Let’s face it: Not all of us can be the next Andy Reid or Vic Fangio. If you’re a weekend warrior playing two-hand touch at the local park or a coach looking for creative strategies, AI tools like Google Gemini 2.0 can provide new insights for offensive and defensive plays.
A few months ago, I began to look deeper at the scrimmage games my friends and I play at our local park in Toronto. This two-hand-touch game isn’t played with pads, timed quarters or referees, but we take it seriously enough to dream up plays on both the offensive and defensive ends in order to feel more immersed in the sport. But why refer to Madden 2025 or our collective memories to come up with innovative plays? Why not leverage the vast resources available in an AI system such as Google Gemini 2.0?
Recently, I’ve been leaning on the chatbot to design impressive (and sometimes never-before-seen) plays I could bring into the real turf, er, I mean snow-covered grass.
I’ve been leaning on the chatbot to design impressive (and sometimes never-before-seen) plays I could bring into the real turf, er, I mean snow-covered grass.
What really wowed me was how Gemini didn’t just lay out the pre-snap and post-snap motion (what all players should do before and after the ball has been “hutted” to the quarterback) but it often included “Keys to Success” or “Additional Considerations.”
These notes helped me determine if the personnel on the field were right for the play. For example, one play noted how the QB in question needed to have “excellent timing and accuracy to deliver the ball to the receivers in tight windows.”
Let’s check out two examples of an offensive and defensive play I learned about through inputs I typed into Gemini.
Offensive plays
Longtime Chiefs defensive coordinator Steve Spagnuolo is known for his exotic blitzes, which proved too much for the Buffalo Bills during the AFC championship. The Eagles offensive line proved more than capable in the Super Bowl, but when a defense blitzes the QB, a smart offensive play should allow him to quickly get the ball out of his hands. In some of these plays, having two tight-end receivers run short routes can be savvy.
My prompt wrote: “Design a football play with a two tight-end set with crossing patterns that will perform well against a blitz.”
Gemini gave me a strong offensive play with detailed instructions, writing, “The QB fakes a hand-off to the RB, drawing the linebackers’ attention and potentially opening up gaps in the defense, Tight End 1 runs a short crossing route across the middle of the field, aiming for the area vacated by blitzing linebackers. Tight end 2 runs a deeper crossing route behind TE1, providing a second option for the QB.”
The play recommended one wide receiver run a straight-line “go route” while the other receiver performs a quick “out route” in case the tight ends are covered. What a shrewd idea! Interestingly enough, the Eagles ran a very similar play during Super Bowl LIX to go up 33-0 over the Chiefs.
As with other plays, Gemini shared why this play can be successful against a blitzing defense:
Exploits aggression: Blitzes often leave areas of the field exposed as defenders vacate their coverage responsibilities to rush the QB. The crossing routes are designed to exploit these open zones.
Quick release: The short crossing route by TE1 provides a quick and easy target for the QB, allowing him to get the ball out before the blitz arrives.
Multiple options: The play offers multiple targets for the QB, giving him flexibility to make the best decision based on the defense’s reaction.
What is disappointing, though, is how Gemini fails to visualize this play on an actual football field. Instead, it outputs this kind of bland and confusing visualization:
Defensive plays
In our scrimmage games, if someone wanted to cosplay as Patrick Mahomes (too soon?) and launch a long pass to a receiver, the defense has to be prepared. In the snow, especially, we can’t keep up with some of the faster plays on our team.
I asked Gemini to “design a football play to prevent a deep passing threat with fast wide receivers and a mobile quarterback.”
It churned out a highly detailed formation were one safety acts as the “robber” player, which means he’ll initially look to stop a deep threat but will then read the quarterback’s movements and have the freedom to “rob” any intermediate routes or even come up to help against the run.
The play also instructed the slot corner player to cover the slot receiver but he “has a dual responsibility: helping with deep coverage if the slot receiver runs a vertical route, and covering short/intermediate routes in the middle of the field.”
What is also helpful is how Gemini revealed some of the weaknesses of this play design. It noted how “this defense can be vulnerable to quick, short passes if the robber safety is caught too high.”
Final thoughts
Some of the concepts that Gemini referenced aren’t exactly NFL-caliber — modern NFL offenses and defenses are much more complex — but for more casual players, bringing these plays to games can be an effective strategy for getting creative with running routes and defensive packages that make football even more thrilling.
Admittedly, it can get boring calling the same plays over and over, and for amateur players/coaches, harnessing the vast knowledge of AI systems to design critical games during crunch time can be exciting.
With tools like Google Gemini, armchair coaches and analysts can generate new play strategies in real time, using AI-driven insights to predict opponents’ tendencies and optimize formations. By integrating AI like Gemini into training and game-planning, teams can stay ahead of the curve, ensuring a more dynamic and unpredictable approach to every game — at least in your backyard.
