http://www.amazon.com/dp/4871878198
http://search.barnesandnoble.com/books/product.aspx?ISBN=4871878198
I am planning to publish a revised version of The Soviet School of
Chess. Actually, I have been planning this for the last two years
since June 2009 when I assigned an ISBN number to it before the book
was ready.
It is still not ready. The reason is I plan to provide every game
referenced in this book in Algebraic Notation in PGN Format in the
appendix in the book. So I have been going through the book and then
searching for the same games in the chess databases.
This is a big job because Kotov often includes only part of the game
such as the opening, middle game or the end game. He does not include
the entire game, only the part that he finds interesting or
noteworthy. However, I am trying to find the entire game.
Also, the book contains a lot of mistakes, such as the wrong year or
the wrong opponent. For example, on page 34 he cites Tarrasch-Chigorin
1906. However, there is no such game. The correct game is Marshall-
Chigorin 1906.
After going through the entire book, I have found 172 games in the
standard chess databases, but there are 14 games that I have not
found. So, I am calling for volunteers. If anybody here can find any
of these 14 games please post the complete game here in PGN notation.
Here are the 14 games I am searching for. Here is the name of the
player of white and black, the year that the game was played and the
page in the book where the game is referenced.
White Black Year Page Number
Kogan Tolush 1937 44
Zubarev Riumin 1931 50
Em. Lasker Nenarokov 1924 63
Konstantinopolsky Lilienthal 1936 85
Bondarevsky Ufimtsev 1936 169
Bondarevsky Anderson 1954 176
Ragozin Bondarevsky 1946 247
Rovner Tolush 1946 269
Dubinin Novotelnov 1948 282
Bykova Kogan 1954 359
Ignatieva Baine 1952 372
Zhelyandinov Voitsik 1953 374
Volpert Malinova 1953 376
Zvorykina Heemskerk 1952 377
Sam Sloan
Not true, unless you just mean there is no such game on that page.
Chigorin did play Tarrasch in 1906, at Nuremberg.
> The correct game is Marshall-Chigorin 1906.
Chigorin played Marshall twice in 1906, at Ostend and Nuremberg.
Kotov gives part of both games on page 34. His mistake is in labeling
the Ostend game as being against Tarrasch at Nuremberg. The Marshall
and Tarrasch games are the same up through White's 6th move.
The bigger mistake on page 34 is that Kotov claims Chigorin was the
"founder" of the King's Indian Defense, based on his three games using
that line at Nuremberg 1906. This overlooks (or deliberately
disregards) the fact that Louis Paulsen was playing the KID long
before then. Just one example of how this book is an exercise in
propaganda, trying to portray the Russian/Soviet school as the sum of
all chess virtues and greatness.
Have you combed through the Soviet yearbooks? Shakhmaty v ____ g.
Louis, must you encourage him. And what's with those annoying 7s?
But Kotov gives the entire game! Why you don't just input the moves?!
(as I have done time ago...)
According to Divinsky's "Batsford Chess Encyclopedia" the Russian
edition came out in 1951, and the elder Yudovich was Kotov's co-
author.
But this partial score situation is not innocent on Kotov's behalf.
Soviet School is the most propagandist chess title ever published, and
many of the lines claimed as 'current theory' were far from that, in
fact deliberated deceits. While I applaud remaking the title in
algebraic, take a good look at what the [anonymous] Dover chess
editors wrote about the title in their preface. Phil Innes
>
> But this partial score situation is not innocent on Kotov's behalf.
> Soviet School is the most propagandist chess title ever published, and
> many of the lines claimed as 'current theory' were far from that, in
> fact deliberated deceits. While I applaud remaking the title in
> algebraic, take a good look at what the [anonymous] Dover chess
> editors wrote about the title in their preface. Phil Innes
>
I respectfully disagree.
The Dover version says: "Therefore, his works tended to be rather
critical of (and occasionally somewhat dismissive toward) American
players. Russian players, on the other hand, were presented and
described in a particularly favorable light."
However the anonymous author of the Dover version was simply wrong.
The American players were simply much weaker than the Soviet players.
With the exception of Reshevsky, none of the top American players
could compete against the Soviets at the time this book was written.
