So another week has passed, and we may have just witnessed one of the best UFC cards in history. PickGPT went 7/13, which was better than last week, but still disappointing compared to expectations. It did however pick another huge underdog in Dricus Du Plessis, which once again kept PickGPT from having an abysmal betting night. If only I knew someone with a weekly show where they confidently picked an underdog to win. They could call it something like “Guarantee of the half-fortnight,” or something like that. Such a person could really benefit from PickGPT. Until I find that person, however, let’s do our usual post-mortem analysis and look for trends that may help us decide whether or not to trust PickGPT while it is still in development.
The first red flag that jumped out to me on fight day was that Jalin Turner missed weight by over 2 pounds. As an avid fan of the sport, I’ve seen plenty of fighters miss weight and look like total crap, but I’ve also seen the opposite. My personal observation is that missing weight tends to take a negative toll on a fighter’s cardio and aggressiveness. This makes sense since weight-cutting is very unhealthy, but obviously necessary for professional fighters. As a scientist, and the creator of PickGPT, I thought about how I might use this information algorithmically. The challenge of course is finding historical data on fighters that have missed weight, as this is not easily gathered data. Until I find a solution to this, I think the safe bet would be to avoid placing bets on fighters that miss weight, because PickGPT is currently not taking this into account.
Second, and probably more importantly is pre-fight injuries, and specific to this card, Kamuela Kirk’s back injury. It was widely reported pre-fight that Kirk had an injury to his L5 vertebra that postponed his previous booking. As someone with back injuries, I couldn’t imagine how that must affect a fighter. I’m sure he wouldn’t have fought if he didn’t think he wasn’t ready, but this is information that PickGPT doesn’t take into account, and it’s challenging to think how it could. I would need to find historical data on fighters coming back from injury, and this type of information often isn’t even revealed at the time of the fight (looking at you Dillashaw). So the wise betting move would be to use caution when betting on any fighter returning from injury, no matter what PickGPT says.
Finally, thank you to Angelo for buying my wife and me some very expensive drinks at UFC 290, when we clearly already had quite a lot to drink. It was my first time being at the fights live, and what a fight card it was. There were several thousand too many people for my liking, but it was still a great experience. And for those keeping track, if you had bet a unit on each fighter PickGPT took through the last 5 events, you would still be up 10.28 units using the BetMGM closing lines, despite some truly lackluster performances over the past two weeks. This performance will only improve over time, so don’t forget to follow the journey. PickGPT is the first of its kind, and it is funded by you, the Premium Members at We Want Picks!
