GOE Calibration on Quads

[Sjs5572]

Watching the men's sp splatfest has confirmed for me that GOE needs to be recalibrated on quads to effectively reward those well done and penalize those that aren't.

It really is a travesty that a fall on a fully rotated 4s or 4t (even with a -1 fall deduction) is rewarded with .3 more points than a 3z with a +0 GOE (6.3 vs 6.0).

A 1.5 per unit GOE calibration would be more appropriate IMHO. Thereto, a perfectly executed 4s, for example, would accrue 14.8 points (10.3 + 4.5). A fall, however, would now be worth 4.8 points (10.3. - 4.5 - 1).

The risk/reward ratio would now better reflect the importance of quad execution.

Thoughts?

[drivingmissdaisy]

I don't mind that a fall on a quad is worth more than a 3Z, because
(a) a 3Z that gets +0 GOE isn't one that is well done.
(b) it encourages risk taking.
I just feel that if the system encourages putting the difficult jumps in, the men will attempt them more and hopefully master them. I agree with the unit calibration for +GOE would be good, but not for -GOE because I would think the penalty for a fall should be the same regardless of the jump attempted.

[Sjs5572]

I don't mind that a fall on a quad is worth more than a 3Z, because
(a) a 3Z that gets +0 GOE isn't one that is well done.
(b) it encourages risk taking.
I just feel that if the system encourages putting the difficult jumps in, the men will attempt them more and hopefully master them. I agree with the unit calibration for +GOE would be good, but not for -GOE because I would think the penalty for a fall should be the same regardless of the jump attempted.

So, you don't even agree with the current system that has a .3 GOE calibration for doubles, a .7 calibration for triples, and a
1.0 calibration for quads and 3 axels? In essence, the current full deduction for a fall is 1.9 on doubles, 3.1 on triples and 4.0 on quads and 3a.

[Mathman]

A 1.5 per unit GOE calibration would be more appropriate IMHO.

If I remember correctly, it was like that a couple of years ago. GOE was 1.5 on quads and 1.0 on triple jumps. (I believe the original version of GOE was 1.0 on all triples and quads.) The latest changes peg it at 1.0 for quads and 0.7 for triples. I can't remember if this is called the Lysacek rule or the Buttle rule , but people were upset that skaters were winning contests without quads, so the ISU deliberately reduced the risk in order to encourage more attempts.

I agree (with Sjs5572) 100%. Practice your quads on your own time. Put them into your program when you can execute them properly. You already get a huge reward for a properly done quad.

(b) it encourages risk taking.

I don't agree. It takes away the risk. Where is the risk, if you get a gob of points whether you succeed or fail?

I am not a fan of any scoring system that rewards and encourages badly executed elements.

[drivingmissdaisy]

I don't agree. It takes away the risk. Where is the risk, if you get a gob of points whether you succeed or fail?

I am not a fan of any scoring system that rewards and encourages badly executed elements.

There is absolutely risk. By doing a quad instead of say, a 3Z, you are risking the loss of +GOE on a well executed triple. A good 3Z would score better than a fall on a 4T. You expend more energy on a quad than a 3Z, whether you fall or land it, so there is the risk of being more tired at the end of your program. There is also the risk of a pop or UR, which would score really low vs a triple, and one is more likely to pop a more difficult jump that they haven't mastered. So there is plenty of incentive to master elements before you put them in a program.

[Mathman]

^Let’s put some numbers up.

Suppose you can rotate your quad but have only a 10% chance of landing it. But you have a 90% chance chance of landing your Lutz. Let’s say you typically get +1 GOE when you land your quad and +1 GOE when you land your Lutz. Ten per cent of the time on your Lutz you get an edge call, a UR, or some other negative feature so that you average say, 4.00 on the rare occasions where your Lutz is unsuccessful.

Expected (average) value of quad attempt, taking into account both successes and failures: .10x11.30 + .90x6.30 = 7.80

Expected value of triple Lutz attempt, averaged over all attempts: .90x6.70 + .10x 4.00 = 6.43.

The current rules say it is better to attempt the quad (if you can get the rotations) even if your success rate in holding the landing is only 10%.

Yes, you might get tired, etc. Still, something seems out of balance here.

[Sjs5572]

ITA, Mathman (btw, I was originally a math major in college, so we seem to look at COP through the same prism). Furthermore, the reward for poorly executed quads is even more extreme when compared to 3sal and 3t.

How can we expect casual fans to remain interested when flopfests can win bronze, and in the case of Chan, gold medals? Figure skaters should produce complete programs; not pseudo xtreme sports half-pipe-like events where participants throw caution to the wind and hope something sticks.

[drivingmissdaisy]

^Let’s put some numbers up.

Suppose you can rotate your quad but have only a 10% chance of landing it. But you have a 90% chance chance of landing your Lutz. Let’s say you typically get +1 GOE when you land your quad and +1 GOE when you land your Lutz. Ten per cent of the time on your Lutz you get an edge call, a UR, or some other negative feature so that you average say, 4.00 on the rare occasions where your Lutz is unsuccessful.

Expected (average) value of quad attempt, taking into account both successes and failures: .10x11.30 + .90x6.30 = 7.80

Expected value of triple Lutz attempt, averaged over all attempts: .90x6.70 + .10x 4.00 = 6.43.

The current rules say it is better to attempt the quad (if you can get the rotations) even if your success rate in holding the landing is only 10%.

Yes, you might get tired, etc. Still, something seems out of balance here.

I don't think it's as simple as the math you put up. First, 100% of quad failures are not going to get full credit for rotation, so downgrades have to be taken into account in the .90x6.30 figure. Second, A person who merely lands a jump 10% of the time likely isn't going to average +1 GOE on the attempts he does land, so the .10x11.30 should be adjusted down as well. These two factors alone would adjust your 6.8 figure below the 6.43 figure for the lutz.

I like that people are trying the hard things because it makes the moments when they put it all together so special. I love Carolina's programs but even when she's clean I don't think "Wow, she really upped the ante for ladies skating" like I did when Yuna won the Olympics or Midori won Worlds in 1989. It's a sport and skaters should aspire to do things other than skate beautifully.

[drivingmissdaisy]

How can we expect casual fans to remain interested when flopfests can win bronze, and in the case of Chan, gold medals? Figure skaters should produce complete programs; not pseudo xtreme sports half-pipe-like events where participants throw caution to the wind and hope something sticks.

It's the needless transitions and spastic footwork that dampens the interest of casual fans. Women are falling on jumps that all the top ladies have done since 1991, and the men since 2002; I (and I think some casual fans) would prefer to see jump content that surpasses what one might see in an exhibition.

[gkelly]

I don't think it's as simple as the math you put up. First, 100% of quad failures are not going to get full credit for rotation, so downgrades have to be taken into account in the .90x6.30 figure. Second, A person who merely lands a jump 10% of the time likely isn't going to average +1 GOE on the attempts he does land, so the .10x11.30 should be adjusted down as well. These two factors alone would adjust your 6.8 figure below the 6.43 figure for the lutz.

Very true.


I'd like to see the minus GOEs for quads go back to 1.5 for each minus as was the case a few years ago. I think that would even things out reasonably well.
I don't have strong feelings about whether the plus GOEs should stay at 1.0 or be raised. I think 1.0 is fine but I'm willing to be convinced.

Perhaps Mathman would be happy if the base value were lowered a bit but the plus GOEs were raised so only good quads would get the really big rewards. But I suspect that the guys who are thrilled to land a fully rotated quad at all would want that feat to be rewarded well.

I like that people are trying the hard things because it makes the moments when they put it all together so special. I love Carolina's programs but even when she's clean I don't think "Wow, she really upped the ante for ladies skating" like I did when Yuna won the Olympics or Midori won Worlds in 1989. It's a sport and skaters should aspire to do things other than skate beautifully.

Are you talking about the way Kostner is skating this year, or her career as a whole? As a teenager she was one of the first skaters to do both 3Lz+3T and 3F+3T in the same long program and one of the first to do those combinations in a short program. Later, wasn't she the first woman to earn level 4 for a step sequence? So she did contribute to raising the standards for the field as a whole -- and now she has to compete against others who can live up to that standard, so good for her if she can earn more points in areas other than jump content.

[Mathman]

Perhaps Mathman would be happy if the base value were lowered a bit but the plus GOEs were raised so only good quads would get the really big rewards. But I suspect that the guys who are thrilled to land a fully rotated quad at all would want that feat to be rewarded well.

I would be happier to raise the base value (to 13.5 for a quad toe), but not to award full base value unless the jump is landed.

I, too, am thrilled when a skater lands a fully rotated quad, and I would want it to be rewarded well.

A little bit off-topic, but here is another peculiarity that came up in the men's short program. (Imaginary Pogue and SkateFiguring have commented on this in other threads.)

Chan intended 4T+3T, then 3Lz for his solo jump. His landing on the 4T was not strong enough to do the combo, so he tacked a 3T onto his Lutz instead. Big points for both passes, with only a small negative GOE on the quad and big positive GOE on the Lutz combo.

Takahashi went do or die on the 4T+3T. He did the 4T OK, but badly underrotated the 3T and almost fell. He ended up with quite a bit less for the combo (8.74) than he would have got for just the 4T alone (10.3). This is because the second jump was downgraded to a double, but he got negative GOE at the rate of the higher jump.

If Takahashi had done 3Lz+3T, even if the 3T had been downgraded with all -3 GOES, he still would have come out ahead because -3 GOE is only -2.1 points for a Lutz combo.

There is no incentive at all, and in fact a negative incentive, for any skater to attempt 4T+3T, 3Lz, rather than the safer 4T, 3L+3T.

[ImaginaryPogue]

Except that for many skaters, a quad out of steps is challenging. No steps before the solo jump, by definition, is automatic -3.

[Mathman]

I would be happier to raise the base value (to 13.5 for a quad toe), but not to award full base value unless the jump is landed.

Maybe it could be done in the same way that they do underroations. 0% of base value for a total belly flop, 25% for a typical fall, 50% for saving yourself with hands down, 75% for staggering around trying to maintain your balance, and 100% if you actually land it. (-GOE if you come to a complete stop, step out, no flowing edge, etc.)

Except that for many skaters, a quad out of steps is challenging. No steps before the solo jump, by definition, is automatic -3.

Hmm. Chan didn't get any -3's for his solo 4T in the Worlds SP. Am I wrong in thinking that he intended the quad to be his combo and changed his mind during the program?

[drivingmissdaisy]

Hmm. Chan didn't get any -3's for his solo 4T in the Worlds SP. Am I wrong in thinking that he intended the quad to be his combo and changed his mind during the program?

I think "steps before the jump" is interpreted liberally. I think even doing the turn into a toe jump would qualify under the rules

[Sjs5572]

Mathman, I don't think you are wrong about Chan's intend layout. He did a 4t3t at Canadian nationals. He only did a 4t at Four Continents, but he fell on that jump.