How do you penalize falls?

Kwanford Wife said:

I think - you fall, no points. Is that harsh? But I still don't understand all the nuances of CoP...

Click to expand...

Isn't it odd that the title of this thread could just have easily been:

How should the CoP reward falls

Here is a crazy idea. First fall -1, second fall -2, third fall -3, etc. Instead of each fall being -1.

Is it required that falls be treated the same way in the SP as in the FS? IIRC, the SP was, originally, the "technical program" that was "all about required elements"; the FS allowed more room for other things. So, unless otherwise forbidden, I would suggest that a fall on a required element in the SP means 0 points on that element (plus the deduction), but this rule would not apply in the FS -- where skaters could take more risks, if they so chose.

I would also combine this with "progressive deductions" (1 point for the first fall, as done currently; a second fall is a 2 point deduction, etc.) in both programs.

attyfan said:

Is it required that falls be treated the same way in the SP as in the FS? IIRC, the SP was, originally, the "technical program" that was "all about required elements"; the FS allowed more room for other things. So, unless otherwise forbidden, I would suggest that a fall on a required element in the SP means 0 points on that element (plus the deduction), but this rule would not apply in the FS -- where skaters could take more risks, if they so chose.

Click to expand...

In the 6.0 system, it was closer to the other way around.

In the long program, skaters could try as many elements as they liked, and they only got credit for what they completed. Jumps "landed on two feet" (weight on both feet, not just a light touchdown) and jumps with falls would get no credit, other elements and jumps with lesser errors would get partial credit, and a fall itself was not officially penalized and specifically "not a bar to winning." Nor was there any penalty for trying a failed element again, as doubleflutz suggests in post 16 above.

Of course, judges were free to reflect in their marks their impression of the program as a whole, including falls, however they chose.

In the short program judges set the base mark for the Required Elements mark based on what the skater attempted, and then they took the mandatory deductions for errors or failures.

So, for example, suppose you're looking at two long programs under 6.0 that are otherwise comparable.
Andy lands 8 triples successfully and doesn't attempt a quad.
Bill lands the same 8 triples successfully and also attempts a quad that ends in a fall.

Who has the advantage? Officially, they both landed the same jumps. So it would be up to each judge to decide whether the fall on the attempted quad disrupted the performance enough to penalize, or whether it showed extra risk and extra content attempted and should be rewarded as such.

Or more likely the decision would be based on factors other than jump content (since the successful jump content was pretty much identical) or on fall vs. no fall.

In a short program, suppose both skaters did the same jump combo (3Lz+3T) and same solo jump from steps (3F), but Carl did a successful but not special double axel and Don fell on a rotated triple axel. Don gets a 0.4 deduction, but Don also starts with a higher base mark. Theoretically the Presentation mark was supposed to be independent of the Required Elements mark, but in practice they were usually within a couple tenths of each other, before deductions -- the theory probably being that a more difficult program presented equally well should be worth more for presentation.

So say a judge thought a triple axel should be worth at least 0.4 more in the base mark than a double axel and also thought that the fall had a small negative effect on the presentation of the program. Maybe she gives Carl 5.2/5.2 and Don 5.2/5.5 (would have been 5.6/5.6 if he'd landed the 3A). Don comes out ahead thanks to trying the harder jump.

I think one fall champion is fine as long as everything else is good and the program includes the hardest jump/jumps. Two falls against a clean skate from others might be too much to be accepted for the highest honor in a competition. But it should be allowed in silver or bronze position depends on the details of skaters' skating. Three falls should take the skater off the podium but might be able to stay in front of a no-fall skate other skaters had depends on the details of the skating. Example, a three fall skating from the last group skater against a no-fall skating from the first group skater. Four falls should probably be the last.

CoP should consider the tolerance level for falls, and penalize the falls accordingly. Falls should be allowed. The first fall should not be penalized any more than we have already. Figure skating is first and for most a sport. It should encourage the risk-taking moves, not give chance for finding other easier ways to compensate the athletic elements.

JMO.

janetfan said:

Isn't it odd that the title of this thread could just have easily been:

How should the CoP reward falls

Click to expand...

