Chess will never be solved, here's why

@playerafar

And the colour?

Fools this fools that. You seem a little confused about who said what. Maybe you should just stick to sourcing your posts directly from chatgpt.

I thought I made it plain that Wiki is not good enough for me, in particular because it didn't agree with my claim (I didn't see any body else agreeing with it either, so I don't know who are the other fools you refer to).

I did explain why in detail, but instead of answering the points you just posted a long summary incorporating exactly the same flaw that I complained of.

Given that you didn't answer @Elroch's question it seems you think I should have posted an anonymous definition.

Just admit it @Optimissed, you're out of your depth with any reasoning over one line.

I'm hesitant to involve myself in this dumpster fire that's spread to the whole town and burnt down the water truck. I wasn't there when the fire started, however... a few things are obvious to me.

Some people seem to have a fundamental misunderstanding of the relationship between models and reality. Models can't define themselves as being correct, they're correct insofar as they accurately model reality. When there's an incongruence between the game of chess as it's played in reality, and your abstract model of the game, it's time to reassess the model. It doesn't matter whether wikipedia says your model works in an established way, all you've learned is that models must evolve and this one needs work.

This shouldn't be a groundbreaking proposition, since this is always how fields evolve, it happens all the time. Every time a new game theorist comes along and extends the body of literature this is what they're doing, or trying to do. Models are always imperfect and evolving. This includes mathematical models...

I don't understand this reference to "abstract truth". Abstract truth is a meaningless concept. We could only ever consider an abstraction true in the most metaphysical, tautological sense. It's a very rare set of conditions under which you could possibly argue for such a thing having meaning. In this conversation what you're really saying is "I imagine this thing to be true". To which the only correct response... sure, you can do that, but it doesn't matter. Your model doesn't model chess, your conclusion doesn't mean anything.

I think in terms of solving the game, what is the significance of perfect information regarding the solution...

For a game that contains perfect information we can produce a solution that guarantees a certain theoretical result. For a game that hides some information this hidden information will continue to cause variance in result even with a solution at hand... We can merely come up with a strategy that provides us with the optimal odds at achieving the optimal result.

For this reason the model that currently defines games of perfect information is significant.. It defines the potential of optimal play resulting from a solution.

On the other hand what would be the purpose for defining all games that are not yet solved as games as imperfect information.. The definition would provide no value. We can just call that an unsolved game, right?

You would clarify distinctions such as those when you extended the model. The model isn't extended yet and you're trying to mash criticisms of it within its existing framework. That doesn't work, and neither does your argument.

On the question of ChatGPT and its progress vs. its limitations...

Well it wasn't directed at you. But my assumption was that his response was directed at me, however I could have misread what he was saying.

Anyway, I could probably spend longer trying to read this thread and write out a better post but... is it worth it? Probably not.

As you said, the model has to serve a real life purpose. The real life purpose the existing model serves is it defines the potential for a solution. What purpose that is more significant does your alternative model serve with a different definition for perfect information? Thats the question one has to answer if you want to offer a better or more effective definition.

I didn't say there was a single color.

I tend to agree with @MARattigan on most things.
Opto doesn't like to admit. He prefers to pretend.
He acts like criticism of his posts is trolling.
Another pretense of Opto's.

If it's finite, it's solvable within a finite amount of time. That's the only meaningful way to define solvable.

See, he understood me. Which I'm surprised by, but it's good enough.
@Optimissed perhaps you missed my post earlier, or maybe I wasn't clear enough, but I don't think I've undermined or questioned anything you've said, actually it's the opposite.

Oh look, it's the guy who said he was removing himself and not coming back because he was above all this.

Practically insoluble cannot be determined but finite is clear for anyone. Finite means theoretically possible. Clarity should be the goal for the definition. The definition is on shaky grounds if its based on an undefined term such as practically insoluble or even worse, "maybe practically insoluble"...

No, the model needs to accurately model reality. In this case... the existing models are assuming a perfect player, and defining perfect information based on that. What I'm saying is, in reality, no such thing exists. Yes, your conclusion is true within your irrelevant model, but there's no point to this. You don't create a model for its own sake. Hence the game theory model needs to be extended - the definition of perfect information needs to be modified - to more accurately model reality.

And that's always the way models works, that's how fields always evolve. I guarantee there is room for a game theorist to come along, and look at these problems, and extend game theory to account for them.

Basically you'd just have to create a model inwhich, instead of assuming a perfect player, you can define certain limits on the players ability to handle complexity. And you could probably model that mathematically using fields of probability and stochastic modeling. You could then tune the model to match the real life game... or increase / decrease player strength along a spectrum however you like. Infact, your model would sort-of resemble an AI model.

Self-awareness continues to elude you.

It absolutely is meaningful. Thats the difference between a game of imperfect information and a perfect information. The information is theoretically accessible by a solution...