Height on jumps

Cherry blossom said:

They dont? thats such BS. The skater is still exacuting and landing the jump even if its the other way around. such a dumb rule.

Click to expand...

I don't believe it's the JSF that didn't allow this.

Rinks in Japan are very crowded so Satoko learned her jumps in Japan in the other direction due to space and safety. Someone can correct me if I'm wrong, but I believe she started off learning jumps in her natural direction when she started skating in Texas.

Spirals for Miles said:

I think the tiny jumps are because she's naturally a clockwise jumper, but since the JSF doesn't allow this, had to learn counter-clockwise jumps. Thus, they're smaller.

Click to expand...

When did this rule take affect? Alissa Czisny was a Clockwise jumper and I know there have been others.

Yeah the procedure for measuring jump height isn't actually all that hard -- the main limitations are 1) getting good quality video (in this case, high frames per second) and 2) having the patience to sit there and count number of frames many times (i.e. for multiple jumps and for multiple skaters). In theory it can be crowdsourced if people are interested in it because the procedure is fairly straightforward. Incidentally, European broadcasts are sometimes interlaced (when there's a lot of horizontal movement, you'll see a comb-like effect at the boundaries of objects), so it's possible to look at the fields separately, and in effect take what was 25 FPS video and convert it into 50 FPS video. So even if 60 FPS video isn't available, 25 FPS interlaced can work almost as well. Whether or not a video is interlaced is one of the things I look for when doing video analysis because of this.

I don't really count jumps that are in combination coming out of another jump, because those jumps tend to be lower but rotate faster. This is why when I compiled my list of jump heights from the 2014 Olympics ( http://www.goldenskate.com/forum/showthread.php?53491-Measuring-Skater-Jump-Height ) I annotated whether or not the jump was in combination.

One thing in their methodology which I'm not sure about is how they did the counting of the number of frames in the air. For the section of text before the triple lutz, they said (via google translate):

"All this suggests that if the authors' calculations for the picture is reliable, then the frame-by-frame method does slightly understate the real height for all skaters. Here the whole problem is when to start counting. Basically, we counted 2-3 frames with the tooth still (or already) on the ice, depending on the size of the angle between the edge of the ridge and the surface of the ice. In general, our figures for the same figure skaters went a little higher than those of danya17 with FSO, but, most likely, they are still below the real height by ~ 5 cm."

Counting the number of frames should be very unambiguous. It is the number of frames starting at when the skater is first in the air to when the skater first touches the ice. (Or when the skater is last on the ice to when the skater is last in the air, which gives the same numbers.) The reason is because the skater leaves the ice at some point in time between the last frame when she is on the ice and the first frame when she is in the air, and lands at some point in time between the last frame when she is in the air and the first frame when she is back on the ice, and so "first in air" to "first on ice" is the closest approximation. I'm not sure, but it seems like the authors are counting "last on ice" to "first back on ice", which would overstate the air time, and thus lead to inflated (i.e. larger than actual) values. This problem is worse if the videos used were 30 FPS rather than 60 FPS, and indeed, the higher ones seem to have been measured using 30 FPS video instead of 60 FPS. So these numbers may be a bit higher than other sources.

I'm pretty sure I measured rotation speed (revolutions per second) at some point somewhere on these boards, though not as a collected list of stats for different skaters (more as one-off measurements for posts). However, it's very inexact, because people rotate roughly 4-5 revolutions per second, so that's not a lot of frames when counting at 30 FPS or even 60 FPS. My measurement technique is slightly different from theirs though; I look at when the skater pulls into backspin position to when they release to prepare for the landing, and then choose in between those frames for when the skater is at 1 or 1.5 revolutions from a previous frame. This is slightly different than their way of looking at a set number of revolutions in the air, which may be imprecise if the skater is a bit slow at getting into backspin position.

There are other quantities which can be measured from video. One is jump distance, which can be measured if there is a video showing the skater's location on the ice. This can be done via fancam of the skater which encompasses the entire rink (or half of the rink if the centerline is clearly visible). Others include how quickly the skater snaps into backspin position, how quickly the skater springs into the air (useful for toe jumps), and even the amount of rotation -- except this last one takes some time and effort to do, as it combines multiple techniques.

Regarding Carolina Kostner's triple lutz, they say it was taken from the Sochi Olympics.

Is there a method for measuring the distance though?
That would also be interesting.

PS.:I meant if someone has a link to a similar analysis.

champs said:

Maybe I didn't fully comprehend the translation of the article, but the simple free falling formula the author used doesn't take into account that the axis of rotation in the air is not perfectly vertical hence there's more acceleration in the vertical component than just the 'g'.

