A few days ago I saw a table from one of the Wikipedia pages dedicated to Yuzuru Hanyu republished on Twitter. The table did not surprise me, it is something I had already noticed on my own, but I think it is worth revisiting it:

For three seasons on numerous occasions the judges recognized the perfection of what Hanyu did on the ice by assigning him the maximum mark. How often has this happened? I counted the number of international competitions that Hanyu has participated in and the number of elements that are part of the programs. Until the 2017-2018 season, 20 elements were performed in the competitions, 7 in the short program and 13 in the free skate. I don’t care that some element was canceled, in my checks I also counted the no value elements. In that case there was a Hanyu’s mistake but, until he performed the element in question, the possibility of receiving a high mark existed. From the 2018-2019 season, 19 elements can be performed in the competitions, since a jump has been removed from the free skate. Let’s see how often Hanyu performed perfect elements (according to the judges):

Season	Competition	N. Elements	max. GOE	frequency
2015-2016	5	100	12	8,33
2016-2017	7	140	6	23,33
2017-2018	3	60	9	6,67
2018-2022	12	228	1	228

In the 2015-2016 season out of (just over) eight elements, one was perfect. In the following season he struggled harder, only one of (just over) 23 was judged as perfect. In the PyeongChang season, however, it was extraordinary, almost one in 7 received the highest mark. And then? Talking about seasons is complicated because some seasons have been incomplete, but based on the number of competitions, in the last four seasons only one element out of 228 has been judged as perfect. Let’s forget his mistakes, those are there now as they have been in the past. Really he is no longer able to perform perfect elements? Has his skating really gotten that bad?

I didn’t save that post, but one person commented saying it’s normal for Hanyu to get lower marks because everyone gets lower marks, so me and others who were pissed off by this table were complaining about nothing. And, a few days later, someone managed to claim that the new scoring system actually helped Hanyu. In this case I have saved the tweet.

I didn’t answer. I did not do it because even if I some time ago wrote on this topic, I was already preparing this post, even if between my commitments outside internet and all the checks it took me longer than I would have liked. However, I would advise all those who make a statement not to limit themselves to the surface, because if it is true that I too use SkatingScores very often for my checks, it is also true that we must know how to use the tools we have available, and a control like that of these tweet shows us the tip of the iceberg but tells us nothing about the rest.

Let’s proceed in order. Is it true that the number of +5 awarded by the judges is lower than the number of +3 in the old scoring system? Yes, it’s true. The Wikipedia table includes all international competitions. I have reduced the field a bit, even if this means excluding programs such as the short program skated by Hanyu at the Autumn Classic International 2017, the historic world record. I have checked all the Grand Prix competitions (senior only), the European Championship, the Four Continents Championship, the World Championship (senior), the World Team Trophy and the Olympic Games that have taken place from the 2010-2011 season onwards. For these competitions I counted the number of short programs and free skate skated, calculated the number of elements that could be executed and, after counting the number of +3 (or +5, depending on the period) awarded, I calculated the percentage.

For three seasons, between 2015-2016 and 2017-2018, those in which Hanyu received the highest number of perfect GOEs, a very high number of +3 was awarded to several skaters. By comparison, in the following three seasons the number of +5 is remarkably low, only in this last season the percentage has risen. So if we look at the surface, the critics are right. But I have asked myself several questions, starting from the moment in which there was this incredible increase of +3. What happened? Is the number of +3 increased equally for all skaters?

My check covers all +3s, analyzed in a file that looks like this:

As you can see, the data is structured so that I can put it in order according to the criteria I prefer. As you can see even from a little bit of my file, the data are a lot. In the competitions I analyzed, in the 2017-2018 season, 38 different skaters (Aaron, Aliev, Aymoz, Baldé, Besseghier, Brezina, Brown, Bychenko, Cha, Chan, Chen, Fernandez, Firus, Ge, Hanyu, Hayrapetyan, Hendrickx, Hochstein, Jin, Kolyada, Kvitelashvili, Majorov, Messing, Miner, Mura, Nadeau, Ponsart, Reynolds, Rippon, Samarin, Samohin, Tanaka, Ten, Uno, Vasiljevs, Voronov, Yan, Yee and Zhou) in 90 programs received at least a +3. The total number of +3 was 409. In order to have a readable graph, I excluded all skaters who did not receive at least in one season the 6% of +3 for his elements.

The skaters who appear in all seasons are few. I combined the two parts of the graph relating to Patrick Chan with a thin line, because in the 2014-2015 season Chan decided not to compete. In fact, it is evident that in the first four seasons, those culminating in Sochi, Chan was the skater capable of giving greater quality to his elements. Chan never lost that quality, when he returned after the break he received better marks than he received before, even though Jason Brown and Javier Fernandez got close to him and, in the end, surpassed him, albeit slightly. But what stands out most are Hanyu’s last three seasons.

In the first two seasons Hanyu did not receive very high marks, but we must not forget that he was very young. In March 2012, when he won bronze at the World Championship in Nice at the end of is second senior season, he was only 17 years and three months old. Shortly thereafter he moved to Canada, and his work with Tracy Wilson paid off. Already in the 2012-2013 season he received a number of +3 higher than 6%, the threshold that I set in order to control only a small number of skaters, and in the following years he has only improved. In the Sochi season he was second, behind Patrick Chan and ahead of Daisuke Takahashi, after that season he became the best. Clearly the best. True, in the 2014-2015 season there was a small decline, probably, if Chan had competed, he would have scored a better percentage than Hanyu. At the beginning of that season Hanyu was the protagonist of a terrible collusion in China, the free skate he skated that day, and the two programs of the next NHK Trophy, are programs that he should not even have skated since his conditions were so precarious, and obviously those 33 elements out of phase all the percentages. But, over the next three seasons, Hanyu has reached levels of perfection that others can only dream of. To better understand what happened, I looked at the competition-by-competition details of the best skaters (over 6%) in the 2015-2016 and 2016-2017 seasons.

In the 180 marks Hanyu received at the 2015 NHK Trophy there were 69 +3, something that is unprecedented. The previous record was Chan’s 35 +3 at the 2014 Olympic Games. Almost double. A number that no one else has ever been able to replicate. No one else, because Hanyu has outdone himself. In the Grand Prix final the +3 became 120, two thirds of the total marks. Chan, Fernandez and Shoma Uno, authors of some of the best performances of their careers, are a long way off. And even if in the 2016 World Championship, where he performed an extraordinary short program but an imprecise free skate, conditioned by the Lisfranc Injury, the number of +3 received by Hanyu dropped to 62, it is still a higher number than the number of +3 ever received by anyone else in his best performance (for Fernandez it is 57 at that World Championship, for Uno it is 54 at World Team Trophy 2017, for Brown it is 50 at the Four Continents Championship 2018, for Chan it is 49 at Team Olympic Event 2018).

In the fall of 2015 Hanyu showed that he was able to perform extremely difficult elements, and to execute them perfectly, inserting them in a program that is not a simple succession of technical elements, but a whole. He destroyed the scoring system, and the ISU decided that this should not be allowed. The marks awarded to the others have been raised. Really, why else would everyone start getting higher marks? Watch the last two graph I published. Did everyone suddenly improve in the same period? Is it really possible that by pure chance they all improved in the same period? Hanyu has perfected his skating, and immediately after the marks of the others have been raised to keep him from dominating like he had just done. And of course his marks were lowered.

Of course, after those competitions Hanyu made several mistakes. But Hanyu always made some mistakes. He skated some perfect programs, he also missed one or more elements in several programs. But a wrong element is a wrong element, and often it does not ruin the rest of the program. Watch the 2017 World Championship short program. We could argue for a long time about which call was correct for the combination, but for now I go on. He got that combination wrong, and the negative GOE was inevitable. But what was missing from his triple axel to not have +3 unanimous? Why did three judges, the Australian Deborah Noyes, the Belgian Francoise De Rappard and the Italian Gloria Morandi, allow themselves to give a +2? And I could give many examples in a lot of competitions.

Instead of watching many scores, I go back to the 2015 Grand Prix final. I am writing years later, and at the time I had no idea that I was going to start writing about figure skating, much less about figure skating judges. Many articles, many documents, are no longer available online. But something remained, and for this I thank Lys and the other people who helped me in the research.

Do you know who Alexander Lakernik is? Since 2016 Mr. Lakernik has been vice president of ISU, in the past, among other roles he held, he was Technical Controller of the women’s competition in Sochi 2014, when Adelina Sotnikova won gold ahead of Yuna Kim, and before was Assistant Referee in Pairs in Salt Lake 2002, the competition marked by the scandal of Marie Reine Le Gougne’s forced vote in favor of the Russians. These are not the only competitions he has judged, but sometimes I like to remember his presence there. Some time after the 2002 scandal, Lakernik became chairman of the ISU technical commission, and his first move (Bianchetti Garbato, Crepe nel ghiaccio, p. 201) after the election was a letter of admonition for Ron Pfenning, the referee of Salt Lake City, which was working to investigate possible Russian involvement in the scandal. I suspect Lakernik did not like the investigation extended to the Russians.

Mr. Lakernik is also a graduate in a scientific subject (mathematics? Statistics? Physics? I honestly do not remember, but he is someone who knows how to use numbers very well, from what I have read he has also written numerous mathematics essays, so we can be sure that he knows what he is doing). We may not share Lakernik’s ideas, but underestimating his intelligence is a mistake.

Not having the original source, I limit myself to post the link to some articles that let us know that Lakernik said something whose significance it was only possible to understand later. Years later.

Andrea Bongiovanni on Gazzetta dello Sport, December 28, 2015:

Il guru russo Alexander Lakernik, referente della commissione tecnica d’artistico della federazione internazionale, in una recente intervista a R-Sport, ha ammesso che, di questo passo, presto (ma non si potrà prima dei Giochi di PyeongChang 2018) si dovranno rivedere i coefficienti di esecuzione dei singoli elementi. Colpa di Hanyu, naturalmente. E di programmi mai così azzeccati

We can translate this lines as

The Russian guru Alexander Lakernik, head of the Technical Committee of the International Skating Union, in a recent interview with R-Sport, admitted that, at this rate, soon (but it will not be possible before the PyeongChang 2018 Games) the performance coefficients of the individual elements will have to be reviewed. Hanyu’s fault, of course. And of programs that have never been so successful

Also Beverley Smith wrote about it:

You knew it was inevitable: International Skating Union technical committees are all abuzz with what to do with the scoring system after Yuzuru Hanyu’s record-breaking exploits from his past two competitions.
In an R-Sport interview in Russia, Alexander Lakernik, a Russian who is chairman of the technical committee for everything but ice dancing, said that changes need to be made to the following:

– Change the levels of grades of execution, maybe go from -5 to +5 or mark it on a scale of 1 to 10;

– Simplify the rules.

[…]

Lakernik says the ISU technical committee will initiate changes to the rules for figure skating at the 2018 Congress. There is a Congress this summer, but he says no radical changes will be implemented because it’s just too close to the 2018 Olympics in South Korea.

So yes, the rules were changed after the 2018 Olympic Games as a reaction to what Hanyu did in the fall of 2015.

The problem is not that the PCS are not keeping up with the TES. In the free skate of Barcelona Hanyu obtained 120.92 of TES, 98.56 of PCS. If he had only received +3 (and with a level four step sequence, while the call was level 3) his TES would have been 125.19. With a factorization of 2.50 instead of 2.00, PCS could have gone up to 125.00 points. A change of this type could be done easily, quickly and would not have upset the scores. Of course, after the 2016-2017 season they would have had to change the factorization again. So? It’s enough to check the program that has the highest base value and set the parameters according to his score. It takes no more than two minutes. And this is a much simpler change than the one they made, a change that would not have distorted the results of the competitions as happened did with the solution chosen by them.

Lakernik did not like the fact that Hanyu dominated the competitions and he reasoned about how to change values. What happened is evident in the next graphs.

Number of +5 is lower than number of +3. This is not without effect in the scores. I took two protocols, those of the short programs of Hanyu and Uno at the Olympic Games in PyeongChang. The work I did is not very precise, in the sense that in the passage from +2 to +3 the score varied by 33.33%, in the passage from +4 to +5 the score varies by 20.00%. The difference is lower than what I calculated but, despite this, the calculation I have made seems interesting to me. These are two excellent programs, for Hanyu it is second highest score obtained with the score code +3/-3 (the second overall, all the four best scores were established by Hanyu), for Uno it is his third best score with that scoring code, the fourteenth overall. Only Hanyu, Fernandez and Uno himself have done better.
Well, I’ve cut the +3 in half. For both of them I replaced the first +3 with a +2, left the second, replaced the third, left the fourth… this is what came out:

I’ve made sure my work is visible. I crossed out half of the +3 with a diagonal line, and replaced them with +2. Then I did a red circle around the highest and lowest mark, excluding them from the average. At this point I recalculated the new score, writing the modified GOEs next to the real ones. I highlighted two elements by placing them in a red rectangle. They are the step sequence and the last spin, the only two elements in which Uno’s score would have been lower, in the other cases it would not have changed. Adding the new GOEs with the BV and the PCS, Uno would have received 0.28 points less. In a very good program, in this case 9.52% of his marks were +3. On several occasions Uno has achieved higher marks, his record being 30.16% of +3 in the Olympic Team Event short program, and on nine occasions he has received a number of +3 exceeding 20% of his marks.

At PyeongChang Hanyu received 42 +3, the 66.67%. Hanyu received a +3 on over 20% of his marks in 21 different programs, on four occasions over 60% of his marks were +3. This is not even the most sensational case, in the 2015 Grand Prix final 77.78% of his marks were +3, and as for his free skate it reached 60.68%, again in the Barcelona final. What would happen to his score by halving the number of +3?

The final GOE is lowest in six of the seven elements, with a loss of 1.97 points. So it is not true that if we decreases the number of higher marks proportionally for all the skaters, everyone’s scores drop in the same way. Whoever received the most high marks – and the number of +3 received by Hanyu is not comparable to the number of +3 received by any of the others – is harmed the most by the increased severity of the judges.

There is also another consideration to be made. The current +4 is a mid mark between the old +2 and the old +3. In the last spin Uno received three +3 and six +2. I guess it’s a really good spin in which probably something was missing to be able to define it perfect. For a judge to give it a +4 is probably the simplest reaction on an element who deserved a mark between +2 and +3 with the old scoring system. In fact, it means raising Uno’s score. Hanyu’s last spin has a lower base value because the level is 3 and not 4, but let’s leave aside the GOE linked to the base value of the element, a rule that did not exist in PyeongChang and that would have no relevance if both spins were level 4. Most of Hanyu’s spins are level 4, not level 3, so he often receives the highest base value. In this case Hanyu received six +3 and three +2. Again, a +4 wouldn’t be surprising. But for Hanyu assigning a +4 means lowering the final score. Uno received 76.00% of the possible GOE for that spin, with all +4 he would receive 80.00%. Hanyu received 90.67% of the GOE for that spin, but if the marks were all +4 he would have received 80.00% of the GOE. When in doubt, the judges tend give a +4, to level out, which hurts the better skaters and helps the good but slightly inferior ones.

This scoring system was supposed to enhance the quality of the elements? Perhaps, if the judges had used it correctly. But with the + 5/-5 scoring code, and the few +5 assigned, any distinction between good elements and perfect elements was removed. The judges may not have done it consciously, certainly many of them did not realize what happened when the GOE scale changed, but in fact the change has enormously damaged Hanyu.

Let’s look at another situation. Hanyu’s triple axel and step sequence in theOlympic short program was perfect. Assigning three marks other than +5 with the new scoring system means considering it of lower quality, thus diminishing its value. And that’s exactly what the judges did. Now not everyone assigns +5, and the element is given a lower value than it deserves. Otherwise, in cases where the +3 are few, even removing them all does not change anything. I stay in the Olympic Games, but I move on to free skate, and on the marks of US judge Lorrie Parker.

Parker was alone in giving Chen two +3. If she had awarded two +2, Chen would have lost 100.00% of his +3, but the final score would not have changed because Parker’s marks, as the highest marks, were excluded from the average.

Starting from these premises, and knowing that if the number of maximum marks decreases equally for everyone, it is Hanyu who is harmed, did the marks of all the skaters have the same type of trend, or did something different happen for someone? In the first graph we have seen that in the 2017-2018 season 4.04% of the marks was a +3, in the 2018-2019 season 1.43% of the marks was a +5, about one third. Have all skaters seen their highest marks drop by two-thirds? Not exactly.

There is another consideration to be made. In the 2017-2018 season, 38 skaters received at least a +3. In the 2018-2019 season the skaters who received at least a +5 were only 20. Several skaters who received a few +3 did not receive any +5, and among those who retired are Patrick Chan and Adam Rippon, two skaters who received a lot of +3, while Javier Fernandez skated only at the European Championship, and was also not very trained. Considering all these factors, all these highest marks not awarded, the best skaters shouldn’t have seen two-thirds of their highest marks disappear. They should have had a lower drop, especially those whose elements are clearly perfect and not in the balance between two different grades.

So how have the marks of the best skaters changed?

Oh. The line that has the strongest downward slope is that of Hanyu, the skater who should have continued to have high marks because his elements are of high quality. The +5 number Hanyu received is just over a third of the +3 he received in the previous season. Also Brown’s and Fernandez’s marks dropped, but not that much. I’m not very interested in Fernandez scores, and not just because he only made one competition. For the European Championship Fernandez trained very little and even though he was still the best European skater, he was no longer the skater who had won the Olympic bronze the year before.

The difference between Hanyu and Brown first decreased, until, in the last season, the values ​​were completely reversed. And I don’t care that Hanyu skated only at the Olympic Games this season. For the short program the judges, the Italian Claudia Brambati, the Chinese Dan Fang, the Israeli Anna Kantor (+3!), The Japanese Masako Kubota, the Belarusian Ekaterina Serova and the Swedish Asa Norback should explain what the 4T+3T combination was missing in order not to assign +5. Brambati, Fang, Kantor and Canadian Cynthia Benson should explain what was missing in order not to assign +5 to the step sequence. I leave the components for another post, this is already long enough. And, with the exclusion of the quadruple axel and the quadruple salchow, what was missing from the jump elements of the free program? There is not a single +5 among those marks. Why?

As for Chen, the +5 percentage is higher than the percentage of +3 he received in the past. A little in the first season, then it increased more decisively. Last season the numbers were lower, but not much lower than his best season under the old scoring system, and this year the percentage has risen again. The difference between Hanyu and the other skaters, very large before, with the transition to the new scoring system has been almost canceled.

For now I have looked at the +3 or +5 of all the skaters, now I look at some detail. Which skaters have won the most in recent years? I have looked at the most important competitions, Olympic Games, World Championship and Grand Prix Final, starting from the moment Hanyu became the strongest in December 2013. Since then there have been 17 competitions. 7 wins and 8 podiums (in 10 participation) for Chen, 2 wins and 6 podiums (in 9 participation) for Fernandez, 8 wins and 13 podiums for Hanyu (in 14 participation). Fernandez has only skated in one competition with the scoring system +5/-5, so I ignore him. The others always seemed at a lower level, including Uno, the only skater beside Chen and Hanyu to win medals with the +3/-3 system and the +5/-5 system. Sure, he just won a medal ahead of Hanyu, but, Hanyu’s bad luck aside (the hole and the ankle), Uno can thank the fact that the technical panel forgot to check the rotation of his jumps. As for Yuma Kagiyama, he has never competed internationally with the scoring system +3/-3, so it is not possible to make comparisons. So I focused only on Chen and Hanyu.

I compared the scores of all the competitions. I started looking at the percentage of +3 (or +5) from the 2014-2015 season. I ignored the first few seasons because Chen was junior, and because the number of +3 increased in a way that was hard to predict in the 2015-2016 season. I wanted to see Hanyu’s marks before that growth.

I did a competition-by-competition comparison, only on senior competition, in the Grand Prix competitions, Four Continents and World Championship, World Team Trophy and Olympic Games. I have ignored only the Challenger Series. If Chen and Hanyu were present at the same Grand Prix competition (NHK Trophy 2016, Rostelecom Cup 2017), I indicated the name of the competition, otherwise, being a distance comparison, I wrote a generic indication: Grand Prix 1 or 2.

Immediately before the free skate of the 2014 Cup of China, Hanyu was involved in a scary incident. After, he shouldn’t even have skated. It is normal that his marks in that free skate are low, as it is normal that they are also low in the next competition, when he was not yet healed. When he felt better – saying he was fine, with him, is often a stretch, but at least he wasn’t too injured – the +3 count increased. Let’s look at the trend of the lines. From the short program from the NHK Trophy 2015, on only one competition his short program not received at least 30% of +3. Nine times he has achieved higher percentages, sometimes even a lot higher. In his second to worst competition he got 30.16% of +3, a bad result for him. But 30.16% is also the highest number of +5 he was ever able to get. That is, what used to be a bad percentage is now the best percentage. If this is not a claim that, according to the judges, the quality of his elements has deteriorated, I don’t know what it is.

For the first four seasons the elements of Hanyu’s short program were markedly better than the elements of Chen’s short program, for the four last the values ​​were very close. And no, I don’t care how many mistakes Hanyu did. Even with the +3/-3 systems he did several mistakes. When he did a mistake he correctly received low marks, but if the other elements were performed perfectly, Hanyu received the marks that his elements deserved. This does not happen anymore. His triple axel of the World Team Trophy 2021 did not deserve a high GOE, not with that landing, but the quadruple salchow and the combination yes, and instead in 14 marks there were only seven +5 and even a +3. Why?

With the free skate (and, obviously, the sums of the two programs), things got worse. For four seasons, Hanyu’s elements were valued of much better quality than Chen’s elements. Now, according to the judges, Chen’s elements are of better quality than Hanyu’s elements.

I notice a detail. I look at 2017, but the same thing happened at the 2019 World Championship and the 2022 Olympic Games (in the 2019 Grand Prix final the judges simply decided to give the marks they preferred, regardless of what the skaters did on the ice). Nothing of the kind happened at the 2018 Olympic Games, I think, for the simple fact that Chen skated two disastrous short programs.

The competition in which Chen received the most +3 (or +5) is the first that he skated after the National Championship. Isn’t it that these very high marks have influenced the international judges?

December 2016, Grand Prix final (a good program for Chen, better than Uno’s, Hanyu’s, Fernandez’s and Chan’s free skate):

3 +3 (red circle), 41 +2 (orange circle).

January 2017, National Championship:

11 +3, 53 +2.

February 2017, Four Continents Championship:

With two elements clearly wrong, even considering only 11 elements and not 13 as at the Grand Prix Final, the numbers are:

18 +3, 31 +2.

Clearly the US national championship, as the US championship always does (this is not the case for all nations, if you’re interested you can read here and here), has influenced the judges’ marks.

I made a final check, this not on the highest GOEs, but on the GOE percentage obtained by the skaters. With the old scoring system, +1 was equivalent to 33.33% of the maximum possible GOE, +2 to 66.67%, +3 to 100.00%. Now +1 is equivalent to 20.00%, +2 to 40.00%, +3 to 60.00%, +4 to 80.00% and +5 to 100.00%. The calculation on the percentages allows me to compare the two scoring system, but also to combine elements from different GOE, for example a sequence of steps of level 3 with one of level 4. In this way the final score does not emerge, but the behavior of the judges.

The first graph is dedicated to the step sequence (SP and FS) and the choreographic sequence (FS).

The graph speaks for itself, however, to make the checks easier, I publish the percentage data and a table in which I have entered the number of elements that fit into a certain quality range. And if in the four-year period that culminated in Pyeong Chand Hanyu received very high marks quite often, now his marks have dropped, exactly the opposite of Chen’s marks.

For the jumps I made two graphs, one for Chen and one for Hanyu. I had tried to make only one for both skaters, but it was too difficult to read it. I was also in doubt as to how to proceed. The only jump that both of them always did is the triple axel in the short program. No, not always, in his first Grand Prix competition in 2016 Chen did a double axel. For the rest, the jump does not change. But the other jumps? Both have changed both the solo jump and the combination in the short program, and in the free skate there is not a single jump element that both have always executed. So I opted for a quality check, regardless of the jump. I checked the percentage of GOE compared to the maximum possible GOE for that jump on the two best jump elements of the short program, and on the best four of the free skate. In this way I looked at the overall quality of the program, beyond a possible single mistake, even if some strange situation appears anyway. In the short program of Skate Canada 2015 Hanyu managed to have two jump elements deleted, in which case there are only the values ​​of the jump that he performed correctly.

I focused on the elements done well. Sometimes the skaters – it happened to both of them – made several mistakes. There are cases where even among the best jumps there is some jump with a negative GOE. This is something that now interests me little. Negative GOEs are part of the score, but if I want to know if an element has been performed almost well, well or very well, I don’t care to know, when there is an error, how serious that error is. Therefore, to better see the values of the jumps from the positive GOE, I have almost completely eliminated the part with the negative GOE from the graphs. We can see that there is a negative GOE, but I don’t care if the percentage of GOE is only slightly negative or very negative.

Chen is… remarkable. In the last four seasons he has improved not only on the steps, but also on the jumps. I’m not talking about consistency, if I had been interested in consistency I would have looked at all the jumps. Looking at only the best jumps, I’m focusing on quality. And, according to the judges, Chen adds much more quality to his jumps than ever before. And Hanyu?

According to the judges, the quality of Hanyu’s best jumps deteriorated. Hanyu continues to receive several high marks, but these marks are worse than what he received until PyeongChang.

There is another detail to consider. Usually Chen starts from a higher base value than Hanyu’s, and a higher base value corresponds to a higher GOE. This is an example of the solo jump that Chen and Hanyu performed in the short program of the 2019 Grand Prix Final:

Chen has a higher base value and, despite lower marks, a higher final GOE. They are all small things, but combined with marks in the components that I will look at on another occasion, they are more than enough to give the competitions the results we know.