• chrisbdbets

UFC 251 Recap

Updated: Sep 8

Machines are the worst. Not really but night one for the model wasn't necessarily stellar. In testing, the algorithm was 75% accurate. In reality, the model was 50% accurate and the winners were heavy favorites. Put another way - in testing, Chris is the man but in reality, Chris is homeless.


Let's look at the misses and see what we can take away.


Martin Day v. Davey Grant


The model preferred Day who had not fought in the UFC since 2018 in which he lost a split decision. I believe 2 things occurred here that caused the machine to miss:

  1. The machine assumed Day would produce the same amount of activity, which did not occur.

  2. The machine did not believe Grant would improve as much as he did. While his significant strikes were slightly over par in round 1, his production in rounds 2 and 3 saw a significant increase. This could either be due to his own improvements or the fact that he was facing an opponent with much less experience.

Takeaway:

The model needs to do a better job with experience. While the simple answer may be to analyze the number of wins and total fights each combatant has, its not quite that simple. Fighters on the decline lose to new up and comers all the time so looking at straight totals may be counter productive.


Elizeu Zaleski dos Santos v. Muslim Salikhov


I don't think the machine did anything wrong here. It was a close fight. On paper dos Santos landed slightly more significant strikes but both fighters seemed to be practicing caution, which is completely reasonable given the match up.



Alexander Volkanovski v. Max Holloway


Like the last example I am not really sure what the machine did wrong here. Max very well may have been robbed. General consensus is that he won rounds 1 and 2 and he should have had one in the back 3. All robbery speculation aside, he was outperformed by Volkanovski in the last 3 rounds in terms of significant strikes. Looking at the fight strictly in terms of numbers, Max's production fell off in the last 3 rounds. Maybe Zoom calls aren't the best training medium?

Or maybe, (conspiracy theory coming) every established trainer in the world needed Max to lose so that Zoom did not blow up their entire business model. This resulted in a mass bribing of the judges in an attempt to restore order and prevent UFC fighters from hiring Instagram influencers before their next fight.


Probably option A but more to come.



Data provided by www.UFCStats.com


0 views

Subscribe to Our Newsletter

  • White Facebook Icon

© 2023 by TheHours. Proudly created with Wix.com