Sam Sloan
Sorry, my post didn't show up. But here is an analysis of opening
theory, where Adorjan is conducting a conversation on the real origins
of the Queen's Indian, correcting impressions by others, it is both
technical [not shown here] and historical. As you will see, and
similar to the claims we currently investigate in 'Soviet Chess'
origins are not as first published or now thought:—
Dear Peter [Boel],René
A few remarks concerning "Who is the Author?" First: the so - called
Botvinnik System was played first and some
more games (avaliable) by Klaus Junge, who died young in World War
II.
There is the 'Hubner Variation" in the Nimzo.But it was employed by
Lajos Portisch earlier in 7 games! (Hell knows where did I mispaced
the scoresheets) Moreover there is an excellent game Johner -
Nimzovitch (1926)
as the very first game of Nimzo (Hübner) Defence.
Even this is NOT true. To my knowledge already the move 3....Bb4 is a
brainchild of the great Russian writer Turgeniev. (game avaliable)
To be continued - AA -Greetings
==
Dear Friends,
You are receiving our Queen's Indian article (1 -2) titled ' Two
Forefathers ' who are H.Dobosz and me. The text
comes soon.It is remarkable that Carlsen played it against Aronian
(Botvinnik memorial) in 2011!
Regards: András, Endre
You are overlooking the many very strong players in the Soviet Union
that we do not know about because they never got out of the Soviet
Union to play against us.
Most are unknown in the West or are only known because they have
openings named after them.
Also, many died at early ages in the wars and purges in the Soviet
Union. Riumin, for example, was considered equal to Botvinnik.
Here are a few names:
Ilyin-Zhenevsky
S Belavenets
Rabinovich
V Kirilov
Verlinsky
Veresov
Riumin
Rauzer
Sam Sloan
I've done two descriptive-to-algebraic book conversions (e.g.
http://shop.chesscafe.com/Laskers_Manual_of_Chess.asp) and am working
on a third. There's no magic wand I know of. I use ChessBase. If a
game is in a database that's a big help, but you usually have to enter
the note variations manually. If a game is not in a database, you have
to enter the whole thing, moves and notes.
And of course the verbal text has to be hand-typed. Or in some cases
it can be scanned into a pdf or Word doc with character-recognition
software, which allows you to cut-and-paste, but the conversion is
often unreliable and creates a lot of typos, unlike ChessBase which
never makes notational errors.
> As for not being able to read it...it's nice to know it, and Andy
> Soltis held the line in Chess Rag for years, but we aren't learning
> Greek in elementary school anymore either. All things must pass.
I agree with Mike that it's just about as easy to read descriptive
as algebraic, and any serious chess player should know both. But in
terms of the market, most players just don't want to bother with
descriptive any more, so if you want to sell books, it must be
algebraic.
Well, it's KIND OF just as easy. Actually, it isn't...you have to keep
flipping the board in your head and it's hopelessly
illogical...algebraic gives every square a distinct reference
point...descriptive is cooler sounding, that's about it.
I waited for years for My 60 Memorable Games to come out in algebraic
(the new version, not the bowdlerized one), before finally reading it.
On the other hand, as you note, it's pretty labor intensive to convert
the games and the notes and this probably leads to repagination which
may be a problem with indexes, etc.
I think Sam's approach is a reasonable compromise, although I think
he'd add a lot of value if he'd either include a CD with the algebraic
game database, or at least, provide a link so someone purchasing the
book could download the games. A reader could 'tap' through the
games with the book in hand, and the on-screen presentation would
provide a return reference point when going through the annotations.
It also seems like it should not take long to learn as you say.
> student of mine was having trouble with descriptive -- we just played
> a few clock games where he HAD to record his moves in descriptive --
> no problem anymore.
Whilst I am inclined to agree it seems to me like converting important
games and annotation into modern notation is potentially worthwhile. It
is interesting to set modern engines loose on ancient expert analysis.
Once it is in machine readable form you can show it either way and
engine based analysis can spot things that expert humans will miss.
>
> One can argue the merits of algebraic versus descriptive, but the fact
> is that learning either takes minutes, and getting comfortable with
> either shouldn't take more than an hour or so.
Perhaps a couple of hours but I tend to agree. The main difference is
how you count the squares and K vs Q side disambiguation.
It is also amusing when the games are not in English abbreviations.
ISTR The transition to algebraic was somewhere in the mid 70's.
Regards,
Martin Brown
Indeed it is. In each of the four book conversions I've done so far,
I've added an appendix, detailing analytical errors. In some cases
long-held verdicts have been overturned.
I've done two descriptive-to-algebraic book conversions (e.g.
http://shop.chesscafe.com/Laskers_Manual_of_Chess.asp) and am working
on a third. There's no magic wand I know of.
--------------------------------------------------------------
>
> Indeed it is. In each of the four book conversions I've done so far,
> I've added an appendix, detailing analytical errors. In some cases
> long-held verdicts have been overturned.
Can't be too hard. In less than 24 hours you went from two finished
conversions to four. Way to go Sam
To be specific, Stan: two conversions I worked on have been
published ("Common Sense In Chess" and "Lasker's Manual of Chess").
Another is finished but not yet in print, and another is nearing
completion. I am not at liberty to divulge those titles yet.
Apparently Mischa won them all, and he won them convincingly, only
playing flank openings with speculative combinations that should have
lost. Rumor has it that he didn't even wear a robe when entering the
room, which may have contributed to a psychological edge against his
opponent, an unknown clubber who was rumored to be female.
These are VERY hard to find, since the game scores were pulled from
Shakhmatny at the last minute by a meddling politico from the Kremlin.
Good luck.
This had nothing to do with the Dover Edition. I did not even see it.
I used the original book published in the Soviet Union.
Sam Sloan
As far as I know, all attempts to write a program that will convert
Descriptive to Algebraic notation have been unsuccessful.
Grandmaster and programmer John Nunn tried to do it and failed.
Sam Sloan
>
> As far as I know, all attempts to write a program that will convert
> Descriptive to Algebraic notation have been unsuccessful.
>
> Grandmaster and programmer John Nunn tried to do it and failed.
>
> Sam Sloan
That's trivial. You must be mistaken, he must be an incompetent
programmer, or both. I could do it easily in probably a week's time.
Convert a page using OCR, then replace text such as 1. P-K4 with 1.
e4, using lookup tables. The only complication I see is if
descriptive is not "conventional" in that you can represent the same
move sequence different ways.
RL
OK. Go ahead and do it. I will give you a week !
I will even give you a break. Do it in ten days.
Programmers have worked on this for years and have not been able to
solve it.
For example, in P-K4, on what square did the pawn move start?
Sam Sloan
http://search.barnesandnoble.com/books/product.aspx?ISBN=4871878198
I just discovered that they put my book on Amazon with an interactive
feature where you can click on a link and it takes you to that page in
the book.
http://www.amazon.com/dp/4871878198#reader_4871878198
I have never seen this before.
However, they did not include a link to the last part of the book
where I converted all the games to Algebraic Notation
Sam Sloan
>On Oct 13, 7:41�am, raylopez99 <raylope...@gmail.com> wrote:
>> On Oct 13, 10:03�pm, samsloan <samhsl...@gmail.com> wrote:
>>
>>
>>
>> > As far as I know, all attempts to write a program that will convert
>> > Descriptive to Algebraic notation have been unsuccessful.
>>
>> > Grandmaster and programmer John Nunn tried to do it and failed.
>>
>> > Sam Sloan
>>
>> That's trivial. �You must be mistaken, he must be an incompetent
>> programmer, or both. �I could do it easily in probably a week's time.
>> Convert a page using OCR, then replace text such as 1. P-K4 with 1.
>> e4, using lookup tables. �The only complication I see is if
>> descriptive is not "conventional" in that you can represent the same
>> move sequence different ways.
You're wrong, Sam. I had a cheap shareware program in the late 1990s
that did it easily. As Ray says, the programming to do so is
relatively simple -- evidently, there's simply no demand for this
functionality.
If you ever find this program please let me know as it would be
useful.
The problems in writing such a program would be significant. For
example, if the move is P-K4, the program could be taught that K4
equals e4 and then search for a pawn that can move there.
But what if the move is R-Kt4. There are four possible squares that
could be Kt4. They are b4, b5, g4 and g5. Then the program would have
to find a rook that can move there.
Sam
>The problems in writing such a program would be significant. For
>example, if the move is P-K4, the program could be taught that K4
>equals e4 and then search for a pawn that can move there.
>
>But what if the move is R-Kt4. There are four possible squares that
>could be Kt4. They are b4, b5, g4 and g5. Then the program would have
>to find a rook that can move there.
Assuming that the program has been given a valid starting position
with sufficient context (castling status, etc.) and that the move, as
recorded, was not ambiguous, this would be a relatively trivial
programming problem.
