, 2005
    Messages:
    13,294
    Likes Received:
    636
    It was A,B,C,D before. Then A,B,C. Then it was a terrible complicated mess. George Masin wrote a paper on it.
     
  3. Gav

    Gav Moderator!!

    Joined:
    Aug 19, 2000
    Messages:
    3,398
    Likes Received:
    484
    If he could write a paper on it the mind boggles!
     
  4. erik_blank

    erik_blank Podium

    Joined:
    Jul 28, 2005
    Messages:
    2,001
    Likes Received:
    61
    Assuming that this isn't just smoke, how could a person read this alledged paper? :)
     
  5. DHCJr

    DHCJr Armorer

    Joined:
    Jul 29, 2002
    Messages:
    5,376
    Likes Received:
    200
    There was also a two-tiered approach when fencing mixed. Some A's were considered C's. For example if Zuganis had been fencing then, she would have been considered a C.

    Eric here it is.
     

    Attached Files:

    Last edited: Oct 27, 2010
  6. tchwojko

    tchwojko Rookie

    Joined:
    Jul 9, 2003
    Messages:
    1,820
    Likes Received:
    191
    Wasn't there a period when an A in one weapon was automatically a C in the other two weapons?
     
  7. SJCFU#2

    SJCFU#2 Podium

    Joined:
    Nov 5, 2003
    Messages:
    3,478
    Likes Received:
    126
    Actually there was never a time when it was A, B, C, D - it jumped straight from A, B and C to A, B, C, D, E sometime around 1983.

    However IIRC that only applied to women when fencing in mixed events (what a pain that was!).

    Not just A's - anyone who was rated C or higher in one weapon was automatically considered to be rated two steps below that rating in the other weapons. IIRC that practice only ended some time within the last decade.
     
  8. oiuyt

    oiuyt Podium

    Joined:
    Apr 26, 2000
    Messages:
    10,240
    Likes Received:
    974
    It ended December 31, 1999 - January 1, 2000.

    That season was when we moved from Open NACs to Division I NACs (D and above at the time), and it was argued that the courtesy classifications in potentially non-fenced weapons were not appropriate when we were restricting entry to the top-level NACs.

    -B
     
  9. KD5MDK

    KD5MDK Moderator

    Joined:
    Feb 15, 2005
    Messages:
    13,294
    Likes Received:
    636
    My bad. In any event, read the paper.
     
  10. K O'N

    K O'N Podium

    Joined:
    Aug 14, 2006
    Messages:
    3,593
    Likes Received:
    458
    Mr Masin came up with the idea of Div II and III national events, so I'm permanently grateful to him, but ELO type numerical ratings for fencing are a bad idea. If you can lose points by fencing there will often be a reason to stay home. Bad effect, right off the bat.

    I also think the case that the current system is so totally broken is one that I haven't seen made very clearly yet. Often when there's a cascade of complaints, they go something like:

    There are too many As in my area!

    Yeah, and not only that, it's too hard to get an A here!

    Ok, guys, those two complaints are in the opposite direction. The only real case to be made for a seeding system being broken is that it's doing a bad job in seeding. Is the E-D-C-B-A-points system doing such a bad job of seeding? How is that being decided?

    After an ELO system is implimented we'll hear that it's too hard to get above 1200 in New York, or that there are too many people over 1350 there, or something, the seeding won't be measurably better than it is now, and meanwhile the kid who lucked into a 1250 at a weak A2 or whatever it's called in the brave new ELO world is sitting home nursing his rating, he wouldn't go fence before JOs or SNs for money. Oy. Terrible idea.

    K O'N
     
    Inquartata and Jason like this.
  11. Inquartata

    Inquartata Podium

    Joined:
    Jul 12, 2001
    Messages:
    36,189
    Likes Received:
    1,183
    Preach on, Brother. Preach on...though it fall on deaf ears. :)

    Oh, the sad, sad day when I must rep an epeeist...sigh. :(
     
  12. fdad

    fdad Podium

    Joined:
    Jul 9, 2008
    Messages:
    4,129
    Likes Received:
    214
    I like the idea of having a finer grained rating system than what we have now. But I agree that any system that rewards a fencer for not fencing instead of fencing is a bad one.
     
  13. oiuyt

    oiuyt Podium

    Joined:
    Apr 26, 2000
    Messages:
    10,240
    Likes Received:
    974
    On a related note, Evan Ranes is the chair of a task force exploring options for a new or modified rating/classification system. Several options are on the review list, including ELO-based, non-ELO-based statistical, and non-statistical numeric systems. People that are interested in the project or who simply have feedback with regards to concerns, potential constraints, or desiderata should probably talk to Evan about the process and how best to provide input.

    -B
     
  14. Inquartata

    Inquartata Podium

    Joined:
    Jul 12, 2001
    Messages:
    36,189
    Likes Received:
    1,183
    The problem with such task forces is that "no change" is seldom one of the options they explore. :mad:

    This makes some feedback worth less than the effort it takes to transmit it...
     
