Parable of Alien Chess • Hacker News Board

Re: Parable of Alien Chess

#171

Post by lmm » Wed, Jan 06, 2021, 2:50 PM UTC

jjk166 wrote: post #169

lmm wrote: post #165

Earlier quoted context omitted.

> You'd have to develop quite a sophisticated model of human behavior to infer that they are treating it as a contest and care about the results. Sure. But again, trying to understand chess without having that understanding would be putting the cart before the horse. > Even then, if they've previously seen humans at a casino, they might reasonably deduce that chess is no more a game of skill than craps is. That would…

You are deliberately misinterpreting the analogy to the point of absurdity. The aliens are only observing a black box system that spits out "white wins" or "black wins," they are trying to determine if there's any reason to develop the means of probing the contents of the black box. Regardless of whether you start by modelling the players or the board, you are still assuming the winner is determined by some complicat…

> The aliens are only observing a black box system that spits out "white wins" or "black wins," they are trying to determine if there's any reason to develop the means of probing the contents of the black box. Regardless of whether you start by modelling the players or the board, you are still assuming the winner is determined by some complicated model with numerous hidden variables which you would have no reason to believe unless you start with the assumption that this is not a random number generator - an assumption for which you have no evidence.

If you actually did that, you'd start creating complicated hidden variable theories to explain every random coin flip you could see, despite the overwhelming majority of them actually being random coin flips.

> The author's thesis is that assuming a complex explanation for a black box's behavior is reasonable because that's the only way that leads to experiments to probe the contents of the box.

The thesis only holds if it's actually common to have a complex hidden detail inside a black box that is nevertheless somehow completely impossible to infer from the outside. And what I'm saying is that that's actually absurd; in cases where there are meaningful details to be found out, it will be apparent from the outside that there is detail in there, at least the overwhelming majority of the time.

Re: Parable of Alien Chess

#172

Post by jjk166 » Wed, Jan 06, 2021, 4:03 PM UTC

lmm wrote: post #171

jjk166 wrote: post #169

Earlier quoted context omitted.

You are deliberately misinterpreting the analogy to the point of absurdity. The aliens are only observing a black box system that spits out "white wins" or "black wins," they are trying to determine if there's any reason to develop the means of probing the contents of the black box. Regardless of whether you start by modelling the players or the board, you are still assuming the winner is determined by some complicat…

> The aliens are only observing a black box system that spits out "white wins" or "black wins," they are trying to determine if there's any reason to develop the means of probing the contents of the black box. Regardless of whether you start by modelling the players or the board, you are still assuming the winner is determined by some complicated model with numerous hidden variables which you would have no reason to…

Black box systems are incredibly common. By definition, there isn't anything readily apparent from the outside. You haven't shown that to be absurd, you only showed that if you can see enough of the interior of a black box, it is no longer a black box.

You complain that the author's logic will lead to creating complicated hidden variable theories but that's exactly what the author is advocating. While there will always be some ever more convoluted model to explain results, any given model is testable, whereas assuming there is nothing to model is not testable.

Re: Parable of Alien Chess

#173

Post by karmakaze » Wed, Jan 06, 2021, 4:42 PM UTC

jjk166 wrote: post #144

karmakaze wrote: post #133

Earlier quoted context omitted.

This was the only part of an otherwise great post I objected to. Science may have a side effect of uncovering truth but it shouldn't be the driving force. We should be striving for better predictive ability that doesn't go against our observations. Arguing which is more real is not science.

Drop some heavy objects and some light objects you have lying around. You will see that the heavier ones tend to fall faster than the light objects. If all you care about is predicting which object will fall faster, you could simply say gravity pulls the heavier one faster and call it a day. But if you were to forego predictive power and seek understanding, you'd eventually see that objects fundamentally fall at the…

The second step would be noticing that objects of the same weight fall at different rates. When your predictive power fails you seek more refined models. For everyday phenomena we can guess at truth, maybe 'the wind'. The deeper you go the less clear it becomes as you don't have any direct interaction with the things you're modelling.

Re: Parable of Alien Chess

#174

Post by lmm » Thu, Jan 07, 2021, 2:10 AM UTC

jjk166 wrote: post #172

lmm wrote: post #171

Earlier quoted context omitted.

> The aliens are only observing a black box system that spits out "white wins" or "black wins," they are trying to determine if there's any reason to develop the means of probing the contents of the black box. Regardless of whether you start by modelling the players or the board, you are still assuming the winner is determined by some complicated model with numerous hidden variables which you would have no reason to…

Black box systems are incredibly common. By definition, there isn't anything readily apparent from the outside. You haven't shown that to be absurd, you only showed that if you can see enough of the interior of a black box, it is no longer a black box. You complain that the author's logic will lead to creating complicated hidden variable theories but that's exactly what the author is advocating. While there will alwa…

> Black box systems are incredibly common. By definition, there isn't anything readily apparent from the outside.

Citation needed. The author is trying to claim that this kind of extremely opaque black box system is common (at least, common enough that we should take the possibility seriously), but their only argument is a made-up example that falls apart under the slightest scrutiny.

> While there will always be some ever more convoluted model to explain results, any given model is testable, whereas assuming there is nothing to model is not testable.

The claim that Russell's Teapot exists in any given orbit is testable, whereas assuming Russell's Teapot doesn't exist anywhere is not testable.

Re: Parable of Alien Chess

#175

Post by karmakaze » Thu, Jan 07, 2021, 4:00 PM UTC

jjk166 wrote: post #168

praestigiare wrote: post #156

Earlier quoted context omitted.

You are imagining this situation from the perspective of already having a more powerful / general model and looking backwards, so it seems obvious. But that is not how things work. Without data to support it, you can imagine all kinds of possible theories - that is not science, and it certainly gets you no closer to understanding.

Yes, the entire point is that what was obviously correct in hindsight was not clearly advantageous to begin with, the author's thesis is that exploring models which may seem worse than what we currently have can ultimately lead to better overall results. The author is specifically criticizing the narrow definition of science which says every incremental step no matter how small must improve your models' predictive po…

I didn't read it so specifically. If optimizing for X you may be in a local maxima so deviation looking in other regions seems fine.

Re: Parable of Alien Chess

#176

Post by praestigiare » Mon, Jan 18, 2021, 5:12 PM UTC

jjk166 wrote: post #168

praestigiare wrote: post #156

Earlier quoted context omitted.

You are imagining this situation from the perspective of already having a more powerful / general model and looking backwards, so it seems obvious. But that is not how things work. Without data to support it, you can imagine all kinds of possible theories - that is not science, and it certainly gets you no closer to understanding.

Yes, the entire point is that what was obviously correct in hindsight was not clearly advantageous to begin with, the author's thesis is that exploring models which may seem worse than what we currently have can ultimately lead to better overall results. The author is specifically criticizing the narrow definition of science which says every incremental step no matter how small must improve your models' predictive po…

Neither you nor the OP have provided a reasonable example of a model which is less predictive and also more "true." the OPs analogy has been discussed to death here, and yours is explicitly of a model which is in fact more predictive.