Stockfish is blind!

Sort:
Avatar of Lyudmil_Tsvetkov
MARattigan wrote:
Lyudmil_Tsvetkov wrote:
MARattigan wrote:

Here's an obviously drawn position that Stockfish evaluates as -72.88.

                                                              White to play.

 

You might have put 10 more light-square bishops: some engines and GUIs accept more pieces than that.

I thought I'd stick to a legal position. But even -73 suggests SF doesn't quite understand that kind of position.

 

Otherwise I could have added 24 more light bishops giving an evaluation of -103.60.

 

                                                        White to play and draw.
 

 Of course if we're adding illegal numbers of bishops we could add illegal numbers of other pieces as well. I notice Stockfish evaluates the one below as -127.99. But is it drawn? You could probably tell me that off the top of your head.

 

 


                                                      White to play and ?

 T've inserted how computer level 10 plays it on chess.com anyhow. (I think it's SF). Looks pretty blind to me.

                                                        White to play and win.

With this position, you might have added 10 more black pawns too, to make SF even dumber. happy.png

Avatar of Lyudmil_Tsvetkov
mcris wrote:

About #297: Why others has not created fake games against Stockfish? Is this what you are asking?

Also on the book "The secret of chess" I already seen such a book and it concentrates much on Dutch defense, which for sure is not on AlphaZero openings list.

Some kind of incomprehensible blabber...

Avatar of Lyudmil_Tsvetkov
FBloggs wrote:
Lyudmil_Tsvetkov wrote:
FBloggs wrote:

Stewardandstewardj makes a good point.  Everyone familiar with Stockfish knows it's stronger than any grandmaster. If a grandmaster was able to beat it, that would be major news in the chess world. So why isn't it major news that a Bulgarian candidate master has beaten it? Because it hasn't been confirmed by independent evidence. It's just a claim. People make outlandish claims all the time. It's not news. Anyone can post games and claim they were played against Stockfish.

But not everyone does that, why so?

You got me dead tired with your constant obstructions, are not you(I mean all) going to say something positive?

The fact that not everyone posts games and claims they were played against Stockfish is proof that the games you've posted are legitimate?

No, the proof is that, when you take 100 random positions from 'Human vs Machine' and feed tham to SF, it will assess them ALL wrongly.

If you take someone other's game collection, SF might be wrong on only 20 out of 100 at most, most probably just 10.

That is an irrefutable proof: SF does not understand such positions, so it is quite beatable there.

You might also want to replay the games move by move, and you will be very surprised(for an unbeliever) to see SF chooses the very same moves as in the book, at least in the great majority of cases, no matter the hardware.

Those are simple, irrefutable patterns, no need to search any further.

Avatar of Lyudmil_Tsvetkov
FBloggs wrote:
Lyudmil_Tsvetkov wrote:

I am eagerly waiting for SF to release version 9, so that I start thrashing it with new impetus(I am more experienced now, you know), but they are afraid to release it.

They're afraid to release the new version because they know you will thrash it?  In the never-ending debate over which chess player is the best of all time, names like Fischer, Kasparov and Carlsen are always mentioned.  The name Tsvetkov is never mentioned.  But now we're informed that not only is the greatest player of all time someone who never held the world championship but never earned the title of master either.  Yet he is not only the best ever, he's the only player able to beat the strongest chess engines.  It's remarkable that such earthshaking news has been ignored by the entire chess world except for a couple of forum threads on a chess website. 

A player who is beating the top engines is definitely much stronger than master-level.

Indeed, it is remarkable the news still has not gotten viral, but there is simple explanation for that: just as Kasparov gets 1/10 or even 1/100 the attention Messi gets, because there are 100 times less chess followers, similarly I get proportionally less attention, because very few people on this Earth are meditating instead of simply thinking and this makes my audience even smaller.

But I still hope at some point people will get interested: one person just suggested to me 'The Secret of Chess' is half a century ahead of its time...

Avatar of Lyudmil_Tsvetkov
greekgift_221b wrote:
FBloggs wrote:
Lyudmil_Tsvetkov wrote:

I am eagerly waiting for SF to release version 9, so that I start thrashing it with new impetus(I am more experienced now, you know), but they are afraid to release it.

They're afraid to release the new version because they know you will thrash it?  In the never-ending debate over which chess player is the best of all time, names like Fischer, Kasparov and Carlsen are always mentioned.  The name Tsvetkov is never mentioned.  But now we're informed that not only is the greatest player of all time someone who never held the world championship but never earned the title of master either.  Yet he is not only the best ever, he's the only player able to beat the strongest chess engines.  It's remarkable that such earthshaking news has been ignored by the entire chess world except for a couple of forum threads on a chess website. 

Well it might be Tzvetkov, not Tsvetkov that is mentioned as one of the best ever.

Actually, Tsvetkov is a gradmaster, and he has beaten Alekhine: https://en.wikipedia.org/wiki/Alexander_Tsvetkov  happy.png

Avatar of MARattigan
Lyudmil_Tsvetkov wrote:
MARattigan wrote:
Lyudmil_Tsvetkov wrote:
MARattigan wrote:

Here's an obviously drawn position that Stockfish evaluates as -72.88.

                                                              White to play.

 

You might have put 10 more light-square bishops: some engines and GUIs accept more pieces than that.

I thought I'd stick to a legal position. But even -73 suggests SF doesn't quite understand that kind of position.

 

Otherwise I could have added 24 more light bishops giving an evaluation of -103.60.

 

                                                        White to play and draw.
 

 Of course if we're adding illegal numbers of bishops we could add illegal numbers of other pieces as well. I notice Stockfish evaluates the one below as -127.99. But is it drawn? You could probably tell me that off the top of your head.

 

 


                                                      White to play and ?

 T've inserted how computer level 10 plays it on chess.com anyhow. (I think it's SF). Looks pretty blind to me.

                                                        White to play and win.

With this position, you might have added 10 more black pawns too, to make SF even dumber.

I did originally throw in a few extra black pawns but I edited it because I thought it was too obviously won. The Stockfish evaluation didn't actually increase over the one I settled on, I think the evaluation must be limited to 128.

 

My version of Tarrasch/Stockfish looks even dumber anyway. If you sneak the position past Tarrasch, Stockfisch just falls over as soon as it's invited to play. 

Avatar of FBloggs
Lyudmil_Tsvetkov wrote:

A player who is beating the top engines is definitely much stronger than master-level.

Indeed, it is remarkable the news still has not gotten viral, but there is simple explanation for that: just as Kasparov gets 1/10 or even 1/100 the attention Messi gets, because there are 100 times less chess followers, similarly I get proportionally less attention, because very few people on this Earth are meditating instead of simply thinking and this makes my audience even smaller.

But I still hope at some point people will get interested: one person just suggested to me 'The Secret of Chess' is half a century ahead of its time...

It's not at all remarkable that the "news" of your exploits hasn't gone viral.  The simple explanation is that it's not news.  If you beat one of the top chess engines in a match, that's news.  If you merely claim that you've beaten one, that's not news.

Avatar of DMAlphaZero
Lyudmil_Tsvetkov wrote:
FBloggs wrote:
Lyudmil_Tsvetkov wrote:

I am eagerly waiting for SF to release version 9, so that I start thrashing it with new impetus(I am more experienced now, you know), but they are afraid to release it.

They're afraid to release the new version because they know you will thrash it?  In the never-ending debate over which chess player is the best of all time, names like Fischer, Kasparov and Carlsen are always mentioned.  The name Tsvetkov is never mentioned.  But now we're informed that not only is the greatest player of all time someone who never held the world championship but never earned the title of master either.  Yet he is not only the best ever, he's the only player able to beat the strongest chess engines.  It's remarkable that such earthshaking news has been ignored by the entire chess world except for a couple of forum threads on a chess website. 

A player who is beating the top engines is definitely much stronger than master-level.

Indeed, it is remarkable the news still has not gotten viral, but there is simple explanation for that: just as Kasparov gets 1/10 or even 1/100 the attention Messi gets, because there are 100 times less chess followers, similarly I get proportionally less attention, because very few people on this Earth are meditating instead of simply thinking and this makes my audience even smaller.

But I still hope at some point people will get interested: one person just suggested to me 'The Secret of Chess' is half a century ahead of its time...  HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHHAAHHAHAHAHAHAHAH

YOU ARE GREAT AT JOKING AROUND!

 

Avatar of DMAlphaZero
Lyudmil_Tsvetkov wrote:
greekgift_221b wrote:
FBloggs wrote:
Lyudmil_Tsvetkov wrote:

I am eagerly waiting for SF to release version 9, so that I start thrashing it with new impetus(I am more experienced now, you know), but they are afraid to release it.

They're afraid to release the new version because they know you will thrash it?  In the never-ending debate over which chess player is the best of all time, names like Fischer, Kasparov and Carlsen are always mentioned.  The name Tsvetkov is never mentioned.  But now we're informed that not only is the greatest player of all time someone who never held the world championship but never earned the title of master either.  Yet he is not only the best ever, he's the only player able to beat the strongest chess engines.  It's remarkable that such earthshaking news has been ignored by the entire chess world except for a couple of forum threads on a chess website. 

Well it might be Tzvetkov, not Tsvetkov that is mentioned as one of the best ever.

Actually, Tsvetkov is a gradmaster, and he has beaten Alekhine: https://en.wikipedia.org/wiki/Alexander_Tsvetkov 

on wiki anyone can make things up

Avatar of MARattigan
MARattigan wrote:

Here's an obviously drawn position that Stockfish evaluates as -72.88.

                                                              

 

Conversely SF will also evaluate won positions such as this Ottó Bláthy composition as drawn.

 

                                                      White to play and win.

 

Avatar of Lyudmil_Tsvetkov
FBloggs wrote:
Lyudmil_Tsvetkov wrote:

A player who is beating the top engines is definitely much stronger than master-level.

Indeed, it is remarkable the news still has not gotten viral, but there is simple explanation for that: just as Kasparov gets 1/10 or even 1/100 the attention Messi gets, because there are 100 times less chess followers, similarly I get proportionally less attention, because very few people on this Earth are meditating instead of simply thinking and this makes my audience even smaller.

But I still hope at some point people will get interested: one person just suggested to me 'The Secret of Chess' is half a century ahead of its time...

It's not at all remarkable that the "news" of your exploits hasn't gone viral.  The simple explanation is that it's not news.  If you beat one of the top chess engines in a match, that's news.  If you merely claim that you've beaten one, that's not news.

What do you want me to do?

The stronger players are discussing my concepts, and you want just bare stats.

It's a pity I wrote those books, people don't like original stuff, the majority, of course.

There are fully untitled and even unrated players, who write books and get perfect acclaim: for example Ntirlis for Quality Chess.

There are even persons at the level of 1800 who write books and try to teach chess.

With them and their books, everything is perfect, I am the only criminal.

 

I what way writing a good book should be a bad thing and its author flogged to death?

Pity indeed I started that...

Avatar of Lyudmil_Tsvetkov

Well, just read this review: some people know what it is all about, but you need to be very intelligent and a very good chess player.

https://www.amazon.com/gp/customer-reviews/R1YE78TYAZQ10Z/ref=cm_cr_dp_d_rvw_btm?ie=UTF8&ASIN=1522041400#wasThisHelpful

 

If Kasparov reads the book now, he can also confirm it is very high quality and offers a wealth of new ideas, but he will hate to do that, as this challenges all established theory, and also most of the established hierarchical order in chess.

 

Really, the book should have been received much more enthusiastically, but that only points it is ahead of its time.

 

 

Avatar of Lyudmil_Tsvetkov

Just read that: I used some of these ideas to mate Stockfish twice this week. Full Disclosure - it was running on my phone, not a powerful computer, and I ran out of time both games, so I guess technically I lost wink.png But a mate is a mate, and I'll take it.

 

The guy beat SF twice, using the knowledge expounded in 'The Secret of Chess'.

Does not that prove the knowledge is worth it?

Did Kasparov or any other GM offer you knowledge with which to beat the top engines?

Did he?

So, I simply suggest you buy the book, I will feel a bit more relieved this way too.

 

One way or another, this discussion starts getting ugly, I am not ready to take all those insults.

After all, I did not do anything wrong/bad, apart from writing a good book...

Avatar of DanielWrench

of course stockfish is blind. Its not a living entity.

Avatar of stewardjandstewardj
Lyudmil_Tsvetkov wrote:

Just read that: I used some of these ideas to mate Stockfish twice this week. Full Disclosure - it was running on my phone, not a powerful computer, and I ran out of time both games, so I guess technically I lost But a mate is a mate, and I'll take it.

 

The guy beat SF twice, using the knowledge expounded in 'The Secret of Chess'.

Does not that prove the knowledge is worth it?

Did Kasparov or any other GM offer you knowledge with which to beat the top engines?

Did he?

So, I simply suggest you buy the book, I will feel a bit more relieved this way too.

 

One way or another, this discussion starts getting ugly, I am not ready to take all those insults.

After all, I did not do anything wrong/bad, apart from writing a good book...

 I was on a 4-day trip, so I couldn't respond

How is it that a person CLAIMING to beat StockFish proves anything? I have not found any article on anyone beating StockFish, and a GM lost couldn't do more than tie 1 game out of 4 games WITH THE HELP FROM AN ENGINE!

Since you just love to avoid half of the question being asked, please answer each of the following questions numbered in order.

1. How can you beat an engine that calculates tens of millions of positions a second?

2. How can you beat this engine when no one else in the world has shown proof of such things?

3. How come you have not shown any proof? (posting the game is not proof, anyone can post a game like that)

Avatar of FBloggs
stewardjandstewardj wrote:

 

Since you just love to avoid half of the question being asked, please answer each of the following questions numbered in order.

1. How can you beat an engine that calculates tens of millions of positions a second?

2. How can you beat this engine when no one else in the world has shown proof of such things?

3. How come you have not shown any proof? (posting the game is not proof, anyone can post a game like that)

You know you're not going to get satisfying answers.  You'll get something similar to the following:

1.  It doesn't matter how many positions it calculates per second.  His pattern recognition system is superior and makes calculation unnecessary.

2.  Because no one else in the world has mastered his pattern recognition system.

3.  The system is the proof.  Further proof is unnecessary.

If you argue that his claims are ludicrous, he'll dismiss your argument out of hand and make yet another ludicrous claim.  That's his story and he's sticking with it.

Avatar of MARattigan
FBloggs wrote:
stewardjandstewardj wrote:

 

Since you just love to avoid half of the question being asked, please answer each of the following questions numbered in order.

1. How can you beat an engine that calculates tens of millions of positions a second?

2. How can you beat this engine when no one else in the world has shown proof of such things?

3. How come you have not shown any proof? (posting the game is not proof, anyone can post a game like that)

You know you're not going to get satisfying answers.  You'll get something similar to the following:

1.  It doesn't matter how many positions it calculates per second.  His pattern recognition system is superior and makes calculation unnecessary.

2.  Because no one else in the world has mastered his pattern recognition system.

3.  The system is the proof.  Further proof is unnecessary.

If you argue that his claims are ludicrous, he'll dismiss your argument out of hand and make yet another ludicrous claim.  That's his story and he's sticking with it.

Obviously no need to answer any questions when you have someone to do it for you.

Avatar of KineticPawn

What doesn't make sense to me is why hasn't he already played ann engine under "quiet" but controlled conditions?

1- Money is obviously a motivator. A victory would be a financial windfall.

2-Increase the amount of people reading and buying his books. 

So why no controlled games?

 

Avatar of FBloggs
TheJackalC4 wrote:

What doesn't make sense to me is why hasn't he already played ann engine under "quiet" but controlled conditions?

1- Money is obviously a motivator. A victory would be a financial windfall.

2-Increase the amount of people reading and buying his books. 

So why no controlled games?

 

It makes sense to me.  He wouldn't have a chance.  

Avatar of Lyudmil_Tsvetkov
stewardjandstewardj wrote:
Lyudmil_Tsvetkov wrote:

Just read that: I used some of these ideas to mate Stockfish twice this week. Full Disclosure - it was running on my phone, not a powerful computer, and I ran out of time both games, so I guess technically I lost But a mate is a mate, and I'll take it.

 

The guy beat SF twice, using the knowledge expounded in 'The Secret of Chess'.

Does not that prove the knowledge is worth it?

Did Kasparov or any other GM offer you knowledge with which to beat the top engines?

Did he?

So, I simply suggest you buy the book, I will feel a bit more relieved this way too.

 

One way or another, this discussion starts getting ugly, I am not ready to take all those insults.

After all, I did not do anything wrong/bad, apart from writing a good book...

 I was on a 4-day trip, so I couldn't respond

How is it that a person CLAIMING to beat StockFish proves anything? I have not found any article on anyone beating StockFish, and a GM lost couldn't do more than tie 1 game out of 4 games WITH THE HELP FROM AN ENGINE!

Since you just love to avoid half of the question being asked, please answer each of the following questions numbered in order.

1. How can you beat an engine that calculates tens of millions of positions a second?

2. How can you beat this engine when no one else in the world has shown proof of such things?

3. How come you have not shown any proof? (posting the game is not proof, anyone can post a game like that)

Post one such game of yours and I will be happy.

It's not only me: everyone who follows my approach is able to beat top engines, you are the last couple of guys still unable to do so. happy.png