Page 2 of 2 FirstFirst 1 2
Results 16 to 30 of 30

Thread: Fan Proposal: The New 10.0 System

  1. #16
    Custom Title
    Join Date
    Jul 2003
    Posts
    3,933
    Quote Originally Posted by draqq View Post
    The protocols revealed at the end of each competition will essentially remain the same. It will still have the list of elements, their levels, their base values, and the individual scores for each component. Some of the 10.0-scaled marks I propose can be added for additional comprehensiveness, but the protocols remains the same. As you, I want to see the full breakdown.
    Good.

    This 10.0 system I'm proposing is simply for what's revealed on the TV screens and for the audiences in the ice rink. It gives a general idea of each judge's mark without having to fill the screen with too many numbers, but hardcore fans will know to go to the protocols online for a more complete picture of how every mark is determined.
    My question is whether the scaled-to-10 marks will actually be used to calculate the results, or if they're just fiction to give the fans a benchmark for how the open-ended base mark + GOE scores compare to previous top scores.

    What if, in your Suzuki example, the announcement in the K&C says "Technical base value 28.6" and then lists the technical scores from each judge (31.9 from judge 1, etc., across the board) and the average (7.8) or factored total (e.g., 30.22) component score from each judge. That way you get to see two scores from each judge like we did under the old system, and they would actually add up to the real score that skater achieved. But they would be open-ended scores, not scaled to a maximum that means something different for each discipline and that can theoretically be exceeded by multiple skaters.

    I understand how the 10.0 reference point is meaningful for giving fans an idea of how this performance stacks up against the best technical content ever as of last year, but I don't understand how or whether it's actually used to calculate the results. How much information that is meaningful in distinguishing skaters would be lost in the conversion -- more or less than is lost by averaging and rounding in the current calculations?

    Would anomalies in how the numbers get crunched this extra time ever lead to different results if they were crunched differently?

    Certainly we've seen differences in 6.0 events that were accidentally scored with the wrong tiebreaker programmed into a computer or with majority calculations used for an event that should have used OBO. And we've seen rounding errors affect the results of close contests under IJS. So whatever system is used can be subject to paradoxical results. I just wonder what the effect of the extra manipulation to the IJS numbers you want to add would introduce. Would it be possible for the results on the detailed protocols to be different from the results on the scaled-to-10 matrix? If so, which would count as the official results?

    A part of my reasoning behind this is because of how the 6.0 system worked in the past. A 6.0 in 1970 was much different than a 6.0 in 2000. What the perfect 6.0 means is to say that this particular program is perfect at this place and time. And I believe that the judges would reward that same program with perfect marks at a later competition if it was performed in the same manner. However, the following year, achieving that 6.0 would more difficult as a new standard would be set in place.
    The difference, though, is that in ordinal judging the judges also explicitly had to leave room for subsequent skaters, so the scores didn't always mean the same thing from one event to the next in the same season. In IJS, they should, although we know that different panels might be more or less generous.

    The other thing I'd ask about your plan is how it would apply to competitions where pretty much everyone watching is directly involved in the sport or else a diehard fan who is more likely to be interested in the detailed protocols than an intermediate level of summary scores -- let's say a JGP event or domestic sectional qualifier, etc. -- and where there is no expectation of television coverage and almost no expectation of ticket sales to casual fans. Would there be any value of running the numbers through this process in those events?

    Is it a change to the actual scoring system that should be used at all competitions to determine results, or just a way of presenting fake scores to fans who aren't interested in the details of the real scores but are interested in more detail than one open-ended number for TES and PCS from the whole panel?

  2. #17
    Custom Title
    Join Date
    May 2010
    Posts
    407
    Quote Originally Posted by gkelly View Post
    Good.



    My question is whether the scaled-to-10 marks will actually be used to calculate the results, or if they're just fiction to give the fans a benchmark for how the open-ended base mark + GOE scores compare to previous top scores.

    What if, in your Suzuki example, the announcement in the K&C says "Technical base value 28.6" and then lists the technical scores from each judge (31.9 from judge 1, etc., across the board) and the average (7.8) or factored total (e.g., 30.22) component score from each judge. That way you get to see two scores from each judge like we did under the old system, and they would actually add up to the real score that skater achieved. But they would be open-ended scores, not scaled to a maximum that means something different for each discipline and that can theoretically be exceeded by multiple skaters.

    I understand how the 10.0 reference point is meaningful for giving fans an idea of how this performance stacks up against the best technical content ever as of last year, but I don't understand how or whether it's actually used to calculate the results. How much information that is meaningful in distinguishing skaters would be lost in the conversion -- more or less than is lost by averaging and rounding in the current calculations?

