Artificial intelligence and gymnastics scoring - your opinions please

The Guardian yesterday has an interesting article about the FIG's plans to introduce computer judging to gymnastics.  I suggest you give it a read.

It won't surprise you to know that I have lots of things to say about this; I think that AI can contribute something to scoring, but not to judgement.  If the implementation of this initiative is not managed with common sense and imagination, we could find ourselves with a sport that is even more devoid of artistry and the aesthetic.  

In response to a comment asking for evidence of biased judging, especially in favour of the USA, I commented as follows.  Would be interested in hearing your opinions.  

'There is plenty [evidence of bias]. But the judging can be unreliable in all sorts of different directions, not just for the USA.

The problem arises at Code level, when the grading of moves and the bonus points are determined. Every country has a say in this, but naturally those countries with the strongest political representation in the sport will have the strongest influence. It is something that builds up over the years. So for example in WAG we currently have a Code that on floor and vault emphasises powerful acrobatics to the extent that the aesthetic has gone AWOL on floor in all but a few exceptional cases.

Vaulting requirements have changed and are so physically demanding that very few gymnasts in the world can prepare two competitive vaults at international level - e.g. in Europe in 2013 only 13 gymnasts attempted to qualify for the vault final of eight gymnasts out of a field of several hundred gymnasts.

At world level the USA leads in vaulting and acrobatics and there is a significant gap between the leading Americans and the rest of the world. When combined with what could be considered biased judging - a natural human tendency to overlook errors in those considered to have an almost mystical command of the sport - this adds up to an advantage [which might be considered a bias]. For example, Simone Biles' highly acrobatic work is astonishing in its accuracy and has extremely high difficulty scores, yet the judges seem to ignore failures in the aesthetic quality of her work. Gradually her scores on bars and beam have crept up as belief in her strengths elsewhere make it uncomfortable for the judges to deduct. The 'wow' factor blinds judges and fans to the less than perfect state of artistic presentation in some of Biles' work while others with more grace and less athleticism struggle to find the same certainty and confidence in the judges' evaluations of their work. These are minor, often and usually tiny granules of distinction that build up in one gymnast's and one style's favour over another. They in turn affect the shape of the sport as it progresses and the Code develops, and in performance affect the psychology of the gymnast and the reliability of competition.

A computer system that relies on measurement and quantification of movements will only emphasise these distinctions and detract from the aesthetic side, unless its implementation is carefully managed to allow for the judges' panel to pay more attention to the impression of the whole routine. There would have to be a splitting of the scores to introduce a technical mark (computer) and an artistic score (judges). As far as I can read, the FIG hasn't yet considered this, so unless the target of 2020 is purely a pilot run, they are getting ahead of themselves on every apparatus except vault. If the pilot is on vault only, as this article seems to suggest, then that could be a good thing as vaulting is a single skill and the measurement and judging process already seems to be highly technical and well elaborated.

The role of President of the FIG is at face value a mouthpiece job, yet for decades this mouthpiece has influenced the direction of the sport disproportionately and favour has been cast on his (no female President ever!) national programme. Titov during the Soviet dominated era saw the language of gymnastics favouring aesthetic, innovative gymnastics, Grandi presided over a period of growth for Italian WAG gymnastics [with the introduction of the additive score leading to the only Italian AA World Champion, with a fall], and now Watanabe introduces a technological step forward that is of potential benefit to the Japanese economy, while the JPN gymnastics programme continues to lead MAG and to grow WAG.

Introducing computer judging will only emphasise the growing tendency in both MAG and WAG to favour content over quality unless the FIG considers the whole picture and implements gradually with review of the gymnastics routines favoured [in addition to the calculation of D scores and deductions for faulty execution] and their likely influence on the direction of the sport.'