*** World Team Game Strategy From Move 51, Explained In Full ***
This is based on chess principles and on observations of what people
are saying on the MSNBC Kasparov World Team Strategy Bulletin Board. In 30
steps (yes, despite only 7 pieces this game is still that complex), any
player can understand the state of the game, and any player doing that can
make real contributions to the World Team as shown in point (30). A Key
with info on FAQ, FEN, EGTB, IMHO, sources and useful links, and chess
annotation symbols is at the bottom. White is "he" for Kasparov, and Black
is "we" for the World.
------------------------------
From current position, expected play is 48.Rxb1 Kxb1 49.Kxg6 d2 50.h8=Q d1=Q.
Strategic Principles and Ideas:
------------------------------
(1) Even though Black has an extra pawn, White has the winning chances
because his g-pawn is unopposed and closer to queening than either of
Black's pawns.
(2) If Queens are traded and Black can only get his b-pawn or d-pawn
to the 7th after White queens, Black will lose. (I do not think any
exceptional positions where White has no check or covering or pinning move
can arise here.) Q vs. b,d,e, or g-pawn is a textbook win. Hence most
trades of Queens are bad for Black.
(3) It is possible for Black to have an a-pawn or c-pawn if a pawn
recaptures when Queens are traded. In isolation, an a-pawn or c-pawn on the
7th is a textbook draw if White's King is far enough away. However: (a) if
both pawns are present, the stalemating defense used in the textbooks will
not work, because the other pawn can still move, and (b) if White's King is
close enough, he may get a mating attack. A good illustration is: White
King on b6, Queen on e2; Black King on b2, Pawn on c2. Black to move draws
by 1...Kb1 2. Qd3 Ka1! (or even 1...Ka1) since 3. Qxc2 is stalemate, and
other moves that stop ...c1Q allow ...Kb1 again. But White to move begins
an attack by 1. Kb5. Now 1...Kb1? loses: 2. Kb4! c1Q 3. Kb3!, and Black can
only stave off mate on a2 or b2 by surrendering his Queen. But 1...Ka1!
still draws: 2. Kb4 c1Q 3. Kb3 Qb1+ 4. Ka3, and now besides 4...Qc1+, Black
has a saving move that would work even if White's Queen were on d2. If you
can't find it, try playing through this line with Ken Thompson's database of
perfect play in endgames up to 5 pieces, which is available on-line.
Because the URLs are too long to come out clickable on this BBS, first
"Copy" the FEN code for this position,
8/8/1K6/8/8/8/1kp1Q3/8+w
and Paste it *after* the following URL after you click on it:
http://chess.traveller.com/scripts/chess_kt_endings/
(Ignore the "White has been stalemated" message that comes up with
no board before you click in your browser window's address box to put an
insertion point at the end, paste the FEN, and hit Return.)
(4) Two Queens always win against one queen, *except* for some positions
where Black can give perpetual check. Almost the only such position is the
"Corner Triangle", in which White has King on h8 and Queens on g8 and h7,
and Black's Queen is checking from (say) f6. Whichever Queen interposes,
Black will check from d8 or h4, and the rotation among h8,f6,d4 will
continue indefinitely. The Corner Triangle draw also works from other
squares on the a1-h8 diagonal---all so long as Black's King (or Pawns) do
not interfere with Black's Queen and Black's King is not exposed to an
interposition with check. A line where the Triangle actually comes up is
51. Qh7 d5 52. Kf7+ Ka1 53. g6 d4 54. g7 Qf3+ 55. Kg8 Qd5+ 56. Kh8 Qd8+! 57.
g8Q Qf6+, and here note how Black's d-pawn shields his King from pinning
interpositions on g7.
(5) If Black gets a second queen right after White plays h8Q, the
position is very dangerous, since White to move may have an immediate mating
attack. But if not, then Black can expect to draw.