    Would anomalies in how the numbers get crunched this extra time ever lead to different results if they were crunched differently?

    Certainly we've seen differences in 6.0 events that were accidentally scored with the wrong tiebreaker programmed into a computer or with majority calculations used for an event that should have used OBO. And we've seen rounding errors affect the results of close contests under IJS. So whatever system is used can be subject to paradoxical results. I just wonder what the effect of the extra manipulation to the IJS numbers you want to add would introduce. Would it be possible for the results on the detailed protocols to be different from the results on the scaled-to-10 matrix? If so, which would count as the official results?



    The difference, though, is that in ordinal judging the judges also explicitly had to leave room for subsequent skaters, so the scores didn't always mean the same thing from one event to the next in the same season. In IJS, they should, although we know that different panels might be more or less generous.

    The other thing I'd ask about your plan is how it would apply to competitions where pretty much everyone watching is directly involved in the sport or else a diehard fan who is more likely to be interested in the detailed protocols than an intermediate level of summary scores -- let's say a JGP event or domestic sectional qualifier, etc. -- and where there is no expectation of television coverage and almost no expectation of ticket sales to casual fans. Would there be any value of running the numbers through this process in those events?

    Is it a change to the actual scoring system that should be used at all competitions to determine results, or just a way of presenting fake scores to fans who aren't interested in the details of the real scores but are interested in more detail than one open-ended number for TES and PCS from the whole panel?
    This 10.0 system works like a user-friendly cover for the current system. Everything is calculated using the same points system as the current IJS. It's just that scores are scaled to the 10.0 when they are announced and shown. I think that this is valuable to both fans who aren't interested in the protocols because it's viewer-friendly and fans (+skaters) who will see the protocols anyway since they give more information than just the bare TES and PCS scores.

    Since all of the numbers are reduced by the same factor, the amount of information lost is quite minimal. We're talking about tenths of a point per judge. That said, the resulting Total Score isn't affected by the 10.0 scaling so people still know the precise final score.

    I suppose the system could just use the actual individual scores for TES and PCS without scaling. But then in the SP, you would have technical marks ranging from 20-60 but still have PCS marked from 0-10. A part of this 10.0 system is meant to transform the TES to a scale that the PCS already uses.

    Using the 10.0 reference point is meant to be viewer-friendly. I don't think a range of scores going from 25.8 to 30.8 is that exciting or even that insightful, despite being precise. Of course, I can experiment with different reporting methods. There's always a sacrifice somewhere.

  3. #18
    Custom Title
    Join Date
    May 2010
    Posts
    407
    Quote Originally Posted by yunasashafan View Post
    Can it, though? Would it be possible to compute a maximum possible score for the SP/LP for each discipline, assuming the hardest jumps and highest spin and step levels with +3 GOE? For example, for ladies SP, that would include a 3A, 3LZ-3LO and 3F, all executed in the 2nd half with level 4 steps and spins, and +3 GOEs everywhere. This way, it would be very unlikely to get a relative score higher than 10. For the men, and for the LP it gets a bit more complicated but we can always make assumptions based on the current level of skating. For example a 4LZ-3T and 4F with 3A in he men SP.

    This maximum would be adjusted every year to reflect the level of skating and the changing requirements/point system.

    I love this idea, by the way.
    You're right that potentially, there is an upper maximum to what can be achieved under the current rules. It would have to consider ridiculous combinations like a 4A+4T+4T as a combo. (Or even a 5T?)

    But the main point for setting the world record as the standard is that it's obtainable. Anything ridiculously higher has very little context, and every skater would receive consistently low marks for TES. I mean, the maximum for a ladies TES would be well over 60+, but even at 60, Yuna Kim's SP would just get a 7.4 on the 10.0 scale. It just makes all the marks feel deflated.

  4. #19
    Custom Title
    Join Date
    May 2010
    Posts
    407
    Quote Originally Posted by FSGMT View Post
    Wonderful idea! I'll be waiting for this
    I will certainly try. The problem of course is that this system is meant to be reported on the fly as the skater's marks go up. Unless I'm mistaken, we don't have access to the protocols until the competition is over... so by then, why not just look at the protocols?

    Luckily, this 10.0 scaling system still does reveal bias and adds reference points, so I'll still do it. It should be fun experiment.