Bluebonnet said:

I think one fall champion is fine as long as everything else is good and the program includes the hardest jump/jumps. Two falls against a clean skate from others might be too much to be accepted for the highest honor in a competition. But it should be allowed in silver or bronze position depends on the details of skaters' skating. Three falls should take the skater off the podium but might be able to stay in front of a no-fall skate other skaters had depends on the details of the skating.

Click to expand...

Shouldn't it all depend on the details of the skaters' skating? Falls are just one detail (or three details if the skater falls three times ), but there's the whole rest of the program to consider.

What was the quality of the skating? What was the difficulty of the jumps completed? What was the quality of the jumps completed? What were the difficulty and quality of the spins, steps, spirals and connecting moves, etc.? What was the quality of the carriage, projection, choreography, musical interpretation, etc.?

How much of the 97 other details do you want to ignore just to privilege the 3 details that bother you the most?

Sometimes the difference between a program with several falls and the next best program in the competition is larger than the negative impact of the falls, at least in the eyes of the judges who see most of those 97 details and not just the blatantly obvious ones.

A few examples:


http://www.youtube.com/watch?v=YyMrZOXhrtU
vs.
http://www.youtube.com/watch?v=vVxkIuTKRcs

http://www.youtube.com/watch?v=sKQPQ_9HZSE
vs.
http://www.youtube.com/watch?v=kIIwlmykmRQ

And of course, if all the skaters fall more than once, the winner is going to be someone who fell more than once.
You can't make a rule in advance that will prohibit someone with a certain number of falls from winning.

All you can do is set rules or guidelines for how to penalize falls in terms of points, but not placements.

gkelly said:

Andy lands 8 triples successfully and doesn't attempt a quad.
Bill lands the same 8 triples successfully and also attempts a quad that ends in a fall.

Who has the advantage? Officially, they both landed the same jumps. So it would be up to each judge to decide whether the fall on the attempted quad disrupted the performance enough to penalize, or whether it showed extra risk and extra content attempted and should be rewarded as such.

Or more likely the decision would be based on factors other than jump content (since the successful jump content was pretty much identical) or on fall vs. no fall.

In a short program, suppose both skaters did the same jump combo (3Lz+3T) and same solo jump from steps (3F), but Carl did a successful but not special double axel and Don fell on a rotated triple axel. Don gets a 0.4 deduction, but Don also starts with a higher base mark. Theoretically the Presentation mark was supposed to be independent of the Required Elements mark, but in practice they were usually within a couple tenths of each other, before deductions -- the theory probably being that a more difficult program presented equally well should be worth more for presentation.

So say a judge thought a triple axel should be worth at least 0.4 more in the base mark than a double axel and also thought that the fall had a small negative effect on the presentation of the program. Maybe she gives Carl 5.2/5.2 and Don 5.2/5.5 (would have been 5.6/5.6 if he'd landed the 3A). Don comes out ahead thanks to trying the harder jump.

Click to expand...

I was thinking about the same sort of thing. And in another thread it was pointed out how few ladies are doing all 5 jumps, making falls too costly would just make that problem worse.

^^^
Yeah, it does. However, if one is concerned about pampering a skater, then there should be a general score for a missing part afecting the entire element. A fall affects more than a -1; A wrong edge takeoff loses the name of the jump; an UR is an incompleted jump or a less than acceptable spin - so lets take 3 points off for each of the aforementioned errors and the same for repeated errors.

A skater who can not do a clean routine shoulld be penalized thusly or more severely than what we are getting now. The plus GoEs should be in the PC scores. The Tech scores should be judged soley on technical definitions. There is a SPORT somewhere in this Figure Skating thing. No?






r

Perhaps falls should be deducted based on difficulty. If the fall happens on an easier element, the deduction should be severe. Why SHOULD a senior skater be falling on a three-turn?!

Then again, difficulty is relative. To some skaters, a triple loop is nothing, but Yu Na Kim doesn't use them.

I think what the problem may be is not so much that the falls don't count but that the effect can be erased. I am not an expert, so if I am factually incorrect on any scoring rubrics, feel free to tell me nicely

If you had a mandatory deduction in the short program under 6.0, and let's say you lost 0.3 on a missed jump--that was it. You couldn't make it up. Your score couldn't be higher than 5.7. But under COE, you lose -3 for falling, and then another -1. Fine. But if you do another jump really well, you can get a +3 GOE. This more or less erases the effect of the fallen jump. Am I correct?