(6) Computers have compiled "tablebases" of perfect play for both
sides in almost all endgames with 5 or fewer pieces---Ken Thompson did a
full set, and it is publicly available at the link above. But most 6-piece
endgames seem beyond the ability of today's machines to solve, at least
within (say) a month, and Kasparov himself declared that the 7-piece
position after move 50 "cannot be proved a win for White or a draw for
Black". The lone 6-piece exception we know is that Dr. Eugene Nalimov has
compiled all positions with KQQ vs. KQQ to help with judgments in
(5)---though even then if Black's other pawn is still present it might
change things! (Computer assistance is outlawed in most chess tournaments
and matches but is allowed by the rules of this "correspondence" match.)
(7) As shown by the tablebases (and known in textbooks), Black would
draw from the position after 50. h8Q d1Q if he didn't have his pawns, as
shown by http://chess.traveller.com/scripts/chess_kt_endings/ cut-and-paste
7Q/8/6K1/6P1/8/8/8/1k1q4+w after-URL. Try it with Black's King on a2, then
a3, then a4, and finally on a5 as given by the FEN code
7Q/8/6K1/k5P1/8/8/8/3q4+w --- suddenly the ending turns into a loss, in 61
moves!!! Although some "tablebase loss lines" contain sequences where Black
keeps White's Pawn from moving for 50 moves, which could be claimed a draw
in over-the-board-play by the "Fifty-Move Rule", the general sense is that
the moral flip-side of the World's using computers is our having to abide by
their objective verdicts if-and-when that time comes.
(8) The main reason the defender is able to draw some inferior Queen
endings like those above is the ability to give perpetual check when White
moves his King out from in front of the pawn.
(9) In view of (7) and (8), Black's pawns are currently a liability,
and in two concrete ways:
(a) they restrict the mobility of Black's Queen to give checks (this
is being felt now), and
(b) they provide squares behind them on which White's King may go to
hide from checks, while White's Queen escorts the g-pawn to coronation (this
is our main worry later).
(10) Black's Pawns can be converted into assets, however, in two
ways:
(a) catching up in a "queening race" with White's g-pawn---it is
currently two steps ahead, but one step may be spent on getting White's King
off g6.
(b) shielding Black's King from checks that would form part of White's
winning strategy in the analogous positions without Black pawns.
(11) In view of (8) through (10), it is considered generally-good
strategy for Black to advance his Pawns at every reasonable opportunity, not
even caring if White can capture them! (Earlier in this game there have
been lines where Black's only saving move has been to move a piece into take
"Kamikaze"-style, and now we have lines with "Kamikaze Pawns".) This is but
one of many paradoxical aspects making this position hard to play! Doing so
opens lines behind them for Black's Queen to check White's King. The one
exception is that advancing the d-pawn does restrict the mobility of Black's
Queen if it stays on d1.
(12) If Black's pawns are captured and White's g-pawn reaches g7,
the position becomes critical but not hopeless. An indication of how
delicate the difference between draw and loss can be is emphasized by Brian
McCarthy (see the compendium that "steni" is maintaining from Denmark at
http://home.worldonline.dk/~ak749/kasp/latest.htm (scroll 2/3 down); or
compare http://chess.traveller.com/scripts/chess_kt_endings/ cut-and-paste
7K/6P1/7Q/8/8/2q5/8/k7+w after-URL, and
http://chess.traveller.com/scripts/chess_kt_endings/ cut-and-paste
7K/6P1/7Q/8/8/2q5/8/1k6+w after URL. The only difference between them is
whether Black's King is on a1 or b1! Also try interchanging the Queens in
the latter: http://chess.traveller.com/scripts/chess_kt_endings/
cut-and-paste 7K/6P1/7q/8/8/2Q5/8/1k6+w after URL---even though Black has a
zillion checks, it is a 39-move road to doom!