  15. IanSerotkin

    IanSerotkin DE Bracket

    Joined:
    Jul 22, 2004
    Messages:
    647
    Likes Received:
    189
    I noticed that (according to the BOD minutes) Evan is not only the chair of the task force, but also the only member. Is that accurate?
     
  16. Jason

    Jason Podium

    Joined:
    Aug 25, 2002
    Messages:
    2,498
    Likes Received:
    321
    Well, "one guy" is an historically accepted USFA definition of "task force", after all.
     
  17. oiuyt

    oiuyt Podium

    Joined:
    Apr 26, 2000
    Messages:
    10,240
    Likes Received:
    974
    As an independent group (as opposed to a working group or project of the Tournament Committee) its existance dates to the September Board meeting. No members other than Evan were named at that time (or subsequently).

    -B
     
  18. Sean Butler

    Sean Butler Rookie

    Joined:
    Jul 10, 2008
    Messages:
    535
    Likes Received:
    15
    Let me start by saying that I am sorry for this "wall of text" but I feel strongly about this subject and I think that Masin is right.
    ---

    Personally, having competed for many years USCF chess competitions and having played the rating game there, I do like the Elo system (it is Elo for Arpad Elo BTW, it is not an acronym). I agree that the fear of losing rating points might be detrimental in a few cases but overall my experience with the system was entirely the opposite. Mostly people didn't compete out of fear of gaining points. The sections of competition in a USCF event were often the following:

    Under 1200
    Under 1400
    Under 1600
    Under 1800
    Under 2000
    Under 2200
    Open

    Since US Chess competitions nearly always awarded cash prizes, something I wish USFA competitions did, and since it did this on a section by section basis, there was a strong incentive for folks to compete. Once a person reached the cusp of entering a new rating bracket however, that incentive was reduced since often such people would wait until a really large tournament was held with larger cash prizes.

    This later scenario was somewhat rare since most people liked the prestige of earning a higher rating enough that they would compete even when it wasn't in their best material interest.

    In the paper by Masin he was not suggesting that the rating system ever be used for team selection, this is especially important. Since the main criteria for team selection should always be international results and since these events are going to be with opponents who are outside the USFA rating system, one should never use rating for selection. Instead the current point system, as Masin indicated, should be used for this. The Elo rating would be used for seeding in the event that a competitor had no points.

    While seeding is important and a person would want to optimize their initial seed in large tournament, it is unlikely that a person with a very high rating would move much, even with a poor result. This at least was true in chess. The reason for this is that as you reached the 2100+ rating level your K-factor became 24, and by 2400+ rating level your K-factor became 16. Effectively this meant that the movement of your rating for both wins and losses was cut in half for high rated players. If we followed a similar policy in US Fencing, one could reduce the K-Value to anything reasonable at whichever level made the most sense to encourage fencing.

    Since quite often fencers at DIV1 level will be fencing others of the same rating, their chance of gaining or losing a large number of points is quite reduced. With a K-Factor of 16, two equally rated fencers would stand to gain or lose 8 points on any given bout. With a very bad showing a person might lose 10-12 bouts in the seeding pools. This would be an average rating loss of about 80-96 points if they lost every single bout, and this assumes too that we don't create an Elo system that accounts for points scored in the bouts, just the victory/defeat.

    To put it simply, a person who is afraid that they will lose 80 points by losing every bout in the seeding rounds, is probably not a big concern for most fencers. If that guy decides to skip the tournament to hold onto his clearly inflated rating, it's not going to disappoint folks too much. The kinds of people we would concern ourselves with are ones who are actually good fencers who think they need to protect their rating. Those fencers would rarely if ever lose more than a few measily points in a tournament because of their actual skill.

    I think that the fencer who skips a tournament to preserve his "high" rating is a myth. Further I think that without cash prizes you will not see fencers skipping tournments to preserve low ratings either.
     
    IanSerotkin likes this.
  19. Jason

    Jason Podium

    Joined:
    Aug 25, 2002
    Messages:
    2,498
    Likes Received:
    321
    I find the concept that we should want to add more extrinsic motivators to competition equally hysterical and horrifying.
     
  20. Philistine

    Philistine Rookie

    Joined:
    Oct 14, 2001
    Messages:
    1,805
    Likes Received:
    111
    While the current system doesn't have an incentive to not fence due to fears of reducing a classification, it does have the flipside problem that there is often no particular incentive to fence because an increase in classification is effectively impossible.

    This can be a problem at the high-end (an A10 has no ratings incentive to enter any local tournament) and on the medium end in certain circumstances--because of either numbers or limitations on classified fencers, you won't get a tournament able to give out a higher classification than you already have.

    On the low end, it's the dual problem of the unlikelihood of getting a classification as well as the other issue mentioned in the paper--no actual outside indicia of improvement causing frustration.

    I'm not sure which effect would result in more people not fencing--but I don't think the current system is obviosuly superior on this issue.

    Out of curiosity, how do other one-on-one sports handle ranking/seeding? Things like tennis, ping-pong, other racquet sports, wrestling, etc.?

    --Philistine
     
    Last edited: Oct 28, 2010

Share This Page