I’m 3-1 right now and feeling pretty good about my fantasy team.

For those of you who are 0-4 the following is my form of condolences:

But you don’t want to stay there and I think there’s hope for the ones who drafted Saquon Barkley in round 1! And for those of you who have no idea what I’m talking about I promise this gets interesting because today we’re going to test whether or not AI can make better recommendations than what I have set for my current line up and then we will see if the AI’s lineup got more points or if I got more points.

So lets begin. Here is my current match up (my team on the left vs. opponent team on the right):

If you have any problems with my fantasy team selection please let me know what I should have done in the comments below.

So here’s my reasoning for my lineup. If you aren’t aware De’von Achane is currently out for the rest of the 2026 season. On my bench Breece Hall was my next best up RB2 slot but quickly became doubtful as he also succumbed to the injury bug. So what am I left to use? Well it just so happens Ollie Gordon II was available in my league and was highly productive last week on my bench! Thus, I’m promoting him to my RB2 spot.

To me this is my biggest risk because I don’t have proven record of him doing well for a long career path but that’s the nature of fantasy football. It’s all a risk. So how does AI calculate that risk?

As mentioned in my previous article before I do any sort of prompting of an AI, I want to start with the idea on paper and draw it out and as I was writing I had this bigger idea before I just went with a much more simple and direct approach. Head on head with AI. I took that prompt and ran it through Claude to give me a better meta prompt to give ChatGPT (kind of meta in itself). Here’s my prompt:

And as you can see 25 seconds later and responding that, “I am not going to screenshot my entire available free agent database inside yahoo sports,” I got this:

When I asked why it assumed Jacoby Brissett was a better match up instead of Kyler Murray it gave me 3 interesting answers:

1. “I treated Kyler as if he were still Arizona’s QB.” – even when I told it this was the 2026 NFL season (Maybe it confused it since the seasons run through the new year?)
2. “Kyler is officially OUT for Week 5 with a concussion.” – Kyler was cleared off concussion protocol and came back in week 3.
3. “Detroit’s pass defense has been horrendous.” – This is both funny as an NFC fan and something I did not consider in my research.
   
   You can read the full recommendation that it gave me in a link I’m making public here.
   
   Why would you use AI to help you win when it gave me 2/3 outdated pieces of information? I want to affirm that this product is not perfect and as a tool it has a lot of room for growth that I still feel is just not serving our daily purposes very well beyond entertainment. Yes is it dangerous in the wrong hands? Absolutely!!! Regulation is still necessary because this computer that we use to do quick statistical analysis that outpaced the Hubble two decades ago.
   
   Understanding it’s limitations, scope, and how effective it is should be highly considered in its regulation. Because the same tool that someone could use for learning how to make really dangerous materials can be used to upgrade my fantasy football team.
   
   Subscribe to see my update next week on how well my team did vs. the AI’s recommendations.
   
   If I missed anything, comment below and tell me what you think I should have done better.
   
   As always… Strong Side! Left Side!