Click to expand...

That doesn't matter. The skater is assumed to be under ballistic trajectory, with negligible air friction. If the skater is tilted by about the same amount on takeoff and on landing, then it's just 'g'. Doesn't matter if the axis of rotation is not perfectly vertical.

It also does assume that the skater is in roughly the same physical position (i.e. where the limbs are) on takeoff and landing. So if a skater has a very high free leg on takeoff (say for an axel) but has a low free leg on landing (say is still wrapped), then this method can be somewhat inaccurate. But it's still going to be fairly close.

drivingmissdaisy said:

I'd be curious to see how Courtney Hicks measures up to these women.

Click to expand...

2015 Aerial Challenge jump heights (in inches):


2016 Aerial Challenge jump heights (in feet):


If my conversions are correct:

mrrice said:

When did this rule take affect? Alissa Czisny was a Clockwise jumper and I know there have been others.

Click to expand...

The claim was that the Japanese Stating Federation (not the ISU) teaches all their skaters to do it the same way.

Vanshilar said:

It also does assume that the skater is in roughly the same physical position (i.e. where the limbs are) on takeoff and landing. So if a skater has a very high free leg on takeoff (say for an axel) but has a low free leg on landing (say is still wrapped), then this method can be somewhat inaccurate. But it's still going to be fairly close.

Click to expand...

That is an interesting consideration. Basketball players get "hang time" by maneuvering their arms and legs with resoect to their centers of gravity, and seem to hover in the air. I wonder is a skater could get the same optical effect.

Yeah I'm not quite sure what their methodology was. The numbers seem to be abnormally large for some unknown reason. Take Polina Tsurskaya for example, I'm choosing her just because I haven't analyzed her before. The post said the data was from her JGP-16 (SPx2, FS-Tallin). I'm not sure what that means (counting the short program's jump heights twice, maybe? But why? It doesn't make sense to do that at all, all it does is inappropriately decrease the measurement error and make the SP jump heights count twice as much compared to the FS heights for no apparent reason). Well I took the videos from youtube:

https://www.youtube.com/watch?v=axBhQnhuPb4 (short program at Tallinn, 2016 JGP)
https://www.youtube.com/watch?v=bd_jBa2o0qg (free skate at Tallinn, 2016 JGP)

Those videos are at 30 FPS, so they're not ideal for measuring jump height. However, I made two measurements with the videos.

1. The first measurement is the standard one: first frame where the skater's in the air to first frame where the skater's on the ice.
2. The second measurement is the maximum, essentially "rounding up": last frame where the skater's on the ice, to the first frame where the skater's on the ice again.

The measurements are of the jump air time (in seconds). They are then converted to jump height via 980 cm * air time * air time / 8.

Now it's important to note that the second measurement will always overstate the jump height; it basically rounds up in both directions, and so the actual jump height is always less. So it sets an upper bound on the measurement. Also, because I used 30 FPS videos, the second measurement is always bigger than or equal to if the same method was applied to 60 FPS video, because with 60 FPS, the rounding is less. In other words, there should be no reason why a measurement is greater than the second measurement, unless there was fundamental disagreement as to which frame counts as the skater being on the ice versus in the air.

I don't know how to do tables, but the data I got, listed as jump, time in video, standard air time (seconds), max air time (seconds), standard height (cm), max height (cm), and the blog's given height (cm) is shown below:

Short program:
3Lz, 1:00, 0.600, 0.634, 44.1, 49.2, 55.2
3T, 1:01, 0.533, 0.567, 34.8, 39.4, 39.5
3Lo, 2:21, 0.534, 0.567, 34.9, 39.4, 46.0
2A, 2:33, 0.567, 0.600, 39.4, 44.1, 41.8

Long program:
3Lz, 0:45, 0.634, 0.667, 49.2, 54.5, 55.2
3T, 0:46, 0.500, 0.534, 30.6, 34.9, 39.5
2T, 0:48, 0.367, 0.401, 16.5, 19.7, (not given)
3Lz, 2:15, 0.601, 0.634, 44.2, 49.2, 55.2
3T, 2:16, 0.500, 0.534, 30.6, 34.9, 39.5
3F, 2:34, 0.567, 0.601, 39.4, 44.2, 49.2
2T, 2:36, 0.367, 0.400, 16.5, 19.6, (not given)
2A, 2:45, 0.501, 0.534, 30.7, 34.9, 41.8
3S, 2:56, 0.534, 0.567, 34.9, 39.4, 37.1
3Lo, 3:08, 0.534, 0.567, 34.9, 39.4, 46.0
2A, 3:22, 0.534, 0.567, 34.9, 39.4, 41.8

Basically, on all jumps except the 3S, a height was given which exceeded the maximum which could have been measured. Some of the differences are quite significant; for the 3Lo, for example, a height of 46.0 cm would imply an air time of 0.613 s when the measured air time was 0.534 s and the max possible air time was 0.567 s. Now, height is related to the square of the air time, so in this case, a 15% increase over the measured air time (or 8% over the max possible air time) results in the calculated height being overstated by 32% (or 17% over the max possible air time). That's a pretty significant error here.

Not sure why they included Anna Pogorilaya's disastrous FS from WC-17 into the data collection...That would definitely skew the data for her. They picked pretty much the best performances or at least decent ones for everyone else.

Blades of Passion said:

Miyahara's numbers, ooph.

I'm very skeptical about some of the data and the selections they used, however. They stated that for Carolina Kostner, only competitions last season were looked at. Yet, they give a number for her Triple Lutz...she didn't do the Lutz last season. Then for Polina Tsurskaya, they used only her best Long Program and her best Short Program twice.

So basically Tsurskaya was given a "best case scenario" result for her numbers (even specifically skewing it by including a SP twice, where she did a better execution than in her LP) while nobody else was.

Click to expand...

Or, you know, they might have been limited by the quality of the videos they found. I'd expect you don't always find a video that is suitable for this for every competition and sometimes have to make due with what you have.

Additionally, they say FS from Tallin and only 2x JGP SP without Tallin/another location mentioned - my assumption would have been they used the SPs from both her JGPs. More consistent with them saying later that they looked at 4 different 3Lz's of hers as well, since as far as I know, Polya isn't that bad a skater yet that she repeats the 3Lz in her SPs.


Of course data like this isn't 100% fail and fault proof, and there is actually discussion about this in the text and the authors say as much. I don't understand why people can't just take this as the interesting read it is, that must have taken a lot of work, and be fine with what it is. It's not as if this has to be your new order of personal faves or as if the judges will be given this list to adjust their GOE rankings accordingly. It's for fun and interest. And I love these things for that.

well well... i read this... and i thought... wow... this is actually more or less how i perceive most of these skaters... i wasn't surprised with any of the results... perhaps karen chen placed a bit higher than i thought.... but good for her, i will look again.

we see pretty much two types of skaters... those who jump high and to control the jumps will rotate slower, and the ones who don't jump as high but rotate faster...

i prefer the first kind .. but i can see why some prefer the second kind of skaters.

it was a fun read and I appreciated that the authors didn't make any hard claims about this....

people may laugh at this but i appreciate that in moguls skiing, there is a component of the score generated with the speed of the skiers.
I think figure skating could benefit from technology to assess objectively some of the qualities of a jump for instance... let's say that to get a bullet for height or distance, there could be software looking into the jump... not judges... but software... average speed of a skater could also be measured... time of skating on one foot.. all of these things could be evaluated. I am not suggesting that we replace judges with computers, but there could be a couple items where scores would be completely objective. Anyways... just a thought.

This is really interesting and that person must have a lot of free time in his hand and patience to look at all those frames!

Miyahara jumps super low but rotates super fast

Sotskova jumps high but rotates in super slow speed

looks like the the most ideal jumpers are

1. Daleman, Osmond, Tukhtamysheva- great height and good speed

2. Chen, Tsurskaya, - great height and OK speed

3. Zagitova- Ok height and great speed

4. Higuchi, Kostner, Pogo, Med- Ok height and Ok speed

**Sakhanovich seems to have OK height and great speed but still manages to fck up her jumps.

It would be interesting to know the data for Mirai

Poor Satoko Miyahara! Still, I like her a lot and she makes up a lot with her other assets. Height isn't everything, you know. Rotate faster and add some distance while making the jump and you don't need to jump super high. That said, I know Satoko's jumps are often pre- or under-rotated, so it makes sense that she's always in the bottom position in the rankings for every jump.

The list is a bit odd as I would have put Tuktamysheva on top for all her jumps among current skaters, especially her axel and lutz (I don't think Tsurskaya has a triple axel so it seems logical to assume that Tuk jumps higher than her). I would never have thought that Tsurskaya would beat her in jump height. But I'm saying this just based on how it comes across by watching the skaters jump on TV. With Tsurskaya I noticed the delayed rotation and the huge distance she travels over the ice with her lutz, but with Tuktamysheva it's like she has springs in her skates or something, the way she catapults herself up into the air.

When going over the rankings for jump height, Medvedeva doesn't "jump out" either (pun intended). This comes as no surprise, to be honest.

To be fair, Evgenia and Alina backload their jumps, it is almost a necessity to reserve their energy for the epic jump drill. Size is a cost of it. Problem is figure skating is an athletic endeavor. They are already well compensated for endurance following this strategy, but they should be deducted for showing lesser strength, power and even speed compares with others efforts, or others should be rewarded for able to display more of them.

Had they space out their jumps a little to create a balanced program, they can afford to have bigger size jumps. I blame the system though... COP does not reward big jumps as much as an extra tano or backloading.

Elsie said:

Poor Satoko Miyahara! Still, I like her a lot and she makes up a lot with her other assets. Height isn't everything, you know. Rotate faster and add some distance while making the jump and you don't need to jump super high. That said, I know Satoko's jumps are often pre- or under-rotated, so it makes sense that she's always in the bottom position in the rankings for every jump.

The list is a bit odd as I would have put Tuktamysheva on top for all her jumps among current skaters, especially her axel and lutz (I don't think Tsurskaya has a triple axel so it seems logical to assume that Tuk jumps higher than her). I would never have thought that Tsurskaya would beat her in jump height. But I'm saying this just based on how it comes across by watching the skaters jump on TV. With Tsurskaya I noticed the delayed rotation and the huge distance she travels over the ice with her lutz, but with Tuktamysheva it's like she has springs in her skates or something, the way she catapults herself up into the air.

When going over the rankings for jump height, Medvedeva doesn't "jump out" either (pun intended). This comes as no surprise, to be honest.

Click to expand...

Medvedeva is definitely an average jumper aside from her Salchow which is pretty good. During her 1st year senior her axel was actually bad although it seems its getting better but still average.

Tsurskaya's delayed Lutz is a wowzer but can be prone to UR. That delayed rotation can get her great GOE but I think its better to wrap the rotation faster and just invest in the in & out transitions for the GOE.

Just as I thought: Medvedeva's jumps are only average. Consistent, yes, but not impressive in terms of height, speed or distance. I think she rotates faster than average, perhaps.

I still vote for Tuktamysheva as the best jumper in the current ladies field, regardless of what the numbers say. Too bad she's not very consistent (which is probably due to her massive jumps as they are much riskier). I could have sworn that Lutz of hers when she won Worlds was the highest ever (not counting Midori Ito's). Or was it at Europeans?

also speaking of Polina's Lutz, that has reminded me of this video. Pretty good editing as well

I understand that skaters who don't backload have more energy to produce high jumps with great distance. That's why you often see the most difficult jumps in the beginning of the program (quads for men, triple lutz and 3-3 combinations for women). Neither Medvedeva nor Zagitova strike me as high jumpers so it makes sense that they're not higher on the list as they both do most if not all of their jumps in the second half (although it doesn't seem as if they're wasting much energy in the first half anyway, so the energy argument is only half true).

What baffles me is how Polina Tsurskaya ended up with such an exemplary lutz (clear outside egde, great height, speed and distance) whereas Medvedeva, also one of Eteri's students, can't get this jump right no matter how hard she tries, even when she does it in the first half of her program. Didn't Eteri teach them both the same thing then? I don't get it.

Elsie said:

I understand that skaters who don't backload have more energy to produce high jumps with great distance. That's why you often see the most difficult jumps in the beginning of the program (quads for men, triple lutz and 3-3 combinations for women). Neither Medvedeva nor Zagitova strike me as high jumpers so it makes sense that they're not higher on the list as they both do most if not all of their jumps in the second half (although it doesn't seem as if they're wasting much energy in the first half anyway, so the energy argument is only half true).

What baffles me is how Polina Tsurskaya ended up with such an exemplary lutz (clear outside egde, great height, speed and distance) whereas Medvedeva, also one of Eteri's students, can't get this jump right no matter how hard she tries, even when she does it in the first half of her program. Didn't Eteri teach them both the same thing then? I don't get it.

Click to expand...

Tsurkaya got her jumps from a different coach before moving to Sambo 70.

Ender said:

Tsurkaya got her jumps from a different coach before moving to Sambo 70.

Click to expand...

Oh, thanks. I didn't know that. That explains a lot.

On the other hand, one would think that basic jumping technique would and should be the same everywhere. I don't skate myself but it's hard to imagine that the exact same jump could be produced in totally different ways. The only differences I can think of are lack of height which is compensated by rotating faster and jumping further.