Read on Twitter
Reading using automated translation, it seems quite accurate when you rearrange the paragraphs as it is meant to be laid out (bc when you copy-paste some sentences will be cut in the middle). He really hits the nail on the head about the limitations of the current judging system.
This is more exciting and distracting to me than Blinding Lights 
He has definitely read our judging discourse. There is literally a sentence directly alluding to it (although not mentioning fans specifically).
Shit he also went after PCS scoring
He also mentions that in big competitions such as Worlds, judges need to score a large number of skaters with the same standards for the sake of fairness.
He is being really candid. Questions whether judges can really evaluate everything in such a short time (and basically asserts that it's hardly possible). Mentions the frequent rule changes. Acknowledges that the burden on judges is immense, yet a lot is left to their discretion.
Legit the most intense Introduction section I've ever read in my life. Go off, King.
He also mentioned that skaters have complained about the judging in interviews. (My own addition: Javi at Euros 2018, anyone?)
The Methods section starts with technical specifications, the PC he used, wifi with 10K mAh battery, battery dimensions, weight, etc. Quite easy to read even from the Japanese text.
He decided to study the loop and flip jumps, because these two jumps have completely different take-off mechanisms.
He also studied the axel (duh) as it is the only jump with a forward take-off. He studied single jumps, triple jumps, and the triple axel.
He performed and recorded each of these jumps twice: 1Lo, 1F, 1.5A, 3Lo, 3F, 3.5A.
From the results it's clear that this is a feasibility study, describing mostly the capabilities and limitation of the system and results from the capture of the above jumps. There is no large body of numerical data.
So he first tested his apparatus by doing a stsq. Sensors worked, but the sensor did not measure horizontal movement as it perceived the skating foot to be in an inertial state (i.e. as if it didn't move). But this seems not to be a problem when jumping.
His equipment displays sensors on body parts in contact with the ground/ice in red, so it's possible to tell where/on which part the foot has contact with the ice, and when/during which phase of the jump.
He says that this will be useful when judging jumps, as it can tell exactly at which point the foot leaves the ice, where the pressure on the foot is at take-off (i.e. edge), whether the foot rotates on the ice prior to take-off.
He writes about prerotated jumps that are often uncalled due to a lack of standard to judge the take-off, which some coaches may take advantage of (no mention of names).
By displaying the center of gravity & quantifying the angle of the edge it would be possible to judge based on numerical values instead of by mere human eyes. He says currently judges can only see from 1 angle, leading to skaters placing F/Lz jumps in the judges' blind spot.
He then proceeds to mention the system's application in calling edges - he demonstrated this by calling the outside edge on a 3F depicted by the stick figure in the picture (i.e. himself).
He also discusses edge assistance on toe jumps and how skaters tend to prerotate this way, which can be recognized by the sensors. Also mentions how a flip done using edge assistance looks similar to a loop (a.k.a. floop 
).
).
He also demonstrated that the more rotations are involved, the more prone the jump is to prerotation. The method he used can tell to what extent this is done.
Oh he also suggests recording and averaging time spent on the toepicking foot in the ideal jump to come up with an acceptable reference range (of rotation on the picking foot).
