Luke skywalker has done it again.

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Harvey Williamson, bob

Daniel Shawul
Posts: 3429
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Luke skywalker has done it again.

Post by Daniel Shawul » Wed Apr 04, 2012 6:06 pm

Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:I do not know what you are talking about but I was always under the impression that this was the april fools prank. Do you really think I belived it when I post it as a Luke Skywalker and sucking you Vas hater fools in ?

Clearly you are deluded to not see that Rybka is the one used not a dumbed down version which does alpha-beta only that is required for a proof.
It is a probabilitic proof that was claimed to have been done to 99.99999% accuracy. You can bring down chess to any level you want (say that of checkers) with prunings. No one said it is a proof like the checkers case so it indeed lame to say that was what is meant.
No you can't! I don't give a damn what some may think is or is not possible.
Why do you think it is claimed as "probablistic proof" if they have to use dumbed down version i.e alpha-beta gives same score as min-max? That is only required for a proof as in checkers. Even the 5.12 termination is not required to show that it is possible with the current Rybka as it is.
I don't think it is. I've already explained and I'm not going to explain again. Got it?

It's hard to believe you're a programmer. Then again...?!
Well that is your problem then. They even admitted that there could be flaws , which has been discussed here many times already. I thought the whole point was the strange 99.999999% assurance so I was disapointed to see today in their explanation, that they said it was impossible based on an exact proof which was never the point. I think they worded the question wrong to Ken thompson and vas as well. Maybe something like:
Can chess be solved ? No ofcourse not easy.
But you can claim that you are 99.999999% lost if you are 2 pieces down ?
They even gave that example to demonstrate the feasibility of the experiment.
You really are an impossible person. I claimed nothing at all other than the article is bogus and you can't solve any opening with current technology. Not even weakly solved. That's not my quote either.

Only Quantum Supercomputers if and when they exist have a chance and they don't exist!
Sure that is for the exact proof. But they claimed probabilistic from start to finish. They say the result could have flows but could be very true as well. So after going to great detail explaining the situation with an example, ,you can't go back to asking a full proof which is 100% not 99.9999999%

About my _weakly solved_ quote, you understand that when it is "weakly solved", it is saracstic and that it means I don't believe in it. Sure you understand sarcasm or you dont :) Right after that even I mentioned as their 5-sigma could be considered as a discovery statistically speaking.

Daniel Shawul
Posts: 3429
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Luke skywalker has done it again.

Post by Daniel Shawul » Wed Apr 04, 2012 6:09 pm

ATOMICC wrote:
Daniel Shawul wrote:...
But you can claim that you are 99.999999% lost if you are 2 pieces down ?
...
What if the two pieces are Knights? :)
Hehe that falls under 0.00001% . They said it not me.

Terry McCracken
Posts: 15798
Joined: Wed Aug 01, 2007 2:16 am
Location: Canada

Re: Luke skywalker has done it again.

Post by Terry McCracken » Wed Apr 04, 2012 6:36 pm

Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:I do not know what you are talking about but I was always under the impression that this was the april fools prank. Do you really think I belived it when I post it as a Luke Skywalker and sucking you Vas hater fools in ?

Clearly you are deluded to not see that Rybka is the one used not a dumbed down version which does alpha-beta only that is required for a proof.
It is a probabilitic proof that was claimed to have been done to 99.99999% accuracy. You can bring down chess to any level you want (say that of checkers) with prunings. No one said it is a proof like the checkers case so it indeed lame to say that was what is meant.
No you can't! I don't give a damn what some may think is or is not possible.
Why do you think it is claimed as "probablistic proof" if they have to use dumbed down version i.e alpha-beta gives same score as min-max? That is only required for a proof as in checkers. Even the 5.12 termination is not required to show that it is possible with the current Rybka as it is.
I don't think it is. I've already explained and I'm not going to explain again. Got it?

It's hard to believe you're a programmer. Then again...?!
Well that is your problem then. They even admitted that there could be flaws , which has been discussed here many times already. I thought the whole point was the strange 99.999999% assurance so I was disapointed to see today in their explanation, that they said it was impossible based on an exact proof which was never the point. I think they worded the question wrong to Ken thompson and vas as well. Maybe something like:
Can chess be solved ? No ofcourse not easy.
But you can claim that you are 99.999999% lost if you are 2 pieces down ?
They even gave that example to demonstrate the feasibility of the experiment.
You really are an impossible person. I claimed nothing at all other than the article is bogus and you can't solve any opening with current technology. Not even weakly solved. That's not my quote either.

