Should the ISU have separate scoring systems for men and women?

[gkelly]

Actually, I think these examples make my point.

I guess I'm still not sure what your point is.

I never bothered to learn a Salchow or a loop, and all I can do is flutz.

Therefore you're at a disadvantage compared with skaters who can do all the jumps. You don't have enough different triples to fill all the available jump slots, so you have to repeat double axel instead of including a jump that would be worth at least 1.5 more.

Plus, it's probably not a case that you "never bothered to learn" a triple salchow or loop. More likely you did try, especially with the salchow, but didn't succeed in mastering it for some reason.

Double salchow is almost always the first double that skaters learn and very often the first triple they attempt, right after they learn the single or double axel respectively, because of the similarities between the salchow and axel technique and the fact that it requires the least actual rotation in the air to count as a clean double or triple. Once they get the hang of rotating 2 or 3 times in the air, though, the toe loop will probably be a more consistent jump for most skaters, and for a rare few the flip and/or lutz will come easier.

Tonia Kwiatkowski come to mind as a skater who struggled with the triple salchow in juniors and dropped it from her repertoire for most of her senior career.

What might be more of a concern would be that skaters for whom the toe jumps come easier would not bother to continue training their inconsistent triple salchow or consistently not rotated triple loop once they have several other triples they can use. Especially when the lutz and flip are explicitly worth more points and the number of jump passes allowed is limited.

But if the skater can do triple loop and can do triple flutz but not true lutz, there's no advantage to repeating the flutz and omitting the loop. Unless the loop is just as likely to get downgraded and/or have other -GOE errors as the flutz is to get the edge call and -GOE.

But I just won the bronze medal, losing only to a skater who also did a four flutz program but had a triple-triple (silver) and an imaginary skater who has both a 3Lz+3T and a 3F+3T (imaginary gold.)

Note that the silver medalist likewise omitted the loop, and the gold medalist would have won anyway if she omitted the loop and replaced it with a double Axel.

So what's the problem? The skaters with better jump content placed better.

And if someone had come out and done the same content as you, with distinctly different takeoffs on the lutz and flip, all else being equal, they would have beaten you too.

Is the concern that you, with five clean (aside from the lutz takeoff edge) triples of three different kinds, one of the repeated types flawed on the takeoff in a way that made it virtually the same as the other repeated type (i.e., four flip-like triples and one triple toe), ahead of other unmentioned skaters who also landed five clean triples but of four or five distinct kinds?

I don't think you can beat them just by planning higher-point-value jumps and fudging the takeoffs of one of them. You'd also have to be better at actually landing those jumps or better at skating or spinning (or spirals, since you're a senior lady in this hypothesis) or presentation.

I think the thrust of the proposal is to prevent skaters from loading up on high scoring toe pick jumps to the extent that edge jumps become irrelevant.

There are other ways to achieve that goal without eliding the difference between a correct lutz and a correct flip.

[steyn]

It seems to me that the two suggestions are contradictory to each other.

In the first suggestion, raising the base value of 3A is because there are few ladies to do it. So it would encourage ladies to learn 3A.

In the second suggestion, combining flip and lutz is again because there are few ladies to do them properly. So it would discourage ladies to learn the true flip and the true lutz.

I don't get the rationale behind them.

If you dislike 4 flutzs would won the game, it would be more consistent to suggest to give more negative GOEs to flutz/lip than now rather than discouraging hard working ladies. Or you can reward skaters doing proper flip AND lutz by other means, for example by awarding an extra bonus point.

I think 8.2 for 3A is not small. It is 4.7 point higher than 3.5 for 2A. That is, if you can do 3A, you will have 4.7 point advantage, which is not small at all. In Mao's case for example, she did not have much benefit because she was not able to do a 3-3. If 3A is such a big thing which is technically so superior to other elements, it is her fault that she failed such an easy element like 3-3 or 3S. I would say if you have a well-balanced program then the advantage of 3A is quite big even with the current base value. (I am not saying that 8.2 is the right base value of 3A. The proper value might be higher or lower, I don't know. I am just saying that the advantage is not negligible even with the current value.)

As for the issue of awarding extra points, say 5 points, to skaters with complete set of triples, should a failed triple (DGed or fall) be counted or not? If it is counted, every skater would try that, resulting in many horrible performances. If not counted, just one mistake would make a huge difference, which I don't think people would be happy about. So my opinion is that we just award extra one point for each different kind of triples. So for example if a skater did 3 different triples, the she would get 3 extra points. If 5 different triples, 5 points and so on. I think this would be enough to encourage skaters to do complete set of triples. This will also effectively raise the base value of 3A by one point without sacrificing the virtue of well-balanced programs.

[prettykeys]

But the loop and the Salchow are very different jumps, never mind edges. As I understand it, the Salchow takes off from the opposite leg from the landing leg, and the loop takes off from the same leg as the landing leg (both from back inside edges.)

Yeah, I did get the feeling that they were...I can differentiate them only by looking to see if the skater takes off on the leg on the inside of the direction of rotation (Salchow), or the leg outside of the rotation (Loop). Otherwise, superficially, they look quite similar to me. On the other hand, the 3Flip is always easy for me to pick out (and note it's not a Lutz) because the skater almost always "flips" directions. And the Lutz is unique because of its counterrotational mechanics. I think all of these jumps need to remain as separate elements. It's not just a "cop out", shrinking the repertoire to 4 different types of main triple jumps as per your suggestion is still narrowing the forms of skills that skaters could be able to showcase, and I don't know why that is desirable. In fact, the example with the ToeLoop/Walley that you gave sounds regretful for me...do we ever see 3ToeWalleys?

I think your suggestion does the opposite oif its intention. A male champion typically might get 85 TES and 80 PCS (8's across the board.)

A ladies' champion might get 65 TES and 80 PCS (8's across the board.)

By your suggested method the man's total score is 82.5, weighted 52%/48% in favor of TES. The lady's total scpre is 72.5, weighted 45%/55%.) This would codify the principle, ladies are pretty, men are athletic.

Oops, you're right. My intention was something like Total Score = TES x (1 + PCS/100) or Total Score = TES x (1 + PCS/50), something like that, in some balanced manner so that the PCS matters but is still related to the TES. My first example still gave TES and PCS as independent marks...although yeah, in principle I prefer that to the arbitrary fixed differences in PCS factoring that we currently have.

I look at it this way. According to the ISU scale of values, the 2A is grouped with the triple jumps as a baby triple, rather than a super single. Likewise, the 3A is grouped with the quads, not the triples.

The gap between the highest double rotation jump (2Lz = 1.9) and the "baby triple" (2A = 3.5) is a whopping 84%.

In contrast, there is only a 37% increase between the triple Lutz and the triple Axel. If the 84% rule were applied consistently, the proper value for the 3A would be 11.0 points. (Just saying...)

Yeah, I've crunched the jump base value numbers before...I came to a slightly different conclusion by trying to lessen the value of the 2A. What I ended up doing is inflating the value of the rest of the triple jumps (+0.5 for each triple and +0.8 for the 3A); however, I also asked for factored GoE's (rather than absolute GoE's of +1, +2, etc.)

http://www.goldenskate.com/forum/sho...ht=#post484703

2A: 3.5

3T: 4.5
3S: 5.0
3Lo: 5.5
3F: 6.0
3Lz: 6.5
3A: 9.0

4T: 11.5

I don't think the reason for the value of the 2A being so high is because it's considered a "baby triple", I think it's in there because it's still a required jump for the ladies (e.g. the SP) and if they miss it, it should be a pretty big deal. Again, however, as a thought experiment, if you want to consider an 84% value increase (I don't know why it has to be done in percentages, but let's run with it) for an additional half-rotation, then allow a 4T to have a base value of 20.2. A little too ridiculous in my book.

37% increase in base value (for optional jumps) between the 3Lz and 3A remains the highest separation for a half-rotation's difference. The difference between a 3A and a 4T is 20%. All this points to be the quads being undervalued, not the 3A.

[prettykeys]

As for the issue of awarding extra points, say 5 points, to skaters with complete set of triples, should a failed triple (DGed or fall) be counted or not? If it is counted, every skater would try that, resulting in many horrible performances. If not counted, just one mistake would make a huge difference, which I don't think people would be happy about. So my opinion is that we just award extra one point for each different kind of triples. So for example if a skater did 3 different triples, the she would get 3 extra points. If 5 different triples, 5 points and so on. I think this would be enough to encourage skaters to do complete set of triples. This will also effectively raise the base value of 3A by one point without sacrificing the virtue of well-balanced programs.

I would count a DG'd jump that is landed decently, but not a fall. It would still be penalized for being DG'd. On the other hand, an edge call (e) wouldn't count towards doing 5 different triples because you are changing the type of jump you are supposed to be doing...

Also, your jump-by-jump point addition wouldn't work. Someone like YuNa who already does 4 different types of triples (Lutz Flip Salchow Toe) would get negligible benefit for taking the risk of trying a 3Loop in competition...if successful, she'd only gain +0.5 total points (5.0 BV +1 for the extra jump, instead of her 2A + 2GoE = 5.5)

Nahh, I'd rather see a clear bonus for doing 5 different types of triples.

[steyn]

Also, your jump-by-jump point addition wouldn't work. Someone like YuNa who already does 4 different types of triples (Lutz Flip Salchow Toe) would get negligible benefit for taking the risk of trying a 3Loop in competition...if successful, she'd only gain +0.5 total points (5.0 BV +1 for the extra jump, instead of her 2A + 2GoE = 5.5)

Theoretically yes. In practice, more than 0.5 point because successful 3Loop would get about +1GOE. So it would be more like 1.5 point difference. I feel this would be enough.

I would count a DG'd jump that is landed decently, but not a fall. It would still be penalized for being DG'd. On the other hand, an edge call (e) wouldn't count towards doing 5 different triples because you are changing the type of jump you are supposed to be doing...

Nahh, I'd rather see a clear bonus for doing 5 different types of triples.

So will you be happy if a questionable edge call makes 6-7 point difference?

[prettykeys]

Theoretically yes. In practice, more than 0.5 point because successful 3Loop would get about +1GOE. So it would be more like 1.5 point difference. I feel this would be enough.

I don't know if YuNa would get +1GoE for her 3Loop. But besides that, I still don't feel it is enough. I know that YuNa still practices her 3Loop, and I don't believe that she "doesn't have it" like some of her critics say. I think she has it, but maybe it's not as consistent as her other jumps...or maybe she just gets nervous doing it in competition. It's probably the combination of those things, plus, why use precious time and energy to seriously practice it when she doesn't need it? All that trouble for 0.5-1.5 extra points? C'mon, we know YuNa and her team are smarter than that.

So will you be happy if a questionable edge call makes 6-7 point difference?

Yes, I will be happy.

[gkelly]

If the intention were specifically to encourage edge jumps, then it might make sense to leave the value of the 3S as higher than the 3T but to increase the value of 3Lo to be equal to that of 3F.

Right now for doubles the values are
2T and 2S 1.3
2Lo 1.5
2F 1.7

If the goal were to encourage edge jumps, why not change that to something like
2T 1.3
2S 1.5
2Lo and 2F 1.7

Already a flutz is worth less than a flip of the same rotation if the proper GOE reductions are taken. With these changes to the scale of values, flutz would also be worth less than a clean loop, and a "lip" would remain worth less than a clean salchow.

If it's just a matter of choosing what to work on, any skater who now thinks it's more valuable to do two flips and two lutzes and no loops, even with the edge calls and -GOE, would have to think again. They could choose to put more training effort into the loop than the lutz and include that instead or in addition. They could put more training effort into the lutz and get the takeoff correct so that they can get positive GOE for it and avoid accusations that they're doing more than two triple flips. Or if they really can't get any reliable triples besides the toe loop and flip, and can get their flips from a lutz approach called as 3Lz e or 3 Lz ! instead of 3F*, then they can stick with that limited jump content and expect to lose to skaters with more variety of triples.

As for the issue of awarding extra points, say 5 points, to skaters with complete set of triples, should a failed triple (DGed or fall) be counted or not? If it is counted, every skater would try that, resulting in many horrible performances. If not counted, just one mistake would make a huge difference, which I don't think people would be happy about.

Yes, these are issues that would need to be considered in developing a proposal to explicitly reward variety of jump takeoffs.

How big should the reward be, and how successful does a jump have to be to count toward earning it?

My hypothetical proposal was 2.0 for 6 takeoffs as doubles or better and 4.0 bonus for 6 different triples or better. (For ladies either make it 5 different triples or include the 2A for ladies but not for men, maybe for both sexes make it 3.0 bonus for 5 kinds of triples, which could include 3A if one of the other triples is omitted.) Any jump with an e call or with average GOE lower than -1 doesn't count. Downgraded triples count as doubles for this purpose, downgraded quads count as triples, downgraded doubles count as singles and therefore don't contribute even to the smaller bonus.

But that still leaves the problem, what if a skater does one of a needed triple successfully and then falls on or otherwise significantly fails a second jump in the combination that isn't needed for the bonus? The GOE of the element as a whole will be too low for the needed jump to count.

Also, if two skaters both attempt the same jump content and both have one big mistake, whether the mistake occurs on a repeated jump or a nonrepeated jump will affect whether they get the bonus, even if the content is comparable.

E.g., let's say we're talking about ladies who get a bonus for 5 different triples plus double axel.

Skater A:
3Lz+3T
3F+3T< (fall)
3Lo
3S
3F
2A
2A<+2T+2T

Skater B:
3Lz+3T< (fall)
3F+3T
3Lo
3S
3F
2A<
2A+2T+2T

Under today's rules, the same attempted content with same number and kinds of mistakes results in the same base mark for the jumps, the same 1.0 fall deduction, and let's assume practically the same average GOE across the jump elements. So it will be non-jump scores that determine the winner.

Now let's introduce a bonus for all kinds of takeoffs if performed with -1 GOE or better.

For A, the combination with the too-low GOE that doesn't count toward the bonus happens to include the two repeated jumps, which were each performed successfully in one of the other jump elements, so skater A does get the bonus.

For B, the failed combination was failed because of the 3T, but the 3T was completed successfully elsewhere. However, that was the only instance of 3Lz in the program. The 3Lz was completed successfully. However, because the second jump was failed, the GOE for the element is too low for the lutz to count toward the bonus.

It's a paradox that jump content that would otherwise be considered practically identical under 6.0 or under IJS without the bonus would give several extra points, perhaps as much as the value of an easier triple jump, to one of these skaters and not to the other.

So my opinion is that we just award extra one point for each different kind of triples. So for example if a skater did 3 different triples, the she would get 3 extra points. If 5 different triples, 5 points and so on. I think this would be enough to encourage skaters to do complete set of triples. This will also effectively raise the base value of 3A by one point without sacrificing the virtue of well-balanced programs.

Huh, that's another interesting way to handle it. Might solve some of my remaining problems. But again, how do you determine which jumps count toward the extra points and which ones don't?

[gmyers]

Yeah, I did get the feeling that they were...I can differentiate them only by looking to see if the skater takes off on the leg on the inside of the direction of rotation (Salchow), or the leg outside of the rotation (Loop). Otherwise, superficially, they look quite similar to me. On the other hand, the 3Flip is always easy for me to pick out (and note it's not a Lutz) because the skater almost always "flips" directions. And the Lutz is unique because of its counterrotational mechanics. I think all of these jumps need to remain as separate elements. It's not just a "cop out", shrinking the repertoire to 4 different types of main triple jumps as per your suggestion is still narrowing the forms of skills that skaters could be able to showcase, and I don't know why that is desirable. In fact, the example with the ToeLoop/Walley that you gave sounds regretful for me...do we ever see 3ToeWalleys?


Oops, you're right. My intention was something like Total Score = TES x (1 + PCS/100) or Total Score = TES x (1 + PCS/50), something like that, in some balanced manner so that the PCS matters but is still related to the TES. My first example still gave TES and PCS as independent marks...although yeah, in principle I prefer that to the arbitrary fixed differences in PCS factoring that we currently have.


Yeah, I've crunched the jump base value numbers before...I came to a slightly different conclusion by trying to lessen the value of the 2A. What I ended up doing is inflating the value of the rest of the triple jumps (+0.5 for each triple and +0.8 for the 3A); however, I also asked for factored GoE's (rather than absolute GoE's of +1, +2, etc.)

http://www.goldenskate.com/forum/sho...ht=#post484703


I don't think the reason for the value of the 2A being so high is because it's considered a "baby triple", I think it's in there because it's still a required jump for the ladies (e.g. the SP) and if they miss it, it should be a pretty big deal. Again, however, as a thought experiment, if you want to consider an 84% value increase (I don't know why it has to be done in percentages, but let's run with it) for an additional half-rotation, then allow a 4T to have a base value of 20.2. A little too ridiculous in my book.

37% increase in base value (for optional jumps) between the 3Lz and 3A remains the highest separation for a half-rotation's difference. The difference between a 3A and a 4T is 20%. All this points to be the quads being undervalued, not the 3A.

The 4T is the most commonly done quad but the quad salchow is worth more. 9.8 Vs 10.3. Do you think that is undervalued too?

[steyn]

I don't know if YuNa would get +1GoE for her 3Loop. But besides that, I still don't feel it is enough. I know that YuNa still practices her 3Loop, and I don't believe that she "doesn't have it" like some of her critics say. I think she has it, but maybe it's not as consistent as her other jumps...or maybe she just gets nervous doing it in competition. It's probably the combination of those things, plus, why use precious time and energy to seriously practice it when she doesn't need it? All that trouble for 0.5-1.5 extra points? C'mon, we know YuNa and her team are smarter than that.

I am not saying that team YuNa will go for it but that I think complete triples is worth such amount of points. It's up to her to do the loop in competition or not. We don't adjust the values to push a particular skater to do a specific element.

Yes, I will be happy.

I see. But then just a quesetionable edge call would be worth more than a well executed lutz. I bet that a lot of people would blame the mighty power of technical panel.

[prettykeys]

The 4T is the most commonly done quad but the quad salchow is worth more. 9.8 Vs 10.3. Do you think that is undervalued too?

Actually, yes. However, my post was mainly replying to the notion that the 3A is undervalued, and then the ensuing calculations/comparisons (base values, percentages, etc.) that were used as support; if the same calculations and percentages proposed for increasing the 3A were used for quads, I think we'd enter into the realm of ridiculousness. As it stands now, I think the 3A is just fine, and that the current values for quads are OK, although they could stand to be boosted a little bit. My point being is this: if there is any jump to be considered undervalued, it's the quads.

I am not saying that team YuNa will go for it but that I think complete triples is worth such amount of points. It's up to her to do the loop in competition or not. We don't adjust the values to push a particular skater to do a specific element.

It doesn't have to be 5.0 points for 5 different triples. It could be 3.0 points flat, or 4.0 points flat, if 5.0 is too much. What I don't want, however, is a step-wise approach for doing things half-way, you know? I want a bonus for completion, full versatility.

I see. But then just a quesetionable edge call would be worth more than a well executed lutz. I bet that a lot of people would blame the mighty power of technical panel.

We can fix those in other ways. As someone else in this thread proposed, maybe change the dynamics of the tech panel and how downgrades are determined. Or, skaters could practice so that there is no doubt as to the edging of a jump. e calls are almost never completely unjustified.

ETA: I hope no one thinks I'm picking on or being hard on YuNa at all. Like I said, I am actually convinced she has a decent 3Loop, but she doesn't put it in because the current benefits aren't worth the risks/time/effort to make it solid. What I want is for the system to provide good incentive for her to do it. She's an excellent example to use because clearly she's not lacking in the TES department and wins competitions without the Loop, but from a skating perspective I think it's desirable to see. Also, I believe Joannie did a full set in a 7-triple program, and I feel that should be recognized.

[steyn]

Huh, that's another interesting way to handle it. Might solve some of my remaining problems. But again, how do you determine which jumps count toward the extra points and which ones don't?

Well, I guess it is most practical to award +1 extra point to each different triple as long as it is ratified without downgrade in the protocol. Then, I would reflect the average GOEs to the extra point for each different triple. So if two 3Loop with -1, +2 GOEs respectively earn 11 points with 10 base values in total, then 11/10 = 1.1 extra points is awarded to 3Loop.

Now I think we have two issues to solve: edge calls and combination jumps as shown in your example. For ! mark, I would multiply 0.5 to the corresponding bonus point and 0 for e (i.e., no bonus for e). For a successful combination jump with positive GOE, I apply the GOE to both jumps when computing the bonus points. Then the GOE is double counted but this small added bonus (~ 0.1 point) will be a welcome feature because doing combination jumps is more difficult than doing two separate jumps anyway. For a combination with negative GOE, I would associate 0 GOE with the first jump because doing 3-3 will need at least adquate first jump. Then the entire negative GOE will be for the second jump.

[steyn]

It doesn't have to be 5.0 points for 5 different triples. It could be 3.0 points flat, or 4.0 points flat, if 5.0 is too much. What I don't want, however, is a step-wise approach for doing things half-way, you know? I want a bonus for completion, full versatility.

I understand your viewpoint and we can agree to disagree. But I feel it would be fairer to distinguish, say, 4 triple skaters and 2 triple skaters as well. I don't like the all or nothing appraoch because a minor mistake can potentially change the entire result of a competition.

ETA: I hope no one thinks I'm picking on or being hard on YuNa at all. Like I said, I am actually convinced she has a decent 3Loop, but she doesn't put it in because the current benefits aren't worth the risks/time/effort to make it solid. What I want is for the system to provide good incentive for her to do it. She's an excellent example to use because clearly she's not lacking in the TES department and wins competitions without the Loop, but from a skating perspective I think it's desirable to see. Also, I believe Joannie did a full set in a 7-triple program, and I feel that should be recognized.

Well I think everybody knows that you are a huge fan of Yu-Na

[tarotx]

I think the current rule for Flips and lutz are right. Though after Sochi I would have a mandatory -3 goe for a jumping pass that contains said flutz or lip and for skaters to have a least a 2 season warning that this will occur. Flats would have a mandatory -1. I do think a flutz isn't usually a flip and nor is a lip a lutz but they are borderline and need to be highly discouraged.

I think the point value for 3A's are right. It's the downgrade penalty that's wrong. I want either some kind of mid way credit or for there to be less scrutiny on the rotation. Also I would like for the ladies to be able to replace the required 2A with a 3A in the short.

I would love for combo's to have their own point totals and for second jumps to have a little less scrutiny on their rotation. And also for sequences to have full credit.

I would also love for their to be a bonus for skaters who can do 5 or 6 different types of triple/quad jumps. But only if they get at least base points on the jumps.

I have always considered there to be edge jumps, toe jumps and axel jumps. I think all 3 should be required in the short. Though I wouldn't count a back end combo jump unless it was a triple/quad. I mean as the type of jump representative.

[bekalc]

They could do just one extra point for every triple (done correctly) and then an extra two points if they are 5 different kinds....

[ImaginaryPogue]

Should doing a full set of triples be worth an extra triple?
I don't think so. The more triples you have (and have consistently), the better chance you can create a program with a high base value, you have more flexibility with jump layouts/combos etc. Using the Olympics, Joannie Rochette would have a higher jump value (before GOE) than Yu Na Kim (if she had landed the 2A-2A sequence ), for example. Additionally, essentially a five or six point bonus lessens the importance of GOE, and I don't want to see that.

I still think they should be recognized though, so one way to do that would be using a percentage based bonus, a la the "back end" 10% bonus. Let's assume that the bonus is 10%. That would mean

Triple Toeloop 4.0 = bonus 0.40
Triple Salchow 4.5 = bonus 0.45
Triple Loop 5.0 = bonus 0.50
Triple Flip 5.5 = bonus 0.55
Triple Lutz 6.0 = bonus 0.60
---
Triple Axel 8.2 = bonus 0.82

So a five triple program would have a bonus of 2.5. You're still getting credit for doing the more difficult stuff, but doing it well matters more. If you focus on jumps exclusively, the levels/GOE lost on other elements would be enough to bring you back down a little

So, for worlds 2010

Kim: Lutz, Flip, Toe = 1.55
Asada: Axel, Flip, Toe, Loop = 2.27
Ando: Lutz, salchow, loop, toe = 1.95
Phaneuf: Lutz, salchow, loop, toe = 1.95
Kostner: flip, salchow, loop, toe = 1.90
Lepisto: lutz, toe = 1.00
Suzuki: : flip, Lutz, salchow, loop, toe = 2.50
Marakova: : flip, salchow, loop, toe = 1.90
Flatt: lutz, loop, salchow = 1.55
Helgesson: Loop, toe, salchow, flip = 1.90

Yes, Ando and Phaneuf would leapfrog over Lepisto.