I started to write a post several days ago, but I don’t know when I’ll end it, there are always something that catch my attention. After the Japanese media started to questioning the base value of the quadruple Axel, I decided to do something. At the ISU they must be busy in writing the new rules, or the calendar of the events, or also the health measures for the competitions (and as I seen today, in choosing the best judges for the competitions), so I do some stats in favour of them. Perhaps, with a little aid, they will find easier to take the right decision.
After the 2017-2018 season, the base value of several jumps was lowered and the GOE passed from a scale to +3/-3 to a scale of +5/-5. The reasons for these changes are two. A wider scale can reward better the quality, a good intention, a subject on which I’ll write another time. The lower base value of the triple and quadruple jumps was meant to balance the importance of the TES and the PCS. A better balance was needed, we all know it at least from 2015 NHK Trophy.
For the components a skater can’t earn more than 100.00 points, but that day in Nagano Yuzuru Hanyu earned 118.87 points only with the technical element, and with a program in which he jumped only three quadruples (in 2015 it was a huge number, now not anymore), two toe loop and a Salchow. How the ISU reacted? Lowering the base value of the jumps.
I’m not in the technical commission, but I’m sure that something different could be done, and I’m not the only one to think so.
At the 57th ORDINARY CONGRESS of SEVILLE – 2018 (April 30, 2018) the Netherland did a proposal that I would have approved instantly:
This factorization would have brought the PCS for the Men’s short program up to 60.00 points, for the free skate up to 120.00 points. It would not been enough, Nathan Chen Technical score at the Olympic Games was 127.64 points, but the imbalance would have been smaller. And they could have used different numbers, as 1.3 and 2.6.
What happened of this proposal?
According to the ISU, in 2018 it was premature to discuss this change because the +5/-5 system was new, and none can say how it will work.
Now, beside the fact that the ISU could have changed the factorization and not the PCS, I did a stat for them. I needed only few minutes to do it, and if I can do a things as this, using simple a protocol, the rule book and a file Excel, anyone can do the same.
I took the planned layout of Yuzuru Hanyu’s free skate in the 2016-2017 season. Why this program? Because it was way before the Olympic Games and the 2018 Ordinary Congress, so they could have done this in advance. And this is not the most extreme program. At the World Championship Hanyu’s base value, with 4 quadruples, was of 103.43 points, Nathan Chen’s base value, with 6 quadruples (and among them two 4F and one 4Lz), was of 112.87 points. But let’s see the program that fixed forever the historical World record.
On the left there’s Hanyu layout for the short program and for the free skate. I’ve written the base value of every element, the highest possible GOE and the sums. In bold I’ve written the highest possible TES and PCS. The difference is big. At the center I used the same layouts but with BV and GOE for the 2018-2019 season. According the new rules, I give to him the bonus for the second half only for the last jump in the short program and the last three in the free skate. I also deleted the last jump, because now in the free skate there is a jump less. Note that, in this way, the jump who I deleted was the 3Lz, but it was possible that he deleted instead a jump with a lower BV. Mine is an hypotesis, in truth the best skaters planned layout different, with an higher base value, but we could not say in advance what they would have done, so my hypotesis is a good point to start the analysis.
For the short program the TES is higher because now the GOEs are higher, for the free skate the TES is lower but only because there’s a jump less. In the sum, the TES is higher now. And the base value were changed to equilibre the two aspects of the programs?
At the right I did three summary tables. Firts there’s the program with the +5/-5 system and the real factorization. The TES could be over 50 points higher than the PCS. It’s right so? Under I used the proposal factorization at 1.2 and 2.4. The difference is over 20 points, less than half but still really high. At the bottom I used an hypotetical 1.3 and 2.6. The TES remain higher than the PCS, but the difference is little. And, before someone can say to me that none never earned the maximum possible GOE, none also never earned the masimum possibile PCS. The higher values ever awarded for the short program is 49.14 (Hanyu at the 2015 GP Final), with a 1.3 factorization it will became 63.88 points, not 65.00, and for the free skate is 98.56 (again Hanyu at the 2015 GP Final), that became 128.13 and not 130.00.