Only Quantum Supercomputers if and when they exist have a chance and they don't exist!
Sure that is for the exact proof. But they claimed probabilistic from start to finish. They say the result could have flows but could be very true as well. So after going to great detail explaining the situation with an example, ,you can't go back to asking a full proof which is 100% not 99.9999999%

About my _weakly solved_ quote, you understand that when it is "weakly solved", it is saracstic and that it means I don't believe in it. Sure you understand sarcasm or you dont :) Right after that even I mentioned as their 5-sigma could be considered as a discovery statistically speaking.
You're hopeless. That's why Dan Honeycutt gave up.

One last time. Computers can't give you a probablistic win or draw in an entire opening like the King's Gambit. It's far too complex. It's far too deep.

A crude idea, maybe but no more.

Same for the Benko Gambit.

All you've managed to accomplish is pull me off track a bit with your tangential arguments.

The point was I never claimed anything other than the Chessbase article was a prank.
Terry McCracken

Daniel Shawul
Posts: 3429
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Luke skywalker has done it again.

Post by Daniel Shawul » Wed Apr 04, 2012 6:54 pm

Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:
Terry McCracken wrote:
Daniel Shawul wrote:I do not know what you are talking about but I was always under the impression that this was the april fools prank. Do you really think I belived it when I post it as a Luke Skywalker and sucking you Vas hater fools in ?

Clearly you are deluded to not see that Rybka is the one used not a dumbed down version which does alpha-beta only that is required for a proof.
It is a probabilitic proof that was claimed to have been done to 99.99999% accuracy. You can bring down chess to any level you want (say that of checkers) with prunings. No one said it is a proof like the checkers case so it indeed lame to say that was what is meant.
No you can't! I don't give a damn what some may think is or is not possible.
Why do you think it is claimed as "probablistic proof" if they have to use dumbed down version i.e alpha-beta gives same score as min-max? That is only required for a proof as in checkers. Even the 5.12 termination is not required to show that it is possible with the current Rybka as it is.
I don't think it is. I've already explained and I'm not going to explain again. Got it?

It's hard to believe you're a programmer. Then again...?!
Well that is your problem then. They even admitted that there could be flaws , which has been discussed here many times already. I thought the whole point was the strange 99.999999% assurance so I was disapointed to see today in their explanation, that they said it was impossible based on an exact proof which was never the point. I think they worded the question wrong to Ken thompson and vas as well. Maybe something like:
Can chess be solved ? No ofcourse not easy.
But you can claim that you are 99.999999% lost if you are 2 pieces down ?
They even gave that example to demonstrate the feasibility of the experiment.
You really are an impossible person. I claimed nothing at all other than the article is bogus and you can't solve any opening with current technology. Not even weakly solved. That's not my quote either.

Only Quantum Supercomputers if and when they exist have a chance and they don't exist!
Sure that is for the exact proof. But they claimed probabilistic from start to finish. They say the result could have flows but could be very true as well. So after going to great detail explaining the situation with an example, ,you can't go back to asking a full proof which is 100% not 99.9999999%

About my _weakly solved_ quote, you understand that when it is "weakly solved", it is saracstic and that it means I don't believe in it. Sure you understand sarcasm or you dont :) Right after that even I mentioned as their 5-sigma could be considered as a discovery statistically speaking.

One last time. Computers can't give you a probablistic win or draw in an entire opening like the King's Gambit. It's far too complex. It's far too deep.

A crude idea, maybe but no more.

Same for the Benko Gambit.

All you've managed to accomplish is pull me off track a bit with your tangential arguments.

The point was I never claimed anything other than the Chessbase article was a prank.
By now you have asked about three or more questions that you think I belived in it and also showed that you didn't understand the discussion at all even though the debate was 99.99999% figure.
You're hopeless. That's why Dan Honeycutt gave up.
The point is you simply replied that it is an april fools without reading anything like every rybka hater would naturally do. Dan tried to referee without looking at the context of the posts.. probably belivied that I belived in it too. Any more questions ?

You are not hopeless. You learned some today :)

Terry McCracken
Posts: 15798
Joined: Wed Aug 01, 2007 2:16 am
Location: Canada

Re: Luke skywalker has done it again.

Post by Terry McCracken » Wed Apr 04, 2012 7:03 pm

:roll:
Terry McCracken

Uri Blass
Posts: 7993
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Luke skywalker has done it again.

Post by Uri Blass » Wed Apr 04, 2012 9:36 pm

Terry McCracken wrote:
One last time. Computers can't give you a probablistic win or draw in an entire opening like the King's Gambit. It's far too complex. It's far too deep.

A crude idea, maybe but no more.

Same for the Benko Gambit.
I also believe that computers of today with some reasonable evaluation function cannot prove +5.12 for king's gambit or benko's gambit but it is only an opinion and not a known fact.

I believe that you are not going to be able to prove it(similiar to another opinion that is that white does not lose in chess with perfect play).

Daniel Shawul
Posts: 3429
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Even More

Post by Daniel Shawul » Wed Apr 04, 2012 9:54 pm

Infact their claim is that 10^18 unique positions (state space complexity) are possible to solve not that many number of variations which I thought it was. The game tree complexity for checkers is actually about 10^30 to 10^40. For chess using Rybka with a BF of 1.8 , the number of variations ( game tree complexity ) is 10^18 already, the state space complexity would be much lower. So it is not only probabilisticaly solvable but also by a huge margin at that . It is stated 10^18 unique positions are possible which is taken from checkers effort.

If you talk about real solutions there is no need to mention Rybka at all. They suggested using something much weaker than even TSCP ... Many people were asking which version of Rybka was used for the experiment which shouldn't have mattered if you were after exact proofs. So it is either no 99.99999% claim or if you do that don't use that lame explanation which assumes use of dumb alpha-beta.

Daniel Shawul
Posts: 3429
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Luke skywalker has done it again.

Post by Daniel Shawul » Wed Apr 04, 2012 9:57 pm

Uri Blass wrote:
Terry McCracken wrote:
One last time. Computers can't give you a probablistic win or draw in an entire opening like the King's Gambit. It's far too complex. It's far too deep.

A crude idea, maybe but no more.

Same for the Benko Gambit.
I also believe that computers of today with some reasonable evaluation function cannot prove +5.12 for king's gambit or benko's gambit but it is only an opinion and not a known fact.

I believe that you are not going to be able to prove it(similiar to another opinion that is that white does not lose in chess with perfect play).
Yes I do belive that too but they were after a probabilistic proof using Rybka not an exact proof.

Daniel Shawul
Posts: 3429
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Even More

Post by Daniel Shawul » Wed Apr 04, 2012 10:30 pm

Quesion:
So this means that the result is not 100% certain, it is just a hypothesis.

Answer:
That is technically correct, similar to the assertion that a position where one side is more than two pieces down, without any compensation, is considered lost, even if you cannot calculate it to a forced mate against any defence. Sure, there theoretically might be a way to save the game, but if Rybka is displaying +5.12 or more the outcome is 99.99999999% secure. That is approximately the confidence number we give to our King's Gambit results: 99.99999999%. It might be that there is a flaw somewhere, but if there is it will not be discovered in the course of this universe – that would require more computational power than could ever be provided. And of course it is possible, and in fact very, very likely, that there is no flaw.
For those who doubt that it was not a probablistic proof that was implied at first.


The second line also clearly states that it would require a huge amount of work more than course of universe. So why would anyone think that it was an exact proof this is implied is beyound me.

mcostalba
Posts: 2679
Joined: Sat Jun 14, 2008 7:17 pm

Re: Even More

Post by mcostalba » Thu Apr 05, 2012 4:40 am

Daniel Shawul wrote: The second line also clearly states that it would require a huge amount of work more than course of universe. So why would anyone think that it was an exact proof this is implied is beyound me.
It was an 1st April joke and you've been fooled. Stop, nothing more. Live with that. There are worst things in life. There is no need to write 100 posts just to keep highlighting this to us. Even an 8 year old kid would have gave up and passed to something else.

Locked