You brought up an awfully interesting point!
Originally Posted by usethis2
Given that I’m fantasizing about a new format (there are others who discuss the improvement of existing system by means of technology on this thread), that’s what I thought: what if the solution would be exactly the opposite? Imagine that skaters have their strongly technical routines and music is there basically to test their timing: why shouldn’t they skate the same routines to different pieces of music each time then, so testing also the flexibility of their timing? Plus those random pieces could be chosen by computer right before the competition. The skaters might be just given the overall theme to prepare their costumes
I like this idea!
Size 7 Knife Boots
? If a skater does a lot of hard triple jumps he gets high TES scores. If he interprets the music well, displays interesting and effective choreography, and (as the rule book puts it) "generates energy that establishes an invisible connection with the audience" then he deserves high scores in Interpretation, Choreography, and Presentation, respectively. This does not seem controversial to me. If he does a lot of Transitions, with variation and high quality, then he gets a good score in Transitions.
Originally Posted by CanadianSkaterGuy
In the best of all possible worlds, the winning skater would be the one who does the best on both sides. Sometimes this doesn't happen. If no one lands a bunch of hard jumps AND presents something of value on the performing arts side, then the judges have their work cut out for them. If this NEVER happens, then figure skating will disappear as a spectator sport (and deserve to, IMHO).
Last edited by Mathman; 03-13-2014 at 06:23 PM.
Last edited by Mathman; 03-13-2014 at 06:25 PM.
Lots of good points there. However, regarding the numbers, I think it's important to also keep in mind the purpose of the technology or any other measurement method.
Originally Posted by os168
Ultimately at the end of the day it's about how to rank skaters. For whatever psychological reason, people want to see that (i.e. who was "best" or 1st place, 2nd place, etc.). Scores are the means rank the skaters. So the discussion is really about, what is the best (or at least, what is a better) way to do the scoring to rank the skaters. As mentioned already technology is one way but it's not the only way and it can't measure everything (i.e. things like artistry). However, even with technology, there's something else that has to be done by humans: the scoring. What I mean is that technology can at most only measure a quantity. It cannot provide value (i.e. score). At most the technology will be able to say for example "during that jump the skate was -83 degrees away from three full rotations." It is up to the humans, i.e. the experts, to say "well then it should be worth X points." Sure, once there's a measured quantity, then the computer can calculate a score based on that quantity, but how to calculate the score -- the relationship between the measured quantity and the value in points -- is still given by a human, i.e. expert.
The advantage of using technology is to reduce the error in human judgment about those measured quantities -- rotations, edges, and so forth. It cannot replace human judgment about what's valuable and what's not. This thus increases the transparency of the process. Under the current system, far as I know, wrong edges and such are called by the technical panel. Although we know their guidance via the rule book, their deliberations are opaque to us (we don't know the specific reasons for why a wrong edge was or wasn't called). Thus the process is not only prone to human error, but it's also prone to influence (i.e. politics/corruption/bias), and we are unable to distinguish between the two. By having technology make the measurements, the process reduces that problem -- we could have something like "that jump was -103 +- 2 degrees away from a full 3 rotations". Presumably the technology is better than the human eye at making those measurements (otherwise it's not useful) -- hence using it where appropriate would increase the reputation of the judging process. The expert opinion then would be on how difficult each move was (i.e. how many points it would get) -- backed up by the actual numbers to ensure that human error, bias, politics etc. wasn't introduced to unfairly skew the results.
I don't think potential ticket sales is necessarily a good way to rank skaters. In the (classical) piano world, Lang Lang is by far the biggest moneymaker. But many pianists have serious reservations about what he's actually contributing to the field compared with current pianists such as Hamelin, Tsujii, or perhaps Yundi Li. He's certainly talented, but in classical music he's known more for his facial and physical antics (and good enough technique to play the right notes while doing them) rather than breaking new ground in expressiveness, artistry, conveying emotion, musicality, etc. The sad reality is that for many classical pianists, if money was the primary metric, then they should just play background music for voice shows like American Idol and forget about classical music.
Originally Posted by Anna K.
To a certain extent, yes, any endeavor needs to consider whether or not it's able to bring in money, for survival if nothing else. But I don't think the ability to get money is necessarily the same as quality.
Although the technology is available in a smartphone, I don't know if it's robust enough to work inside of a skate, which is subject to quite a few big shocks. I also don't know if skates are the best place for the technology, since it is in the skater's possession and thus more easily tampered with. I'd lean more toward an external solution such as cameras with image-recognition and measurement software. Although this would cost money, I'm not sure it would be all that much. After all, cameras are already in use at a lot of places, especially for more high-profile events. Just consider the cameras that the technical panel has (actually, I don't know how many they use). So it's not a matter of going out and getting new hardware.
Originally Posted by I♥Yuna
The main difficulty is in the software. There's a fair shot I could write it myself if I had the time (I have some background in image processing and wrote an optical character recognition program to analyze a game before). The video analysis I would think is fairly straightforward; edge detection to identify the skate location in the image, then some processing knowing the location and some other parameters of the camera(s) to calculate the skate's location in space. Probably even easier if the camera can resolve the ice, for example the figure that the skate is leaving behind as it moves, as a reference point. That would be a hardware requirement on the camera though, and also somewhat a software requirement on the video quality (for example, it likely couldn't do this on videos from Youtube since they're heavily compressed). Then it's just a matter of doing this for multiple frames and eventually you can work out the skate's location in space over time, and thus calculate the relevant quantities (such as edge and rotation). It seems like an undergrad project, and once completed, it could of course be copied to wherever it's needed. Maybe I'm understating the difficulty of this, but I don't particularly see where there might be major problems.
If the system doesn't work because of a breakdown or whatever, then you just revert back to how it's done now -- humans eyeballing it from replays. The system is meant to be an improvement, not the only way that's allowable. I would hope that even if the system is in place then the judges are still competent enough to be able to judge events if needed, just that the system would provide for more accuracy (and reduce judge workload) when available.
Also, although the system would have a cost, so does the current system of judges. Devoting manpower to measuring objective quantities has a cost in itself, not necessarily in terms of money (the judges don't get paid for it, right?) but in terms of the judges you need to have available, their own transportation etc. costs, and the opportunity cost of them not being able to do something else. The business case then is a tradeoff in paying a one-time cost of getting the software written (perhaps licensing fees if it's outsourced) and possibly making sure the right camera equipment is available, compared with the expenses for the manpower involved in manually measuring the quantities with judges, as well as the potential error and bias issues (i.e. legitimacy issues) with using humans as the measuring sticks. I don't think it's that unreasonable. Again though, it comes down to whether or not the organizers would actually be willing to relinquish the influence that they have over the outcome of competitive events. As I said above, the potential for human error is one way to cover up bias, corruption, etc.
I'm sure there may be ways to cheat the system. The software's acceptance will depend on its reputation as being able to accurately and neutrally evaluate different skaters. Coincidentally or not, judges have the same issues -- with human factors thrown in.
Originally Posted by usethis2
I wouldn't say so. I'm not a figure skating expert but I would say it's more of an art form. The technique should be used to convey something, rather than being the end in and of itself, just like with any art. Take painting for example: pointillism (where the painting was made up of lots of little dots) was exciting not only because it was a new technique, but because it was able to convey a different visual effect than previous painting techniques. (Incidentally, it foreshadowed the technological use of pixels many decades later.) In piano playing, playing scales is a technique that everyone learns, but no one other than the pianist's parents will want to hear them play scales. But playing scales quickly does have its place as part of a piece in adding to the excitement; take for example, Hamelin's playing of Liszt's Hungarian Rhapsody 2 (http://www.youtube.com/watch?v=pIMzL2-4bjg at around 7:45; that's 21 evenly-spaced notes in 1 second each run that you're seeing/hearing, or a new note every 0.05 seconds) where scales are used as part of the "exciting" section of the piece. As someone elsewhere mentioned, if figure skating were just about the jumps, then it just needs to be a 10-second thing of the skater doing a jump. The choreography and artistry is what ties everything together and weaves a bigger story with more emotional impact than just the technical elements.
Originally Posted by usethis2
Very cool post, Vanshilar.
It depends if we're talking about individual success or the development of the branch. For an individual, doing eveything to increase ticket sales might well be destructive because it may lead to the decline of personal performance level. For the branch, it can't develop without attracting more money and investing it in the quantity and quality of events. Or, can it?
Originally Posted by Vanshilar
^ The argument about money seems to be between the people who long for the good old days when a tip-top skater could become a millionaire, and the "realists" who are content for the sport to settle into its natural niche as mainly a participatory sport. Is it a sport for the athletes and their families or for the spectators?
Sorry I didn’t answer this post earlier!
It’s so much better than a Luddite post because you hit the point – two points! – by stating where technology in FS stands today.
Originally Posted by Mrs. P
FS is a sport with a history and I believe that many on this forum wouldn’t mind to travel back in time to enjoy it. However, here goes why I think the days of real time judging are over and it’s basically the question about how many pains it will take to admit it.
Corruption and other aspects that are not related to sport aside, every sport is a physical activity; hence in every sport there are borders of what a man/woman can physically do. It takes some period of time until these borders are reached. Till then, the sport develops rapidly but when it’s there it’s only about a detail, a hundredth of a second, that decides the victory and results of top athletes are very close. Please correct me if you think that FS isn’t there – because I sincerely think it is - and for FS it means tons of details that can influence of decide the outcome.
In fact, it was because of one of forum members admitting it in a different thread:
that actually inspired me to start this discussion. FS is crying for technological aids; however, the discussion what technologies and to what degree should be involved is obviously still open. I’d say that the use of video replays has many imperfections and I’d vote for more radical technology involvement.
Originally Posted by gkelly
To be fair to gkelly, here is the post in another thread where she explains the abovementioned post and her real opinion about technology:
I’d say that every novelty, including technology innovation, causes the interest of the public. So, if the use of technology is inevitable, then why not to make an advantage of it and let it draw the modern profile of FS?
Originally Posted by gkelly
By the way, other than technological innovations on the sport itself, another aspect to consider is how the judging is done. There are likely software improvements that could be done under the current system. For example, under the current system, the GOEs are supposed to be based on a well-defined set of criteria, such as height, unexpected entry, etc. If this is the case, what if for each element, the judges input which ones they observed? Then the software automatically counts up the number of observed criteria and calculates the GOEs based on them. Then it would be more transparent as to what criteria the judges specifically used in their scoring, even if the judging were still anonymous (i.e. no changes to the rules of the current system). Otherwise under the current way of reporting the judging results, it's still a bunch of opaque numbers that don't provide much meaningful feedback as to where a skater could improve, not to mention allow for a lot of politicking.
Well, I have one better. Sometime in the future computers should be able to recognize every human move relative to the surroundings hence they will automatically output GOEs according to set criteria. We have facial recognition and retina recognition today. It will require some serious CPU/GPU power and serious programming but one can dream. (Sadly even that system will leave PCS to the judges. )
Originally Posted by Vanshilar
To be precise, technology already is able to recognize every move according to the set criteria, like, facial recognition.
Originally Posted by usethis2
And, such system would leave only part of PCS to judges. Let’s see the criteria that exist under CoP and give a yes to what can be done by a machine and no to what can’t:
* Balance - yes, rhythmic knee action – yes, and precision of foot placement -yes
* Flow – yes, and effortless glide –no, it’s subjective what’s effortless
* Cleanness and sureness of deep edges, steps and turns - yes
* Power/energy and acceleration - yes
* Mastery of multi directional skating - yes
* Mastery of one foot skating - yes
* Variety - yes
* Difficulty - yes
* Intricacy – yes, if comparison with programs of other skater counts; no, if intricacy is subjective or undefined
* Quality – no, it’s undefined what’s quality; in case it’s defined - yes
* Physical, emotional and intellectual involvement – no, it’s subjective
* Carriage - yes
* Style and individuality/personality – no, it’s subjective
* Clarity of movement – no, it’s undefined; in case it’s defined - yes
* Variety and contrast - yes
* Projection – no, it’s undefined; in case it’s defined - yes
* Purpose (idea, concept, vision, mood) – no, it’s subjective
* Proportion (equal weight of parts) - yes
* Unity (purposeful threading of all movements) – purposeful no, threading yes
* Utilization of personal and public space – no, it’s undefined; in case it’s defined - yes
* Pattern and ice coverage - yes
* Phrasing and form (movements and parts structured to match the phrasing of the music) - yes
* Originality of purpose, movement and design – yes, if comparison with programs of other skater counts; no, if originality is subjective or undefined
* Effortless movement in time to the music (timing) - timing yes, effortlessness no
* Expression of the music's style, character - yes, if comparison with programs of other skater counts; no, if originality is subjective or undefined, and rhythm - yes
* Use of finesse to reflect the nuances of the music – yes
Except a half-point of SS, two points of five in PE, one-and-a-half point of CH, and a half-point of IN everything else are yes or possible yes. Think about it!
Could such recognition programs measure these qualities on a scale of 0-10 (with decimal increments), if the qualities are properly defined?
For the No answers that you've put in blue (except maybe "effortlessness"), and also Originality and Expression of the music's style, character, those tend to be the aspects that casual fans are most interested in, the least technical areas and the ones that make skaters stand out as individuals with individual personalities. And of course hardcore fans love those aspects as well, although fans of specific skaters tend to be heavily biased.
Should those qualities be removed from the scoring entirely so there would be no more judges, and individuality and charisma would become irrelevant to who wins?
Let judges address only the subjective aspects? They'd have less to do, so they could do it better? Although it would all be subjective.
Let fans vote on those aspects, turning those parts of the scoring into a popularity contest, but often overridden by all the machine-measurable objective aspects?
- of course, a program could award points but should it? It’s also possible that the rank 1 in category A and 5 in category B could make rank 3 in total. Imagine that you have complete data about a just skated program with its every nuance: it would be much easier and more objective to rank skaters according the data. A computer could compare the line left by footwork of one skater to that of another, or put silhouette on silhouette. It would be also an interesting material to view if available on the internet.
I honestly think that skaters would have more creative freedom if their artistry wasn’t predetermined by judges and I also believe that a technology judged competition could exist without the subjective part of judging. However, that might depend on the discipline. If it’s about the deliverance and development of a certain style that doesn’t prescribe much freedom anyway, then the judgment of style experts might be applicable. Like you said, judges might have less to do and so they might do it better. Instead of nine anonymous judges there might be three or four personalized judges who would be given an opportunity to express and defend their opinion. It would be still subjective but it would be transparent.
Plus, technical knowledge won’t be necessary to be a judge so celebrity judges might be invited for commercial or less serious events.
The fan involvement is another aspect that is possible because of development of other kind of technologies, the media. Of course, a sports competition is not a popularity contest but the Gala could be: why not to invite to the Gala only those athletes who prove to be popular according to an internet vote? I think that would make a sense.