I wrote about Thinking, fast and slow by Daniel Kahneman here. It’s not a complete analysis of the book, is only the explanation of some passages from the first 160 pages, and their application to figure skating. The topic are the bias, his influence, and how the bias can born. Kahneman wrote not about figure skating but about human brain, but what he wrote is really important also for figure skating’s judging. I repost here only some passages from that post. After, I’ll add a doubt of mine that I haven’t written about yet.

Kahneman recall a conversation he had with the director of a financial company. The man had just invested a lot of money in Ford stock without asking about their market value, because his intuition had told him it was the right thing to do. As Kahneman discovered during the conversation, the man was recently hit by the Fords he had seen at a recent cars show. He liked the product, he hadn’t questioned the company. According to Kahneman, when the man wondered if it was right to invest in Ford stock, his brain had not been able to provide an answer because he did not have the necessary information. On the other hand, the brain had communicated his appreciation of that cars. From the question “should I invest money in Ford stock?” the brain had switched to the question “do I like Ford cars?”. The man had not noticed that the internal question had changed and when the answer was “yes” he had made the investment. The brain replaced a difficult question with an easy one, and the man did not realize he had received the wrong answer.

Now let’s go to skating. A skating judge must evaluate Chen’s skating skills. The correct question is “are his skating skills excellent?”. If the answer is yes, the vote must be between 9.00 and 9.75, otherwise it must be lower (there is also 10.00 for outstanding, but let’s not joke, even if someone in the US championship managed to go so high). This is the free skate at the 2019 Grand Prix Final.

What happened? The judges didn’t mark Chen’s skating skills, the program was full of crossovers and in the choreographic sequence, which only received +4 and +5, there was also a stumbling. They marked his past results. “Is Chen a strong skater?” “Yes, sure, he won two World Championships (in 2018 the Olympic gold medal was absent due to injury, the silver medal was injured there, the bronze medal was absent because he retired, but these are details, in 2019 his only real opponent was injured again, but this is also a detail) and two Grand Prix finals (in both cases Hanyu was absent because he was injured, but these are always details), so he is strong “. Result? The lowest mark was 9.00, and even that mark is too high for what Chen did on his free skate. As Kahneman says, when we have to face difficult problems, we often respond to an easier problem, usually without noticing that a substitution has been made.

I skip a lot of things that I’ve written in the post that I linked, an automatic translator can make you understand the meaning, and for me it’s important, but now I’m focusing on another possible (I fear really probable) problem.

To judge a competition is difficult, I’m not denying this. The judges has to judge a lot of things in a short time, so one of ISU’s priority must be to give them better technology in order to make their work easier. In the competitions there are several judges – nine in the most important competition – for minimize the effect of an eventual mistake made by a judge. This works only if the judges operate independently from each other.

When a lot of people must make a difficult valuation, they make mistakes. It’s inevitable. But if

the mistakes that individuals make are independent of the mistakes made by others and (in the absence of a systematic bias) tend to have zero mean. However, the error reduction only works well when the observations are independent and the errors are not correlated. If the people share a bias, the aggregation of judgments does not reduce it. Allowing them to influence each other greatly reduces the accuracy of the group estimate.

This, paraphrased, is Kahneman’s explanation.

If all those who give the answer to a difficult question (in which who must answer don’t have enough information) do so independently of each other, and many give the answer, the average is correct. However, if a part of those who give the answer does so on the answer given by someone else, right or wrong, the average is flawed.

If we apply this idea to figure skating, we can say that

if each judge assigns his marks independently (and if they are honest and competent, but for the general discourse I pretend this is always true), without knowing the marks of the other judges, in the end the result will be correct.

So, if a judge talk with another judge, or look at his marks, he distorts the result of the competition. This is the reason for which, as is written in the rules, each judge must work independently, and for which in 2011 Walter Toigo was judged guilty of misconduct and violation of the duties of judges and ISU Code ot Ethics and suspended.

I suppose that Toigo didn’t liked the suspension. For what I know, he isn’t among the judges who was suspended several times, as a suspension were nothing for them. Sviatoslav Babenko, for example, was suspended at least two times, I wrote about him mostly here and here. Here I published a list of judges suspended in the ’70s, and some name is written two times. Probably for them a suspension wasn’t a serious thing, so they don’t changed their behavior and were suspended a second time. But I don’t know all the disciplinary procedures. If you know others, please inform me, so I’ll add to the list.

For now we can let Toigo with his suspension and go on with Kahneman’s book. Kahneman wrote about anchors. In Italian is ancore, I suppose that it’s a simple translation of the English word.

For him, the anchoring effect occurs when people, having to assign a value to an unknown quantity, start to do so from a certain available value. It is one of the most established and recognized phenomena of experimental psychology: the estimates remain close to the number from which the subjects started, and this is why the image of the anchor was evoked.

He made several examples to explain it. One example was about the height of a sequoia (pag. 165). First he asked to several people if, according to them, the highest sequoia was higher or lower of 365 meters. After he asked to them which was, according to them the height of the highest sequoia in the world. He repeated the same two questions with other people, only this time the number that he said to them wasn’t 365 meters but 54 meters.

For the average of the first group of people the highest sequoia was high 257 meters, for the average of the second the high was 86 meters. The difference is of 171 meters.

Impressive.

Now, if a skater earn high PCS in the National Championship, this is an anchor that makes get up his PCS also in international competitions. This isn’t a speculation, this is proven by facts.

These are Chen’s marks in PCS in the 2016-2017 season, his first season among senior.

Noticed how much the marks have risen in the National Championship? Sure, afterwards they lowered, but they didn’t return to the level they were before. The anchor has had its effect. If in the National Championship Chen has mostly obtained votes above 9.00, and even when he remained below him he did not go too far, how is it possible in the international competitions to give him marks below 8.50? The anchor has taken hold in the minds of the judges and the marks after the National Championship were high.

To better understand how much the judges are influenced by the previous competition, I look at his last three short programs in the 2016-2017 season. The competitions are Four Continents Championship, World Championship and World Team Trophy. The Four Continents Championship was the first after the National Championship. I limit myself to the short program because it’s in the short program that the expectations on the skaters born from the previous competition have an effect, at the free skate the skaters arrive with a more immediate response, the score they have just scored in the first segment of the competition. That day the only small problem in Chen’s short was on the quadruple Flip, which received three -1.

Chen performed the best short program, helped by Hanyu who in the combination performed a double Salchow instead of a quadruple. With also the second free skate, he won the competition. Obviously Chen was among the favorites at the next World Championship, also because there are few skaters capable of overcoming 300 points. In Helsinki, however, things did not go well for him, starting with the short program.

He fell on the triple Axel, and this mistake relegated him to sixth place. In the free skate there were two falls, on the quadruple Lutz and on the quadruple Salchow, and three other jump elements received negative GOEs. The result was a fourth place in the free skate, which prevented him from moving up from sixth overall. In the minds of the judges, Chen’s odds must have dropped a bit. The last competition was the World Team Trophy.

On a technical level it was the program that created fewer problems, yet for this program Chen received lower marks in PCS than those of the two previous competitions. The judges no longer expected an extraordinary performance from Chen, and this lowered their marks.

So the expectations, and the anchor, the last result of the skater, and the last marks he has received, are important. For this reason judges that are too generous in their marks in the National Championships must be banned from judge international competitions, they false the international results even if they aren’t present at a specific competition.

And, besides facts (anchor and marks too high, two things that can be proved), and besides my desire for fair competitions that can be obtained, among other things, with a strict control over the judges’ operate, there is another concern.

This is only my speculation, I’m not saying that this is true. It’s a… possibility. A concern. Perhaps the competition in that day was judged with honesty, even if there were a lot of mistakes. And that there were a lot of mistakes is a fact. I’m writing about the Men’s competition at 2019 Grand Prix Final.

An indisputable fact? During the choreo sequence, Chen stumbled. With a stumble the sequence isn’t effortless, and the skater deserves also a deduction. For his sequence on that day, at best, Chen should have received +2. He has received only +4 and +5. This is a huge mistake, done by all the judges.  There are other mistakes, but a real analysis is too long for now, so I will postpone it for when I’ll have more time.

Ok, what I wrote until now is a premise. Think of the anchors. Think of a judge that isn’t so sure of which marks deserves every skater. The easier thing is to look at the marks of other judges. But if you don’t want to do it? Now I go in the speculation territory. I’m not saying that the things are gone in this way. I’m saying that this is a possibility, a possibility present at every competition. If I use the marks of a precise judge for my hypothesis, isn’t to accuse him of bad behavior, but because his marks are perfect for my example.

This was the result of the short program. First Chen with a huge advantage on Hanyu, who has made a big mistake in the combination. If we would like to be precise, both of scores are wrong. Chen deserved to be first, but with a lesser score, Hanyu deserved to be second, but with a higher score but, as I’ve said, I don’t watch the whole competition. In fact, after the short program, the World champion (no, for now I don’t watch neither Saitama) was first, the World silver medallist was second. This is a fact.

Now imagine… you aren’t really able to say which skater is best. You know that both are strong. Perhaps you’ve already has had problems with the Code of Ethics and the ISU’s rules, perhaps some of your friend has had this type of problem. How you differentiate two strong skater, if you can’t honestly say which differences are among them?

Probably the easy way, the way with which you think that will save you from any trouble, is to find an anchor. The World champion is first? Surely, if he do a clean program and he win, there will be no problems. So, your marks will be impeccable. None will contest them. Only +4 and +5 for the World Champion, no +5 and few +4 for his opponent that surely, surely, will arrive second. And PCS given in the same way.

For Hanyu his marks in GOEs are a few over the average, but for Chen is marks are really stunning. And in PCS, his marks for Chen are the highest – in fact he has given always the same really high mark, as for him every component is the same as the others, for Hanyu are the lowest.

The anchor, the need to put Hanyu in second place, was so strong that he is the only one to have managed to place Hanyu in third place in the components, behind Aymoz. He has done it giving to both of them the same marks in SS, TR, PE and IN, 8.75 to Hanyu and 9.00 to Aymoz in CO.

The absurdity of the marks is underlined by the 8.75 given to both in performance. Hanyu has done 5 quadruples. True, he has done only a single Axel, and his 3F was underrotated, but Aymoz has done only 2 quads, fell on one of them and two of his jumps were underrotated. Same performance?

The high marks in the National Competitions are a very dangerous precedent because they set the level of the skaters in the minds of the judges, and if a judge isn’t so sure of his marks, he is influenced more. If ISU want fair competition, must monitor closely all the judges, ban from his competitions the most biased – even if the bias is done in the national competitions – use better technology and train better all the judges. A training that must be done also on a psychological level, making them read books as Thinking, fast and slow.