Fantasy football progress report: 2026

I’m 7–0 in fantasy football this season. Last year, I went 11–3, then lost the championship by two points.

I’ve learned a few things since then. I wanted startable players, even on my bench. So I spent significant time preparing for the draft.

I’ve also been lucky. Except for quarterback Jaxson Dart, I’ve avoided major injuries. And in our 10-team league, finding another decent quarterback isn’t that hard.

Preparing for the draft

Our draft was in person. I wanted something I could scan quickly, not a bunch of browser tabs.

I asked ChatGPT to prepare a cheat sheet based on my preferences for certain players. I also ran several practice drafts and asked AI to grade the results.

Those practice drafts helped me refine my strategy. I focused on wide receivers and running backs, knowing I could find decent quarterbacks later.

In fact, I waited until round 10 to draft my QB. I drafted only one QB and one tight end (TE), leaving more picks for wide receivers and running backs. I didn’t pick a kicker or defense until the last two rounds.

I was flexible. If another manager picked my first choice in a round, I needed alternatives. And if a highly rated player was still available in a later round, the cheat sheet helped me recognize the opportunity.

A little luck during the draft

I had two lucky breaks.

With the fourth pick in round one, I could have taken Puka Nacua, my favorite player from last year. Instead, I picked Jaxon Smith-Njigba (JSN).

Later, someone grabbed Brock Bowers just before my third-round pick. I had targeted Bowers, so I went with Trey McBride.

Both Nacua and Bowers suffered injuries early in the season.

Week by week

The draft gave me a good start. Since then, I’ve used AI to help choose starters, manage bye weeks, and evaluate waiver pickups.

I’ve built a roster with eight startable wide receivers and three running backs. That gives me options, especially when players get hurt or have a week off.

But that leads to a different kind of problem: who do I start?

I compare recent performance, opportunities, and matchups. I also ask AI to explain its recommendations, especially when they don’t match what I’ve seen.

Sometimes I disagree. I want to know why AI favors one player over another, not just see a projected point total.

I also look ahead at bye weeks. I’d rather plan for a missing starter than scramble for a replacement at the last minute.

For kickers and team defenses (DST), I keep things simple. I drafted them last, and I replace them when I see better matchups. I don’t get attached to either position.

Turning the lessons into a Git repo

I’ve started a Git repository to capture what I’m learning. It’s still a work in progress.

I’m building an AI agent that follows consistent rules when comparing players. I want it to explain its recommendations, including the assumptions behind them.

I’m also recording some of my close decisions so I can look back after the games.

One lesson from last year still matters: a good decision doesn’t guarantee a good result. Sometimes the player I bench scores more than the one I start.

I want to know whether I missed something important or just had bad luck.

So far, so good

I’m happy with my 7–0 start. I’ve put more thought into preparing for the draft and managing my team this season.

AI has helped me prepare, evaluate choices, and question my assumptions. But I still make the decisions.

And I’ve had some luck. I remember how last season ended, even after an 11–3 record.

Wish me luck!