(13) Both (7) and (12) illustrate a general principle: the closer
Black's king is to the corner square a1, the better for us. The reason the
exact corner is special is that it is the only square to which White's Queen
cannot give check when interposing along the b-file or 2nd rank, his King
having run to the a-file or the 1st rank, respectively. The squares near the
corner are called the "drawing zone": a2, b2, and b1 are usually safe places
to be, but even a3, b3, c2, and c1 (forget c3!:-) carry high danger. It
deserves its own number:
(14) Principle: Keep Black's King near the "Safe Corner". The
significant exceptions found by analyzing lines of play so far are:
(a) a4 or adjacent squares may be safe if Black's b-pawn is nearby
to help shield Black's King (this was a key principle of why Kasparov
avoided "Endgame K", a variation that would have left a position like ours
after Move 50 but with Black's King on b3 rather than b1); and
(b) in a queening race, Black's King may be needed near the d-file
to assist the d-pawn or to escape a two-Queen checkmate trap in the a1
corner.
(15) Principle: Queens should be centralized. Unlike Rooks they gain in
mobility from being in the center, and Black's Queen in particular needs
every help it can get in finding squares to check from.
(16) Principle: Against a pawn on g7, the a1-h8 diagonal is key for
Black to control---unlike a defending Rook which would best be stationed
behind the passed pawn on the g-file. This and (15) are illustrated by the
positions in (12) and by many others that you can find by exploring at the
Chess Archives' Ken Thompson tablebase.
(17) Principle, a bit less definitive: The defending Queen is better
pinning the g-pawn from behind on a diagonal than from the side along a
rank, whether the pawn is on g5, g6, or g7. This is supported by the 5-piece
tablebase endings and analysis of early moves in out current 7-piece
position, e.g. when Black plays ...Qe6 to pin g6 against h6. "Pin from the
side, he'll have your hide; pin from behind, more chances you'll find" (I
made that up:-).
(18) Principle: Just because Black has a zillion checks now, doesn't
mean he can check forever. The textbook and tablebase endgames show many
"spacewalks" by White's King that seem to defy logic but eventually work!
Perpetual-check resources have to be checked extremely carefully---and
computers have already been shown to be often unreliable at judging them
because they work forward with a limited move horizon. Perfect ten-move
lookahead can bring a 3000 rating in the middlegame and squat in a Q-ending.
(19) White's generic winning strategy is to advance his g-pawn, move
his King away from the queening square g8 (supporting the pawn with his
Queen if necessary), and have his King dance away from checks by Black's
Queen. Basic tactics on the way to this goal include:
(a) inducing Black's Queen to less-active squares,
(b) inducing Black's King onto exposed squares---even a1 is exposed
in lines such as 51. Qh7 d5 52. Kf6+ Ka1 53. Qf5 Qd4+? 54. Qe5!+-,
(c) co-ordinating his Queen and King so that all Black checking squares
are covered,
(d) "building bridges" across which both the King and White's
interposing Queen step hand-in-hand to the left or down the board,
(e) gaining tempos by checking Black's King to cover squares Black
wanted to check on or to activate White's Queen, and
(f) running behind Black's Pawns with his King.
This last tactic is the special danger in this game (it was also a
large reason "Endgame G" proved to be winning for White, i.e. why 47...Nh8
48. g6 d2 49. g7 d1Q 50. Rxd1 Kxd1 51. gxh8Q b1Q+ would have lost for
Black).
(20) Black seems to have three basic defensive strategies, plus a
fourth:
---------------------------------------
(a) "Passive Strategy": stay as close as possible to positions and
moves that are known to draw in the analogous "tablebase" positions without
Black's pawns. Keep Black's Queen active and King in the safe corner.
(b) "Active Strategy": try to catch up in the "queening race" by
advancing one of Black's Pawns whenever possible. Opportunities for
catching up arise when Black can drive White's King in front of his pawn
with checks, when White needs a move to re-position his Queen and cannot do
it with check or without allowing an interposing pawn advance, and when
Black's queen is already guarding the next square for White's pawn and White
isn't. A surprise way to gain a crucial tempo in the race is by offering a
Queen trade when a Black pawn can recapture, as happens in the line 51. Qh7
d5 52. Kf6+ Ka2! 53. Qf5 Qd4+ 54. Qe5 Qe4!---compare to the line in (19).
(c) "Delaying Strategy": play to hold up White from advancing his
pawn further than g6, or maybe even holding it on g5 for awhile. This
strategy is prominent in lines with 51. Qh5 Qc2+. It often goes
hand-in-hand with the Active Strategy, as a device for making White spend
tempos trying to break a "holding box"---and of course, it and the "Passive
Strategy" become the same if White plays g7 and Black has not evened the
queening race.
(d) "Strategy K": Run Black's King to a4 together with playing the
b-pawn to b5 or b4. This strategy was found to be effective in "Endgame K",
seeming even to allow Black to hold White's Pawn on g6 indefinitely. An
example of it is a line analyzed on the BBS today (9/23): 51. Qh7 b5 52.
Kf6+ Ka2 53. Qa7+ Kb3 54. Qf2 Qa1+, and later in this line Black was playing
...Ka4 in response to a check.
---------------------------------------
(21) The "Conventional Wisdom"---not agreed by everyone but the general
tenor of this BBS as it seems to me---is:
(a) The "Passive Strategy" by itself will lose---ultimately because
of White's King having more opportunities to hide. However, it is the
ultimate drawing fallback, covered by wings of EGTB angels.
(b) The "Active Strategy" is promising, but can get Black into trouble
if we neglect activating Black's Queen while pushing one or both pawns. It
is currently favored by most analysts, because (i) it has good hopes of
succeeding---see (22) next, (ii) many lines reach positions where Black has
caught up in the queening race and no longer stands worse, and (iii) it is a
concrete plan that is easy to recognize and often limits White's options in
attempting to stay ahead. Also (iv) it involves more-"forcing" lines that
are usually shorter and easier to analyze than many other ones.
(c) The success of "Active" cannot be proven, however---we have
Kasparov's own word quoted above on this! (Of the 100 hours Kasparov
declared that he spent on this game in August, many hours must have gone
into his choice of 38. h6, and once he proved that he could win endgame "G",
he could basically spend the rest of that time on this endgame---and this
was in August!) On this basis, it seems that a successful Black defense
will have to combine "Active" and "Delaying".
(d) "Delaying", however, seems to be much harder to analyze. At many
turns, in many slightly-different possible positions, we have to evaluate
whether White's King can escape the perimeter of perpetual-check tries set
by Black's Queen. The lines are "amorphous"; it can be hard to tell if
White is making real progress or not. IMHO, the success of "Delaying" will
require extremely careful minute analysis of a number of "Critical
Positions", to verify that White cannot indeed make a winning jailbreak with
his King---and this level of analysis has only barely been started...!...at
least in the public forum. My "Move Tree" is intended specifically to
enable /comparative/ rather than /separate/ analyses of these positions, and
so unlike the FAQs is/will-be organized with long lines from these positions
footnoted at the bottom rather than embedded in the main text.
(e) "Strategy K" seems too risky to /initiate/, because it costs time to
run Black's King across the "no-man's land" of risky squares a3/b3/c3. But
it should be kept in mind, because there are some cases where White tries to
inch his Queen in via checking on h1 then g2, or from f6 and f7, and Black
may find it tactically as well as strategically justified to run up the
board in response. One "surprise" tactic supporting all these strategies is
that Black may be perfectly happy to expose his pawns to capture, even with
check ("Kamikaze Pawns!")---an example I missed originally found by Otto ter
Haar (and others?) is 51. Qh5 Qc2+ 52. Kh6 Qc1!? 53. Qg6+ Ka2 54. Qf7+ d5!
(22) How and Why the Active Strategy Can Succeed: Black stands to
gain back one of the two tempi we are behind in the queening race almost
right away, because White needs time both to unblock his King and activate
his Queen. Indeed, this is achieved by 51...d5 (or ...b5) in response to
any move except 51. Qh5, which is why on grounds of principle the Russian GM
School has considered 51. Qh5 the most dangerous move from the very start of
looking at this endgame. (Note: Current feeling on 51. Qh7, regarded as the
other major try, is that 51...Ka1 and 51...b5 are less nerve-wracking
responses than ...d5. And 51. Kf7? is simply met by ...Qd5+, forcing the
King back to g6, and Black Does Not Stand Worse.) That leaves one other
tempo to gain.
(23) General Reasons to Prefer Starting "Active" with ...d5:
(a) ...d5 offers better central control than ...b5;
(b) it immediately clears the diagonal to a possible White hiding
place on b8;
(c) it sets the pawn on a path obstructed only by Black's queen,
which hopes to move out with check, and not by Black's King, which may be
stuck on b1.
(d) more?
(24) General Reasons to Prefer Starting "Active" with ...b5:
(a) ...b5 is less restrictive on Black's Queen---in particular,
(b) a d-pawn on d4 shortens Black's ways of occupying the a1-h8
diagonal, with reference to principle (17);
(c) it supports Strategy "K".
(d) more?
Conventional Wisdom is that (23) outweighs (24), including the
not-so-much-early-but-later consideration that even gaining 2 tempi in the
race with the b-pawn may not be enough because White may win one tempo back
by driving Black's King to b1. (If Queens are traded and Black has K on b1
and even P on b2 when White plays g8=Q, Black loses.) Hence most analysts
including myself have favored it. However, the option 51. Qh7 d5 is getting
some rough sledding in the early going, leading to interest in 51. Qh7 Ka1
and 51. Qh7 b5.
(25) How to Gain the Second Tempo: Black might gain it by a tricky
checking move that also covers g7 or g8, but White also has chances to hold
up Black's racing pawn this way. There is one organic feature of this
ending that gives special hope to Black. White's King can try to
run-and-hide in basically two directions: down the f/g/h files or across
ranks 6/7/8 to the Queenside. The former ought to hold no special danger
for Black, because his pawns are not on that side, and because the
corresponding lines in EGTBs are known to be draws. Indeed, my sampling of
EGTB wins says that most winning lines seem to go across anyway---the
symmetry around the a1-h8 diagonal is not perfect because g8 not h7 is the
focus, and a larger factor is that when a1 is unavailable, Black's King gets
to choose between b1 (less danger down) and a2 (less danger across). Now
the "Delaying Strategy" often seems to force White to play his King to the
h-file---where it is away from Black and often covered by White's Queen
too---in order to evade checks and advance the g-pawn. Hence Black may end
up pinning it along the a1-h8 diagonal with a pawn on b4 or d4, or pinning
it along b1-h7 with a pawn on b5 or d5. Now in order to get to the
Queenside, White may need to cross in front of his pawn once more on the
g-file, thus blocking it for one move again. This would be the equalizing
tempo gain!
(26) Dangers That May Make (25) Fail---three that I see now:
(a) White may be able to make an "End Run" by emerging with Kh7(-h6),
getting to g6 or g5 behind his pawn after a vertical or diagonal check, and
then squitchee his King to f7 or f6 and thence to the queenside. Black's
most effective counter would be to set up "opposition checking" two squares
away to the left, e.g. e6/e5 vs. g5/g6, or d7/d6 vs. f7/f6. But White's
Queen may be patrolling some of those squares---and Black's d-pawn may be in
the way of the latter option.
(b) White may be able to counter with a tempo-gaining move of his
own that holds up Black's pawn(s) from advancing, or forces Black's King in
front of one (the b-pawn in particular).
(c) Since Black may need to have his King on a1 or b1, and our pawns
would be no further than b4 or d4 at that stage, White may be able to check
and capture the advanced one, in a position where the other one is still
enough of an obstacle to change a tablebase draw into a present loss!
(27) Working Hypothesis: Black needs to follow some kind of non-passive
strategy to draw---"Delaying" or "K" if not "Active" alone. This follows
from the first "Conventional Wisdom" item under (21). A consequence is that
just because a move gives check doesn't mean it's good or even
harmless---the old adage "Patzer sees a check, patzer gives a check" is in
force here. Especially when White's King is in front of his pawn, Black may
need the move to undertake something more constructive.
