Anticomputer chess live and kicking

Originally posted by @paul-leggett
OK, I read GM Smerdon's blog, and it has changed my opinion from my original reaction, which was poor.

I am going to get the kindle edition in the next few days, and attempt a superficial review ( I work for Disney, and the holidays are crazy for us).

Based on Smerdon's review, I suspect that those of us who have read Kmoch and Berliner will sa ...[text shortened]... hild, and his name is Tsvetkov!" -which I intend to be complimentary and funny at the same time!

Thank you, Mr. Leggett.
I suspect this is the right approach: to judge things/people on their merits.

Mickey Mouse is always funny.

Originally posted by @lyudmil-tsvetkov
Thank you, Mr. Leggett.
I suspect this is the right approach: to judge things/people on their merits.

Mickey Mouse is always funny.

The only part I don't understand about you playing here is this: It appears you say you can beat engines. If so why don't you tackle the engine users here directly? We have time controls of 1 day with no added time or 3 days no added, or 3 and 3 or 3 and 7 or 7 and 7 and so forth.
I would love to see you trounce the engineers here🙂

Originally posted by @lyudmil-tsvetkov
Thank you.

No problem.

Originally posted by @sonhouse
The only part I don't understand about you playing here is this: It appears you say you can beat engines. If so why don't you tackle the engine users here directly? We have time controls of 1 day with no added time or 3 days no added, or 3 and 3 or 3 and 7 or 7 and 7 and so forth.
I would love to see you trounce the engineers here🙂

I might play at some point, I am just very busy right now.

Good Luck with the book. Sadly it does not have the word 'Opening'
in the title so it will get ignored by the majority of punters.

----

Hi Paul,

I read the review as well and the first thought I had was Hans Kmoch's 'Pawn Power'.
Smerden is 33, the book was published in 1959 but it is pretty famous. I am surprised
he never spotted the link or maybe he has never heard of it.

Just some of Kmoch's terms.

Buffer duo Two opposing duos facing each other with one rank in between.
Candidate Unopposed or half-free pawn.

Conditional backwardness A pawn that is backward in only certain respects.

Faker A half-free pawn with inadequate helpers.

Frontspan Vertical distance between a pawn and the forward edge of the board.

Mute chain lever A chain lever in which the bases of the opposing pawn chains are not attacked, e.g., a5, b4, c3 vs. a7, b6, c5. Doesnt produce passers.

Ram Two deadlocked, directly opposing pawns.

Telestops Squares beyond the stop square in the frontspan of a pawn.

Tight duo A duo in contact with an opposing pawn(s), whose axis forms a ram.

The list goes on and on....and on....

Don't know why people are so interested in being able to beat a computer.
Why bother? The game is human v human and exploiting human blunders.
A machine cannot and does expect a blunder. It has no hope.
The thing does not even know it is playing a game.

They would lose games from positions 1300 players have won because
they cannot set a basic two move trap.

Originally posted by @greenpawn34
Good Luck with the book. Sadly it does not have the word 'Opening'
in the title so it will get ignored by the majority of punters.

----

Hi Paul,

I read the review as well and the first thought I had was Hans Kmoch's 'Pawn Power'.
Smerden is 33, the book was published in 1959 but it is pretty famous. I am surprised
he never spotted the link o ...[text shortened]... lose games from positions 1300 players have won because
they cannot set a basic two move trap.

I'm glad I'm not the only one to see the link with Kmoch- although I suspect we both date ourselves somewhat! Hans Berliner's My System also has the same context about it- less lingo but more chutzpah and somewhat arrogant tone.

My desire to play a computer in chess ranks up there will my desire to test my sprinting skills against a car- or even a bicycle for that matter. Even as practice, I get more..."practical" practice against the guys at the club, or on here.

The book is cheap, and my main motivation is simple- if a Grandmaster says he found interesting practical ideas in the book that he had not seen expressed elsewhere in chess literature, that is a rare thing, and worthy of attention.

Whether I'll get the same value at my level of play is another matter, but the price is right and worth a shot.

Originally posted by @greenpawn34
Good Luck with the book. Sadly it does not have the word 'Opening'
in the title so it will get ignored by the majority of punters.

----

Hi Paul,

I read the review as well and the first thought I had was Hans Kmoch's 'Pawn Power'.
Smerden is 33, the book was published in 1959 but it is pretty famous. I am surprised
he never spotted the link o ...[text shortened]... lose games from positions 1300 players have won because
they cannot set a basic two move trap.

Hi Green Pawn,

thank you very much for your attention.
'Pawn Power' is a great book and a reference in computer chess, though I find
some of its concepts a bit outdated and imprecise. Even though an IM, Hans has not had the option to utilise the opinion of top engines, as well as the access to very large game databases. Some people find it awful, others think it is a gem, for me it is more of a gem, but a bit outdated.
There is a single reason why I am doing this: looking for original ways of thinking, unchartered waters, and the goal to understand the perfect chess player with the help of engines gives me precisely that.

Originally posted by @paul-leggett
I'm glad I'm not the only one to see the link with Kmoch- although I suspect we both date ourselves somewhat! Hans Berliner's My System also has the same context about it- less lingo but more chutzpah and somewhat arrogant tone.

My desire to play a computer in chess ranks up there will my desire to test my sprinting skills against a c ...[text shortened]... t the same value at my level of play is another matter, but the price is right and worth a shot.

I guess the place where I first visited and this one are 2 quite different places. 🙂

Hi Lyudmil,

'Pawn Power' is like 'My System' and Marmite. You love them or hate them.

All for original thought and ideas. Kmoch outlandish terms are known to work as
a memory aid and the term 'Passer' has been firmly established in chess lore.
(Thankfully a lot of the others have failed to grip the public.)

Not sure if learning how to beat to beat or play against a computer is beneficial
to OTB play v human. But if the reader gets just one thing from your book that he
can point back to later years and say: "Yes that bit won me games' then you have succeeded.

This diagram shows the basic flaw in modern computers. (White to play)


It will see 1.Kh2 is mated in 3 moves and 1.Kh4 is mated in 8 moves.
It will always play 1.Kh4 and yet 1.Kh2 sets the stalemate trap 1....f1=Q =.
lots of won game have been chucked due to a thoughtless automatic promotion.
The computer will of course see it but can never play such a move as 1.Kh2.

Again another example from a human v human game. (White to play)


A computer will always play 1.Kc3, 1.Kb3 or 1.Ka3.
The human played 1.Ka1 Kc1 stalemate.

They simply cannot roll the dice to hope for a blunder.
Yet in human v human chess this can and does work.

Take it down a level and instead of looking for mates look for evaluations.

It's the computers move, it has two choices. one moves evaluates at 0.01
to opponent the next choice evaluates it at 0.09. The computer no matter
the position will always take the 0.01 option even if playing the 0.09
moves sets the human incalculable problems OTB problems to solve.
They have no idea what an incalculable position for a human is.

Chess history is littered with examples of humans knowingly not playing
the best move because of the OTB problems (and traps) it sets their opponent,
especially in lost or poor positions. This is what humans do.

Originally posted by @greenpawn34
Hi Lyudmil,

'Pawn Power' is like 'My System' and Marmite. You love them or hate them.

All for original thought and ideas. Kmoch outlandish terms are known to work as
a memory aid and the term 'Passer' has been firmly established in chess lore.
(Thankfully a lot of the others have failed to grip the public.)

Not sure if learning how to beat ...[text shortened]... nd traps) it sets their opponent,
especially in lost or poor positions. This is what humans do.

I remember once in an otb tournament where a friend of mine had the opportunity to liquidate into a bishop and knight vs king ending.

He did not have the confidence to do it, so instead he gave a piece back and transformed the position into a clearly better but not necessarily winning position, although he converted in the end.

Of course we gave him grief, resolutely refusing to admit we would have done the same thing back then!