  5. #20
    Custom Title
    Join Date
    Jul 2003
    Posts
    3,933
    Quote Originally Posted by draqq View Post
    This 10.0 system works like a user-friendly cover for the current system.
    Ah, OK. As long as it's clear that's all it is. It will take a little longer to read scores in the K&C and give some extra information about which judges gave higher aggregate scores to each skater and how the score compares to a benchmark score.
    To keep up the suspense, maybe it would need to be introduced something like "The 10-scaled technical scores for this program by judge were [list of numbers] and the averaged component scores by judge were [list of numbers]. With a total Technical Elements Score of X and a Program Components Score of Y, that puts Skater A in first place."

    I still think it's important in long programs to announce the standing for the long program alone as well as the overall placement. That's an area where the reporting has been incomplete and confusing under both systems.

    Since all of the numbers are reduced by the same factor, the amount of information lost is quite minimal. We're talking about tenths of a point per judge. That said, the resulting Total Score isn't affected by the 10.0 scaling so people still know the precise final score.
    You mean the scores could be off by as much as 0.1 per judge? That's almost a full point with 9judges, and some contests have been known to be decided by closer margins than that.

    What if two skaters are so close that a tenth of a point or so in margin of error shows Skater A with a slightly higher 10-scaled score but Skater B slightly higher in the real scores that are used to calculate the real results?

    If it's that close, they're probably both equally deserving. But the system has to draw the line somewhere, and even if we decided that hundredths of points meaningless nitpicking and final scores should be rounded to the nearest tenth, accepting a somewhat greater frequency of ties, we probably do still want to maintain tenth-of-point precision. But the scaling is only accurate to a full point?

    The announcers would need to be are very clear that the 10-scaled scores are just derivations of the real scores that are reported for entertainment value and are not the actual numbers used to determine the results. Otherwise, if the 10-scaled scores are announced, avid fans write them down and add them up believing that they are following along with the results, and then the real scores show that someone else won, they will believe that someone pulled a fast one.

    No more so than if the total scores were reported without decimal places or without mentioning deductions for falls, etc., or if the PCS factoring for disciplines other than the men aren't explained.

    Arena announcers need an efficient (preferably) and consistent format for announcing scores and placements. The more they streamline it to save time, the more possibility for confusion with any system. So TV announcers should be prepared to anticipate and explain results that might seem incorrect or paradoxical at first glance. This is already an issue without introducing another layer of numbers between the real calculations and the public.

  6. #21
    Custom Title
    Join Date
    May 2010
    Posts
    407
    So as a test run, I'm going to choose the ladies' short program and long program at TEB. It actually takes 2-3 hours of manual tabulation (which a computer can do in a half-second) to get each judge's scores in this proposed 10.0 system, so I'm only limiting it to the ladies event.

  7. #22
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    28,663
    I would still vote to scale the scores to a six point scalle rather than ten. That way, figure skating could keep its "brand." Everybody uses "rate these from one to ten" -- gymnastics, diving, how much pain you are in when you go to the doctor, how pretty is Bo Derek.

    Six-point-oh means figure skating. We shoiuld try to keep it, if only as a marketing gimmick.

    Anyway, one interesting thing that might happen is that the majority of judges could favor one skater but the point totals give the prize to another.

  8. #23
    Custom Title
    Join Date
    May 2010
    Posts
    407
    After much manual tabulation, which would take a computer about a half-second, I was able to convert the ladies SP into the proposed marks for this 10.0 system. It has been posted here:

    http://www.goldenskate.com/forum/sho...10-0-System%29

  9. #24
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    28,663
    ^ Very cool.

  10. #25
    Custom Title Mathman's Avatar
    Join Date
    Jun 2003
    Location
    Detroit, Michigan
    Posts
    28,663
    It would be even more cool if the ISU didn't stymie the fans by mixing up the judges' marks. When one judge is way out of line, we don't know whether that judge is scoring extra tough/extra leniently across the board or whether that judge is playing favorites.

  11. #26
    Simply the best. l'etoile's Avatar
    Join Date
    Mar 2011
    Posts
    1,056
    Wow, that's really cool, draqq!!!

  12. #27
    Down With It
    Join Date
    Jul 2003
    Location
    USA
    Posts
    13,590
    First off, this is a really interesting concept. Thanks for sharing with us.

    However, if I may nitpick: in the way the scores are presented, I see pros and cons to both the way it's currently done and the way you are proposing it be done. The way it is now, I can clearly see that the TES and PCS add up to the total score. However, under your system, at first glance- although it is far easier to see any scoring disparities between judges, the numbers don't appear to add up to the score presented below and that's a bit of a Sounds like the individual marks are being scaled to 10.0 - why not scale the final score as well?

    Also- even under your system we STILL don't know which judge gave which score...so although transparency is improved a little, it is still not 6.0 where your mark was associated with your country.

    Not matter what judging system is used, the REAL problem is not being solved- and that is, inconsistency in judging among events and within an event. The order you skate in should not matter (it does). The PCS you receive (and I'm talking the skating skills scores, etc.) should not vary from SP to FS for some skaters (it does). Changing the system once again isn't going to fix that...

    A proposal to help with skating order judging problem: Mix up the entire field among groups like USFS does for Nationals in the SP. For the FS - you can mix up the field like this: for GPs, mix up the entire field. For Championships, have the top 12 mixed up among the final 2 groups and the bottom 12 mixed up among the first 2 groups.
    Last edited by R.D.; 11-18-2012 at 02:54 PM.

  13. #28
    Custom Title
    Join Date
    May 2010
    Posts
    407
    Quote Originally Posted by R.D. View Post
    First off, this is a really interesting concept. Thanks for sharing with us.

    However, if I may nitpick: in the way the scores are presented, I see pros and cons to both the way it's currently done and the way you are proposing it be done. The way it is now, I can clearly see that the TES and PCS add up to the total score. However, under your system, at first glance- although it is far easier to see any scoring disparities between judges, the numbers don't appear to add up to the score presented below and that's a bit of a Sounds like the individual marks are being scaled to 10.0 - why not scale the final score as well?

    Also- even under your system we STILL don't know which judge gave which score...so although transparency is improved a little, it is still not 6.0 where your mark was associated with your country.

    Not matter what judging system is used, the REAL problem is not being solved- and that is, inconsistency in judging among events and within an event. The order you skate in should not matter (it does). The PCS you receive (and I'm talking the skating skills scores, etc.) should not vary from SP to FS for some skaters (it does). Changing the system once again isn't going to fix that...

    A proposal to help with skating order judging problem: Mix up the entire field among groups like USFS does for Nationals in the SP. For the FS - you can mix up the field like this: for GPs, mix up the entire field. For Championships, have the top 12 mixed up among the final 2 groups and the bottom 12 mixed up among the first 2 groups.
    The main reason why the final score isn't scaled is that the total score is still want counts at the end, and some competitions can come down to a tenth or even a hundredth of a mark.

    However, I can see how the total score could be announced as both a scaled score and a points score. Something like "TSS 8.0/64.00". Of course, I don't want to have too many numbers are starts to look like the protocol itself.

    And you're right, this 10.0 system doesn't fix the fact that the judges are completely anonymous. That is what leads to bias, and as you can see, some judges give crazily high or low marks in technical or artistry. It's just running wild now. On the good side, if they actually revealed the judges' country, this 10.0 system wouldn't need to be changed at all. Any changes they make to the current system automatically adapts to this 10.0 system.

  14. #29
    Custom Title
    Join Date
    May 2010
    Posts
    407
    Quote Originally Posted by Mathman View Post
    It would be even more cool if the ISU didn't stymie the fans by mixing up the judges' marks. When one judge is way out of line, we don't know whether that judge is scoring extra tough/extra leniently across the board or whether that judge is playing favorites.
    In the protocols? do you know if they mix up the judges marks for each skater or for the whole sheet? In other words, is Judge 1 for the first skater the same Judge 1 for all the skaters?

    I actually think that might be true. Not certain.

  15. #30
    Custom Title
    Join Date
    Jul 2003
    Posts
    3,933
    My understanding is that for international senior competitions, for each skater each column of numbers is from one judge, but the order in which the judges' columns are listed changes with every skater. So the numbers in column 1 for the first skater will be from a different judge than those in column 1 for the second skater, etc.

    This is the anonymizing process.

    Last I heard, they don't do this for junior scores internationally.

    I know for a fact they don't do it for domestic scores within the US, including at Nationals. All the judges' columns remain in the same order throughout the protocol, and in the same order (judge 1, judge 2, etc.) as listed on the cover sheet of the results.

    I have no idea about domestic competitions in other countries.



    My understanding is that the point of mixing up the scores for senior internationals (where the stakes are highest and outside influence, e.g., from federation officials, most likely) is to make it impossible -- or at least much more difficult -- for anyone who wants to influence a judge to score certain skaters in certain ways to know whether the judge actually scored as the outside influencer wanted, giving judges the freedom to judge according to their own conscience rather than to how anyone else pressures them. How well it actually serves that purpose, I have no idea.

    The judges should be able to recognize their own marks after the fact, at least for the skaters where they gave at least one memorable score that no one else gave. They wouldn't know which other judge gave which marks unless the other judge told them.

    And meanwhile, the skaters and the public are also unable to know which judge gave which scores, whether they want to ask for explanations from specific judges or look for patterns or just satisfy their curiosity.

Page 2 of 2 FirstFirst 1 2

Posting Permissions

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