They will just give every close round to the fighter that is supposed to win for their payment and have him build up a lead.
No. I'll take human error and corrupt boxing judges over AI. As much as those things really suck the life out of boxing they're still human.
Then the fighter who would be “losing” Could go for a knockout - or be content to lose the decision- but at least they would know where they stand after every round & know what they need to pick it up or try harder to put the other guy down.
I doesnt take away cheating though. Judges need to be held accountabke trough a rating system. Only those with high ratings should get quality fights. Im also in favour of 5 judges instead of 3 and them being punished if they put in a card like Moretti and having to explain giving a round 10 to Haney.
Yeah cheaters gonna cheat - it is after all boxing. but at least the guy getting cheated knows it’s coming & can adjust the game plan accordingly Meaning go for the KO in the 12 even if it means getting KOed yourself. Won’t fix cheating but it will give the fighters a fighting punchers chance.
Tell me you don't know **** about scoring without telling me you don't know **** any scoring. None of that **** matters unless the striking is dead even, which it never is, but you could keep judges to score ring generalship (lmao) if that is a tie breaker. I get where you're coming from. Those things would matter if we were judging hard sparring. But boxing is effective punching. We should look at doing away with the must system if scoring is going to be incredibly subjective anyways. Think of it this way, would anyone watching deny that Haney lost the *real* fight? Who was the defeated man at the bell? Who lost confidence in themselves? It doesn't matter, we always make up reasons to back our fighter but all that matters is effective striking and that is theoretically what the judges are scoring.
Just like if ydksab and you shouldn't talk about it, you shouldn't talk about AI and Machine Learning if you don't understand how it works. The programmers develop aspects of the end to end solution from data ingestion, data transformation including feature engineering, and model, as well as training and testing the model with the available data. There can be bias in the solution, but the programmers can not selectively predetermine the result of a fight. A model of this complexity would evolve over time and the weights would be tweaked to align with the scoring methodology, ie the relative differences between different types of landed punches. A model could be trained to score effective aggression, albeit with a lot of training and a high degree of complexity, if that can be visually determined. To me, that would include the boxer moving forward, holding the centre of the ring, and culminating in positions where they throw punches. It is a non-trivial implementation but would be much better and more objective than human judges.
This. Both boxers know where they stand and what's needed for either of them to win the fight. Massively negates any judging bias.
I'd like more technical info, if anyone knows where to find it. Are they using a video interpreter of some kind so that the AI is "watching" the match live, is it a prompt given to the AI by writers, or something between the two? Will it score live, semi live, or need the context of the fight to score the rounds within it? AI kind of works backward from us. We start with 0 knowledge and build up to a picture. It starts with more information than it needs to know and limits down to a picture. Or text or video or whatever. It is interesting Ring is claiming to be able to do something Open, Microsoft, and Google can not but maybe Turki actually brought on programmers or maybe it's a bit BS and someone's just typing what happens into GPT to get a score.
I think we should give AI a shot, considering how many truly horrendous cards we get almost weekly. I am sorry, but how the **** does one judge score a fight 116-112 one way and then another has it 111-117 the other way? AI is a learning tool. You just have to feed it good information.
I think it would have been better to put it to the test on a non-title fight first. A big fight like Usyk vs Fury 2 to start with might be a bit risky considering what's at stake.
Something new and different, interesting from a technical perspective. But it does put the human judges under even more scrutiny. Which I don't think leads to better outcomes. Judges are already under enough pressure to perform