But your question underscores a deeper problem with the older
descriptive literature, namely that moves often WERE ambiguous, but
human insight could disambiguate. On some of the small scale
tournament bulletins, the games wouldn't even play. In fact, I
remember an ad by Jack Spence, claiming that in his bulletins, the
games all played.
The shareware program I used enabled clicking on a board
representation to record one's own games, one at a time, and keeping
them in a file (technically, a database, but the program lacked most
of what we normally associate with db functions). I can't remember if
it even supported a paste function. I don't think it did, so the
software could always count on a clean game score. It ran on Windows
3.1, long before there was much question of it being linked to a
playing engine.
For larger conversions, you'd probably need a facility which enabled
human intervention somewhere in the process to correct faulty game
scores.
Mike, how did one enter moves into this conversion program?
You clicked on a board representation. You may have been able to
enter a move textually from a command line -- don't remember.
You could switch from descriptive to algebraic (short or long) at
will. The score was always visible in one segment of the panel.
I can't remember all that much about the program -- it was like $5
shareware -- as I mentioned to Sam, I can't remember if it even
supported a paste function.
Yes, good points made by a fellow programmer. Like you say, if you
start from a legal position then the program can "track" the moves as
they are played, so Sam's question about "which square was the piece
on before the current move?" is easily solved. For ambiguous moves of
course you need some sort of expert function to take a guess, or
simply to reject the game and let a human manually edit it.
RL
Link did not work for me. Book at $24.32 is overpriced for a
reprint. $10 would be more reasonable. You seem to like the $24
figure as I see this quoted for your other reprints. Like a tourist
trap, you depend on the occasional sucker buying your book, and you
keep some in inventory just for that reason. Nice strategy assuming
the mice, rats and silverfish don't eat up your stock.
More books by American authors below, at $21 and $15. Better value?
Hard to say. You can get (and I've gotten) an e-book from
PirateBay.org of Eade's book. It's OK as a beginner's text.
RL
Chess For Dummies James Eade
$20.96 BN.com Price
The Complete Idiot's Guide to Chess by Patrick Wolff The Complete
Idiot's Guide to Chess Patrick Wolff
In that case it sounds useless for converting old books, or more
precisely, no more useful than Fritz or ChessBase. You still have to
enter all the moves manually, unless of course you already have the
game score from a databse, in which case the old software is not
needed at all.
It seems to me the only way such software would be useful is if one
could scan an old book using character-recognition software to put the
games into Word doc or pdf format, and then cut-and-paste the game
score into the program. That would expedite conversion of games found
in a book but not on database.
> You may have been able to
> enter a move textually from a command line -- don't remember.
>
> You could switch from descriptive to algebraic (short or long) at
> will. The score was always visible in one segment of the panel.
That would be useful if you wanted to convert from algebraic to
descriptive then, but there's no demand for that.
Nonsense, there is no descriptive move R-Kt4. There is only R-KKt4 or
R-QKt4 and if it was blacks move then the algerbraic would be Rg5 and
Rb5.
I wasn't implying the old program itself was useful for larger batch
conversions., just that software to switch from descriptive to
algebraic and the reverse has been around a long time, and involves
no particular problems when the game scores are correct. Sam claimed
this itself to be problematic.
So most of the building blocks are all there, or not particularly
difficult to construct.
Scan+character recognition --> recognize the instances of game scores
--> convert to algebraic and replace
The jokers in the deck are incorrect or ambiguous game scores. As Ray
has noted, more problematic software (or human intervention) is
required to attempt correction (or to decide when correction is
hopeless).
Sure there is. Imagine, say, a position where there is a white rook
on b1, and it can move anywhere on that file. It's the only rook on
the board. In that case, Rb1-b4 would be R-Kt4 or R-N4 in descriptive.
Or if the rook was on a4, with the whole 4th rank open except for a
white pawn on g4. In that case Ra4-b4 would again be R-Kt4 or R-N4.
> There is only R-KKt4 or
> R-QKt4 and if it was blacks move then the algerbraic would be Rg5 and
> Rb5.
The convention in descriptive is always to record the move as
briefly as possible without incurring ambiguity. Thus in the
situations I described R-Kt4 would suffice. Or after 1.P-K4 P-K4 you
can write 2.B-B4 instead of B-QB4.
My book is 512 pages, about 30% larger than the original book, plus
the pages are bigger.
You are forgetting that I have added 200 games in figurine algebraic
notation with diagrams that the original book did not have.
You cannot compare my books with the low-grade re-prints that Dover
puts out.
I am competing on quality, not price. It would be nearly impossible to
produce a quality product for less than I am charging now.
I am surprised that you are comparing my book, "The Soviet School of
Chess" to beginners books like Chess for Dummies and An Idiot's Guide
to Chess. Do you think those Soviet players were beginners?
Sam Sloan
For a known colour or rook there are only two N4 squares and if there
was a rook able to reach both of them it would be normal to resolve the
ambiguity in descriptive notation. Mistakes were made though.
>
> Assuming that the program has been given a valid starting position
> with sufficient context (castling status, etc.) and that the move, as
> recorded, was not ambiguous, this would be a relatively trivial
> programming problem.
If it is dealing with a scanned game text (and a known alphabet of
characters used for the pieces and squares) then the problem is a lot
more constrained than general OCR text recognition.
The next move must always be legal from the current board position.
Plenty of heuristics exist to help an engine decide on the right path:
The king must always move out of check (or be mated).
Pieces are not often left en prise in GM level games etc.
>
> But your question underscores a deeper problem with the older
> descriptive literature, namely that moves often WERE ambiguous, but
> human insight could disambiguate. On some of the small scale
> tournament bulletins, the games wouldn't even play. In fact, I
> remember an ad by Jack Spence, claiming that in his bulletins, the
> games all played.
Simplest method is to generate all possible moves from the position and
if there is an unresolved ambiguity tag it with the original notation as
a comment for human intervention. Select only those with the right piece
and destination square and see if it is in fact unique.
Most books do get it right, and a computer analysis can easily track
forwards or backwards and spot that some later moves are impossible and
so infer where the error was made go back and retry.
> For larger conversions, you'd probably need a facility which enabled
> human intervention somewhere in the process to correct faulty game
> scores.
I think an engine could make a pretty good job of this provided that the
game score only contained a modest number of ambiguities.
I am sure you could construct a horror game that was unparsable with
ambiguities, but most games would be reconstructable by OCR and engine
guided analysis (even basic move generation would get most of them).
Regards,
Martin Brown
>If it is dealing with a scanned game text (and a known alphabet of
>characters used for the pieces and squares) then the problem is a lot
>more constrained than general OCR text recognition.
I believe Sam was concerned with games embedded in general text, so
the first problem identifying the beginning and end of the annotated
score.
>Simplest method is to generate all possible moves from the position and
>if there is an unresolved ambiguity tag it with the original notation as
>a comment for human intervention.
Yup. Seems as if we're in a similar ballpark to a style analyzer.
> I am surprised that you are comparing my book, "The Soviet School of
> Chess" to
It's not your book. You stole it.
Even for that the grammar for descriptive notation is actually more
regular than that of algebraic (which uses abbreviations for pawn moves)
so looking for consecutive move numbers followed by "x" and "-" with
plausible syntax around them would probably be good enough.
>> Simplest method is to generate all possible moves from the position and
>> if there is an unresolved ambiguity tag it with the original notation as
>> a comment for human intervention.
>
> Yup. Seems as if we're in a similar ballpark to a style analyzer.
That was how the engines did it when descriptive notation was common.
Regards,
Martin Brown
How can programs convert "Black should capture the Rook Pawn and then
the Queen Rook"?
They don't need to. The associated textual annotations around the game
should stay as annotations unless they explicitly spell out an
alternative line in valid syntax.
I can't recall ever seeing a game described entirely in free form
sentences without move numbers. Some of the explanatory text will need
clearly alteration since algebraic consistently counts ranks from Whites
bottom left corner whereas descriptive is from side to move.
OTOH we still talk of a having rook on the seventh rank (even when in
practice it is on the second rank when we are playing as black).
It should be sufficient for a first pass to translate into algebraic
anything that matches the general syntax of a classic descriptive move.
I have assumed at least one tab or space before a valid move for
simplicity but the following Regex should match English descriptive
moves (subject to typos and failure on special cases like promotions O-O
& O-O-O). Some other languages are a lot more cryptic.
[KQ\s][BNPR*]-[KQ*][BNR*][1-8]
[KQ\s][BNPR*]x[KQ*][BNR*]
And then permutations to distinguish ambiguous cases with qualifications
eg
/[1-8]
/[KQ*][BNR*]
/[KQ*][BNR*][1-8]
Regards,
Martin Brown