After my last post about Ayako Shimode (or Shimoide, I’ve find her name written in both ways, but I don’t know Japanese so I don’t know which transliteration from the kanji is correct), a lady in twitter has written:

My little consolation is that a single judge can’t mess up the overall scores because the highest/lowest values don’t count.

Unfortunately, this is only partially correct. A single judge can’t completely mess up the overall scores, but what if the judges are two? And even only one judge can lower the score.
I’ve made a little experiment. I’ve took my file excel with the transcription of the protocol and I pretended that Shimode didn’t exist. It’s all visibile, I’ve put a line over the marks, but I haven’t deleted them, so you can see which marks I’ve put out of the protocol.
All Shimoda’s marks are deleted, so all the calculation are made on eight judges. I know that in most of the competitions the number of the judges is odd, but it isn’t mandatory. In the recent past, there was several important international competition with only eight judges on the panel: look, for example, at the protocol of the short program of the Pair’s competition at Skate America 2016. In this competition there was eight judges, so I can do this type of calculation.

After deleted Shimode, I’ve deleted the highest and the lowest mark for every element and every components, then I’ve made the sums and calculated the average of the six remaining judges. This is the result:

I don’t know if you can see the results from all the devices, this is a montage of the detail of the scores:

Without Shimode, every mark grow. The GOE grow from 29.56 to 30.63 points, a growth of 1.07 points. The PCS grow from 97.22 to 97.52 points, a growth of 0.30 points. Without Shimode, with only eight judge, the score of Hanyu would be 1.37 points higher. In this competition this difference wasn’t important, but what could happen when the gap among skaters are low?