Page 2 of 7 FirstFirst 1 2 3 4 5 6 7 LastLast
Results 16 to 30 of 93

Thread: Olympic judging changes ( 5 judge results)

  1. #16
    Custom Title
    Join Date
    Nov 2006
    Posts
    604
    The cost issue is a red herring. Last season you had 12 judges, three of which were discarded and the remaing nine were used. If they gave up the fiction that the discard judges accomplished anything, they could have used all of nine judges and the costs to the competition would have been reduced by the cost of the three officials left home -- with no decrease in scoring judges.

    Also, at Worlds this year, more money was spent entertaining OC, his guests and the Council than would have been spent on the judges that were cut. Apparently parties are more important than the quality of the scoring.

  2. #17
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    27,929
    I think the time to protest was last October, when the rule was changed for all competitions except the Olympics. In that case, I think the financial angle might have been legitimate.

    Here is the quote from the article:

    Although the issue was not brought up to members at a congress last summer, the ISU council passed a similar rule last October to reduce the size of the judges' panel to nine from 12 for ISU championship events.

    However, the October change didn't include Olympic panels. ISU vice-president David Dore said council made the change to make the rule similar to the world championships.
    Quote Originally Posted by gsrossano View Post
    The cost issue is a red herring. Last season you had 12 judges, three of which were discarded and the remaing nine were used. If they gave up the fiction that the discard judges accomplished anything, they could have used all of nine judges and the costs to the competition would have been reduced by the cost of the three officials left home -- with no decrease in scoring judges.
    For some reason the random draw (and anonymous judging) seems to be more important to the ISU than either saving money or improving the quality of the judging. They could seat twelve judges and use all the marks if they wanted to -- that would be even better.

    In fact, the random draw is worse than just accomplishing nothing. It accomplishes nothing AND creates in the mind of the public the perception that the ISU is pulling a fast one somehow.

  3. #18
    Tripping on the Podium
    Join Date
    Apr 2009
    Posts
    70
    I really can't put my finger into what they're doing. In the back of my mind, I think it'll be easier to manipulate the judging system if there are few judges to control. In my opnion, I wish they would just do their job and stop stirring up all these changes. Poor skaters, they just ahve to skate their best and then wait for a mediocre result / judging of their good/best work. I could feel their frustration.

  4. #19
    Custom Title Joesitz's Avatar
    Join Date
    Jul 2003
    Location
    New York City
    Posts
    20,185
    Quote Originally Posted by gsrossano View Post
    The cost issue is a red herring. Last season you had 12 judges, three of which were discarded and the remaing nine were used. If they gave up the fiction that the discard judges accomplished anything, they could have used all of nine judges and the costs to the competition would have been reduced by the cost of the three officials left home -- with no decrease in scoring judges.

    Also, at Worlds this year, more money was spent entertaining OC, his guests and the Council than would have been spent on the judges that were cut. Apparently parties are more important than the quality of the scoring.
    That's really shocking but there probably are budget for entertaining purpose. He should not have abused that budgetary item. Don't know if the budget is public. If not, some other official must check it.

  5. #20
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    27,929
    This is what I think they should do.

    Seat nine judges, no random draw, then trim the mean by discarding the bottom two and the top two for each line.

    If the goal is to make it harder for people to cheat, that would be a step in the right direction. Let's say there are two judges in collusion. In the random draw, although it is possible that one of the conspirators is thrown out, it is more likely that it will be two of the other judges, thus strengthening the hand of the conspiracy.

    But if instead you kept all nine votes, then it would be very likely that the two conspirators' votes would be the extremes that would be discarded, as the culprits tried to inflate one skater's marks and low-ball another's.

    Also in the case of one biased judge actiing alone, by trimming the mean more severely there is a greater likelihood that this judge's vote would be high or low and thus discarded.

    This method does not have any statistical downside, and it even has an extra bonus (over the method of averaging seven or nine scores.) Dividing by five is easy, there are no rounding errors, and we can do the arithmetic without a calculator.

  6. #21
    Custom Title
    Join Date
    Nov 2008
    Posts
    156
    ISU just wanna have controllable system.
    That's all.
    Need more discussion?
    They want another Sarah 'Cheated' Hughes.
    All the current change is for that.
    ISU has ideal Olympic's Gold members and podium pitures, Maybe...beneficial ones.

  7. #22
    Custom Title
    Join Date
    Oct 2006
    Posts
    1,560
    Quote Originally Posted by Mathman View Post
    ... it would be very likely that the ... conspirators' votes would be the extremes that would be discarded, as the culprits tried to inflate one skater's marks and low-ball another's.
    But the ISU is so out of touch and has so much power. How can you make them listen?
    Last edited by jennylovskt; 04-22-2009 at 11:55 PM.

  8. #23
    Tanguera feraina's Avatar
    Join Date
    Mar 2007
    Location
    San Diego, CA
    Posts
    2,149
    Quote Originally Posted by Mathman View Post
    Seat nine judges, no random draw, then trim the mean by discarding the bottom two and the top two for each line.
    Good idea. Random drawing accomplishes nothing. Trimming highs and lows is a good idea. It's a compromise toward median from mean, and when there are extreme outliers (either from cheating or noise), the median is more stable than the mean.

    But there will be no new discussion of this, nor of the downgrade rules, until after 2010 Olympics, right?

  9. #24
    Custom Title
    Join Date
    Apr 2008
    Posts
    215
    Sorry to beat a dead horse, but this idea that "only five" people decides who wins isn't, I don't think, quite correct. 9 judges form the panel for the short, but of those 9, only 5 are seated for the long. Four "new" judges come in. Each panel has: 1. Two judges' randomly selected out, and 2. the high and low marks thrown out. So while only the decisions of five judges "count" in determining the standings of the short and long, in actuality, some 13 judges figure into the process. That's in addition to the three-person technical panel. So actually 16 people are involved, plus of course referees, data entry operators, etc.

