Well, it's inherent to evaluation of figure skating since the beginning, made slightly more precise and certainly more granular since the beginning of IJS, but not to the point of matching the precision of scoring in less complex or less qualitative sports.
6.0 short program judging had mandatory deductions (many of which included a numerical range, as do many of the GOE reductions), but other than that the guidelines for scoring programs under 6.0 were much less specific than they are under IJS. Judges go through years of trial judging and judging lower level events as they work their way up to international and if possible ISU judging appointments. During that process -- and before, if they were skaters themselves -- they get a sense of what is "average" or "good" or "very good" for each kind of element and for each overall program-wide quality or criterion (program component). What to expect as average for novices or juniors vs. seniors. For men vs. women.
That standard is always going to be a work in progress as judges get more experience watching skaters at all levels, and as the field of skaters as a whole tends to improve dramatically at some kinds of skills and to let other skills that are no longer considered as important to fall by the wayside.
And then they're supposed to consider all the good points and all the bad points in coming up with their scores.
Under 6.0 there was one technical score for everything in the whole program -- judges had to weigh the good qualities and the errors or weaknesses and the difficulty of the content and balance it all out to come up with one score. There were no explicit guidelines about what was more important and what was worth less. Each judge had to decide that for themselves, influenced by what they learned from more experienced judges through the trialing and early career judging process.
With IJS, it's just the good qualities and errors or weaknesses of one element at a time -- with tech panel/scale of values taking care of the difficulty determination. So that's a lot more precise than 6.0 judging. But there are still caveats that the range of quality from very poor to outstanding is continuous but the scores available for GOEs are discrete integers, errors can occur with varying degrees of severity (which is reflected in the rules where a range of deductions/reductions is suggested), and positive qualities can also occur with a range of quality from above average to good, very good, excellent, and outstanding. (The +GOE bullet point rules either +3 or +5 do not explicitly allow for rewarding one point for "good" and more for "excellent" execution of the same quality, as the earliest GOE guidelines ca. 2003-05 did -- perhaps calling the bullet points "guidelines" is a way to allow for that flexibility.)
True, the 6.0 system was also very vague. Sure, they had mandatory deductions in the short program on the technical side, but judges could be quite lenient or severe, or some judge might have caught a flutz or double-footed landing while others didn't. I think the IJS system is overall an improvement; I just think that it needs to take a step further and provide data to clarify some of the adjectives used to assess GOE.
Clarity in the form of spelling out ranges of positives, or giving video examples of poor, fair, average, good, excellent? Stating explicitly where judges have room to exercise judgment or to balance out strong positives, mild positives, strong weaknesses/errors, and mild weaknesses, vs. where there is only one correct final GOE?
Up to a few years ago there used to be a number of errors that required negative GOE or required -3 GOE especially in short programs, but then the rules/guidelines were explicitly loosened to allow judges to offset reductions for errors with positive qualities.
Now the only mandatory GOE is for doing only one valid jump in the SP jump combination slot.
Providing video examples would be a good start for clarity (for the viewers too), not only for qualitative GOE purposes but also for assigning levels to elements. For instance, what are the judges actually looking for when assigning levels to footwork sequences and the key points in the compulsory portion of the (now) rhythm dance?
As for the webinar itself, I would have liked to have seen more video examples of negative GOE elements (or elements that have both positive and negative GOE), as opposed to explanations of each positive GOE bullet point.
If you're going to measure height and distance on jumps or rotational speed on spins, should the cutoff between "good enough to earn this bullet point" or not be the same for a 4'10" woman or a 6'2" man? Or even skaters of the same sex and skill level but with a full foot of height difference?
Or should the technology just tell judges the number of inches/centimeters or the number of RPMs and let judges draw their own mental guidelines – to which they could then be more consistent in drawing their own lines as to when to award the bullet point.
Spin centering is less dependent on the size of the skater, so that could be more absolute. But could the technology account for the inevitable change of circle size that occurs with a switch to a forward edge while spinning, so that change-edge spins aren’t penalized for a slight change of centering that would be penalized in a backward-edge-only spin?
Similarly, could the technology have different settings to assess the rideout on the first jump landing of a combination depending whether the subsequent jump is a loop or a toe loop?
Etc.
I’m sure it would theoretically be possible to account for all these variables. But at what point does the gain in precision justify the cost of developing and implementing the systems and technology?
Would it make sense just to determine certain aspects of a performance that can be measured absolutely and just plug those numbers directly into the scoring with no consideration of the size of the skater or the exact element being performed, and let the judges concentrate on judging all the qualities that can’t be measured objectively at all?
The point of the technology in my mind would be providing precise information and data to judges (and the viewers), not as a sort of mandatory requirement where this skater has to, say, achieve X meters to get a +2 on a jump, or something like that. The judges will indeed have to translate the given numerical info into the GOE, accounting for gender, size, and such. It's not that I think the judges aren't knowledgeable or all that inconsistent really, but it's just to provide some factual, measured grounding and context to some of the GOE marking. I'm the kind of person who loves to see replays of top-scoring elements where it's clear just how far a lady went in a throw jump or how much height a skater got on a 3A.
I think that ultimately this kind of technology will help the skaters improve on the GOE side of every element. All of them already know that they need to work on more speed/ice coverage, fast spins, higher jumps, etc. but the data would help them (and us) figure out what it will actually take to get those coveted +2s, +3s, and so on. And with enough data over time, the judges themselves will come to determine the general numerical ranges that they expect to see for a "fast" spin, "good" ice coverage, etc.
) but the video was very detailed and comprehensive. I like the speakers - they are really kind to skaters with negative features, and they try going for some humour so it's not too dry and monotonous. 
