Analysis on element "efficiency"

[Alex Fedorov]

Of course, I know that the women's tournament at the Grand Prix has already been discussed, especially since I myself participated a bit in this. But this discussion was more emotional than constructive. In any case, I do not intend to talk here about the final, which can hardly be called anything other than a disaster. The question is different - was this catastrophe an ordinary accident, or did some negative processes naturally lead to it? I will try to answer this question by analyzing the protocols of performances of skaters at all six qualifying stages of the Grand Prix of this year and similar protocols for last year. Of course, this set of data will not be enough for a full-fledged statistical study, but what is available may give rise to certain reflections.

A literal comparison of the scores obtained by skaters at one or another stage of the Grand Prix is probably the most dubious way to determine whether the level of competition has increased or decreased during the year. In addition, before the start of the current season, the rules for scoring have changed. Therefore, I propose to use a different, hypothetical, criterion - the effectiveness of performing of the main elements of skating.

By efficiency, I mean the percentage of successful attempts to complete an element. In this case, the attempt is considered successful if the skater received a score higher than the base value of the element, or at least a score equal to the base value. It doesn't always mean perfect performance. For example, at the Grand Prix stage in Finland, the protocol contains the following lines:

“3Lz+3Tq q 11.11 x 0.17” (free skate Mai Mihara)

“3Lz!+3T ! 10.10 0.08” (free skate Mana Kawabe)

I will consider both of these attempts successful, despite the marks of the judges, because in both cases, points were obtained that exceeded the base value of the cascade. At the same time, if the protocol, for example, says “2Lz” where the skater planned to perform 3Lz, then I will consider such an attempt unsuccessful even with a positive GOE (such cases sometimes happen), since the score will still be lower than the base value of 3Lz .

Sources:
http://www.isuresults.com/results/season2122/gpusa2021/
http://www.isuresults.com/results/season2122/gpcan2021/
http://www.isuresults.com/results/season2122/gpita2021/
http://www.isuresults.com/results/season2122/gpjpn2021/
http://www.isuresults.com/results/season2122/gpfra2021/
http://www.isuresults.com/results/season2122/gprus2021/
http://www.isuresults.com/results/season2223/gpusa2022/
http://www.isuresults.com/results/season2223/gpcan2022/
http://www.isuresults.com/results/season2223/gpfra2022/
http://www.isuresults.com/results/season2223/gpgbr2022/
http://www.isuresults.com/results/season2223/gpjpn2022/
http://www.isuresults.com/results/season2223/gpfin2022/

I will start the analysis with the effectiveness of solo triple jumps. It is clear that any skater taking part in the Grand Prix must be able to perform these elements - otherwise, he will doom himself to defeat even before the start of the competition.

The statistics for each of the triple jump attempts in 2021 (with the exception of the triple Axel, which, like other elements of ultra-c, will not be counted here) can be shown in the form of the following table:

Table J2021

The number of successful attempts is given in parentheses. Each of the lines corresponds to one of the stages of the Grand Prix (chronological order is observed). The last two lines sum up - the total number of attempts, in parentheses - the number of successful attempts, as well as the efficiency, expressed as a percentage.

A similar table for the 2022 Grand Prix looks like this:

Table J2022

The difference between 2021 and 2022 is pretty obvious and hard to miss. And the first thing that catches your eye is the sharply reduced efficiency of 3F execution. However, the performance of 3Lz also decreased markedly.

 

[Alex Fedorov]

Of course, such a direct comparison may seem biased to someone, since in 2021 9 Russian figure skaters competed at the Grand Prix stages. Anticipating such objections, I prepared the third table, which displays the statistics of the performance of solo triple jumps in 2021 by non-Russian skaters:
Table J2021WR

When comparing tables J2021WR and J2022, it is easy to see that although the gap between the efficiency values in 2021 and 2022 has slightly decreased, it still remains very significant. At the same time, the decrease in the efficiency of 3F execution is still most noticeable. In addition, the number of attempts to execute the simplest and cheapest element, 3T, has tripled.

Here the question arises - what caused this phenomenon? Perhaps a significant change in the composition of the skaters? Such an assumption at the first moment looks logical, because in addition to the nine skaters from Russia absent this season, many other high-ranking athletes did not participate in the Grand Prix this time - for example, Wakaba Higuchi or Alice Liu. To test this hypothesis, I compiled a list of twenty skaters who took part in the Grand Prix both in the 2021 season and this season. Based on this list, I made another table that displays the statistics on the performance of the main solo jumps in 2021 and 2022 by those who went on the ice two seasons in a row.


Table JN20212022

As you can see, the efficiency of performing 3Lz by such athletes in 2022 not only did not decrease, but even slightly increased. Thus, the drop in this indicator in the J2022 table can indeed be attributed to the appearance of “newbies” (newbies, of course, are conditional, most of these skaters have extensive competitive experience, they just weren’t at the Grand Prix in 2021). But the sharp decrease in the efficiency of 3F has not disappeared, this mysterious phenomenon is fully preserved even when using such sample statistics.

 

[Alex Fedorov]

Next, I will try to compare the effectiveness of performing 3 + 3 cascades in 2021 and 2022. To do this, I made another table:
Table C2021-22

Table C2021-22 actually consists of three parts: combination performance statistics in 2021, combination performance statistics in 2021 excluding the contribution of Russian skaters, and combination combination performance statistics in 2022.

There is no 3Lz+3Lo cascade in this table, because in 2021 this element was performed only by Anna Shcherbakova (two successful attempts out of three). In 2022, Isabeau Levito was the only performer of the 3Lz + 3Lo combination (one successful attempt out of two). In this case, it is hardly possible to draw any conclusions.

Analyzing the C2021-22 table, it should be immediately noted that in 2022 the efficiency of performing the key combinations 3Lz + 3T and 3F + 3T increased (excluding Russian skaters in 2021). But this is largely an illusory improvement. In 2021, 42 figure skaters performed at six stages of the Grand Prix, of which 9 were Russian. In 2022, 45 figure skaters competed at six stages, among which there were no representatives of Russia. And if you look at the data in the table, it is easy to see that in 2022, 45 athletes made exactly the same number of 3Lz + 3T attempts as 33 athletes in 2021. And the number of 3F+3T attempts increased by only one.

For comparison, you can go back to tables J2021WR and J2022 and look at the number of solo jump attempts. You can immediately see that the number of 3Lz attempts increased from 46 to 70, the number of 3F attempts from 64 to 82 and 3Lo from 56 to 82.

In other words, the most “expensive” 3 + 3 combinations have become rarer in 2022, which directly indicates a regression. For those 20 skaters who took part in the Grand Prix in both 2021 and 2022, I made a separate table:

Table CN20212022

As you can see, the conditional “veterans” completed almost the same number of 3Lz + 3T and 3F + 3T attempts in 2022 as in 2021. There is still a difference, but a small one. Thus, it can be argued that the conditional “newbies” tried to perform these combinations much less often than the conditional “veterans” or than those 13 skaters who were in 2021 and absent in 2022. The reasons for this are not clear, because many “veterans” are not the favorites of the competition at all, and among the “newbies” there are, for example, Rika Kihira, Rinka Watanabe, Isabeau Levito and other famous athletes.

 

[4everchan]

The first year of a new olympic cycle is often less good is the hypothesis that comes to mind.
1) some veterans retire
2) some veterans take a longer rest period during the summer and are not as fit / ready when GP season starts
3) some skaters find it the perfect time to experiment with new layouts, series (and there is a brand new possibility with the + 2a series, which is very rewarding)
4) new skaters come to the senior ranks - If some are strong, confident and do well right away, some will struggle with nerves and lack of experience.

Interesting analysis you have provided but I wouldn't say I am surprised. I would suspect that by the times worlds happen, there will be less of a difference...

At the same time, in order to verify the hypothesis of the new olympic cycle, one would need to keep doing this data collection for a few cycles to have a large enough sample. The data is there (previous olympics from the start of IJS). In general, what we may see if that the harder jumps and combos are more present than 20 years ago... but their success rate in a first year of the quad, is much lower than at peak shape during the Olympic season.

Even without the Russians, what impressed me, was the number of triple axels done this year... My perception is that the jump was attempted by way more skaters than before and succeeded too by more skaters... so is there really a regression when more skaters attempt the hardest triple?

I would explain this with point 3) of the 4 points I put above.. and it may also contribute to explain some unsteadiness in other jumps... For instance, the new inclusion of the triple axel means a completely different layout for the skaters... they cannot start with the big combo anymore... and a lot of time is needed to set up the 3a, which perhaps rushes the rest of the elements.

Alaine Chartrand comes to mind : for a season, she did try to include the 3a in her program. When she finally decided to remove it, if I recall properly, she included way more choreo in the first part of the program... and it was a completely different feeling, with a much better success for her for the rest of the program.

To conclude, there are good and bad seasons in figure skating. The olympic cycle certainly has an influence on that. But in the end, the sport doesn't really ever regress. It's not a clear up or down trajectory ever. It just fluctuates a lot more in the first season after the Olympics (for all sorts of reasons (as mentioned in the 4 points - and there are probably other points too) than it does in the season just before the olympics.

 

[Alex Fedorov]

Even without the Russians, what impressed me, was the number of triple axels done this year... My perception is that the jump was attempted by way more skaters than before and succeeded too by more skaters... so is there really a regression when more skaters attempt the hardest triple?

Here you are seriously mistaken. This year, at six stages of the Grand Prix, 7 attempts were made to perform the triple axel, of which one was successful (Rinka Watanabe). Last year, at six stages of the Grand Prix, 27 attempts were made to perform the triple axel, of which 10 were successful. At the same time, non-Russian skaters made 14 attempts, two of them were successful.

But in this case, I can give a partial explanation. A significant part of the 3A attempts last year were made by Wakaba Higuchi and Alysa Liu. This year they simply were not there, and Rinka Watanabe was only able to partially replace them (although all attempts by Alysa Liu in 2021 were unsuccessful). But there is one not very good detail here - last year, one of the successful attempts of the triple axel was performed by Mana Kawabe. This season she also competed at the Grand Prix and never attempted a 3A.

Unfortunately, Amber Glenn, despite successive attempts to complete 3A in 2021 and 2022, has not yet been able to achieve success. In any case, she strives for this, which can only be supported.

 

[4everchan]

Perhaps I have the impression that more triple axels were attempted because I looked at the JGP/Challengers as well. I watched a lot of events... I saw more 3As... but it's quite possible they all didn't happen on GP. Still, I do believe I have seen more skaters attempt them, and I wouldn't certainly call this a regression.

 

[Alex Fedorov]

At the same time, in order to verify the hypothesis of the new olympic cycle, one would need to keep doing this data collection for a few cycles to have a large enough sample. The data is there (previous olympics from the start of IJS). In general, what we may see if that the harder jumps and combos are more present than 20 years ago... but their success rate in a first year of the quad, is much lower than at peak shape during the Olympic season.

...

I would explain this with point 3) of the 4 points I put above.. and it may also contribute to explain some unsteadiness in other jumps... For instance, the new inclusion of the triple axel means a completely different layout for the skaters... they cannot start with the big combo anymore... and a lot of time is needed to set up the 3a, which perhaps rushes the rest of the elements.

...

To conclude, there are good and bad seasons in figure skating. The olympic cycle certainly has an influence on that. But in the end, the sport doesn't really ever regress. It's not a clear up or down trajectory ever. It just fluctuates a lot more in the first season after the Olympics (for all sorts of reasons (as mentioned in the 4 points - and there are probably other points too) than it does in the season just before the olympics.

As you know, the 2020 Grand Prix was actually disrupted, and it is not possible to use the results obtained that year. But, on the other hand, you can use the data recorded in the ISU protocols in 2017 and 2018, when there was also a transition from one Olympic cycle to another.

You can compare the performance of solo triple jumps before the Olympics in Pyeongchang and after this Olympics by looking at this table:

Table J2017-18WR

This table does not show results obtained by Russian skaters. The number of athletes from other countries in 2017 and 2018 was the same - 31.

As you can see, not much has changed here. Attempts of 3T was too few, and the statistics for this column are therefore insignificant. Information about cascades 3 + 3 looks like this:

Table C2017-18WR

In this case, the only noticeable change is a drastic reduction in the number of 3F+3T cascades.
Thus, the hypothesis that the regression may be associated with the transition to a new Olympic cycle is generally not confirmed. It is interesting to compare the latest table with the cascade statistics in 2021-2022. It is immediately noticeable that five years ago there were more simple 3T + 3T cascades. In addition, some skaters tried to perform combinations that are practically not used today (3Lo + 3T, 3F + 3Lo, 3Lo + 3Lo). There is indeed progress being made here.

What confuses me a little is the fact that the total number of attempts to perform 3 + 3 combinations has changed little since then. In 2017, 31 skaters made 75 attempts to perform 3 + 3 combinations, in 2021 33 skaters made 69 attempts (that is, even less!), In 2022, 45 skaters made 83 attempts. That is, 14 skaters added compared to 2017 made only 8 additional attempts (mainly 3T + 3T).

As for the experiments - maybe I wasn't attentive enough, but I haven't noticed them this season. Unless sequences have become more frequent, but this is not due to the new Olympic cycle, but to changes in the rules. Unfortunately, due to these changes, it no longer makes sense to compare the 2021 and 2022 sequences.

The analysis probably needs to be continued by comparing the performance efficiency of the simpler 3+2 and 2+3 cascades, but that will come later.

Sources 2017-2018:

http://www.isuresults.com/results/season1718/gprus2017/
http://www.isuresults.com/results/season1718/gpcan2017/
http://www.isuresults.com/results/season1718/gpchn2017/
http://www.isuresults.com/results/season1718/gpjpn2017/
http://www.isuresults.com/results/season1718/gpfra2017/
http://www.isuresults.com/results/season1718/gpusa2017/

http://www.isuresults.com/results/season1819/gpusa2018/
http://www.isuresults.com/results/season1819/gpcan2018/
http://www.isuresults.com/results/season1819/gpfin2018/
http://www.isuresults.com/results/season1819/gpjpn2018/
http://www.isuresults.com/results/season1819/gprus2018/
http://www.isuresults.com/results/season1819/gpfra2018/

 

[Alex Fedorov]

Perhaps I have the impression that more triple axels were attempted because I looked at the JGP/Challengers as well. I watched a lot of events... I saw more 3As... but it's quite possible they all didn't happen on GP. Still, I do believe I have seen more skaters attempt them, and I wouldn't certainly call this a regression.

Well, you can always study the protocols yourself and check my information. Links were given in the first post. 3A is in any case a rather rare element, and it is not difficult to check this.

The Junior Grand Prix as a whole made a more favorable impression than the senior competitions. But we do not yet know how many of the 3A shown there will "survive" in a few years.

 

[4everchan]

We will indeed see about the juniors. I am not finding the discipline is regressing. Statistics are always interesting but they do not tell the full story.

For instance, one could think that pairs have regressed since 2018, because there are pretty much no more quads being performed since then (i believe only the quad twist of Sui-Han at the Olympics)... but were pairs weaker from 2019-22.. not quite ready to say that.. the BV of quads was too weak so skaters had different strategies.. This year, with the +2a sequences, the SBS jumps are getting used in a different way, some skaters landing very rewarding series. (young pairs at Japanese nationals landed 3s-3t+ 2a) Is the sport regressing? Right now, there are many new teams. Only two very seasoned teams are still competing (K-F and Riku and Ruichi) We will have to wait and see .. but at first sight, it could look like the pairs are regressing with so many doubles being performed... but I doubt it will stay like this for very long.

Back to women : what is needed to take into account here is that a few weaker athletes were given assignments this year. The void created by the Russian skaters not being able to compete left spots to fill... (and some others who have retired or are injured) I know you have removed Russians from your stats to alleviate for this, but the ratios will be off no matter what because the data cannot be completely compared. And then, if you compared only the data from athletes who participated in previous years, does it really tell that the sport is regressing or simply that they are not at their peak after the Olympic year? I mean, look simply at Kaori Sakamoto.. She's far from being at her peak so far this season.

This is the fun and flaw of statistics. It is very hard to have a meaningful and objective analysis with small samples not comparing exactly the same things. I mentioned pairs, because there, one can see it without stats : many teams not only don't do quads but don't even do triples... even at GPF... But with women, I wouldn't say the sport has regressed, as a matter of fact, I personally think it is getting better in more countries than before.

That's my perception.. I respect your analysis nevertheless. Thanks for bringing this up. If anything, my observation will be stimulated at the ISU championships.

 

[decafjava]

Very interesting discussion. It will take a few read throughs to grasp everything if I ever but thanks.

 

[ladyjane]

I am sorry, I love these statistics, so don't get me wrong. Bu't in my (maybe very personal view) we are looking at figure skating, not figure jumping or ice jumping, It might be difficult to take other elements than jumps into account, while judging PCS as an efficiency thing is even more difficult. I like the discussion though. And I follow it.

But I follow just one sport and that is Figure Skating. I don't really care who is fastest, highest or furthest. I wish to be moved. I wish to see new things, whether it is a new jump, a new spin or otherwise. I wish to be touched by a story. That's all, and it's probably my bad.

 

[Alex Fedorov]

I am sorry, I love these statistics, so don't get me wrong. Bu't in my (maybe very personal view) we are looking at figure skating, not figure jumping or ice jumping, It might be difficult to take other elements than jumps into account, while judging PCS as an efficiency thing is even more difficult. I like the discussion though. And I follow it.

But I follow just one sport and that is Figure Skating. I don't really care who is fastest, highest or furthest. I wish to be moved. I wish to see new things, whether it is a new jump, a new spin or otherwise. I wish to be touched by a story. That's all, and it's probably my bad.

You see, jumping in this analysis is not some kind of "winning key". It's more of an indicator. In figure skating, almost everything is subjective, because it is half an art. Or maybe art is even in the first place, and sports in the second. And even if we talk about jumps and not mention the PCS - for example, two skaters did a triple axel in a row - one received a GOE of +1.6, and the second got +2.4 - and the fans look and do not understand what was the difference. I do not touch on this problem, I have a very simple criterion in my analysis - whether this jump was made technically correctly or not.

One can use a simple analogy with a symphony orchestra: the evaluation of how music is performed is ideally a matter of the individual taste of the listeners. But if the violinists in our orchestra do not know musical notation and have never held a bow in their hands, then we will not hear anything good even if there is an absolutely virtuoso flutist in the same orchestra.

And when I see that figure skaters for some reason lose their triple jump skills in a year, then this, of course, alarms me. Yes, jumping is not everything. But such elements are necessary, otherwise there will be no program, no holistic impression will arise, and the result is unlikely to touch someone's soul.

As for the criteria for evaluating non-jumping elements: in general, such a criterion can be, for example, the average level of an element. For example, 12 skaters performed 12 step sequences, four of them were given level 4 by the judges, five - level 3, three - level 2. Averaging all this, we get that in our competition the average step sequence level was 3.08. The disadvantage of this method is that assigning the level of a sequence of steps is also subjective.

I have not yet made such statistical calculations, but I assume that here, if there are changes over the year, then they are small.

 

[noskates]

I can appreciate your hard work and dedication, but my brain is not willing to analyze all this. I'm on statistic overload.

Carry on....

 

[rain]

I simply disagree with your basic premise that the grand prix final was some kind of "disaster" or "catastrophe" (which it wasn't). This bias colours everything that comes after. It's flawed from the start.

 

[AxelLover]

I simply disagree with your basic premise that the grand prix final was some kind of "disaster" or "catastrophe" (which it wasn't). This bias colours everything that comes after. It's flawed from the start.

I think that the GPF as a whole was quite good, it's just the ladies free skates that might be called a disaster. There was no quad attempt, just one triple axel attempt (didn't work out) and just one +3Lo combination (not clean either), so the jumping content was very weak overall. I would be OK with that if it lead to many clean (or almost clean) skates, but that didn't happen. There were lots of URs, lots of bad landings and everyone except for Kaori suffered a fall. Kaori, on the other hand, popped three of her jumps and made a severe mistake on her 3Lz too.

Anyway, I'm glad for threads like this. It takes a lot of time to summarize statistics like this, but it's worth it. It's definitely interesting to read, at least for me.

 

[noskates]

Why does there have to be a quad attempt or a triple axel attempt to declare a competition "good" or "great"! This is the part of figure skating that is really a turnoff to me these days where the whole sport isn't considered and only the jumps are important. To me (and of course everyone is entitled to their opinion) that's like saying a diving competition wasn't good because the water was too wet. I think ALL the elements should be considered and how they're executed makes the competition successful. Again - JMO. We've all seen competitions where the skaters were more consistent - that's for sure - but the medal winners earned their medals.

 

[radkat]

There has been some decline but imo much of it can be chalked up to normal post-Olympic slump. It's probably exaggerated by burnout from many skaters having pushed themselves so much technically last year, compared to previous Olympic seasons, in order to be competitive against skaters with some questionably attained jump content.

I think there's also some level of delusion among fans about what female bodies are even capable of (with current coaching methods anyway). We don't *know* if they can do quads without being chemically assisted. We don't know that very many of them can consistently do a 3A or 3Lz3lo combos. It's just so weird to me to be disappointed by not seeing a level of athleticism that might not be physically possible within the very discipline you've chosen to be a fan of.

 

[Alex Fedorov]

I simply disagree with your basic premise that the grand prix final was some kind of "disaster" or "catastrophe" (which it wasn't). This bias colours everything that comes after. It's flawed from the start.

well, I don't think so. I can call what happened in Turin a disaster, I can call it an accident, I can call it an ordinary tournament, I can even call it a true triumph of mature women's skating - this will not change the information recorded in the protocols at the Grand Prix qualifying stages anyway. And when I evaluate efficiency, I strictly adhere to the same rules when examining data from different years. Actually, this is one simple rule: if the score given by the judges for an attempt to perform an element is greater than or equal to the base value of this element, then such an attempt is considered successful.

I can say that the systematic problems with the 3F this season that were revealed through the study of the protocols came as a surprise to me. Initially, I assumed that if a regression is found, then it will be reflected first of all in a more difficult 3Lz or in cascades.

 

[Alex Fedorov]

I think that the GPF as a whole was quite good, it's just the ladies free skates that might be called a disaster. There was no quad attempt, just one triple axel attempt (didn't work out) and just one +3Lo combination (not clean either), so the jumping content was very weak overall. I would be OK with that if it lead to many clean (or almost clean) skates, but that didn't happen. There were lots of URs, lots of bad landings and everyone except for Kaori suffered a fall. Kaori, on the other hand, popped three of her jumps and made a severe mistake on her 3Lz too.

Anyway, I'm glad for threads like this. It takes a lot of time to summarize statistics like this, but it's worth it. It's definitely interesting to read, at least for me.

I would not like to discuss the final in this thread. Yes, at the very beginning I expressed my own opinion about this event, but there is no need to go into detail. In addition to my subjective fears that such a discussion will lead to the closure of the topic, there are objective reasons not to do this:

1) only six skaters participate in the final, and the fewer participants, the more random factors play a role.
2) there is literally nothing to compare the final with - the last time a similar event took place in 2019, and there three out of six skaters were Russian.

 

[lilimum]

I think it is difficult to compare last and this season. Many skaters are taking the opportunity to jump the +2A combo and perform instead a 2nd 3T a 2nd 2F in their free program. If 7 triple are the maximum they have to jump doubles anyway than why risk to mess a 3+3 combo ? This option changed completely the "jump strategy" in the free program and it is perhaps the reason that the "veterans" kept their 3+3 combinations. In addition to that this season the "poor take off" especially of Flip and Lutz should get more negativ GOE than during the last seasons.
In addition to that the total number of triple attempts increased over the years. I think this is not a sign of regression