UFC 251 Predictions
Updated: Sep 8
Machines are pretty sweet. They allow us to interact with strangers thousands of miles away, wire money to people in time of need, and blog about getting catfished by a Nigerian prince (detail coming soon). Most importantly though, machines allow humans to completely deny accountability for major decision making. Stop me if you have heard this one before,
"I didn't want to shut down the global economy but this computer said 2 million people were going to die."
You see how easy that fucking was? How can anyone get mad at this:
So that's what I am doing today- taking MMA data, creating a machine learning algorithm (also do you capitalize Machine Learning? I feel like it comes off douchey), and conveying the results. Do I know what the computer is doing? No. Do I understand why the computer is making these choices? Hell no. But that's not going to stop me from acting like I accomplished something.
Martin Day -162
File this one as Exhibit A in "I don't know what the computer is doing".
While Day certainly looks to be the more active fighter, he is 0 for 1 in the UFC. You would think a more experienced opponent would be able to handle him.
Karol Rosa -278
This one makes a little more sense.
With Melo being 0-2 in the UFC and Rosa being the more active fighter, the model preferred Rosa.
Elizeu Zaleski dos Santos +108
Elizeu is the more experienced fighter with a more diverse set of wins.
Rose Namajunas -200
At this point the pace metric looks to be deciding these fights.
Petr Yan -240
Max Holloway +188
This is the only instance where the algorithm chose the less active fighter.
Data provided by www.UFCStats.com
Lines provided by www.DraftKings.com