Measuring Skater Jump Height

[hurrah]

You forgot these numbers are only the height. Ice coverage and distance of the jumps count as well...

Well, the data here doesn't allow for a comparison of distance, but I've seen Mao's jumps live versus other skaters jumps, and it did not look like her jump distance was noticably smaller than other skaters. Just look at the photo under my poster name for proof that Mao's triple-axel jump does cover distance as well.

In any case, the present data proves that Mao has the highest average jumps out of recent ladies skaters, and no one can say for sure whose jump distance is greater. As of now, it could possibly be Mao just as much as it could be Caro or Yuna.

 

[Meoima]

Well, the data here doesn't allow for a comparison of distance, but I've seen Mao's jumps live versus other skaters jumps, and it did not look like her jump distance was noticably smaller than other skaters. Just look at the photo under my poster name for proof that Mao's triple-axel jump does cover distance as well.
In any case, the present data proves that Mao has the highest average jumps out of recent ladies skaters, and no one can say for sure whose jump distance is greater. As of now, it could possibly be Mao just as much as it could be Caro or Yuna.

It's interesting... I wonder about Mao's UR problem. She rorates quite fast, still some of her jumps got ding. Maybe because she delayed the rotation?
On video it's hard to tell about the distance of the jumps. So true.

 

[taltaltel]

Well, after all the years and years of unending talk about Yuna's and Caro's huge jump versus Mao's small jumps, it turns out that Mao's triple-axel jump is the queen of all jumps, and she has the highest average jump height out of everyone!

Yuna's 3ltz+3f+3sal+2axel jump height average is 35.75
Caro's 3ltz+3f+3sal+3loop+2axel jump height average is 36.4
Mao's 3ltz+3f+3sal+3loop+2axel jump height average is 37
Mao's 3axel+3ltz+3f+3sal+3loop+2axel jump height average is 40.3

Seems like you just took the peak value of each of those jumps to come up with the averages, and also for some incredibly strange reason neglected the triple toe. Based on what they actually presented in the programs and how the judges would have judged cumulatively, taking all jumps with a BV greater or equal to 2A,

Yuna Kim 3Lz + 3T + 3F + 3S + 3Lz + 2A + 3S + 2A: average of 35.125 cm
Carolina Kostner 3Lz + 2A + 3T + 3F + 3Lo + 3T + 3S + 3S: average of 34.5 cm
Mao Asada 3A + 3F + 3Lo + 3Lz + 2A + 3T + 3S + 3F + 3Lo: average of 37.333 cm (without the triple axel, 34.875 cm)

To make this more interesting,
Adelina Sotnikova 3Lz + 3T + 3F + 3Lo + 2A + 3T + 3F + 3S + 2A: average of 36.111 cm

So yes, Mao's triple axel is the queen of all jumps, that is undisputed. That alone accounts for her higher average. If we take that one jump out of the equation and focus on the standard repertoire (which would be unfair for Mao, but just for comparison's sake), Adelina had the highest jumps of the night in the Sochi FS, followed by Yuna, then Mao, then Carolina.

 

[hurrah]

I took out the triple toe because I did not consider it reliable to compare a triple toe done as a single or first jump to one done as a second jump in a combination.

Adelina did jump very high that night, though, for sure. I guess the Russian air suited her.

 

[taltaltel]

I took out the triple toe because I did not consider it reliable to compare a triple toe done as a single or first jump to one done as a second jump in a combination.

I agree that you would normally expect the second jump to be lower (actually not necessarily true either, for example, Julia's 3T in all her 2A-3T combos is higher than the initial 2A). But that comparison is still highly relevant - the OP mentioned how jump height is intricately related to the GOE given to the jumping pass, so yes, both jumps in the combo are important. So I don't really see the rationale of deliberately leaving the 3T out of the equation. I will also maintain that you can't take the maximal jump height because we are comparing jump heights in programs, not just as isolated jumps.

 

[hurrah]

I was using the data to make the point that the years and years of constant, relentless talk on skating boards that suggested that Mao's jumps were smaller was not based on facts, and the fact of the matter is, the data shows tha Mao' jumps are not at all small.

I don't really care about the GOEs of that night. As I said, the Russian air that night probably suited Adelina really well.

 

[taltaltel]

I was using the data to make the point that the years and years of constant, relentless talk on skating boards that suggested that Mao's jumps were smaller was not based on facts, and the fact of the matter is, the data shows tha Mao' jumps are not at all small.

Indeed they are not small. The triple axel in particular was huge. I think it's interesting to note, however, that both triple jumps that she put as the second jump in combination (3Lo in the 3F-3Lo, 3T in the 2A-3T) was significantly lower than any of her other triple jumps (both at 28 cm). Could explain the UR call she got on both jumps.

I get that you are frustrated with the criticisms Mao's been getting over the years. This is not intended to bash her at all though. :confused2:

 

[CarneAsada]

Midori Ito too

Yeah can we do a comparison of all ladies' 3Axels? I'm sure Midori and Tonya are much higher than anyone else's but I am curious about where Nakano (one of her clean 3Axels) and Nelidina fit in. Maybe do Miki's quad, too.

I imagine the frame rate might be different but I have no idea about FPS of video from 20 years ago. Direct comparisons might be a bit harder because of less accuracy in older video.

I was using the data to make the point that the years and years of constant, relentless talk on skating boards that suggested that Mao's jumps were smaller was not based on facts, and the fact of the matter is, the data shows tha Mao' jumps are not at all small.

I didn't want to say it myself but it definitely needed to be said. Finally.

 

[Vanshilar]

I posted Olympic Ladies Free Skate results for the top 12 scoring skaters in the free skate. Let me know if you think I should bother doing them for (any of) the bottom 12 skaters, or if I should move on to the Olympic Short Program.

Oh, right. I forgot about using fancam video. That definitely is possible, though I do wonder about how much the camera angle might complicate things, as well as the size variation between rinks (aren't Asian rinks often a bit bigger?). For speed into the jump, would you be looking more for the glide speed going in or the speed retained up to takeoff? On some jumps a lot of speed is lost during the jump entry.

I account for the camera angle, using some perspective geometry math. Since my focus was/is on Olympics, I assume an Olympic-sized (heh) rink of 30 m by 60 m. The rink dimensions would need to be verified for each venue, since they are used to calibrate the x-y coordinates. For speed of the jump, I could look at the glide speed into the jump (i.e. during the gliding part prior to the takeoff for each jump) and the horizontal speed of the jump itself (measured via x-y coordinates and the jump duration).

PS: Vanshilar, can you help us with the short program? I think jumps in short programs might be different to jumps in the long. Thank you.

I took a look at the GPF short programs, and I'll post the results when I get a chance but basically the camerawork was atrocious in terms of figuring out jump duration -- there were things like the camera being panned out during some jumps (so I couldn't clearly see when the skater took off and landed), the feed switching cameras mid-jump (so I'm left wondering if I want to trust that the timing between cameras was properly synced up), not having the skates in the field of view (to see when the skater left the ice and hit the ice), etc. The videos were all natively 50 FPS which why I chose them to download, but the quality for jump duration turned out to be not so good. (Obviously they're not capturing the video just so somebody can measure jump duration though.) So yeah basically the results won't be as reliable.

It think it would be interesting to see where several "big" jumpers fit in there - Surya Bonaly and Tonya Harding are two possibilities.

Um, sure. If you know of any good Youtube videos that show those jumps clearly, I can download them and take a look.

In any case, the present data proves that Mao has the highest average jumps out of recent ladies skaters, and no one can say for sure whose jump distance is greater. As of now, it could possibly be Mao just as much as it could be Caro or Yuna.

Uh, let's not get ahead of ourselves here. Keep in mind the precision in the jump measurements can be up to 4-6 cm, so differences of fractions of a cm or even up to 1-2 cm may not mean much. Also there's other things such as if some jumps naturally tend to be higher than others and also happened to be preferred by some skaters, if those jumps were done earlier or later in the program, etc. I'm not saying there's no differences between skaters in terms of jumping height, but that it's a bit early to be drawing too strong of conclusions.

It's interesting... I wonder about Mao's UR problem. She rorates quite fast, still some of her jumps got ding. Maybe because she delayed the rotation?
On video it's hard to tell about the distance of the jumps. So true.

Actually when I get around to rotations I'll be discussing some of those issues in more depth. For example, there's a difference between being under-rotated within the rules (i.e. landing with more than 90 degrees off from backwards of the direction of flight) and what judges would actually mark as being under-rotated. For example, Mao's Olympic FS 3A was definitely significantly under-rotated (i.e. probably something like 150 degrees under-rotated when only 90 degrees is allowed under the rules) yet wasn't marked with a UR. So there's other considerations that come into play (ability of judges to determine UR, ability of skater to recover or "cover up" the aftereffects of UR in the landing glide, whether or not judges give more leeway in different situations, etc.) that I think should be discussed, but are often ignored in internet partisans that freeze-frame and then cry "UR!" for skaters they don't like. But anyways, yeah, the jump (horizontal) distance and rotation will eventually be collected for different skaters as well. (I guess jump distance is easier to measure than rotation so will probably be posted first. Measuring rotation will require knowing the distance, while the reverse is not true for how I measure them. Discussing jump rotation will also inevitably be contentious.)

 

[Meoima]

Vanshilar, I have just found this documentary by Japanese TV which might be useful to you, but what a pity there is no English translation. They talked about Yuzuru's SP jumps at Sochi. There's a part they say Yuzuru's quad in his short program had 3.72m ice coverage or something like that.
They also discuss his speed as well.
Part 1: http://www.dailymotion.com/video/x2do2d1_データ-前半_sport
Part 2: http://www.dailymotion.com/video/x2dnf8h_データ-後半_sport

It seems Yuzuru and Plushy have technically the same rotation speed, which Plushy is a bit faster, about 0.008 sec.

They also analyzed some jumps. But only in Sochi.

I think we might ask some Japanese poster to help, this documentary is very interesting.

 

[Vanshilar]

I just posted some stats for Midori Ito's jumps. If I weren't looking at the videos frame-by-frame myself, I wouldn't have believed them. Her triple axel has a duration of about 0.7 s or 60 cm, which is basically the duration (or height) of men's quads. She has pretty good height on the other jumps as well. Sadly, the measurement errors on them are fairly large (25 or 30 FPS videos).

There are two videos of Ito that I looked at but didn't post. The first is of her 1992 Olympics FS here: https://www.youtube.com/watch?v=polwvMNVgFU . She fell on her first triple axel attempt, but tried another one over 3 minutes into her program...and lands it. Looking at this second one, it seems like it should be 18 frames long at 25 FPS (i.e. 0.72 s). Over 3 minutes into the program. Sadly, the video has a lot of "repeat" frames so the measurement error is multiple frames, so I decided not to use it. Also, there's a clip of her doing a triple axel - triple toe combination here: https://www.youtube.com/watch?v=QQgR0cBsK4Y where it looks like the triple axel was 22 frames long at 30 FPS (i.e. 0.73 s), followed by a triple toe that was 16 frames long (0.53 s). (The second combo is a 3A-2T, and the 3A was 21 frames long or 0.70 s.) So the measurements are consistently at around the 0.7 s mark.

One thing I noticed about her jumping technique though is that she never really gets into a good spin position in the air. That is, she never does a really good scratch position. Instead, her left skate goes up to her right knee and then gradually moves downward toward her right skate, and there's always a gap between her knees. Contrast this with Mao Asada's technique where she's pencil-straight with the knees together for the last two revolutions or so of the spin. Thus Ito really relies on her height for her triple axel. It's not as if she can't get into a good scratch position quickly either; she gets into a pretty good position for her triple toe loops pretty quickly.

Vanshilar, I have just found this documentary by Japanese TV which might be useful to you, but what a pity there is no English translation. They talked about Yuzuru's SP jumps at Sochi. There's a part they say Yuzuru's quad in his short program had 3.72m ice coverage or something like that.
They also discuss his speed as well.
Part 1: http://www.dailymotion.com/video/x2do2d1_データ-前半_sport
Part 2: http://www.dailymotion.com/video/x2dnf8h_データ-後半_sport

It seems Yuzuru and Plushy have technically the same rotation speed, which Plushy is a bit faster, about 0.008 sec.

They also analyzed some jumps. But only in Sochi.

I think we might ask some Japanese poster to help, this documentary is very interesting.

Thanks, unfortunately I don't understand Japanese. It looks like they use the skater's height as a ruler to measure the horizontal distance. However, this doesn't really take into account that, relative to the camera, the skater is not only moving from side to side, but also forward and backward (i.e. away or toward the camera). Measuring the distance in that direction is a lot more difficult since it depends on things like how much the camera is zooming and such. But I'll assume their measurements are fairly accurate. Yeah obviously they use fancy software to do stuff like show multiple images of a skater during the jump; I'm pretty sure people have done some pretty sophisticated analysis in this area (I wouldn't be surprised if most of the elite-level skaters and coaches are pretty familiar with seeing their jumps in slow motion, if nothing else so that the coaches can point out the flaws in the skaters' techniques).

 

[yyyskate]

finally, read this thread, that was amazing work. One good way to judge under-rotation, is to look at the ice tracing, kinda hard if you only have broadcast footage,if by chance, you got the photos of skater's ice tracing, or a good close-up, then you can draw a tangent line at both the curve of take-off edge and landing edge, and measure what is the degree of the angle between these 2 tangent lines.
I intend to do some examples here, but kinda lousy at edit photos, I will try.
think about this method, it is kinda hard to determine pre-rotation...

 

[yyyskate]

Vanshilar maybe you can also do Patrick and SoYoun's jump height too. SoYoun seems got very big jumps. and Patrick's data perhaps is very useful when compare to Yuzuru to study the efficiency/economy of jumps.
also from your data, Yuna's both lutz are of equal height, therefore her tight landing in her second lutz is indeed caused by a slight tilt of body axis in the air. Air position could be a factor that affect jump landings as well. maybe by study that, we can find out how skaters without perfect air positions managed to land their jumps.
also can you do Mirai's jump height as well, and figure out, what really is the cause of her under-rotation, height? or others.

 

[dorispulaski]

Vanilishar, I would love to see how large Tonya Harding's triple lutz and flip from her 1994 Olympic LP were, and her 3a from Skate America 1992-1993 season were.

This is very interesting work to me!

 

[giulia95]

Vanilishar thanks for your amazing work!!
it's sooo interesting!

 

[Vanshilar]

I've updated the thread with results for some of the Olympics ladies SP. I've only gotten through the first 9 so far (in order of how they scored in the free skate). Overall the video was pretty good quality since it was 25 FPS interlaced, so I was able to extract the individual frames via Avidemux and effectively view it at 50 FPS (albeit at half the resolution). The replays, when usable, were often at 1/3 the speed so those jumps were analyzed at 150 FPS which is, needless to say, very good.

The jump layouts in the different SP's were fairly similar to each other (probably due to requirements), so they allow for a better direct comparison between the different skaters.

finally, read this thread, that was amazing work. One good way to judge under-rotation, is to look at the ice tracing, kinda hard if you only have broadcast footage,if by chance, you got the photos of skater's ice tracing, or a good close-up, then you can draw a tangent line at both the curve of take-off edge and landing edge, and measure what is the degree of the angle between these 2 tangent lines.
I intend to do some examples here, but kinda lousy at edit photos, I will try.
think about this method, it is kinda hard to determine pre-rotation...

Well when I get around to rotation it will be by directly looking at the skates in the video. But yes, the ice tracing can give some pretty good measurements, when available. However, there's some stipulations to this:

1. According to these forums, the way to measure pre-rotation and under-rotation is by the skate position at the time the skate leaves the ice and hits the ice. Usually, on takeoff a skater will rise up onto the toe pick, and then the skate will continue rotating on the toe pick before it leaves the ice. Similarly, on landing usually the skater will land on the toe pick first, then continue rotating until the edge touches the ice, at which point the skater begins traveling to generate the tracing. Thus, the tracing only measures the angle of the edge when it was moving, not the actual position of the skate at the moment of takeoff and landing.
2. One way to use the tracing to see an under-rotation is to look for a "hook" or really small curve at the landing tracing. However, this only shows that there was definitely an under-rotation; if there isn't one, you still don't really know if the jump was under-rotated or not.
3. For an edge jump (salchow, loop, axel), it's fairly easy to assume the flight path is from the end of the takeoff tracing to the beginning of the landing tracing. However, I'm not sure what the flight path should be assumed to be for a toe jump, however, because there's the edge tracing as well as a divot for the toe pick. Should the end of the edge tracing, the toe pick divot, or the midpoint between them be considered as the starting point for the flight path? This is actually a general question for when I get around to measuring jump rotations.

Vanshilar maybe you can also do Patrick and SoYoun's jump height too. SoYoun seems got very big jumps. and Patrick's data perhaps is very useful when compare to Yuzuru to study the efficiency/economy of jumps.
also from your data, Yuna's both lutz are of equal height, therefore her tight landing in her second lutz is indeed caused by a slight tilt of body axis in the air. Air position could be a factor that affect jump landings as well. maybe by study that, we can find out how skaters without perfect air positions managed to land their jumps.
also can you do Mirai's jump height as well, and figure out, what really is the cause of her under-rotation, height? or others.

I'll get around to SoYoun's jumps in a bit. However, what do you recommend for Patrick's and Yuzuru's data? I'm wary of using the Olympics for that because it was such a splatfest so I don't think it's as representative of their abilities. Is there another event where the video is readily available (and hopefully good quality) for them? Similarly, what are some good events for Mirai?

I should note that jump height is only one component of a successful jump. There's also rotational (angular) momentum, or how much rotation or spin energy (actually, momentum) the skater can generate when taking off; moment of inertia, or how tightly the skater can straighten their body in the air to spin more quickly; snap time, or how long it takes the skater to transition from their takeoff position into their back scratch position in the air; and other factors like the amount of tilt (both perpendicular to the flight path and in the direction of the flight path) in the axis of rotation as they spin in the air. So I don't think it will just be an easy one-solution thing; her coaches most likely have done a lot more in-depth analysis of her jump technique and would have fixed it already if it were so easy.

 

[yyyskate]

yes it is kinda difficult to measure pre-rotation in one scenario, that is when the skater lifted his entire blade off except for the toe pick, but pivot too much (more than it suppose to be) on the toe-pick, then it is kind hard to determine the extent of pre-rotate, if only depend on ice-tracing. but bear in mind though, ice tracing is the ultimate reflection of ice/blade relationship. it tells lots of information, but one needs really detailed photo to debacle that (I took some myself, will try to analyze a few if I have time..). I also think analyze rotation using footage also is sufficient, could easily figure out when skate lift off ice/in air/landing, also easily shows skaters' body, but some detailed tricks may not as telling as ice tracing. (e.g. by studying take-off ice tracing, I found out that Yuna and Liza's lutz jump technique is somewhat different, both fully-rotated of course)
for toe jump, lutz, flip, and toe only looked at the edge of jumping leg not the toe-pick leg.
Patrick could use his cleaning programs at 2013 TEB, patrick and Yuzuru dont have an event together when both did relatively well. if this does not matter (I assume it should not because you are using frames to measure height ) Yuzuru dont have an event when he is absolutely clean, 2014 worlds LP is clean, but jump quality is not his best. I think Japan national is his best jump so far(except for 4S fall, but could use 4S in GPF), 2012 worlds they both did relatively well(yuzuru jump all clean, Patrick fall on 2A, but big jumps all good) Maybe you could download footage from youtube, usually great skates from skaters have high quality videos there. while trying to find good skates for you, I re-watched Patrick's a few programs, Gosh, her is blade work, really nobody can touch him

 

[yyyskate]

yes, all you said about Mirai and jump physics in general is true. I just suggest that some Mirai jump height data would be good to know compare to others or herself.
Mirai's under-rotation problem is most problematic recently, so some of her recent "clean" program would be a good example. this year's SA or COR lp are both good examples (actually, could do both, since Mirai seems think that she did better in COR than SA, would be interesting)

 

[yyyskate]

https://www.youtube.com/watch?v=3rJX0_QqCTE
mirai at 2010 national... what a talent
mirai at 2013 national...
https://www.youtube.com/watch?v=GvfwxRMCLCw
wish you good luck this year, Mirai.

 

[Interspectator]

If, and only IF you have time, I'd like to see what height Mura gets on his Jumps. They seem HUGE.
His best performance was Skate Canada this year. :thumbsup: