Not logged inRybka Chess Community Forum
Up Topic The Rybka Lounge / Computer Chess / Human versus Machine
1 2 Previous Next  
- - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-08 05:48
Anyone interested in human-computer matches might check my new book: (paperback) (ebook)

also available on by author and title),, etc.

Amply commented and diagrammed games.

Seems like the first book with extensive coverage of a large number of winning games against the top engines.

Kasparov, Carlsen and Nakamura still have not written one. :)
Parent - - By Kreuzfahrtschiff (***) [de] Date 2017-10-08 09:12
u r so funny  :)
Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-09 10:30
Thank you.

I am funny, but the games are real.

You can judge that even only by looking at the arising positions:
no one plays like that, neither humans, nor engines, you can not find similar
games in any database, so someone should have contrived the system.

You think it is an extraterrestrial?
Parent - - By APassionforCriminalJustice (***) [ca] Date 2017-10-08 13:18
Lol, Kasparov's book would be more like him discussing how Stockfish would utterly destroy him. Heck he even thought that Rybka was an extremely strong engine. Deep Blue was junk. He wouldn't waste his time talking about that. And Fritz... who cares.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-09 10:32
Kasparov is weak, man, learn from people who can beat Stockfish. :wink:
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-10 03:08
One thing I wonder is why routine books like for example a book treating the Budapest Gambit in the most usual of ways, a book entitled
something like 'Tactics in the Budapest Gambit', or 'Winning Tactics in the Budapest Gambit', that actually just takes ready-made samples
out of some game database, filters the games, and then shows some very obvious tactical solutions, shallow at that, would get much more
attention than a book treating a completely new, original and unsurveyed subject, like the way a human can beat the top engines?

After all, the book about the Budapest(which, btw., might be altogether lost with perfect play) is extremely routine and unoriginal, one could change it
for any good database, while the other book treats topics that have not been treated before.

Why would anyone prefer the first book, any guess?

Has the modern world become so zombied into following routine and repetitiveness, that it would not like anything new?

In the past, people used to cherish new and unchartered waters, but not any more?

In the past, writers who offered something new were highly respected and sought after, but not now?
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-10 03:09
Comparing 'The Secret of Chess' and 'Human versus Machine: How to beat Stockfish and Komodo',
I wrote the latter much quicker, the former took whole 4 months, but the interesting thing is
how notions presented in 'The Secret of Chess' are visible in the games showcased in 'Human versus Machine'.

For example, the games exhibit patterns and notions like:

- twice backward shelter pawn on f7
- pointed chains
- white and black KID structures
- fully closed sides of the board, etc., etc.

all of which could be found in 'The Secret of Chess'.

Of course, it is actually the other way round: the many thousands of games(over 50 000, to be clear)
I have played against engines and top engines and the knowledge I derived from them are reflected
in the knowledge presented on the pages of 'The Secret of Chess'.

That is how I verified that knowledge: by playing an infinite number of games against the very top,
and it seems to work.

If anyone would like to consider the games in 'Human versus Machine' as fake ones, well,
you simply don't have a point, looking at the specific positions, you will not find even a single one
that even distantly resembles any human or engine game you could find in any database.

There are simply no such games and positions, so who came up with the concept and system?
Also, checking evaluations, you will easily see the games are for real. Current Stockfish development version
still does not understand most of them.

Again, why would beating Stockfish and Komodo be less interesting than reproducing a routine game from a
public database?
Parent - - By Labyrinth (*****) [us] Date 2017-10-10 15:39 Edited 2017-10-11 14:33

>Again, why would beating Stockfish and Komodo be less interesting than reproducing a routine game from a public database?

What exactly are you claiming here? That you can, from the starting position, reliably defeat Komodo and Stockfish due to a weakness you found, or you found a weakness in certain given, unusual positions?
Parent - - By MarshallArts (***) [us] Date 2017-10-10 17:40
I think he's saying he can steer the game into those types of positions the engines are having trouble with.

Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-11 02:56 Upvotes 1

Play 1. e4.
Stockfish frequently replies with 1. e6, the French, than play
2. d3,
Stockfish will play d5(doubtless the best move)
than continue with 3. Nc3,
Stockfish will kick the knight with d4(this is already a positional inaccuracy, that will
allow white to take the initiative on the king side)
4. Ne2, and then play g3, Bg2, f4, and you easily start outplaying Stockfish.

That is the King's Indian Attack(KIA), that Fischer played for a reason, but I treat it
in a bit different way.

Similarly, you can choose the Dutch Stonewall, or the Stonewall Attack for white, or different
pointed chains setups(for example, similar to those arising in the French Winawer).

For precise trreatment, though, you should check my book.
Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-11 02:47
You are right, I can reliably defeat Stockfish and Komodo(never owned any Houdini version since 3.0)
from the starting position, probably any 1 in 10 games or so, but also draw a big portion of the rest.

Stockfish and Komodo don't have a single weakness, they have multiple evaluation and search weaknesses.

Btw., if you play through the games of my newest book, and after learning the patterns,
you might be able to do the very same, maybe not in the very first game you play, not in the second,
but after you practised 10 games or so.

there are just patterns one has to learn.
Parent - - By MarshallArts (***) [us] Date 2017-10-10 17:44
look what just came out:

They never give up :wink:
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-11 03:01
I was not referring to this particular book, but you are right again.

Scrolling down the page, I see Moskalenko scored 3 wins and 8 draws from 11 games:
not very convincing, is it, and he probably played lower-rated opposition?

btw., it is almost certain that the Budapest is altogether lost with perfect play:
I have checked that many times with Stockfish.

I could provide you with the lines now, but really don't have the time.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-13 01:03
I see a real danger, though, in investigating this and similar books.

People will start winning too many games against Stockfish, and Stockfish
will badly need another 400 elo to ward off competition attempts successfully. :)
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-27 06:08
Part II of the book is out:

(both paperbacks and ebooks included, just switch between versions)

Also available on other amazons, uk, de, etc.(search by author and title)

In this edition, games against Stockfish DD, Stockfish 5, 6 and Komodo 10 are represented.

7 or 8 different openings featured, basically boiling down to 4 main winning pawn structures:

- Stonewall Attack(Stonewall Defence): pawns on d4-e3-f4, d5-e6-f5 for black
- King's Indian Attack(King's Indian Defence): pawns on d3-e4-f5, d6-e5-f4 in the standard KID for black
- Central chain structure, arising out of the Queen's Pawn Game, Torre Attack(or out of the possible Slav for black): pawns on c3-d4-e5, c6-d5-e4 for black
- Central bind structure, arising out of the English Opening(Sicilian Defence for black): pawns on c4-d3-e4, c5-d6-e5 for black

It is not easy to beat the top engines, so take a look at the games and explanations.
Maybe, you will find that interesting.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-28 01:45
Here just reposting a brief reply of mine to a disbelieving member from a chess forum
(so that people are aware in what way a human may get an advantage against the top engines in the 4 mentioned structures):

Forum member wrote:
It does not make much sense to play king's Indian defense as the computer is especially strong in sharp positions
Playing the stonewall as white is known to be bad as it almost guarantees black a way to trade off his light squared bishops leaving white permanantly crippled and as black it is supposed to be good only under certain circumstances, which you will not get very often.
It's incredibly difficult, if not impossible to get the c6, d5, e4 pawn structure as black in the opening if white doesn't play badly and even if you do get it, you get positions similar to the Caro-Kann and French, which are both well respected and it's hard to see any fantastic advantage you obtain in getting these structures.
The 'central bind structure' is playable as white, but doesn't give much advantage and with white playing properly, it's very difficult to get as black.
However, the computer does have difficulty in playing against the king's Indian attack-like setup you mentioned before (with e4, e6, d3, d5, nc3, d4) and I think that you can get an advantage against it. But an advantage is all and I fail to see how anyone besides another engine can convert it into a win against such powerful defenders.

The king's Indian Defence involving d6-e5-f4 pawns is a closed one and far from sharp, so that is precisely
the position a human would like to get.
KID=KIA with black, so if the KIA is good, the KID is good too. One tempo is not of such a critical significance
at the current level of top engines.
Concerning the Stonewall Attack, indeed, white has fully equal, draw, at most, if black plays Bf5 early on to trade
light square bishops, but, fortunately, even current Stockfish development still prefers e6 and Bb7/a6(not Komodo though).
c6-d5-e4 is not that hard to get, both Stockfish and Komodo like a line like 1. d4 d5 2.c4 c6 3. Nc3 Nf6 4. Nf3 g6!(this is the trick,
definitely strongest continuation) 5. e3 Bg4!(again, best) 6. h3(that is how top engines play) Bg4 7. Qf3(bishop pair lacking,
but the queen is very displaced here) e6, then Bd6/g7, and at some point e6-e5 break is pushed.
de5 is rarely good, so there are excellent chances black will push e5-e4 later on, getting the abovementioned structure.
Of course, as the game is closed, engines see nothing, black will get decisive advantage only 20 moves later after a lot
of regrouping.
On the contrary, the central bind is best possilbe disposition for white at all, as 1. c4 is definitely white's best possible move.
For example, 1. c4! e5 2. Nc3 Nf6 3. e4!(g3 first, followed by Bg2 is also possible), and white gets big advantage, not sure if
winning though.
With black, you can get that for example from the Sicilian, Nimzovich-Rossolimo variation(see the game I just posted), as top engines
still prefer 1. e4 c5 2. Nf3 Nc6 3. Bb5 g6 4. Nc3? Later black plays Bg7 and e5, and the bind is there. Very simple.
Parent - By MachineLearning (*) [au] Date 2017-10-30 00:53
Lyudmil, please stop trolling about your nonsense. You aren't a strong player.
Parent - - By MachineLearning (*) [au] Date 2017-10-30 00:51
A 1600 player is writing a book??


Please don't buy the book. Dr Kevin Bonham explains the reasons in the link.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-30 06:47
Where did you get that from?
If you had tried to google me, you would have known that I am much stronger than that:
- Bulgarian candidate master since 20 years
- Bulgarian master level rating, over 2200
- over 2100 FIDE elo

And all that, dating back 12 years ago, when I had been playing chess just occasionally.

The last 5 years I have spend all my time playing an analysing chess, so calling one of the best
chess players a weak one is a big distortion of the truth, to say the least.

Anyway, thanks for the feedback, at least it sends me back to limelight.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-30 06:52
I have always wondered why there are people who always take such a negative stance?
Why not read what is available about the book, there is a preview, you can read at least the
introduction and the first game.
Did you do that?
What did not you like about it?
Did you find the game uninteresting?
Was it poorly commented?

Please, be specific.

I have always tried to do that, as that is the right approach.
If you are not specific, your words are dust in the wind.
Parent - By Labyrinth (*****) [us] Date 2017-10-30 23:26

>I have always wondered why there are people who always take such a negative stance?

No idea. It's completely unnecessary.
Parent - - By MachineLearning (*) [au] Date 2017-10-31 02:57
This is not negative stance. Dr Kevin Bonham has explained clear and loud why nobody should buy the book. He has given many strong reasons.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-10-31 05:28
For people who have not been following talkchess forum, I just want to
bring back to life an intriguing thread on talkchess, involving a live chess game between me and Stockfish, played in late 2014:
This is just to show how much analytical effort has gone into developing the right strategies to overpower the top engines.
With each move consistently analysed for half an hour, and Stockfish using 16 threads, the amount of knowledge one gets from similar sessions is certainly tremendous.

And that is only one of maybe more than a thousand similar analytical threads on talkchess during the last 5 years.

Some might try to raise cheating allegations against me, but I am worth
what I am worth.
Parent - - By MachineLearning (*) [au] Date 2017-10-31 05:51 Edited 2017-10-31 05:55
Your post is useless. Dr Kevin Bonham has already pointed out in:


you were playing an unfair odds match. Your results meant nothing. It wasn't a fair match, so your results are null. You cheated because you had manipulated the conditions and the moves.

If you are a 2500 rated player, I would be a 5000+ FIDE player. Carlsen would have been a 1000000+ rated player. Let's do the scaling!
Parent - - By free Willy (*) Date 2017-10-31 08:20
@ Lyudmil Tsvetkov
Is this your game against komodo?
huge bullshit, please stop posting such nonsense.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-01 05:37
Dear free Willy, what did not you like about the game?
Too unadvanced for your strength?
Parent - - By free Willy (*) Date 2017-11-01 07:48
Hello, Lyudmil! I have nothing against selling books.
Please don't post such games any more, because that makes you untrustworthy.
Eval of first move e4 is  1.23 at depth 19. Never would Komodo give such an eval.
You have ten times more time.
huge bullshit, please stop posting such nonsense.
peace and out
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-02 05:07
This is because Komodo had been using large contempt factor, just check the whole thread.
And, this was done due to specific insisting on the part of Larry Kaufman, the co-author
of Komodo, that when playing against humans, Komodo should use contempt.
Time control was not 10 times more, for this particular game was 2 min + incr. vs 10 min + incr.,
but only because the engine used contempt, and that usually gives you more tactical puzzles
to solve.
In the majority of games, I am using only twice as much time, and some are even played with at same TC.
So, before making any claims, just check better.
Parent - By free Willy (*) Date 2017-11-02 08:09
sorry this is huge bullshit, please stop posting such nonsense.
Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-01 05:33
You did not even read the thread I posted: this concerns completely different game played
online and followed by a sufficiently large audience.
Stockfish had been showing +200cps in its favour, when I announced mate.

O, sancta simplicitas!
Parent - - By Labyrinth (*****) [us] Date 2017-10-31 17:12
Seems pretty negative to me, and I wasn't impressed with Bonham's arguments, a lot of which seem to be targeted at Tsvetkov's playing strength which is mostly irrelevant. The strongest players are not necessarily the strongest against computers which can require a very specialized understanding, nor are they necessarily good communicators of chess principles meaning that they might not make ideal authors for a book.

Yes it appears some of Tsvetkov's examples are handicap games, but you still have to win such games which is not easy. As computer playing strength increases steadily it becomes harder and harder to find positions in which the computer plays sub-optimally. I would expect someone who finds such things to be someone who spends a lot of time looking for them playing against that particular engine and not simply an overall strong chess player.

I don't know if Tsvetkov's books are good, but I don't see much in the way of direct criticism of the books themselves, I guess because someone has to actually buy a copy and look through it when it is much easier just to write him off as a loon.

Also I don't understand this comparison with ARB at all. I've had many dealings with ARB, and he is not even in the same universe as Tsvetkov. ARB is severely mentally ill and it's a struggle to even hold a coherent conversation with him.

Anyways, if you don't like Mr. Tsvetkov just ignore him, and don't buy his books, or at most, complain about him over-advertising. Anything more is imo uncalled for and reflects poorly on the person doing so. It's perfectly fine to ask questions about his work, but that's not what I'm seeing for the most part.
Parent - - By MarshallArts (***) [us] Date 2017-10-31 21:57
The exhortation not to buy his books is very mean-spirited and basically unfounded.
I can understand if someone does not like the playing conditions which these games
were conducted under, but the criticism is going too far.

His books are quite good actually. I only skimmed through the Secret of Chess, but the newer
human vs machine game books seem packed with good and crisp explanations that can help elevate
a reader's play even against other human players. The games themselves are H vs M masterpieces,
regardless of whatever handicaps were used by the author. A very high level understanding of chess
transpires when looking at these well commented games. I was positively surprised by the quality of
these games and their annotations.

Parent - - By MachineLearning (*) [au] Date 2017-11-01 00:14
No. Lyudmil Tsvetkov has been constantly lying about his computer chess "skills". In the book, website, and even here, he's been telling everybody he's a super strong player, someone with "at least 2800" FIDE strength.

This is obviously nonsense. Unless you believe he's a better player than Magnus Carlsen, everything in the book must be false. If he can't even introduce himself, who would you think the book is worth for anyone to buy?

Many chess book writers without a GM title such as the popular Gary Line, don't claim what they don't have. Lyudmil Tsvetkov is claiming he's the strongest player ever in the history. This is not true, and we can conclude his books are pure rubbish.

You're welcome to waste money on his trash, but he should at least admit what he's before selling his books.

Remember, false advertising is illegal in many countries. That's what he's doing now.
Parent - - By Labyrinth (*****) [us] Date 2017-11-01 01:37

>Unless you believe he's a better player than Magnus Carlsen, everything in the book must be false.

This is logically incorrect.

It doesn't even follow if you improve it by saying "Unless he's a better player than Magnus Carlsen, everything in the book must be false."

(Such illogical statements remind me of ARB more than anything Tsvetkov has said!)

>he's been telling everybody he's a super strong player, someone with "at least 2800" FIDE strength.

I actually heard him say he was more like 2200 or something, but doesn't really matter as I explained. This is a specialized area. It's quite different than overall chess strength as it is generally measured.

>Lyudmil Tsvetkov is claiming he's the strongest player ever in the history.

Also haven't seen this, but w/e.

>Remember, false advertising is illegal in many countries. That's what he's doing now.

Melodramatic hyperbole.
Parent - - By MachineLearning (*) [au] Date 2017-11-01 02:08 Edited 2017-11-01 02:10
No. It's very important. False self-introduction make all his arguments invalid. This is like sending a fake resume for a job application. If a candidate lies on the resume, the interview would be pretty much immediately over. Same logic here. One example - In 2012, Yahoo's former CEO Scott Thompson was fired after the resume scandal. Lyudmil Tsvetkov doesn't have the exception - he has no rights to claim something that is obviously not true for making profits. This is absolutely unethical and illegal.

Your logic is incorrect because you're advocating a scam. Why would a buyer believes there's a value in the book if Lyudmil Tsvetkov claims he's has "at least 2800" rating? Why would an interviewer hire a candidate with a fake resume? Would you do that? Yahoo didn't and they did the right thing by terminating Scott Thompson.

If you want a link, I give you one:

I don't know when people on this forum and elsewhere will finally learn that my objective current chess strength, especially under quiet conditions, is at least 2800?

His books only have value if Lyudmil Tsvetkov is really a respected computer chess writer, who knows what's talking about. Otherwise, it's just rubbish or imagination.

Lyudmil Tsvetkov is as ill as ARB, they could be the same person.
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-01 06:08
Machine learning, obviously you have learned nothing about machines.
The 2800 statement was only within the context of the talkchess forum, where
people know me and we have our own little communication tricks.
I was 2200 back in 2006, with some 2400 performances(what concerns my Bulgarian rating).
Then I was professionally busy and did chess only about 1-2 hours daily on average.
Stockfish and Komodo were not available then to train.
I have not competed after 2006, even a single game, but I have played and analysed 3 times
more games than before.
In the last 5 years, I even quit my job to fully concentrate on chess.
In that period, I have been busy with chess 16 hours per day.
So, it might make sense my current chess strength is better than that more than 10 years ago.
How much better? I don't know.
My personal assessment is a know 3 times as much theory and general chess knowledge
than when I stopped competing, and am 3 times stronger, but that is just my personal assessment.

Are not people allowed to improve their playing strength without competing?
And lastly, as other people also rightly point out, I am not selling myself, but my books.

Parent - - By MachineLearning (*) [au] Date 2017-11-01 06:33
Oh, "3 times stronger"?? 2200*3 = you have a FIDE rating of 6600!!!!

I thought you were just above 2800. But I'm wrong, I'd underestimated you.

No. When you sell your books, you're also selling yourself because people chose who to buy from.
Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-02 05:11
well, I started at 2000, so 200 points to my highest rating,
multiply that by 3, some meagre 600 points, is not much,
is it? :cool:
Parent - By Labyrinth (*****) [us] Date 2017-11-01 09:06

>False self-introduction make all his arguments invalid.

That's an ethical appeal as opposed to a logical one. The problem with ethical appeals is that if the world's most hated or least intelligent person tells you 1+1=2, it's just as true as if Euler himself had told you.

That's why your statement that unless he is of x playing strength than everything he writes must be false is logically flawed. It goes deeper than that though, because Elo is not what's needed to write about computer chess. It's a specialized area!

>Lyudmil Tsvetkov claims he's has "at least 2800" rating

Everyone knows that's absurd though, it's kind of tongue in cheek, and was said on a chess forum and not anywhere in the amazon description to any of his chess works, not in the author's bio either. If on amazon he had written "I am an international grandmaster rated 2800 FIDE", then sure, I'd call false advertising, but he has not done that at all.

>Oh, "3 times stronger"?? 2200*3 = you have a FIDE rating of 6600!

Actually no, x times stronger doesn't work that way with Elo. For example 2300 to 2500 increases the small probability of winning against a 2800 by about 3x.

>His books only have value if Lyudmil Tsvetkov is really a respected computer chess writer,

He's a new author, so how can he be a 'respected computer chess writer'? Is there even any such person in such a niche market?

>who knows what's talking about.

Sure, but your methods of determining that are what appear to be flawed.

>This is absolutely unethical and illegal.

If you want to be this melodramatic and hyperbolic, I could just as well say that slander is unethical and illegal.
Parent - By MachineLearning (*) [au] Date 2017-11-01 03:33
Lyudmil Tsvetkov is charging a commercial rate for his books. We, as the consumers, have the expectation on Lyudmil Tsvetkov because he's not providing free services.
Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-01 05:54
Dear MarshallArts,
thank you so much!

How I wish I had similar review of Human versus Machine on Amazon. :)

Thank you very much again.
Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-01 05:45
Dear Mr. Labyrinth,
thank you so much for your stance!
I really appreciate that in a not very welcome surrounding.
Please, write a mail to ltsvet(at)yahoo(dot)com, and I will send you
one of my 3 available books for free as a pdf file, whichever you choose.
Parent - By billyraybar (***) [us] Date 2017-11-09 14:36
Seems pretty negative to me, and I wasn't impressed with Bonham's arguments, a lot of which seem to be targeted at Tsvetkov's playing strength which is mostly irrelevant. The strongest players are not necessarily the strongest against computers which can require a very specialized understanding, nor are they necessarily good communicators of chess principles meaning that they might not make ideal authors for a book.

Spot on
Parent - - By Josef (**) [at] Date 2017-11-05 09:36
Hi Lyudmil,

being interested in the man vs. machine topic for a long time, but unfortunately a weak chess player, I bought all of your books and find it interesting that I may discuss them with the author here.

Firstly I have to say that they are quite inexpensive, I bought all 6 of them on for under 23 €.

About your statement "Seems like the first book with extensive coverage of a large number of winning games against the top engines":
That's definitely not true. Already about 15 years ago I bought this book from Dr. Ernest F. Pecci:

It also was quite an eye opener at that time. Quote from a customer review: "I was having all sorts of problems beating Fritz (6 or 7) as well as Shredder 6, on their highest levels. After reading and studying Dr. Pecci's book - I now kick butt!!".  Unfortunately it didn't work for me for being only a weak player, but I found it enlightening though. Were you aware of this book, and if so, what do you think about it?

I just started reading through your books, but already have 2 questions:

1. In your human vs. machine books you write that the games are played "in the period 2013-2017 against different Stockfish versions and Komodo 10". However, the games are distributed unevenly among dates and versions.

In part I, all 15 games were played between Sept. and Dec. 2013, 14 of them against Stockfish 4 (released Aug. 2013) and the last one against Stockfish DD (released Nov 2013). In Part II, the first 8 games are still against Stockfish DD from Dec. 2013 and Jan. 2014.

Then there are 3 games against Stockfish 5 (released May 2014), 2 games against Stockfish 6 (released Jan. 2015) and 2 games against Komodo 10.1 (released March 2016).

Progress in engine development still happens at a very fast rate. You are certainly aware of this, since you write in the foreword to your Secrets: "On the other hand, chess computers do get improved on a regular scale, adding some 50 elo or so each and every year". So, of the 30 games presented, 23 are against engines 200 elo below current level and even your newest engine is currently about 75 elo behind. Concerning stockfish, I always use the latest development build, since even the about 1 year release schedule is quite long in terms of progress rate.

Do you plan to write another Part III with games against current engine versions?

2. In your Secrets I currently stick at this diagram:
8/6k1/6p1/1p1p1p2/p2PnP2/P3B1P1/1P4K1/8 w - - 0 1
You write: "Black will be able to win the game with far less effort than if it had a dark-square bishop instead of the knight". I fed the position into my current Stockfish 17110313 and let it calculate to depth 66 (which lasts only some minutes even on my 5 year old computer; since you don't tell which side to move I tested with both). But no win for Black is found. There are 3 possible explanations for this:

a) there actually is a forced win for Black, but it doesn't materialize any sooner than after 33 moves. If so, can you elaborate?
b) Stockfish overlooks the winning move sequence even within it's proclaimed calculation depth (could be possible since this engine this engine employs an extremely harsh method of search tree pruning called razoring). Do you support this theory?
c) within the foreword of your secrets you write "The existence of such programs [...] allows us to look deeper into the game of chess. Gems of the past are now debunked, the widely acclaimed play of world champions like Capablanca, Alekhine, Kasparov and Carlsen appears to have lots of gaps, with a multitude of tactical mistakes throughout even their masterpieces". Is it possible that this position falls into this category?

Now, acclaimed, you write "Concerning the diagrams, please bear in mind, that those are not simple fens, so do not check them with engines, they do not even have side to move. Rather, they are supposed to serve as an illustration accompaniying specific features. Their purpose is just that: to illustrate the terms." But, within a book proposing a new approach to chess theory based on the inner workings of modern chess engines (and considering myself an engine freak) I cannot accept a restriction not to check book diagrams with my top engines. Quite contrary, I take it as a quality attribute if a chess author claims his books "engine proven".

Best Regards,
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-07 10:00
Dear Mr. Josef,

and I have been wondering for so long who my benefactor on was. :)
This is the first time someone buys all available editions...
Please, tell me your last name, so I am a bit more aware who my interlocutor is.

Oops, so many questions and such a rich reference material you have linked.
No, I was not aware of Mr. Pecci's work, thanks for giving the link, I read the whole preview
and it seems like a very interesting and well-thought book, I could not have a glimpse at the games themselves, though,
but I suppose they are interesting too.
Well, this is human-machine from more than 15 years ago, involved engines are Fritz 5 and 6, which are supposed to
be some 700 elo weaker than current tops(someone might want to correct me).
Mr. Pecci's was undoubtedly a great achievement, but there certainly is a difference.
For example, nowadays it is already not possible to win with openings like the Alekhine, Trompovsky, most French lines, etc.
The machines are already too strong for that.

It is interesting that Pecci points out the Bird(or similar structures) as the most successful opening against engines(and most of
his games are like that), I can only confirm that, many of my games involve that opening.
Similarly for the English structures. I fully agree with that.
Seemingly, however, he has too few KIDs in relation to my KIDs.

Another interesting thing is the very low number of black wins Pecci presents(obviously he specialised just with white),
but in my second book I have also just very few black wins.

Your query about dates and Stockfish versions.
Well, I simply played most of my engine games earlier on, so naturally I have more wins against earlier versions.
It is not that more modern engine versions started frightening me, or I was unable to score as convincingly as till then,
I simply played much fewer games later on, and probably the main reason for that has been that I got fully convinced that a new
Stockfish version, +50 elo to the current, although a bit stronger, adds nothing special in terms of human handling, and beating it
precisely the same way as the older version was a piece of cake. I just tried some games to be certain the same way of winning
was still possible, and then at some point you simply get bored by the repeated routine...
So, at some point I simply stopped playing such positions against engines and concentrated instead on handicap games, where you
can abundantly train your tactics.
I would very much of course like to see an engine version that I would not be able to beat with a similar approach, unfortunately,
+50/75 elo basically mean nothing, and I am able to crush current dev Stockfish with the very same ease as before in the very same
As said, I just don't play nowadays many games like this, and, of the games I play, not all have the necessary stuff to be published, some
have mutual mistakes, making the games ugly, or absoliutely nothing special about them, etc., so those are just a small pick of the
games I have.

Of course, if I would like to, I can take current Stockfish dev, play for 5 hours with it, and score at least 3 wins(from maybe 30 blitz games),
but what is the point of it, when I know it is easily doable and I have done it many times.
At some point, there will be a continuation, of course, but Stockfish is afraid to release their new version, so I should be waiting for
Stockfish 10 and Komodo 14. :)
Other thing is, I don't own all commercial programs, just Komodo 9/10, but the winning routine is repetitive.

I was also thinking of possibly writing an intermediate version with handicap games against Komodo 10 and Stockfish 8(1 or 2 pawns, knight for pawn, etc.),
until newer engine versions arrive, but I don't know if there will be any interest.

Concerning the diagram, well, black should be able to win, by transferring its king to a5(g5 break does not bring anything at this time), then push b5-b4
to break, penetrate with the king on c4, etc. At the same time, white should guard the weak g3 pawn all the time with the king or bishop, and that restricts white very much.
Only the king walk to a5 takes some 10 moves or so, so I would not be surprised if 60 plies are simply not sufficient for Stockfish to see a win, or
it simply prunes something too much.
But again, those are no fens, as you have rightly observed, those are just illustrations.
I wanted to have all available evaluation material in a single volume(The Secret of Chess is 300 pages A4 fromat, so around 500 standard format),
and simply had no space/option to expand, otherwise, of course, I have and can post material/diagrams/analysed positions(games), etc.
to fill at least 5 times as much space.

I guess, for the time being this is sufficient, sorry, if some me,bers might find this post too long, and thank you very much
for your interest.
If you have any other questions, please don't hesitate to ask.
Parent - - By MarshallArts (***) [us] Date 2017-11-09 02:45

> No, I was not aware of Mr. Pecci's work, thanks for giving the link, I read the whole preview
> and it seems like a very interesting and well-thought book, I could not have a glimpse at the games themselves, though,
> but I suppose they are interesting too.

Well, I still recall an old thread of yours from 2013 where an old friend replied with this:

Parent - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-09 09:24
Carl, is it you MarshallArts?

Otherwise, I can not think how someone could think of this thread?

It was a long thread, I have been focused on playing the engine,
and might simply have missed a post or 2.

Anyway, it is as if I see Mr. Pecci's name for the first time.

Otherwise, Fritz 6 on common hardware was still well under Deep Blue in terms of strength.
On the other hand, current tops on common hardware are much stronger than Deep Blue.

So, for my part, I should be right one way or another.

Mr. Pecci's seems to be a great book though.
Parent - - By Josef (**) [at] Date 2017-11-09 23:36
Hi Lyudmil,

thanks for your comprehensive answer.

What I find interesting about Dr. Pecci's book is that he gained a foreword from Kasparov. As far as I see he has no noteworthy chess rating, so I wonder how he managed that. Maybe you also should send Kasparov your Secrets book, I believe his opinion still counts.

Another historical issue is that in 2003, about 2 years after the release of this book, there were 2 huge MvsM-Events where Kasparov fist played Deep Junior and then, some months later, X3D Fritz (for one coverage of this event see We know that Kasparov was aware of Dr. Pecci's method, and if it had worked and anybody had the skills to perfectly adopt it, it would have been Kasparov for sure. However the match after 4 games ended in a hard fought draw, no sign he could easily outplay the computer (in game 3, the only game won by Kasparov, Fritz looks very weak though). Kasparov in one interview said he was confused by wearing VR glasses, but then he should have tested them in advance.

So, either Dr. Pecci's method did not work or Chessbase closed the holes quite quickly.

In general, the decade from the mid 90's to the mid 00's was a great period for MvsM (and especially World Champion vsM) starting with the Deep Blue matches and ending with Kramniks crushing defeat by Deep Fritz in 2006. It's hard to understand from todays point of view that no one predicted the computer chess is not related to AI until the world champion had been beaten by relatively simple (though highly optimized) search algorithms running on office PCs without any knowledge base or even the transfer of GM brain neurons into silicone. Kasparov once said (prior to his defeat by Deep Blue I think, but I have no date): "If a computer can beat the World Champion, a computer can read the best books in the world, can write the best plays, and can know everything about history and literature and people". Watson admittedly came quite close to the last claim, and already in 2011, but that has nothing to do with computer chess.

Concerning your quest for my last name, I won't disclose it here in the forum, but certainly can provide some infos about myself. I'm a software developer aged 46, hold a degree in Computer Science from an Austrian university, am married with 2 children aged 6 and 9 and live with my family in a small house in the suburbs of a major Austrian city (not Vienna, anyway). In my everyday work I develop software for Linux based realtime industrial controllers used in production lines of, for example, car factories. I mostly write low level C++, some C code for kernel drivers but also lots of Java for the user interfaces (you can imagine our software like chess engines in C++ and chess GUIs in Java). I have been playing around with chess engines for quite some time now (exactly since 86 when I received my first C-64 computer and started with programs like Grandmaster, which was an inappropriate name for a chess program at that time for being much too weak, and would still be inappropriate for one of today for being much too strong, see, I never managed to beat it, it was like magic at that time, as well es MyChess II, Sargon, Colossus and ChessMaster. Concerning open source chess codes there are basically 2 eras, one before and one after Fruit. In the before era there were Gnu Chess and Crafty, and it was accepted that these were decent although not too clean programs, but if you wanted one really strong, like Fritz, Shredder, Junior, Hiarcs or ChessTiger, you had to pay for it. Fruit, on the other hand, was open source, number one and clean, it was everything at this time. I never got into chess programming deep enough to achieve strength improvements for one code, but about 10 years ago I implemented a visualization tool for breaking down evaluation of a slightly improved derivate called Toga II and called it Toga LOG (it's still on the Internet if you search for it, though I lost write access long time ago after a provider change). Then I got children and didn't do much in computer chess, only recently I started again to read some forum threads and have some look on the Stockfish source.

Concerning your book position: Yes, you are right, one should try to understand a chess position and provide reasoning about it like you do your answer. Just claiming "engine x shows value y at depth z, therefore it's a draw" is bad habit, I apologize. On the other hand, every positional assessment has to be converted into tactics at some point in time, since by definition the only way to win a chess game is to deliver mate to ones opponent, and mate is tactical. So, just saying "it's no FEN and there are no lines to be examined" is unsatisfying.

If there were no chess engines, we just couldn't come to a conclusion about this position and discuss it forever or leave it. But since we both "own" at least latest Stockfish (which by incident happens to most likely be the strongest engine available today, it's easy). Firstly, I wonder for what reason you try to prove that you might beat these engines under "fair conditions" and why discussion arises what such conditions would be. I'm not so much interested in your tactical strength or even positional understanding as I want to know the truth about this position. If you can reasonably show me that Black wins this position, by employing whatsoever tools, I will credit you. So, just send me one analysis line how you win the position against your Stockfish version, without variations, but tell me your version number and possibly write down calculation depth for each computer move. Due to multithreading artefacts my Stockfish will not play the same moves, but there is multivariant mode where I can see if your computer moves receive about the same eval then the ones my Stockfish plays and so recognize if it's regular engine moves against which you win the position.

About your remark "Stockfish is afraid to release their new version": I guess that's you misunderstand that. They release a new development version every few days. On the homepage, click the button "Development builds" and you may download not only the latest but every version you like. Making one of these versions the new official one once a year or so is mostly a matter of form.

I recently played around a bit with current Stockfish and I have to say I'm really impressed how strong it has become. Now one might ask why might someone with an estimated 1500 rating need a 3400 engine? Not for playing, for sure. But the truth is that an engine for being able to explain a decent chess book written by a human GM or IM to an apprentice in the beginners rating, it needs to be very strong.

I took a look into one rather old but still hyped strategy book from IM Jeremy Silman:

I bought one issue of this book almost 20 years ago, started reading, didn't grasp it, tried it with engines, but they showed me something different than the author said. Then, some years ago I read more hype about the book and once again bought the reworked 4th edition as an ebook, but still didn't grasp it. Now I tried again with current Stockfish and gave it one position quite from the beginning of the book where the author explains the meaning of rooks on open files:
2qr1rk1/p1p2ppp/1p1p1b2/1P6/P1PP3P/3Q2P1/4RPBK/4R3 w - - 0 1
He writes that Black tries to put a rook on e8 and exchange rooks after which White's advantage evaporated if not White could play Bc6 with a won position. So this is a positional claim like the one we are discussion from your book, I can believe it or not. Stockfish evaluates about +1.5, plays Bc6 and for Black after about 40 s calculation
2qr1rk1/p1p2ppp/1pB2b2/1P1p4/P1PP3P/3Q2P1/4RP1K/4R3 w - - 0 2

Here Stockfish Bxd7 and surrender the e8 square. But I would play cxd5, take the pawn and keep the e8 square. I can hardly believe it, but it seems that cxd8 leads to a drawn position despite of a big material advantage for White, since some moves later this position occurs:
8/2p1bk2/1p1r2p1/1P1P1p1p/2KP3P/6P1/5P2/4Q3 w - - 46 60
White is up 2 pawns but cannot make any progress.

Anyway, after it's mistake with cxd5 and Blacks g6, White might try to trick Black by offering the exchange with Re5:
2qr1rk1/p1p2p1p/1pB2bp1/1P1PR3/P2P3P/3Q2P1/5P1K/4R3 b - - 1 3
If Black take the exchange with Bxe5 here, he is lost.

So, in about 15 min analysis Stockfish learned me more about this position than reading books for several hours.

One central doctrine of Silman is "Talk to the board, and it will talk to you" (it's easy to remember because it sounds like spiritualism).

When I googled for reviews of Silmans book, I found a guy on Youtube who call himself "The Backyard Professor". He is a huge fan of Silman and especially of this book and did a lot of videos about the book. This guy does a great job of explaining in a simple but absolutely expressive way what Silman actually wants to say in his book. Here is one video out of this series. Please tell me if you enjoy it as much as I do:

Best regards,
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-10 13:21
Dear Mr. Josef,

many thanks for the detailed information you provided(certainly better than just
giving me your name :) )

On your points:

1. Yeah, I sent once Kasparov team a copy of The Secret of Chess, no reply still,
I can think of 2 reasons:
- Kasparov is still reading it
- the pdf is already in the recycle bin, much more probable

2. Concerning engine-human matches in the early years of this century, well, Kasparov drew Fritz in 2003, but that has already been
Fritz 8 or something, so 100 elo stronger than Fritz 6; beside that, Kasparov was playing on a hardware station, maybe 8 at least or 16 cores,
while Pecci most probbaly has been using single core, I guess quads became widespread after the year 2000.
This makes another 200 elos in doubling speed, so 300 elos stronger was the machine Kasparov has been facing next to
Pecci's Fritz. That certainly is some strength differential.
The other obvious thing is that Pecci has played 5 000 engine games, to win 100, you see the distinction. :)
He has been a good player, though, despite his rating, if he was able to win games from Fritz.

3. About my constructed position, I will have to apologise, but I am fully unable to analyse now, for the simple reason that,
if I do that now, I will have to:
- abandon my writing activity
- abandon my publishing activity, have to do some corrections, etc.
- abandon other related stuff

This is very important to me, so I simply can not do that.
I am almost certain black is winning, but I wonder why you stick with that position?
If you want to prove I make mistakes, of course, I make mistakes sometimes, sometimes big ones.
Why don't you choose any other position from my book? Have you really checked them all with Stockfish?

As said, and specifically pointed out, those are not fens, if the text says those are not fens, then they simply are not.
The purpose of these diagrams is to show the respective term as clearly as possible.
If you want to analyse the position, you can do so, of course, you might invite other members, etc., engines,
to join you, at some point I could do that too, but really not now, simply out of time.

I have analysed collectively so many much more complex positions during this summer, you might simply check the
talkchess forum, analysing a position to a definitive conclusion is a HUGE task, especially if it is a more complex/even one.
Sometime, that might take full days.

4. About Silman's position, well, pretty much the same, you have been running Stockfish for 15 minutes on it and, as rightly observe,
learned more than from any textbook, but I have been doing so with much more complex and fully uninvestigated positions for
years on end and with extreme dedication.
Stockfish says 150 cps and picks Bc6, my opinion is the position is won for white, but not in a an easy way, this might take
30, 40 or even more than 60 moves. Imagine fully analysing this for 60 moves...
Silman's was just general observations, and he was right the fight for the open e file is a decisive factor.
I can not say for sure which move is the best(Bc6 is one choice), simply because I have not analysed it, but,
at first glance, there are other possible best move candidates:
- Bh3(always strong, gaining a tempo by kicking the queen)
- Qe3(another way of controlling the e file and not allowing Re8), etc.
The position is simply too complicated to be judged hastily.
Concerning if cd5 or Bd5 is better, again, this depends a lot on deeper analysis and tactical possibilities,
intuitively, I would pick Stockfish's Bd5 capture, though(those doubled isolated pawns on the d file are fully depreciated otherwise).
There is an alternative way of handling the position, though, by pushing c4-c5, and after bc5 dc5, black has a passer on d4,
but white's trio on the queen side is even more dangerous.

As said, this is an extremely complicated position, so you can never exhaust it in 15 minutes.
What I observe quickly is that, apart from white controlling the open e file, other weighty positional elements are involved, namely:
- opposite colour bishops, increasing drawing chances
- sufficiently active black bishop on f6, already attacking the d4 pawn, etc.

So that, although white has considerable advantage due to the e open file and general passivity of the black pieces,
black has its assets and drawing chances too, which makes the position highly complicated to play and close to both
white win and draw.
In order to be certain what it is, you will need 80 moves analysis ahead with perfect play.

5. About Silman himself and the video you linked, I will beg to disagree here, as Silman is not one of my favourites.
His books are selling, and one of the reasons for this is there are more weaker players around than there are stronger
What should I like about Silman?
- he lists pretty much the same positional factors everyone else lists: open files, isolated pawns, opposition,
couple more and that is it; in distinction, I am listing 5 times more
- he is using flashy chess diagrams with arrows pointing here and there, so what, I am not a moron, am I?

Obviously, there are people who like that and claim to be learning easier with arrows and a lot of colour available,
but those are usually the worse chess players around, you might want to do a check.
Chess requires an enormous extent of abstraction, so I don't see how arrows in the diagram or intense colours could help.

Concerning the video, well, to be honest, I don't particularly like it. Why?
2 reasons:
- the white pieces are featured in red, not the best colour for me to concentrate
- he is explaining the position amidst a mountainoius terrain, with some rocks and grass vegetation
obscuring the view; again, not the perfect conditions for concentration.

Otherwise, his analysis might be good, I don't know, as I did not follow the whole film.

Many thanks again for your interest, I am fully ready for further discussion, but not
analysing whole games at this point in time... :)
Parent - - By Lyudmil Tsvetkov (**) [bg] Date 2017-11-19 10:49
I simply don't know what to say.
Obviously, I have angered beyond repair some person with my posts about
the books I have published, so that he has resorted to the meanest action possible, posting a 1-star review on Amazon, not once, but twice, for both part of 'Human versus Machine', without even having read the book:

It is the very same person, Serverless, who probably has accounts on all main computer chess forums(and I bet it is Machine Learning here).
He has already posted 3 1-star reviews for my different books, completely untrue and without any substance.
That guy should know that I have reported abuse to Amazon.

Why are people so mean?

I have put an awful lot of effort into writing these books and get almost nothing in return, and still people resort to unimaginable dirty tricks to
deprive me of even the slightest consolation to see my books if not assessed, than at least not denigrated.

What a pity.
Up Topic The Rybka Lounge / Computer Chess / Human versus Machine
1 2 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill