While I was thinking about the marks awarded to Hanyu, Henni147 posted on Twitter a very interesting table in which she compared the marks in the short program of Hanyu and Chen in the last four years, distinguishing the competitions in which they both participated from those in which participated only one of the two.
My approach was different, focused only on Hanyu, although I don’t rule out doing the same type of control for Chen too. I watched the GOEs (only international competitions) of his entire senior career, short program or free skate it makes no difference.
The GOE changes not only according to the mark awarded by the judges, but also according to what was the maximum possible GOE for each element at the time Hanyu performed it. This involved a check on the old scales of value. Not only. In the 2018/2019 season, the ISU moved from the + 3/-3 system to the + 5/-5 system. The marks of the two systems are not comparable. Does a +4 awarded now equal a +2 or a +3 under the old scoring system? We can’t chose one of these numbers for sure. What we know is that a +4 is equivalent to 80% of the maximum score obtainable by the skater, while in the past a +3 was equivalent to 100% and a +2 to 66.67%.
With only one or two marks, it is not possible to make serious statistics, but if we look at an entire season we can check the percentage of GOE obtained by Hanyu with his elements. My first step was to calculate the GOE percentage of all elements. My file is like this:
That’s 1017 lines, taking screenshots is boring, so I don’t post the whole file. At this point I eliminated columns E and F and concentrated on G. I lined up the marks in five different columns: GOE equal to or highest than 66.67% (the old +2), GOE between 33.33% (the old +1) and 66.66%, GOE equal to 0.00 or positive but less than 33.32%, GOE negative and no value.
In this way I was able to calculate, season by season, how high Hanyu’s marks are in each band. Being the positive marks already divided into bands, it is normal for the movements to be small, but it is interesting to note that with the passage of time Hanyu has managed to add quality to his elements.
In each season Hanyu has missed some element (purple line). The worst peak is in the 2014-2015 season but, if we think about it, that season he was the protagonist of a terrible clash before the free skate in the Cup of China. In that program he fell five times, something that, as far as I know, he has never done in any other competition. I found a novice competition, in 2005, in which he fell four times, but after 2007 only once, in Skate America 2012, he fell three times, all other times the number of falls oscillates between 0 and 2. If you take away the data from that Cup of China, the percentage would change.
I inserted the line dedicated to negative GOEs to show that there have always been some, but at the moment it is not the one that interests me the most, so I redid the graph excluding the purple line. In this way, having only the positive values, it is easier to see the evolution from one season to another.
In this graph I have highlighted the 66% line in red and added a red line to highlight the 33%, so you can see the trend of the scores more easily. Vertically I highlighted the Olympic seasons with the blue line, and it is important to remember that after the 2017-2018 season the scoring system was changed.
In general, both the highest scores, those for elements performed perfectly or almost perfectly, and those for elements performed well, have grown. This means that Hanyu is adding more and more quality to its elements. The blue line has two clear dips. The first in correspondence with the 2016-2017 season, the second, a little more accentuated, in correspondence with the 2018-2019 season, and in this second case I wonder if the judges, for elements to which they previously assigned +3, are switched to assigning +4, forgetting that perfection deserves +5. In this sense, on Wikipedia there is an interesting table (and if we all know that Wikipedia cannot always be trusted, two pages dedicated to Hanyu, the main one and the one dedicated to records and results are marked as high quality pages for the completeness and accuracy of the information published):
In the past several elements of Hanyu have been recognized by the judges as perfect, in the last three seasons this has happened only once.
Returning to my graph, the only line that has dropped is the green one, dedicated to scores positive but not so much. This means that Hanyu always adds quality to his elements. If the marks are low, it is probably an element that contains positive bullets, such as a high jump and preceded by steps, but also some little mistake that lowers the final mark.
Looking at these graphs alone, one would think that Hanyu’s scores have grown almost steadily. But there is one thing these graphs don’t say. How often has Hanyu received high marks? I could have done a graph alone, but I think that certain things are better to look at them calmly, in order to see the extent well. I counted the number of elements Hanyu performed in each season, and looked to how many he did well.
The first graph contains three lines, dedicated to elements performed well, therefore with GOE equal to or higher than 33.33%, to elements with only slightly positive GOE and to wrong elements, a category that brings together those with negative GOE and with no value.
Apparently it is easier for Hanyu to mistake an element and get a negative GOE than to just execute one without giving it any value. However, the elements that are done well dominate. If we exclude the season of his senior debut, the blue line is always clearly the highest, and in the last three seasons he has performed almost all of his elements well. Only in the 2015-2016 season he had higher values than now. So, are Hanyu’s current scores the highest in his entire career? Not really. What you find below is the last graph.
I kept the three lines from the graph you just saw, and added other two. One tells us the number (as a percentage) of elements in which Hanyu received a GOE equal to or higher than 66.67%, the other tells us the number of elements in which Hanyu’s GOE is between 33.33 and 66.66%. Look at those lines well. Until last Olympic season the black and orange lines were very far away, the elements with GOE equal to or greater than 66.67% for three seasons (and the 2014-2015 season is a bit peculiar) were more than double those with GOE between 33.33 and 66.66%.
Those lines tell us that a skater who always adds quality to his elements, and a lot, has seen his marks level down, because according to the judges a lot of his elements are good but not as good as until a few seasons ago. Has the quality of Hanyu’s elements really deteriorated?
