GTO Characteristics

Ahh I'm sorry Ceres that was a bit harsh. Guess I woke up on the wrong side of the bed this morning. We appreciate your contributions to the forum

Nah mate - it's fine. I apologise for butchering the point.

That said, after doing some more research overnight I do think the distinction is helping me understand things a bit more deeply.

Primarily, by defining GTO separately as the iterative process of reaching Nash - rather than the final proof itself (Nash) - I think the way solvers/bots actually approach their work might be slightly easier to understand for the layperson.

- i.e. why and how bots are de facto able to compete and improve on each other (dynamic iteration)
- conceptualising the compromised 'solved' end state; given if pure Nash and GTO were truly the exact same thing this would mean the game is technically fully solved
- could also help more broadly appreciate that Nash is only ever approximated, never absolute


Nuanced but i think it breaks things down a bit.

Rather than fusion Nash/GTO being static and somehow both unexploitable and exploitable without asking why. All still opaque to me either way frankly, but I feel moderately less befuddled today having looked briefly into the what.

might finally be ready for some solver articles.

There's definitely value in separating the process from the outcome. In particular, it's important to consider the difference between the fixed equilibrium strategy, and the dynamic exploitative algorithms we pit against each other to find NE.

But no one defines GTO this way. What you're describing, the algorithm that finds NE, is called CFR. (Well there are other algos but this is the most common method).

Fair. But not strictly true afaict.

And if a whole new generation of players are plugging these terms into Google/AI (and they will be), people are going to be having this debate ad infinitum.

ChatGPT - different
Google Gemini - different
Claude - Same
Perplexity - Same
MechaHitler - Same

Ok I give up, life's too short.