Created a MMA fight Picker GPT

[TSaL]

You need gpt plus to read it

Yes you need subscription to use gpt.

 

[Vampire life]

Yeah you need the “plus”

Yes you need subscription to use gpt.

edition too
I don’t want gpt getting me , so I can’t sign up for an account

 

[TCE]

Looking great so far. How it predicted de Lima Tko R2 and Hernandez Sub R2. Crazy good

 

[TCE]

His precision is so precise

 

[DalchaLungiambula]

Important to note that if you had have taken the betting favorite, their shortest priced method of victory, and their shortest priced round to win you would have scored 770 points on Tapology tipping and come 104 out of 6893:

Volkanovski by decision - 0 points
Whittaker by decision - 75 points
Garry by decision - 75 points
Dvalishvili by decision - 75 points
Hernandez by submission, round 1 - 65 points
Lemos by KO/ TKO, round 1 - 50 points
De Lima by KO/ TKO, round 1 - 65 points
Nakamura by KO/ TKO, round 1 - 50 points
Mingyang by KO/ TKO, round 1 - 100 points
Barlow by KO/ TKO, round 1 - 65 points
Elliot by decision - 75 points
Maverick by decision - 75 points

Concluding at least from this card, I don't think it is anything that the world hasn't seen before at all.

It's just it's ability operate in a clinical and accurate manner, which is what the models that sports betting has run off for decades.

Your average Tapology tipper is unlikely to utilize such resources, and just bases selections off the gut, which if you have any experience with sports betting, you would know is a method almost impossible to generate a profit.

One thing that it clearly has over the betting market is it correctly predicts finishes in the 2nd and 3rd round, reasoning that in relatively even fights, the losing fighter will be fresh enough to resist takedown attempts and submission attempts in the first round, but will wilt in the 2nd, 3rd round. Pretty much exactly like what you saw in the Hernandez/ Kopylov fight.

Sportsbooks always price the round the fight will end in descending order starting round 1, which I think the AI has identified is erroneous.

In saying that, I am very excited to see the trajectory of it's success tipping, and think OP has done a phenomenal job, which to me has managed to identify at least 1 thing that the betting market has been asleep to for ever, which is a really big deal.

And I also may be wrong entirely, that it is no more accurate than following the betting lines and taking shortest priced favorites, their method, and the round.

Time will tell I guess!

[Edit: Scored the 2 previous cards to today's PPV, and following this approach would have scored 390 (3579 out of 4892) & 440 (1103 out of 5007), for 1600 points across the last 3 events compared to TSal's 1770; and that's not to mention TSal killed it and came first on Tapology on the card 3 events removed from today's PPV, so his AI is still very likely more accurate and has a clear edge on the public.]

 

[IBeLurkin]

@TSaL can we get some predictions for this weekends fight night?

 

[TSaL]

https://www.tapology.com/ue/107471-3926496.png

Seems like too many decisions to me.

 

[DalchaLungiambula]

https://www.tapology.com/ue/107471-3926496.png

Seems like too many decisions to me.

Got the same results as you this time by asking it every day for 6 days and going with the fighter it states the most - EXCEPT, it told me Bondar all 6 times!

 

[TSaL]

Got the same results as you this time by asking it every day for 6 days and going with the fighter it states the most - EXCEPT, it told me Bondar all 6 times!

Interesting

 

[DalchaLungiambula]

Interesting

Oh, and I actually got Altamirano over dos Santos every time.

I also got Lacerda the first 3 times, but then it evened out with Chairez the final 3 times.

Interesting that it's with the underdogs that I'm getting this discrepancy.

I did get Puelles all 6 times though, who is the only underdog you ended up with.

Looking forward to seeing what happens!

 

[Harlekin]

https://www.tapology.com/ue/107471-3926496.png

Seems like too many decisions to me.

Yair dominates but needs more than 1 round, Ortega has a chin.
Ai thinks the shoulder injury was a regular finish.

Moreno sub Rd 3 I doubt, too.Got him by decision.

 

[sonhow]

https://www.tapology.com/ue/107471-3926496.png

Seems like too many decisions to me.

This week I mostly agree, last week I had Topuria, Kopylov, Barlow and Maveric as major diffference also picked Merab but with little confidence.

This week I dont see big difference but I actually might end up betting Ortega even though Im Yair fan and normal man would fall off a cliff with the way Brians career has turned out so far. Its just that everything about Ortega screams variance to me.

 

[TSaL]

This week I mostly agree, last week I had Topuria, Kopylov, Barlow and Maveric as major diffference also picked Merab but with little confidence.

This week I dont see big difference but I actually might end up betting Ortega even though Im Yair fan and normal man would fall off a cliff with the way Brians career has turned out so far. Its just that everything about Ortega screams variance to me.

It will be fun to see what happens.

 

[TCE]

@TSaL do you have predictions for the card this Saturday, Sherbro?

 

[TCE]

Has anyone used the AI tool for this weekends fight card? If so, can you please share?

I'm in New Zealand, but will pay for the plus again when I get back. Thanks in advance.

 

[TSaL]

https://www.tapology.com/ue/106064-3951545.png

 

[Levi_]

[Edit: Scored the 2 previous cards to today's PPV, and following this approach would have scored 390 (3579 out of 4892) & 440 (1103 out of 5007), for 1600 points across the last 3 events compared to TSal's 1770; and that's not to mention TSal killed it and came first on Tapology on the card 3 events removed from today's PPV, so his AI is still very likely more accurate and has a clear edge on the public.]

It will be interesting to see if the current trajectory will hold up. One of the good things about ChatGTP is you can ask it to explain exactly why it made a decision. In theory you could work to establish the patterns of accurate and inaccurate picks by having ChatGTP write a couple paragraphs on why it made a particular pick (and looking for the patterns that way, comparing correct and incorrect picks).

I would think a well-constructed AI program could pick better than the general public, but not as well as a professional gambler. That remains to be seen though- you take away the capacity for human bias by having an AI analyze the data and give a response.

 

[TCE]

https://www.tapology.com/ue/106064-3951545.png

Thanks for the picks

Interesting the AI picked Rozenstruik. I'm leaning more towards Gaziev in this fight by TKO.

Is this perhaps because the AI doesn't have enough information on Gaziev, do you reckon?

 

[TSaL]

Thanks for the picks

Interesting the AI picked Rozenstruik. I'm leaning more towards Gaziev in this fight by TKO.

Is this perhaps because the AI doesn't have enough information on Gaziev, do you reckon?

Could be.

 

[TSaL]

It will be interesting to see if the current trajectory will hold up. One of the good things about ChatGTP is you can ask it to explain exactly why it made a decision. In theory you could work to establish the patterns of accurate and inaccurate picks by having ChatGTP write a couple paragraphs on why it made a particular pick (and looking for the patterns that way, comparing correct and incorrect picks).

I would think a well-constructed AI program could pick better than the general public, but not as well as a professional gambler. That remains to be seen though- you take away the capacity for human bias by having an AI analyze the data and give a response.

Oh man I have been some conversation with it asking why it does something. Its pretty frustrating when you give it a specific "You must do this" only to have it do something else 2 prompts from now. I'm hopping the addition of gpt memory will help.