Originally posted by @greenpawn34
Hi Lyudmil,

'Pawn Power' is like 'My System' and Marmite. You love them or hate them.

All for original thought and ideas. Kmoch outlandish terms are known to work as
a memory aid and the term 'Passer' has been firmly established in chess lore.
(Thankfully a lot of the others have failed to grip the public.)

Not sure if learning how to beat ...[text shortened]... nd traps) it sets their opponent,
especially in lost or poor positions. This is what humans do.

Thanks GreenPawn.

I appreciate your feedback very much and partly agree with some of your points, though,
I will beg to differ on the usefullness of engines.
For me, especially top engines, are extremely useful when it comes to improving one's chess abilities, so that, even if that was the only good aspect of chess engines, I would whole-heartedly endorse their wider use.
Of course, we are humans and shall remain so, though there is some intriguing 'perfect chess' notion that entices us to dive deeper, even if in the process we lose a part of our humanity.

Originally posted by @paul-leggett
I remember once in an otb tournament where a friend of mine had the opportunity to liquidate into a bishop and knight vs king ending.

He did not have the confidence to do it, so instead he gave a piece back and transformed the position into a clearly better but not necessarily winning position, although he converted in the end.

Of course we gave him grief, resolutely refusing to admit we would have done the same thing back then!

Happy he did not get the pair of bishops vs knight ending, which is also winnable, but requires some good knowledge and technique.
Not to mention a pawnless endgame winnable in 510 moves...

Hi Lyudmil,

I don't mind people disagreeing with me, most do, and I agree the computer is a useful tool.

But I also point to the 1000's of very good players who emerged before 1995
when computers 'got good'.

Most of us are home and casual players. I do not think anyone on RHP
earns their living by playing chess so most of us would be very happy
being just half as good as some of the great players from the pre computer era.

Apart from a few gifted individuals these old masters got there with the book,
board and study method. So to improve a computer is not really needed.
That method has a proven track record. It works if you are willing to put the time in.

In my opinion the computer is overused and over evaluated at the lower level.
Weaker players use them to do their thinking and exploring for them and the
hype around these things makes it seem it is a must have tool. It's not.

At the top level yes. But these guys really know what they are doing
and use it correctly as a tool and not as a crutch.

Also I have faith that one day with some radical new programming
they will design a computer that knows what a difficult position is
for a human to solve and it will be able to set 20-30 moves deep traps
that should the player step into it on move one then there is no escape.

At the moment they can see them but cannot play them because perhaps
on move 15 the eval drops by 0.05. When that day comes I expect some
wonderful new opening variations and middle game sac-sac attacks to appear
as it considers not playing the best or safest move but looks for a human type error.

This is from a game on here (Black to play)


All computers I have tested choose in order 1...Kxb7 then 1...Kb8 or 1...Kc7.

The human played 1...Kc7 setting a 'you can win this trap' which the
White human player fell for playing 2.Ka6 expecting and only seeing
2...d3 3. Ka7 White promotes first and wins.

Instead Black played 2....Kb8 and White resigned.



What is needed is a program that will always choose in such positions
1...Kc7 looking for an error (which is human trait) rather than expecting best play.

Btw., seemingly, the last Alpha-SF match only confirms some of the conclusions from my book:
- that the French is close to lost opening
- that 1.c4 could be best
- and that kingside fianchetto and long pawn chains are indeed very important

Only thing I have fully failed is my claim 1.d4 is weak, as Alpha plays this very frequently, but again, it does not know 1...c5 draws easily.

Openings played a vital role in this match, SF lost mainly due to openings.

Look at number 5: https://www.amazon.com/Chess-Board-Games-Puzzles-Books/b?ie=UTF8&node=4406

Chessbase reprinted David Smerdon's review, and some people have hurried to buy the book, true, for each ebook purchase, I am getting just a bit more than 2 dollars.

Chess Base: https://en.chessbase.com/post/the-secret-of-chess