But I agree with Mr. Lakernik, this change need to be done for all the disciplines. Now we can see what happened with the new system, so I watched the layout of the 2019 Grand Frix Final of Anna Shcherbakova and Alexandra Trusova. As I am considering the maximum score, in my calculation all the jumps are fully rotated (but I left a 2S as Trusova’s second jump, even if we know that she didn’t intend to jump a double), and spins and steps are of level 4.
The Ladies needs a higher factorization, perhaps 1.0 for the short program, in which they can’t jump quads (why?), and 2.4 for the free skate, but a change as this will work only if the scores are given according to what the skaters do on the ice, not according to their wins in the last years, or their nationality, or their sympathy. I need to watch closely the PCS, but not now. In truth I started to write thinking of the base value of the jumps, but as usual my little detours are long.
Ok, the BV of the jumps was lowered. Let’s see how.
In the old system the GOE weren’t linked to the base vaue, but were divided in several steps. In the first there were all the solo jumps and the double toe loop and Salchow. In the second there were the double loop, flip and Lutz. The double Axel was considered different from the other jumps, more than a double, less than a triple, and he was alone. After there were the jumps from the triple toe loop to the triple Lutz. The triple Axel was considered so difficult that was put in the same step of the quadruples. And the quadruple Axel was alone again.
To watch all the numbers can be boring, so I did some graph. The first say to us the difference in the base value.
At the left there are all the data, in the center only the data that I used for the graph. Only the single loop has the BV a little lifted up. Proportionally, comparing them with the jumps that precede and follow them in the graph, it seems that the most penalized jumps are the 4Lo and the 4Lz. Otherwise, the 4F didn’t lost so much, and the 4T and 4S are deemed equal. But the jumps don’t have only the BV, they have also the GOE.
All the single jumps were penalized, but for the Axel the penalization is really little. The doubles go on pairs: toe loop and Salchow the same little growth, loop and flip a loss of some cents (bigger for the loop), Lutz and Axel again the same growth, a little bigger than for toe loop and Salchow. For the triple, toe loop and Salchow are the only ones that lost something. The growth of the loop is the same as for the double Lutz and Axel. But the flip grow much more. With the GOE, a jump for wich the BV didn’t change saw his importance in the score grow a lot. His growth is bigger than the growth of the triple Axel. But it wasn’t the Axel the most difficult jump? Why his importance was lowered? Sure, for a 3A a skater can earn much more points than for a 3F or a 3Lz, but the difference isn’t so big as before.
And then there are the quadruples. The jump who had best growth is, again, the flip, followed by Salchow, toe loop, loop (same growth as 3F), Lutz and Axel (same growth of 2Lz, 2A and 3Lo). On the 4A I’ll come back later, let’s see the other quads.
Toe loop and Salchow are the most common, they were done for the first time in 1988 and 1998, and a lot of skaters, men and now also woman, can jump them. The first quadruple Lutz has been done in 2011 but, until Boyang Jin did it in 2015, the jump was an isolate episode. The first 4F and 4Lo were both jumped in 2016, by Shoma Uno and Yuzuru Hanyu, and several skaters followed the path of a lot of quadruples in their programs.
Which jump is the more difficult? For every skaters there are some differences, a skater can like more a jump, another skater can like more another, but the stats can give to us some indication. Almost a year ago the ISU published an interessant statement.
Don’t watch the first phrase that I highlighted in red, it’s an old screenshot in which I expressed my perplexity on the same technique for flip and Lutz, but now I’m not interested in it. What interest me now is the last phrase:
If you look at the statistics, the quad loop very likely is the most difficult quad
I too did some stats and wrote about them, in Italian, here. I wrote the post in January, now did some more stats, adding the international competitions of the 2020-2021 season. I hope I have not forgotten some skater or some jump, If you find any mistake and you write to me, I’ll fix tables and graphs.
In the first table we can see all the 4Lo, 4F and 4Lz landed with a positive GOE in international competitions (no Japan Open, I can’t take it seriously). The names of the Ladies are higlighted in bold.
The Lutz is, by large, the most common jump, and both for Lutz and flip we see a skater (Jin and Uno) able do land 30 clean jump, and Chen is not far. With the flip he is very close, but also his Lutz is a good jump. In this rank the best loop jumper is only seventh, preceeded by 4 skaters that do Lutz and 2 that do flip. Tree Ladies jumped the Lutz, two the flip, none the loop. After, I added the quadruple with a GOE negative. Be aware, I didn’t distinguish among wrong edge, step out or falls, it would be too much work. I watched only the GOE. The graph is the same, the GOE negative are added above the positive with a lighter colour.
There are two more skaters for loop and flip, three for Lutz. Some skater has a good percentage of good jumps, some skaters not, but they do the jump anyway, a sign that to try the quad knowing that probably they will not land it well, with this scoring system, can still aid a lot their final score.
Again the Lutz is the most common jump, the loop the most rare. The loop is eighth for numbers of try after two flip and five lutz. All this jumps give to the skaters the base value of the jump (sometimes there was a base value lower because the combination was lacking, a detail that did not enter in my graph). But I checked also the attempts. Not the jumps popped, it’s impossible, but the jumps downgraded. The skaters tried to do the jump, but they failed.
I added another name for loop, flip and… Axel. At the 2018 Rostelecom Cup Artur Dmitriev Jr. did an Axel in which he landed clearly forward after four revolutions. For now his is the only attempt (he tried again at the National Championship, and again the jump was downgraded).
Again the Lutz is the most common jump, the loop the most rare. It’s curious that there are only 6 downgraded flip and 11 downgraded among both loop and Lutz, but the 11 loop downgraded are done in 77 jumps, the 11 Lutz in 302 jumps. It seems that the loop is the jump that can end downgraded more easily than the other jumps.
A thing struck me: two downgraded loop were jumped by Yuzuru Hanyu. I remember the jumps, at Skate Canada 2016. He landed forward, the calls were right. But, how often Hanyu’s jumps were downgraded? I counted all the jumps listed by SkatingScores. I watched the rotations, not if the GOE was positive or negative. I divided the jumps by type and call: full rotated, q (only one, the rule is new), < and <<. I counted also the jumps done in the National Championship, I excluded only the jump invalided for the Zayak rule because for these jumps there weren’t real calls. As 1Eu I listed all the jump he did as second jump of a combination of three jumps, regardless of the sign on the protocol.
For only two jumps Hanyu completed the rotation less than the 90% of the time (column JJ); the quadruple loop and the quadruple toe loop, and the difference among the two jumps is big. Of his only six downgraded (one of them the euler he did at the 2018 Rostelecom Cup, when he was injuried, and that he skate that program is stunning, none can do what he did that day), two are quadruple loop. I higlighted in red (column JK) the two jumps in which his percentage of downgraded is higher. So it seems that even for a skater that rarely doesn’t complete the rotation of a jump (over 95% of fully rotated jumps), the quadruple loop is particularly difficult.
Returning to all the skaters, if the difference of how widespread is a jump isn’t clear, I did another graph. The graph is based on the first of the three tables:
So its clear that the number of Lutz is much higher than the number of flip, and the flip are much more than the loop. In fact there are more clear flip, 67 (second table), than loop in total, 66 counting the positive and the negative GOEs. And there are almost as much clear Lutz than flip with GOE positive or negative.
In the first table there’s another detail, in the line “% GOE +”. Among the 77 loop, only 29, the 37,66%, earned a positive GOE. Among the 141 flip, 68, almost the half (48,23%) earned a GOE positive. Among the 302 Lutz, 132 earned a GOE psitive, the 43,71%. So it seems that to obtain a GOE positive on a loop is more thifficult than on the other jumps.
The last little table is the numeber of the skaters that did, or attemped, every jump, and the number increases from loop to flip to Lutz.
So, when the ISU said that “If you look at the statistics, the quad loop very likely is the most difficult quad”, I give to them the credit to have said something true. Which reactions there were at the ISU’s statement? Philip Hersh in his New math: Figure skating’s latest recalculations change skaters’ formula for success notes that Chen’s
two highest-valued jumps, the quadruple Lutz and quad flip, no longer add up to much – or as much – of an advantage.
For Chen an higher base value for the loop isn’t an advantage. I remember that the Russian federation wasn’t happy for the changes, but I don’t remember where I read it. If I’ll find a source, I’ll put the link here. For now I put a last graph, with the jumps divided by a thing of almost none importance, the nationality of the skaters.
Every horizontal line marks 10 jumps. I can’t really imagine which nation(s) can’t be happy with a change that gives the same base value to loop, flip and Lutz. Fortunately the change wasn’t done, the ISU said that it would be no correct toward the skaters, and implemented only the rule of the q, a rule that, with the actual technology, it’s impossible to apply in the right way.
If anyone is wondering… the ISU can change the base value for the next season even now. Little changes are done every year, so they can do it if they want. But they want? Perhaps the ISU nees only to notice that the scores were not done in the right way, so if we talk enough about them… The loop is the rarest jump among quadruple loop, flip and Lutz, and deserves a base value at least equal than the base value of the other two jumps.
And then there is the Axel.
This is the base value of all the jumps.
Have you seen that the big growth is among 3Lz and 3A? And then came the growth of the 4T compared to the 3A, of the 2A compared to the 2Lz, and only then the growth of the 4A compared to the 4Lz. As fairytales say to us, once a time there was a different scoring system, and the Axel was deemed much more difficult.
This is the base value difference from every jump and the jump that precede it according the old (green) and the new (blu) scoring system.
The flip is the only jump that earned something by this change. It’s fascinating to see how a triple Axel is deemed much more difficult than a triple Lutz, but a quadruple Axel dont’s seems so more difficult than the quadruple Lutz. Even the Lutz have lost something, but if there are many skaters that landed a 4Lz and none that landed the 4A, perhaps the 4A is a little more difficult.
A thing that I asked to me is how much grow the difficulty if a skater add a rotation of his jumps. So I watched the difference in the base value from a single toe loop to a double toe loop, from a double toe loop to a triple toe loop and so on.
Again the table on the left is the full data, the graph is based on the other table. For the same pair of jumps the green column is almost everywere thaller than the blu column, the difference among almost all the jumps was lowered. If before for a 4T a skater earned 6.00 points more than for a 3T (beside GOE), now he earn only 5.30 points more. But for the quadruple he deserves anyway a difference of points higher than the difference that exist among a triple and a double. Increasing the number of rotations increases the difficulty, and the difference in the base value is greater. There is only one exception. For the ISU to learn to jump a triple Axel is much difficult (+4.70 points) than to learn a quadruple Axel (+4.50 points).
If we want to see some numbers…
For some inexplicable reason, while a double Axel is considered much more difficult than a double Lutz, with the difference between the two jumps reaching 1.80 points, and the difference between triple Axel and triple Lutz is huge, with a difference of 3.15 points, a quadruple Axel shouldn’t be too difficult compared to a quadruple Lutz since the difference is only 1.50 points.
This is another table that I did some time ago.
Thanks to a friend I noticed another very curious detail. A triple Axel is considered so difficult that an underrotated triple Axel is worth more than a correctly executed triple Lutz. How is it that an underrotated quadruple Axel is worth less not only than a quadruple Lutz but also a quadruple flip and a quadruple loop and is worth little more than a quadruple Salchow?
I’m not the only one to notice the strange value of the 4A, on twitter I’ve find some interesting graphs:
So, if the ISU is wondering if the base value of the jumps is right, the answer is no. Both the quadruple loop and the quadruple Axel deserves a base value higher. A change like this one can be done in every season, so if the ISU want to do something right and demostrate that figure skating is a serious sport in which the difficulty are rewarded, knows what to do.
The real question is: what want the ISU?
Pingback: Sportlandiaより「ISU殿、問題は深刻です」 | 惑星ハニューにようこそ