So if you do another jump really well, and get a +3 GOE, all you've wound up losing is 1 point off of a total of around 75-80 (for a top man).

0.3 is 5% of 6.0. 1 is 1.25% of 80 and 2.5% of 40 (if you separate the PCS and Tech scores). Falls do count for less than they used to. (It would be a little more damaging if a woman fell.)

I think they should get rid of GOE. If a jump is perfect, that's it. Full credit. No bonus points (except maybe the one for doing the jumps at the end). As it stands now, a good jump erases almost all the deductions of a fallen jump, which is why certain people fall all over the place and still win. You only need to land a little over half --5/8--and if you land them really well, it's like the falls never happened.

Poodlepal said:

If you had a mandatory deduction in the short program under 6.0, and let's say you lost 0.3 on a missed jump--that was it. You couldn't make it up.

Click to expand...

Mandatory deductions applied only to the short program.
It was not allowed to make up missed jumps by trying them again or by adding a different extra jump in the short program.

But base marks in the short program were not based only on how many jumps were landed. The difficulty of the jumps attempted counted. So did the difficulty of the spins and steps and spirals. So did the quality of all those elements. So did the quality of the basic skating.

So a skater who started with the highest difficulty and did everything else well and missed one jump in the short program might earn a Required Elements mark of 5.5 or 5.6 or 5.7 even after the deduction, because they started from a base mark of 5.9.

On the other hand, another skater with less overall quality or less difficulty might start from a base mark of 5.5 or 5.6 or 5.7. So even if they didn't make any mistakes and don't get any deductions, their score would still be lower than that of the stronger skater with the one mistake.

In long programs there were no required deductions. If you attempted the highest difficulty, with good quality on all but one or two elements, and fell on one jump, it was still possible to score 5.9 for Technical Merit.

Your score couldn't be higher than 5.7. But under COE, you lose -3 for falling, and then another -1. Fine. But if you do another jump really well, you can get a +3 GOE. This more or less erases the effect of the fallen jump. Am I correct?

Click to expand...

There are lots of areas in which to gain points. In IJS it's spelled out where each point comes from. In 6.0 the judges just evaluated the whole program in their in their heads and then translate that evaluation into two marks. A program with lots of good stuff and one or two big mistakes would often score higher than a program with lots of OK stuff and no big mistakes.

+3 for difficult jumps is really really rare. More likely the skater would make up for the missed jump by earning +1s and +2s on a lot of different elements throughout the program. I.e., it was a high-quality performance throughout except for the element with the fall.

I think they should get rid of GOE. If a jump is perfect, that's it. Full credit.

Click to expand...

What constitutes perfect?
Is a jump that's telegraphed, is entered slowly, rotates with a wrapped or tilted position in the air, and lands less than a blade length from where it took off with no speed and a weak landing position "perfect"? Scott Hamilton might say so. If it's fully rotated and landed on a back outside edge, it satisfies the requirement for the jump. No need for -GOE (or for short program deductions in 6.0), although some judges might ding it a little for the lack of quality.

So why should that successfully landed jump earn the same score as a jump that's entered and exited with speed and flow on strong edges with good height, good distance in the air and on the landing edge, and good body position in the air and on the landing edge?

No bonus points (except maybe the one for doing the jumps at the end). As it stands now, a good jump erases almost all the deductions of a fallen jump, which is why certain people fall all over the place and still win. You only need to land a little over half --5/8--and if you land them really well, it's like the falls never happened.

Click to expand...

Here you're clearly talking about long programs. So deductions were not an issue under 6.0.

It has always been true that a skater with good jumps and good other skills could beat a skater who landed more jumps with less overall quality. Or a skater with more successful content despite a couple of falls could beat a skater who skated an easier program cleaner. This isn't new. We're just seeing the nuts and bolts of how it happens now.

gkelly said:

It has always been true that a skater with good jumps and good other skills could beat a skater who landed more jumps with less overall quality. Or a skater with more successful content despite a couple of falls could beat a skater who skated an easier program cleaner. This isn't new. We're just seeing the nuts and bolts of how it happens now.

Click to expand...

If only that were true.
As I watch on IN or TV I see a score. I never see any "nuts or bolts" that you mentioned.

I watch an SP or LP, see a couple of replays of highlights of the program and then a score pops up. I don't see anything more. I don't see URs or edge calls and simply have to guess what the score means.

I guess 146 means the judges thought it was better than the skater who received 143.
I understand it is possible after the fact to get a look at the score with these nuts and bolts - but it is not always clear to me where to find it.

And usually I would rather watch the next skater rather than go looking for this information.

It seems to me ISU intentionally divulges as little info as possible when they post a skater's score.

Nuts and bolts are fine - but to be honest by the next day I am not as interested. Does that make me a bad fan? I watch alot of skating because I enjoy it. As to the wonder of the IJS - it is not as transparent to me as you describe it.

I think a majority of US fans were left a bit bewildered by the results/scores of the ladies event last season. Finding out the next day more about it was not so satisfying and as a fan it feels like we are left out of this secretive system.

And yes, I watched which judge gave which mark for too many years and still feel like something is missing - because IMO it is missing.

janetfan said:

I understand it is possible after the fact to get a look at the score with these nuts and bolts - but it is not always clear to me where to find it.

And usually I would rather watch the next skater rather than go looking for this information.

Click to expand...

Right, it's not available immediately because it has to be verified first. And in case there isn't time to analyze it in detail while waiting for the next skater to skate.

The detailed protocols are published very shortly after the conclusion of the long program. For short programs you do have to wait a day or two to get access to them.

For international events, they are almost always linked are the main page of isu.org, and then for up to a week or more on the Singles/Pairs/Ice Dance page.

For US qualifying events, they're usually published on IceNetwork. Lately they've even been giving live scores, without the breakdown by judge.

There are often links given in the event threads on this and other discussion boards.

At live events, copies of the detailed printouts are often available for purchase. But it saves money and trees to look for them online.

The summary results, listing total element score, averages of each of the five components, and number of deductions, as well as the total segment score for each skater, is posted on the wall in the arena.

Under 6.0, all that was ever posted or made available for purchase in the 6.0 protocols was the two scores and the ordinal they represented for each skater from each judge. There was never any information given to the public or to the skaters about reasons for deductions, where a skater gained or lost points. Now all that information is available within a day or two, or within an hour or so for long programs, for those who are interested enough to track it down.

It's very educational. But if you're not that interested in the details, it might not be worth your while. Different strokes.

^ This would be an interesting experiment: Get a bunch of "average fans" and let them watch some programs judged under 6.0, then some programs judged by the Code of Points.

Let the spectators guess who won, and then compare that with who actually did win. Which system better satisfies the audience's perception that this skater's performance was better than that skater's?

My impression is that the ISU does not care what the answer to that question is.

I believe it to be in the interests of the sport to treat fans as something more than just a target to be tutored and lectured at. IMO it would be a good start to give 0 credit for a fall on a jump. This would be a far, far better solution than to launch into an explanation of why an under-rotation or a wrong-edge take-off is actually a worse error because of blah, blah, blah. We cannot blame the fan who responds, "Whatever."

And that way Scott Hamilton can yell into the mic, "Now here he goes up for his QUAD! This is a huge element in figure skating. Just think about it! 1440 degrees int he air! This element is worth a whopping 10.3 points!...Aw...well, give that cowboy a big hand anyway, that's all he is going to get for his ride tonight."

gkelly said:

Right, it's not available immediately because it has to be verified first. And in case there isn't time to analyze it in detail while waiting for the next skater to skate.

The detailed protocols are published very shortly after the conclusion of the long program. For short programs you do have to wait a day or two to get access to them.

For international events, they are almost always linked are the main page of isu.org, and then for up to a week or more on the Singles/Pairs/Ice Dance page.

For US qualifying events, they're usually published on IceNetwork. Lately they've even been giving live scores, without the breakdown by judge.

There are often links given in the event threads on this and other discussion boards.

At live events, copies of the detailed printouts are often available for purchase. But it saves money and trees to look for them online.

The summary results, listing total element score, averages of each of the five components, and number of deductions, as well as the total segment score for each skater, is posted on the wall in the arena.

Under 6.0, all that was ever posted or made available for purchase in the 6.0 protocols was the two scores and the ordinal they represented for each skater from each judge. There was never any information given to the public or to the skaters about reasons for deductions, where a skater gained or lost points. Now all that information is available within a day or two, or within an hour or so for long programs, for those who are interested enough to track it down.

It's very educational. But if you're not that interested in the details, it might not be worth your while. Different strokes.

Click to expand...

Thanks for the info. I am interested - but to a point - mostly when I see a decision or podium that I disagree with.

I think you are not being fair to 6.0. I immediatley saw a tech score and then a presentation score. I also saw how each judge scored the skater and which federation they were from.

Maybe you don;t think that is important but not everyone agrees with that.
If it was a close decision the information i really wanted was to see was the ordinals. Was it 5-4, 6-3? Was it so close that better presentation is what won the event?

That's what i grew up with and I am not convinced despite the good reasons you provide that I am so wrong about this.

I think I care more about how each judge marked a skater for tech and presentation than comparing an anonymous mark for a spiral sequence or seeing the varying and subjective GOE.

Not saying the nuts and bolts aren't good. But I still see them as mostly subjective and whether we break skating down into 10 categories or two - it will always be more about what a judge liked. I want to know which judges liked Michelle and which ones liked Irina. I already know Michelle has a better spiral and that Irina jumps a little higher.

janetfan said:

Not saying the nuts and bolts aren't good. But I still see them as mostly subjective and whether we break skating down into 10 categories or two - it will always be more about what a judge liked. I want to know which judges liked Michelle and which ones liked Irina. I already know Michelle has a better spiral and that Irina jumps a little higher.

Click to expand...

To be blunt, the system is not for you. NOT YOURS. NOT FOR YOU.

It's for the ten-year-old skater and her coach, who is going to find it much more helpful to know that her skating have gone up dramatically from a few months ago since she started all those drills, but she still needs work on listening to the music and connecting with the audience, and that her layback is awesome, the sit spin needs work, her double sal is really nice, the axel is coming along well, but the double loop is cheated and the lutz isn't a lutz.

Anonymous judging is a terrible thing, but everyone complaining about the complexity of the system being too hard to understand needs to stop. IT'S NOT FOR YOU.

^ Yes - a good point. Even the quad. Stephane Lambiel had a quad last season - but no triple axel. And Mao has a triple axel - but no triple lutz.

Still, COP does decide which jumps are the hardest - by assigning more points to the harder elements. I think this the most straightforward way to reward risk. Well - not really reward risk - reward ability

It encourages skaters to work toward a quad - but not to just throw in a quad and hope for the best because they won't get penalized too harshly if they fall. No. This is not the way Yagudin or Plushenko won all of their Olympic and World medals. They used the quad because they were so good at quads they were probably going to land them. Same thing with Yuna and her triple triples.

Perhaps it is true that the quad wasn't rewarded enough and maybe it's good that they are giving it more points. I have no problem with that. Because that will encourage skater to practice quads- knowing that they might need it to beat the other guy - but they aren't going to put one in at worlds if they know they have a 70 percent chance of falling. Or at least they shouldn't.


As for whether a fall should get zero points, I am a bit torn. Here's why:

Skater A fully rotates his quad and falls.

Skater B doesn't fully rotate his quad and falls. Shouldn't one fall A be a little less costly than fall B?

doubleflutz said:

NOT FOR YOU.

Click to expand...

Quite so. I have gradually come to realize that this is indeed the stance of the figure skating establishment. Figure skating has made its choice. It is a participatory sport and not a spectator sport. The rules are for the participants not for the audience.

I guess all I can say is, be happy in the life you have chosen (to quote Ebeneezer Scrooge's ex-girl friend. )

Layfan said:

As for whether a fall should get zero points, I am a bit torn. Here's why:

Skater A fully rotates his quad and falls.

Skater B doesn't fully rotate his quad and falls. Shouldn't one fall A be a little less costly than fall B?

Click to expand...

No. Don't look at it in terms of "costs" and penalties. There would be no penalty for a fall. You just didn't do that element.

Keep the shiny side down.

And this is what I don't get, Mathman. You're clearly not willing to apply that reasoning to other things. Get an edge call? No penalty, but you clearly didn't do that element.

It's like saying 1+1=2, but not being willing to say 1+2=3