    Five has sort of been a key number for panels for a long time. A "bloc" of five has often decided who wins, and in the days of 6.0, each judge knew full well what holding up, say, 5.6/5.9 meant, and where it would place each skater. Nowadays, it's far harder for judges to sit there and calculate precisely where the GOEs and PCS scores they enter will put skaters in the order. Next season they won't even know if the jumps they mark have already been downgraded. Of course it's still possible there will be judges scoring PCS marks too high, for biased reasons, but now they won't even know if their marks count and, if their marks are the high or the low, they are thrown out. And anyone trying to "bribe" judges has no way of knowing if the judge they "pay off" will be seated for both panels, or if his or her marks will actually count.

    Perhaps I am naive, but I really think this system, although certainly imperfect, is harder to manipulate than the old 6.0. As always with IJS, though, the trouble is it is rather complicated and most folks don't want to sit and hear (or read!) a half-hour explanation.

  10. #25
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    27,929
    The random draw and anonymous judging did not come in with the CoP. It was the essential feature of the "interim system" that was rushed into place after the Salt lake City debacle and used for one season until the ISU could get the CoP in place.

    It seems clear to me that the purpose of these provisions was to make it harder for the public to detect and to priove cheating on the part of the judging panels.

    Quote Originally Posted by nylynnr
    As always with IJS, though, the trouble is it is rather complicated and most folks don't want to sit and hear (or read!) a half-hour explanation.
    Not only don't most folks want to hear a complicated explanation, but worse, when you try to explain it, after a while folks just conclude you're a con man and a crook.

  11. #26
    Custom Title
    Join Date
    Nov 2006
    Posts
    604
    The bottom line for determining the precision of the results is how many sets of marks are in the calculation -- and under the new panel size there are five sets of marks in the calculation. It doesn't matter how you get down to the five sets of marks. What drives the mathematical "quality" of the scores is the number 5.

    Quote Originally Posted by nylynnr View Post
    Nowadays, it's far harder for judges to sit there and calculate precisely where the GOEs and PCS scores they enter will put skaters in the order.
    Sorry, you don't give the judges enough credit. Yes, in a large group they may not know exactly where they have put the skater in 17th place; but then who cares about 17th place. For the top skaters, and the handing out of medals they know exactly what marks they need to give, and what range of marks will not look suspicious. When I leave a panel I know exactly who I gave my top marks to and who I gave my lowest marks to. (The middle of course is all a blur under IJS.)

    When marks are posted on the arena scoreboard, even if it is only TES and PCS, the savvy judges know how their marks compare to the rest of the panel, and what marks they need to give to help or hinder subsequent skaters. For example, if Brian Joubert skates first and gets a PCS of 75, then his average PC score was 7.5. I know from that if I marked high or low. If Evan Lysacek skates next, I know if I want to help him beat Joubert I have to go above 7.5.

    Even with exactly placing the skaters under 6.0, a judge could only give a skater a nudge in a given direction, and could not guarantee a specific outcome. Smart judges know how to give the skaters a nudge under IJS also.
    Last edited by gsrossano; 04-25-2009 at 12:36 AM.

  12. #27
    Custom Title
    Join Date
    Apr 2008
    Posts
    215
    [Even with exactly placing the skaters under 6.0, a judge could only give a skater a nudge in a given direction, and could not guarantee a specific outcome. Smart judges know how to give the skaters a nudge under IJS also.[/QUOTE]

    True, I just think it is a bit tougher. In the old 6.0, you knew absolutely who you put 1 or 2 or 3, and there was no chance your vote was going to be thrown out. Plus everyone knew your vote absolutely counted, which, it could be argued, opened up a greater possibility of outright bribery. And of course, no matter how high the GOE's are in IJS, to get the highest scores a skater has to have the ratified jumps and Level 3 or 4 elements, and those come from a different quarter than the judging panel. So while both systems are imperfect, one could argue IJS is more difficult (though not impossible) to manipulate. Better? That's another topic.

  13. #28
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    27,929
    Quote Originally Posted by gsrossano View Post
    The bottom line for determining the precision of the results is how many sets of marks are in the calculation -- and under the new panel size there are five sets of marks in the calculation. It doesn't matter how you get down to the five sets of marks. What drives the mathematical "quality" of the scores is the number 5.
    I think throwing out the high and low scores muddles that a little.

    I think we would expect that the standard error estimated from a random sample of size five would be greater than if we took a sample of size seven and trimmed before averaging.

  14. #29
    Custom Title
    Join Date
    Nov 2006
    Posts
    604
    Quote Originally Posted by Mathman View Post
    I think throwing out the high and low scores muddles that a little.
    Only a little.

    (And I never said the standard deviation wouldn't change -- only that the precision was determined by the number of marks in the calculation. For any calculation method the precision for that calculation method will improve with an increase in the number of marks, for nearly all noise sources. There are exceptions to this rule, but they are few and do not apply to marks in skating as far as I can tell.)
    Last edited by gsrossano; 04-25-2009 at 12:42 AM.

  15. #30
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    27,929
    Actually, I was thinking of the extreme case. Suppose you sat 9 judges, "counted" all nine, then took the median. This reduces the magic number from 5 all the way down to 1. But this would be an OK way to go, as far as I can tell.

    And it would be really hard for a single judge, or a small minority of the panel, to boost their favorite.

Page 2 of 7 FirstFirst 1 2 3 4 5 6 7 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •