GOE scoring

[irenar5]

Does a fall from a jump automatically garner a -3 GOE?    Sometimes I see a random -2 among the -3s  for a fall.

How does a GOE get determined?  Is there an official rule, where an element performed in a certain way gets a mandatory negative GOE? 

I am very frustrated seeing GOE scores ranging from -2 to +2 for the same element.  I understand a 0 and -1 spread or -2 and -1.  But  to call the same element "almost perfect"  and "almost failed" for the same execution is ridiculous!

Any thoughts?

[Doubletoe]

It is supposed to be an automatic -3 GOE for a fall.  See page 13 of this document:  http://www.usfsa.org/content/ISU%20Communication%201861%20Singles%20Pairs%20Scale%20of%20Values.pdf

[irenar5]

The judges really need to be administered a test prior to scoring competitions!  Some give -2 not -3 for a fall! 

[nicklaszlo]

Judges do have to take tests.  If the judge decided it was a "touch down" instead of a fall they may award -2.  In addition, a jump with a fall may earn GOE from the list of positive GOE criteria.  The highest GOE possible for a jump with a fall is -1.

[aussieskater]

It is supposed to be an automatic -3 GOE for a fall.  See page 13 of this document:  http://www.usfsa.org/content/ISU%20Communication%201861%20Singles%20Pairs%20Scale%20of%20Values.pdf

Not quite - what is required is a reduction of 3 grades on what would have been given had the fall not occurred (see p13), and the final GOE must be in minus.  This does not mean that final GOE must be minus-3.

ETA - Nicklaszlo beat me.

[fsk8r]

I've sat on the judging panel timekeeping in the past and when the referee was signing off the results sheets, she was looking at them with the judge sitting next to her, and was commenting on how it was interesting how the spread of marks was down the panel (they were listed in orderof where they were sitting). Apparently the angle at which they saw the element affected how they were judging as one end was giving an element +1 and the other end -1 with the 0s in the middle. So you really must remember that the angle at which they watch will affect how they mark things which helps explain some of the spread seen.

[Query]

So is it your theory fsk8r that if they took the time to judge each move after looking at video tapes from several angles, judges would have more consistent scores?

I wonder if the fact that judges aren't supposed to consult with each other (I think) also affects things. If only they could say, pointing at a video, "Look at that edge." (For example.)

Of course, if judges took time to look at multiple video views on every move, it would take a very long time to get the results. In a televised spectator sport, that would never work, so this is pure speculation. Besides, much of the reason for interest in figure skating stems from judging controversies. A truly objective system might kill figure skating! 

I remember being very puzzled, pointing out on this forum that when looking at 2014 Olympic scores, some judges would mark something as -3 or -2 that others marked at +2 or +3. I wondered whether the whole process was completely political, and had very little to do with the actual performance. Someone pointed out what you just pointed out.

The anonymity of the current ISU system makes it very difficult to analyze sources of inaccuracy and bias.

[fsk8r]

I think judging is actually very hard to do well and I have the upmost respect for those who give up their time to do it. Yes, sometimes they'll make mistakes but I genuinely believe they're trying their best to come up with the farest result (at least at the level of competition I attend). And as long as they judge what they see from the angle they see it, the best overall result will come out as the size of the panel should compensate for the different views.

[Doubletoe]

Judges do have to take tests.  If the judge decided it was a "touch down" instead of a fall they may award -2.  In addition, a jump with a fall may earn GOE from the list of positive GOE criteria.  The highest GOE possible for a jump with a fall is -1.

Yes, I think it is -3 GOE subtracted from what the jump GOE would have been without the fall, so if it was an amazing jump with good height, air position, difficult transition, etc. but the skater fell on the landing edge, it could conceivably be a -1 GOE.

But I don't think it's the judges who decide whether a skater fell; i think that's the technical panel, who would call it as a fall (and subtract a point from the technical score for the fall).

[PhysicsOnIce]

So is it your theory fsk8r that if they took the time to judge each move after looking at video tapes from several angles, judges would have more consistent scores?

I wonder if the fact that judges aren't supposed to consult with each other (I think) also affects things. If only they could say, pointing at a video, "Look at that edge." (For example.)

Of course, if judges took time to look at multiple video views on every move, it would take a very long time to get the results. In a televised spectator sport, that would never work, so this is pure speculation. Besides, much of the reason for interest in figure skating stems from judging controversies. A truly objective system might kill figure skating! 

I remember being very puzzled, pointing out on this forum that when looking at 2014 Olympic scores, some judges would mark something as -3 or -2 that others marked at +2 or +3. I wondered whether the whole process was completely political, and had very little to do with the actual performance. Someone pointed out what you just pointed out.

The anonymity of the current ISU system makes it very difficult to analyze sources of inaccuracy and bias.

I sat on the Grand Prix Final Judging Panel as a timekeeper this year in Barcelona. Judges Actually can ( and sometimes do) look at the multiple video views, although not completely different opposite the two views do provide slightly different perspective on the elements. Most judges do at least take a second look through each element, but they are not required to do so, and though there is no specific time limit for for the judges once the technical panel authorises the elements there is often an unseen pressure to get the scores in ASAP.

Also, you have to keep in mind that positive goe givers (for example a difficult air position or difficult entry) don't have to be acknowledge by all judges. A judge for example can say "yeah he/she entered the axel from a spread eagle, but the entry edge was too long or the position was awkward the I'm not giving the +GOE". Take that same situation and add a fall. Automatic -3, but one judge liked the spread eagle entrance and another did not, you have one judge giving -2 and the other -3.  In short GOE is additive, however a failed element will at least have a -1.  Also keep in mind that things like GOOD quality edges and speed can give GOE, but who defines them?

[Query]

So is it your theory fsk8r that if they took the time to judge each move after looking at video tapes from several angles, judges would have more consistent scores?

I definitely don't understand the judging system, but I can do basic math.

If judges looked at every performance several times from several different views, and watched the critical elements in slow motion, zoomed in on skates and again on body elements as well, again from several different views - I guess it would take about 10 - 20 times as long to judge as to do the skate. That would be completely impractical for your average television viewer with a short attention span.

At less high profile competitions, maybe people would accept it (??), especially if everyone there could see all the videos too. But a major part of the cost of running a competition is the rental of the ice and facilities, which would go way up - so your entry fees would too.

Plus, the competitions would last more days, so you'd have to pay judges and maybe others for more travel expenses, again raising fees, and making it harder to rent the longer block of time for the competition.

And you'd need to recruit several times as many judges to accommodate the extra time. Maybe you could do that if they paid judges for their time - but that would raise your entry fees too.

And competitions, even minor ones, would need all that video equipment (some facilities would even need to heavy up the electrical capacity around the rink) - raising your fees again.

And BTW - the video people would have a lot more power, because the way they videoed skaters' performances, and what they selected as critical, would strongly influence scores, so people could still complain about bias.

I think your idea is really great. But the implementation is potentially a little complicated and expensive.

----

From what I've heard, good choreographers and coaches do try to take angles of view for the judges into account when they create programs, to optimize scores (e.g., sometimes they want to hide something - though of course those angles can be a little different at different facilities, especially the smaller ones.

There are lots of other factors that make judging potentially inconsistent too, no matter how competent and honest the judges. Like acoustics - in some cases there can be as much as 0.25 seconds time difference between the music the skater hears and what the judge hears - even more for the audience - that's a huge difference in judging the sync between skating and music, and much greater than the inconsistency at a land-dance performance on a small stage. And music with a fast beat can be very muddy and difficult to distinguish, for skaters, judges and/or audience in rinks that echo a lot.

Maybe we have to live with imperfection?

[icedancer]

I just wanted to butt in here that I don't believe that the judges actually get to look at the videos.  The technical panel does and they do discuss each element before giving an edge call, fall, etc (falls are obvious but touch-down, hand-down, two foot, under-rotated, etc. are hard to see sometimes).

The judges are told (through the headphones and on their screens) about the "mistakes" on the elements.

And please remember: it all goes by very very fast and judges are only human after all.

I encourage anyone who has any interest at all in judging (or any criticism of it) to get involved somehow - whether it be as a volunteer at your rink, volunteer as test chair, volunteer trial judge or just sitting at the sidelines taking it all in and try to make the calls - go back and look at how you scored compared to the actual volunteer judges sitting there on the panel.

[Doubletoe]

And this is exactly why they have 9 judges, not just one.

[karne]

I just wanted to butt in here that I don't believe that the judges actually get to look at the videos.  The technical panel does and they do discuss each element before giving an edge call, fall, etc (falls are obvious but touch-down, hand-down, two foot, under-rotated, etc. are hard to see sometimes).

The judges are told (through the headphones and on their screens) about the "mistakes" on the elements.

The judges can see the videos on their screens, but I don't believe they have access to all the same functions as the tech panel: slow, super slo mo, frame by frame etc (and yes, many tech panels will analyse a close spin or jump frame-by-frame in an attempt to give it to the skater).

The judges do not wear headphones and are not privy to the discussions of the tech panel. The only non-technical people who are privy to this discussion by necessity are the Data Operator and the Video Cutter.

A fall must be -GOE, and -3 must be taken off. Unfortunately, this gives rise to certain people only getting -2s or sometimes even -1 for a fall. Frankly, I think a fall should be mandatory -3GOE - it cannot score anything else.

[icedancer]

The judges can see the videos on their screens, but I don't believe they have access to all the same functions as the tech panel: slow, super slo mo, frame by frame etc (and yes, many tech panels will analyse a close spin or jump frame-by-frame in an attempt to give it to the skater).

The judges do not wear headphones and are not privy to the discussions of the tech panel. The only non-technical people who are privy to this discussion by necessity are the Data Operator and the Video Cutter.

I admit to not knowing what actually goes on in International Competitions.

At our local competitions - when we have IJS - we have screens for inputting scores which will show the edge calls, downgrades, etc and then a space for our GOE and then other screens for the components scores.  There is no video.

We do wear headphones and the panel will go through the list of elements to make sure that we are all on track. We do occasionally hear their discussions.

[Doubletoe]

A fall must be -GOE, and -3 must be taken off. Unfortunately, this gives rise to certain people only getting -2s or sometimes even -1 for a fall. Frankly, I think a fall should be mandatory -3GOE - it cannot score anything else.

So you think any fall on any jump should get a final GOE of -3, no matter how magnificent the jump may have been on takeoff, in the air, and on the landing before the fall (let's say the skater's blade hit a divot or rut in the ice on the landing edge)?  The judges' guidelines for GOE ask the judges to consider the preparation, takeoff, air position, rotation and landing of the jump in determining GOE, but you're basically saying the only important part of a jump is the landing.  If that's the case, why not just give a score of zero for any jump that results in a fall, even if it's a magnificent, fully rotated quadruple lutz?

[karne]

I admit to not knowing what actually goes on in International Competitions.

At our local competitions - when we have IJS - we have screens for inputting scores which will show the edge calls, downgrades, etc and then a space for our GOE and then other screens for the components scores.  There is no video.

We do wear headphones and the panel will go through the list of elements to make sure that we are all on track. We do occasionally hear their discussions.

The judging screens are the same whether you are international or not - if you look at the screen next time, you'll notice a space for video replay. We don't use video either.

I am surprised that the judges themselves are allowed to hear the technical panel's discussion. That's...awkward.

The Data Operator does do a readback of the elements for the tech panel to make sure all are correct.

[karne]

So you think any fall on any jump should get a final GOE of -3, no matter how magnificent the jump may have been on takeoff, in the air, and on the landing before the fall (let's say the skater's blade hit a divot or rut in the ice on the landing edge)?

Yes, I do. Because at the end of the day, a fallen jump is a failed jump. It doesn't matter if fifty transitions were done into it, the rotation was full, the air position divine. "landing on your butt" is not something that should ever be rewarded in the scores.

The judges' guidelines for GOE ask the judges to consider the preparation, takeoff, air position, rotation and landing of the jump in determining GOE, but you're basically saying the only important part of a jump is the landing.  If that's the case, why not just give a score of zero for any jump that results in a fall, even if it's a magnificent, fully rotated quadruple lutz?

Not at all. Of course everything else is important. But a fall immediately wipes out everything else in the jump that was done before it in terms of GOE.  Technically, of course, the jump was attempted, and everything must be looked at, and a base score given - because you can't not give something for the proper edge, the full number of rotations, etc. But that's the technical panel's job.

A fall should be a flat -3 and I'm not averse to some kind of reduction in base value percentage, either. Because if we're not careful we end up with planned falls, like Hanyu's 4S last year. Lunatic situations like where Hanyu fell five times and won a silver medal, or when Chan fell four times and won Skate Canada.

One of Max Aaron's quads at US Nationals got -1s and the odd -2 from the judges. One of Hanyu's falls on a triple jump got -2 at Cup of China. Are you telling me that a fall off a triple jump should be punished the SAME as a slightly pitched-forward, awkwardly landed (but cleanly landed) 4S?

[icedancer]

The judging screens are the same whether you are international or not - if you look at the screen next time, you'll notice a space for video replay. We don't use video either.

I am surprised that the judges themselves are allowed to hear the technical panel's discussion. That's...awkward.

The Data Operator does do a readback of the elements for the tech panel to make sure all are correct.

We use a "mini-system" for non-qualifying competitions here in the US.  I have done it many times. There is no room for video on the screen.  I have only heard a couple of brief discussions on the headphones from the tech panel and nothing earth-shattering or shady going on there LOL  I'm pretty sure that at these non-qualifying competitions they are not using any video replay either but I could be wrong about that too I suppose.

I checked with a friend who is a National judge - she said that they see the video and can do a quick replay but they don't get the detail of slo-mo, etc that the tech panel sees but they don't need to use it much - maybe just to check for rotations in pairs programs (like if one of the team does a triple and the other does a double - just to check).

They don't have headphones.

[fsk8r]

Yes, I do. Because at the end of the day, a fallen jump is a failed jump. It doesn't matter if fifty transitions were done into it, the rotation was full, the air position divine. "landing on your butt" is not something that should ever be rewarded in the scores.

Not at all. Of course everything else is important. But a fall immediately wipes out everything else in the jump that was done before it in terms of GOE.  Technically, of course, the jump was attempted, and everything must be looked at, and a base score given - because you can't not give something for the proper edge, the full number of rotations, etc. But that's the technical panel's job.

A fall should be a flat -3 and I'm not averse to some kind of reduction in base value percentage, either. Because if we're not careful we end up with planned falls, like Hanyu's 4S last year. Lunatic situations like where Hanyu fell five times and won a silver medal, or when Chan fell four times and won Skate Canada.

One of Max Aaron's quads at US Nationals got -1s and the odd -2 from the judges. One of Hanyu's falls on a triple jump got -2 at Cup of China. Are you telling me that a fall off a triple jump should be punished the SAME as a slightly pitched-forward, awkwardly landed (but cleanly landed) 4S?

They already penalise a fall. Any fall. A stupid trip clicked blade on a crossover and a fall on a quad jump. You get a deduction of minus 1. And at the low level the penalty is more than a single jump is worth. I don't need -3 GOE as well. It's hard enough getting a technical score at bronze level.

That's a problem with the IJS system, it's designed to work for all levels of skating, but it creates a situation where the low level skater is really penalised for a fall (it's the same as two single jumps!) but then the points for rotation are such that people feel the penalty isn't so great when the elite level skaters fall.

But I can safely say it's improving the quality of skating for those of us competing under it.

[karne]

That's why I prefer a percentage base value punishment over a -GOE punishment. But we have to work with what we've got.

[Doubletoe]

Yes, I do. Because at the end of the day, a fallen jump is a failed jump. It doesn't matter if fifty transitions were done into it, the rotation was full, the air position divine. "landing on your butt" is not something that should ever be rewarded in the scores.

Not at all. Of course everything else is important. But a fall immediately wipes out everything else in the jump that was done before it in terms of GOE.  Technically, of course, the jump was attempted, and everything must be looked at, and a base score given - because you can't not give something for the proper edge, the full number of rotations, etc. But that's the technical panel's job.

A fall should be a flat -3 and I'm not averse to some kind of reduction in base value percentage, either. Because if we're not careful we end up with planned falls, like Hanyu's 4S last year. Lunatic situations like where Hanyu fell five times and won a silver medal, or when Chan fell four times and won Skate Canada.

One of Max Aaron's quads at US Nationals got -1s and the odd -2 from the judges. One of Hanyu's falls on a triple jump got -2 at Cup of China. Are you telling me that a fall off a triple jump should be punished the SAME as a slightly pitched-forward, awkwardly landed (but cleanly landed) 4S?

I guess we just fundamentally disagree on falls, then.  I don't believe that a fall should negate the total value of what the skater just did.  To me, that's like saying someone's life was worthless because he died at the end of it, LOL!

That's why I prefer a percentage base value punishment over a -GOE punishment. But we have to work with what we've got.

So, basically, the same end result as a mandatory -GOE punishment, since GOE is also a base value percentage?  But not a big fat zero for the final score of the jump?

That's a problem with the IJS system, it's designed to work for all levels of skating, but it creates a situation where the low level skater is really penalised for a fall (it's the same as two single jumps!) but then the points for rotation are such that people feel the penalty isn't so great when the elite level skaters fall.

As an Adult Gold level skater who has to compete under IJS, I agree that a fall hurts my score a lot more than that of a skater who is rotating triples and/or quads.  Then again, I think that the time, effort, pain and risk these skaters put into mastering those triples and quads should earn them that advantage.  If a fall on a quad makes all of that training time, energy and risk worth nothing, then why should anyone even try it?  Why would anyone ever have any incentive to push the technical boundaries of the sport?

[sarahspins]

I guess we just fundamentally disagree on falls, then.  I don't believe that a fall should negate the total value of what the skater just did.  To me, that's like saying someone's life was worthless because he died at the end of it, LOL!

Agreed.. I think the mandatory deduction for a fall should be a percentage of the base value of the jump, rather than a fixed value.  Losing something like 25% of the base value of a jump for a fall would eliminate the "lose more value than the jump was worth" paradox at lower levels, while increasing risk fairly at higher levels - because 1 point isn't much on a jump with a base value of 10+ points, but it certainly is for a jump that is worth less than a point (which would be all single jumps, and a single axel is only 1.1).  If you went into a quad knowing you could lose 2.5 points for a fall, you'd want to know you'll hit that jump before throwing it in your program... but as it is now, skaters put quads in when they know they can't land them because even with a fall, if it's rotated, it's still worth more than a triple.

[Loops]

Agreed.. I think the mandatory deduction for a fall should be a percentage of the base value of the jump, rather than a fixed value.  Losing something like 25% of the base value of a jump for a fall would eliminate the "lose more value than the jump was worth" paradox at lower levels, while increasing risk fairly at higher levels - because 1 point isn't much on a jump with a base value of 10+ points, but it certainly is for a jump that is worth less than a point (which would be all single jumps, and a single axel is only 1.1).  If you went into a quad knowing you could lose 2.5 points for a fall, you'd want to know you'll hit that jump before throwing it in your program... but as it is now, skaters put quads in when they know they can't land them because even with a fall, if it's rotated, it's still worth more than a triple.

This argument sounds very sane and reasonable.  What is the reason this method is not in place?  Why just the fixed point off and not a percentage? 

I'm still trying to wrap my head around the GOE system.  I presume its still evolving?  I've not competed solo under GOE yet, but I can imagine the scores must be demoralizing at the lower levels.  Since by and large that's going to be kids, seems like even just for the psychological factor there'd be an argument for percentage based rather than flat rate deductions.

[Christy]

I sat with some judges at a competition and it seemed like it was an automatic -3 for a fall, however some went back and adjusted a couple of scores - I don't know the reasons why, but it seems like they did the -3 as soon as the fall happened, but then reviewed some information and made a couple of judgement calls.

I think an automatic -3 for a fall seems harsh as it is only part of the overall jump, and a fall does not mean that the rest of the jump was wrong - sometimes it's a result of landing on less than perfect ice.