I’m sorry for my mistake that cost you your victory, but in the end, it’s nothing serious. For me it is not a problem. I make the rules, so you must only accept the injustice and keep quiet.

No, no one said these words, but the concept is this. Let’s start from a recent competition. In the free skate of the European Championship, the Russian judge Olga Kozhemyakina gave a not particularly high mark to Tarasova/Morozov’s triple twist.

Kozhemyakina is an expert judge – she could be a judge in the next Olympic Games – so she is able to understand what mark deserved the skaters. Clearly here she did a mistake: she pushed the wrong button. It’s almost sure that she would have awarded a +5. With the mistake, the lowest mark was hers, so the +4 of the judge 9, Helene Cucuphat, entered the score. Because of the mistake, the GOE for the twist was 0.17 points lower than it deserved. The mistake didn’t influenced the result: Mishina/Galliamov won the free skate for 2.61 points, the competition for 3.39 points.

What’s the problem? The problem is that even if a judge makes an obvious mistake, that mistake is not corrected. According to Alexander Lakernik, vice president of the ISU, there is no way to correct such a mistake. In an interview in Russian that I translated with an automatic translator I found this nice statement:

Can this be repeated in other competitions, including the Olympics? Maybe. But everything will remain as it is, even if there is 0.05 points between the pairs. We have such rules.

These are the rules. But the rules are written by the ISU, and if they don’t work they must be changed. Was it possible to predict that such an error would occur? And how likely is a major competition to be decided by such a small difference?

I checked the difference in scores when there was a medal involved (so I also looked at the difference between third and fourth place, and in some cases also between third and fifth) in the Olympic Games, World Championship, European Championship, Four Continents Championship and Grand Prix final. Only senior competitions, all disciplines. My list only includes occasions when the difference between two skaters or pairs of skaters is 1.00 points or less. I have highlighted with the blue boxes the occasions when the difference is less than 0.17 points, therefore a smaller difference than the mistake made by Kozhemyakina.

There have been several Olympic medals, some even gold, decided by a difference of less than 1.00 points. Several world medals, some with a difference lower than 0.17 points, with the most sensational case occurring at the 2014 World Championship. When there are several scores close together, the names of the skaters are surrounded by a black line. On that occasion Cappellini/Lanotte won gold, beating Weaver/Poje by 0.02 points. If they had mistakenly received 0.17 points less, not only would they have lost the gold. They should have been satisfied with the bronze, as the difference between them and Pechalat/Bourzat was only 0.06 points. The difference between success and not winning any medals was above the point, but not by that much. As for the 2019 World Championship, if Evgenia Medvedeva had scored one more point she would have overtaken Elizabet Turynbaeva and would have won silver, not bronze, if she had scored one point less she would have finished fifth, overtaken by Rika Kihira and Kaori Sakamoto. And at the European Championship 2018 the bronze was won by Zabiiako/Enbert that overcome James/Cipres by 0,01 points.

Apparently we have several competition whose outcome was decided by a minimal difference. How often an error as this happened? We don’t know, there’s no way to know it. If Kozhemyakina had pressed the button that assigned a +4, no one would have noticed anything. We might have considered she was slightly severe comparing to seven judges, but nothing striking. Tarasova/Morozov, however, would still have lost those 0.17 points. How many small mistakes have there been, and are there in every competition? Because there are mistakes, all the time, even if we rarely can notice them.

Contrary to what some seem to think, I don’t look at all protocols. How many really weird marks are there that I haven’t noticed? I don’t know, if someone will point out others I will add them here. Meanwhile, I notice three strange marks, not in the GOE but in the components.

The first protocol comes from Skate America 2003, the first competition in which the ISU Judging System was used. There is an obvious mistake in Sasha Cohen’s marks. We are talking about over twenty years ago. Is it possible that in all this time no one has worried about the risk of having competitions conditioned by errors of this type? Is it possible that no one has found a solution?

The second protocol is of the French national championship of the 2006-2007 season. It is a national championship, not an international competition, but a serious federation should report any problems. Has the French federation informed the ISU that the competitions could be distorted by errors of this type? If yes, was the ISU interested in the information?

The third is the 2020 Junior World Championship, and it proves how, in all these years, nothing has changed.

Mr. Lakernik is also aware of another problem, as demonstrated in another interview:

– At one of the stages of the Grand Prix there were marks from -1 to +4 for the same jump. Such assessments can be?

– Not. The situation you are talking about clearly requires discussion. There is an accepted assessment corridor, and it is rather soft. From the middle, 1.5 points in both directions. That is, if the grade for an element is +1, then upward is acceptable +2, and downward – 0.

So a difference of 5 units between the highest and lowest mark in the GOE is wrong. But that never happens, does it?

We have seen a so big difference already with the old scoring system, the +3/-3. This is Mikhail Kolyada’s free skate at the 2018 World Championship:

Almost all -3, a -2… and a +2. On a jump in which Kolyada fell. I assume Judge 2, Tarja Ristanen of Finland, pressed the wrong button. But what should I suppose with Tarasova/Morozov’s free skate at the 2017 European Championship?

For the Azerbaijani judge Irina Medvedeva (+3) the lift done by the pair was perfect, for the French judge Anthony Leroy (-3) it could not have been worse. The marks are mostly positive, but there is another negative mark and also a 0. How was the lift of the pair? From the marks it is impossible to tell for sure.

There were also some mark that didn’t respect the rules. Back in 2013 in the SP for an element different than the required element, was mandatory to give a -3. It seems that at the World Championship with Javier Fernandez not all judges remembered the rule.

Forgotten rules aside, if in the past I had come across some difference of five units between the highest and the lowest mark, we must remember that figure skating is a creative sport and that we cannot limit the imagination of the judges. Therefore we should not be surprised that now we can find even greater differences.

This is the 2018 Junior Grand Prix held at Ostrava:

If you want to have fun you can also watch the Ladies’ free skate, and see how strict Judge 9 was with almost all the skaters.

Beyond the Kozhemyakina’s mark, the biggest difference I found – it’s almost impossible to do worse – is in Ziegler/Kiefer’s short program at the 2018 Helsinki GP.

In almost all the protocols I have highlighted an unacceptable difference on a single element, in one case the difference occurred on two elements. These are errors that happen repeatedly but are not corrected. Find a solution? For example, make that a light comes on on the judge’s monitor, during the review phase, if one of his marks is far from the average of 5 or more units, so that he can check the mark he awarded and possibly change it.

And what can happen when there are smaller differences, and therefore considered acceptable by the ISU, not on a single element but on all or almost all? For the 2019 Grand Prix Final I look at the protocol not in the ISU version but in the SkatingScores version. The number of wrong marks awarded in this competition is so high that it could be written about for a long time, but at the moment I will focus only on those awarded by judge three, the Austrian Elisabeth Binder, to Yuzuru Hanyu and Kevin Aymoz.

In almost all cases, the mark awarded by Binder remains within a margin of +3 or -3 compared to the marks assigned by the other judges, so there aren’t big mistakes. The only exception is in Hanyu’s first combination, with her mark being four units lower than the less severe mark. Is everything alright? Lakernik would probably say yes, yet if we look at the marks I have highlighted it is evident that something is not right.

The red circles indicate the marks in which Binder was the most severe judge. With Hanyu she was the most severe in five out of twelve elements, almost half, with Aymoz she was the most severe in one element. The orange circles indicate the marks in which Binder and only one other judge were the most severe. One circle for both skaters. The dark green circles indicate when Binder’s mark was higher than the skater’s average mark. In Hanyu’s protocol there is none, in Aymoz’s there are two. The light green circles indicate when Binder’s mark was exactly averaged with the marks of the other judges, and if with Hanyu she never did it, with Aymoz she did it twice. Binder’s marks are almost never sensationally different from those of the other judges, yet with Hanyu they are always downward marks, while with Aymoz there is some low mark, but the high marks are more.

And if we look closely, with Hanyu Binder was particularly strict on all quadruples, the elements from the highest GOE, while with Aymoz she was generous on two jump elements, one of which was a triple axel, and for the other triple axel was in line with the other judges. Only on the first of Aymoz’s quadruple toe loops she was strict (but not really strict, there is a difference of two units with the highest mark, on Hanyu’s quadruples the difference is three units in three cases, four units in the other). On the second quadruple toe loop Aymoz did not complete the rotation and fell. Between the +2 which was the maximum possible starting mark if there are at least two bullets, the -5 for the fall and the deduction between -1 and -2 for the <, the only possible marks were between – 4 (being generous with Aymoz) and -5, so the judges did not have a lot of space on this mark to vary the GOE.

The highest marks that Binder has awarded to Hanyu are three +3, in a spin, in the steps and in the choreographic sequence. On the other hand, he awarded Chen a +5 for his choreography sequence with a stumble – on this element all the judges gave the wrong mark, because there is a clear stumbling from Chen and yet the marks are between +4 and +5 – and a +4 for his quadruple salchow. According to her Aymoz deserved a +4 on the triple loop, on the triple axel and on the choreographic sequence. The other skaters did not receive marks higher than +3 from Binder, but Samarin received a +3 on the starting combination 4Lz+3T, Jin received one on the combination 3Lz+3T, Aliev one on the quadruple lutz. Apparently Hanyu was the only one who failed to give quality to any of his jumps. Yet, as I remember the competition, up to the first combination, actually wrong (triple flip underrotated and with step out), there were no problems.

If we look at the purple circles, Hanyu underrotated a single jump, and on that line I also indicated his step out. Aymoz underrotated two jumps, and in the first case I indicated the fall. It is true that at the end Hanyu was exhausted and failed to perform the combination 3A+3A (by the way, his single axel was preceded by a counter, wide and on the music, and did not contain any errors worthy of a deduction, the GOE correct is +3. A small difference, but all marks should always be correct), but he still completed five quads. His 1A lower the BV, prevent to him to have a higher GOE, but even not counting that element the other elements are more than enough for a high GOE. Aymoz tried only two quadruples and on one he fell. Hanyu’s base value is 5.10 points higher than Aymoz’s. How is it that according to Binder Aymoz deserved a higher TES? Since when does Aymoz manage to put so much more quality of Hanyu on the elements? A Hanyu who proposed only one wrong element, against Aymoz’s two, and with a less serious error than Aymoz’s error on the quadruple.

Binder was stricter on the the GOEs than on the PCS, strict with Chen (who in truth received really generous and undeserved marks from all judges, so it’s impossible to watch his marks in a comparative way. The only possibility to judge what he deserved is to watch the marks and the way in which he skated) and Hanyu, generous with Aymoz. For my check I don’t watch the SP scores (but seriously, +4 on 4S? Why? What bullet was lacking? And… the same mark that for Chen’s 4S in FS? How it’s possible? She really watched the two jumps? Higher marks in the spins for Chen? No, I really don’t want to watch the marks).

I did a closer check on Binder marks and calculated Hanyu’s FS score as if there were only eight judges. I publish the complete protocol only of the technical elements. Above we see which marks came out of the Hanyu’s real score calculation. Below I have deleted all of Binder’s marks, then I deleted the highest and lowest of the remaining marks. Without Binder, Hanyu would have received 1.39 points more for the technical elements. As for the components, in the first of the two tables at the bottom right I entered only the final data, the real ones and those excluding Binder’s marks. Binder lowered Hanyu’s PCS by 0.28 points, the free skate total is 1.67 points. Look up and see how many important medals were decided by a difference of no more than 1.00 point.

Is it normal that there are more severe judges than others? Sure, but there is severity and severity. Both Hanyu’s 4T+1Eu+3F< combination with step out and Aymoz’s 4T< combination with fall received the same -5, but the two jumps are different.

To understand better, I also looked at the overall marks. We must remember that in this competition several judges have awarded strange marks. For Judge 6, Walter Toigo, Chen only deserved +4 or +5, his marks are really inflated. Aside from the stumbling in the choreographic sequence, which shows that there was a general blindness to the quality of Chen’s elements on that day, Chen’s jumps are small, by rules he does not deserve marks higher than +3, but all judges had given him at least one +4, and nearly everyone had given him several +5. We can’t take his marks seriously. Toigo also managed to give Hanyu three marks in components below 9.00, I suppose an underrotated step out jump is a serious mistake, because for Hanyu these are low marks. On the other hand, two underrotated jumps, one with a fall, are not serious errors, since Toigo assigned the same marks to Aymoz (except in composition, where he assigned 0.25 points more to Aymoz), and for Aymoz these are high marks .

The screenshot I publish below contains three tables. I start from the one above. Thanks to SkatingScores I easily checked how many points each judge assigned to each skater for technical elements. These points are written in black. The boxes in red are the result of an operation, as I highlighted they contain the difference between the score assigned by the judge and the base value. In practice, in red we see how high the GOE assigned by each judge is. The last column is dedicated to the average of all judges, the value that we seen on the protocols.

The table in the center contains only the GOEs, I have kept the red to highlight that they are the same numbers. Some boxes have a black border. They are the highest and lowest GOE received by each skater.

In the last table I kept only these marks and I moved the values ​​that were previously in column II to column HY, so as to have two columns dedicated to differences at the bottom. In column II you will find the differences between the lowest GOE received by the skaters and the highest GOE, in column IJ the differences between Binder’s score and the real score.

In three cases the lowest GOE is that of Binder, but if with Aliev the difference is acceptable, in the case of Hanyu (and, to a slightly lesser extent, Chen) it is a huge difference. I remember that a quadruple lutz, the jump with the higher base value ever completed and performed in this competition by Chen, Hanyu, Samarin, Jin and Aliev, has a base value of 11.50 points. A quadruple lutz (11.50) and a triple axel (8.00) are not enough to bridge the gap between the highest and the lowest GOEs, and a quadruple lutz is not enough to bridge the gap between Binder’s GOE and the average. How can such a large discrepancy be acceptable? How can it be, when Binder shows us that a judge can be strict with one skater, generous with another, and turn values ​​upside down?

The ISU must absolutely find a way to prevent errors such as the one that occurred at the recent European championship with Tarasova/Morozov, but also reduce the decision-making power of the judges, preventing such high discrepancies. Otherwise sometimes the best skater will win, sometimes the one who will have more luck with the judges’ mistakes, sometimes the one that is most likeable to the judges who judged that particular competition.