(28) Working Hypothesis: White's most dangerous moves to examine at
any stage are those that directly combat Black in the "Active" and/or
"Delaying" strategies. This hypothesis is at work right now: 51. Qh5
disturbs "Active" immediately by forcing Black to move our Queen, but we
compensate by getting a little jump on "Delaying"---and White's Queen is not
so great on h5. And while 51. Qh7 allows Black to gain the first tempo
right away with 51...d5 or 51...b5, White can reply by playing his King to
the "danger side" of the g-pawn by both Kf6+ and Kf7+, with a discovered
check and initiative to also get a free move with his Queen that makes it
hard for Black to check and drive White's King back. Thus 51. Qh7 aims to
keep Black permanently from gaining the /second/ tempo, and the discovered
check + Queen-move combination is so potent that we are seriously
considering ducking it with 51...Ka1 (corner move!), hoping this leaves
White's Queen on h7 not too lovely either. Almost every analyst considers
these two moves to be the most dangerous tries...BUT, we also have to keep
an eye out for unusual "creeping moves" that seem not to advance immediately
but set up opportunities later. An example after 51. Qh5 Qd4 is 52. Qf3!?,
which tempts Black into 52...d5 53. Kf7 Qe4?!, when Black finds himself
without useful checks after 54. Qf1+ Ka2 55. g6.
(29) Tentative Conclusions: the result may turn on minute details of
particular positions, in (25) versus (26) and/or in "holding patterns" for
trying to delay White. In my field of computational mathematics, this is
called "Chaos"---"which way a butterfly flies in Brazil may affect which way
a hurricane goes in the Atlantic." Whether we pin from c2 or d3 at Move 53
could affect the tactics at Move 63 or 73. And Chaos is ultimate
complexity---this is as complicated as chess can get. If you played through
the analysis of "Endgame G" and thought that was tricky, what we're coming
to could involve a maze of twisty passages to Gs, all not quite alike. Or
maybe the World has an easy draw after all---but that would mean Kasparov's
declaration that we don't would be wrong---and he hasn't been wrong yet.
My own prognostication---without specific analysis, just a "feel"
for the positions and tactics, and IMHO---is that Black will not be able to
gain both tempos needed to equalize the queening race, BUT the coverage
White needs to thwart the race can be kept incompatible with the coverage
needed to escape checks permanently on the Queenside, so Black will draw.
That is, Black's salvation lies in coordinating *all three* basic
strategies, or making "K" work if it arises.
Objectively, however, whether Black draws or loses may ultimately
come down more to geometry than to strategy! THIS IS NOT NORMAL CHESS. This
is War With The Underlying Forces of the Universe! (at least the 8x8
universe:-).
***(30)***If you understand the above points of strategy and can see
them in the context of the quoted lines of play, then you are a good enough
player and analyst to make contributions of value to the World Team. Here
are some concrete things you can do, in no particular order---and no one
person (hopefully not even Kasparov!) can do them all:
(a) Many lines of analysis lead to---and stop in---positions where
White may-or-may-not try an "end run" with the King. Long lines given from
such positions may be only one option of many. Play through the given
line(s), then go back to the position and insert a Queen check for White,
and try following the line again. Does it still work OK for Black? Is there
a difference? Can Black compensate some other way? (Computers will be good
at checking 10+-move wins from these positions.)
(b) Check and query everything. Often a whole bunch of the strongest
players will be under delusion---I certainly have been! (Example, with
reference to the stronger players: how many of us realized before last
Tuesday that the central line of endgame "G" with 55. Qg7! Qc6+ 56. Kd8 Qd6+
57. Qd7 Qb8+ 58. Ke7 Qe5+, along which Peter Karrer found 59. Kf7! and
perceived the danger of the Qd7+ & Kxd5 battery against Black's King, is
/completely unnecessary/--- White can play 57. Kc8!, forcibly transposing
into what /was/ known to be the unique winning line against 56...Qb6+. This
also shows how discovery may come along lines of accident!)
(c) Look for cases where the BBS favors a position in one line that
actually looks worse than a position rejected in another line---having so
many separate lines makes it difficult to keep this perspective.
(d) Explore more of the EGTB Q+gP vs. Q positions---we may need more
knowledge of drawing cases to influence decisions made sooner.
(e) Check whether suggested moves seem to be carrying out a useful
strategy---note the first item under "(21) Conventional Wisdom." Don't be
afraid to suggest a move not given in a FAQ, especially if it seems to
contribute more to a strategy. (I believe that strategy is still the
guiding factor in moves 51-60, and these are the most crucial moves, but
once we get to bewaring of "end runs", tactics and geometry will
predominate.) Keep an open mind.
(f) Actively try to find winning ideas and lines for White. Maybe you're
even rooting for Kasparov!---one man overcoming the whole world and a race
of silicon beasts. That's fine...just please report your findings
/here/:-)!
(g) In the "Move Tree" and elsewhere you will find mention of special
queries whose answers we may need to know. For example, no one to my
knowledge has yet checked whether Black's drawing resources in the EGTB draw
lines given above still work when he has an "extra" Pawn on b7. Can Black
actually keep White's King away from b8 altogether? Another query: in those
EGTB draw positions where White's g-pawn is still on g5 or g6, can Black
actually always prevent the pawn from getting to g7 in the first place?
(h) Check the BBS to make sure your research has not already been done.
/Whether it is being duplicated is less of an issue/---often we may need
independent confirmation of results, and especially in early positions,
/multiple opinions/. Tigran Petrosian would not see a position the same way
as Mikhail Tal, and we may need them both! Finally, please understand that
every day with each move we will unavoidably be throwing away *all but one*
topmost branch of the FAQs and Move Tree! (Not quite---there are many
transpositions here.) Nothing different happens in over-the-board chess
when you think before each move! Hence many contributions will wind up "on
the cutting room floor." This doesn't mean they were worthless---indeed, the
value of analyzing a move is realized *before* the move can be played, not
when it is played, because the analysis is how you judge all your options.
On this scale this is called /preparation/. Andy Soltis said in this
month's Chess Life magazine that Kasparov has 3,920 new opening variations
he has prepared but never been able to play---and the only way we'll find
the things he has prepared in this ending is to trawl them ourselves. GOOD
LUCK WORLD TEAM!
--Ken Regan
Dr. Kenneth W. Regan, Associate Professor Computer Sci. and Engineering,
Univ. at Buffalo (Opinions not < SUNYaB) 226 Bell Hall, Box # 602000
Tel.: (716) 645-3180 x114 Buffalo, NY 14260-2000 USA
regan@cse.buffalo.edu
---- Key: () SmartChess FAQ or "the FAQ" = a source maintained by
MSN analyst Irina Krush and other analysts under the auspices of the World
Wide Chess Superstore, kept at http://www.smartchess.com (click on
"SmartChess Online", then on "Garry Kasparov versus the World", then on
"SMART-FAQ" and scroll down to bottom). It is not so much a FAQ as a
compendium of current game analysis with input from its maintainers, this
BBS, and other sources on the World Team. It is computer-generated in two
formats for reading by most common chess programs, but is not readable as
plain text. An HTML-viewable big tree of the FAQ's variations with a
chessboard, including some comments, is maintained by "99% Energy" at
http://www.comicastle.com/99percent/pgn2.html; this post has been reproduced
by "99% Energy" at http://www.gamersx.com/messages/view.asp?id=xbwfj
() GM School, GM-School FAQ. A shorter plain-text page of analysis
kept at http://www.gmchess.spb.ru/english/kasparov/siciNN.html, where NN is
a number, currently 92. The Russian version with "russian" in place of
"english" in the URL usually comes out a few hours earlier. It is
maintained by members of the St. Petersburg Grandmaster School.
() Move Tree = an updated compendium, part existing already, that is
like the SmartChess FAQ but has the complementary aim of emphasizing all
early options for comparative purposes rather than prioritizing lines to be
analyzed many moves deep.
() ***CRITICAL..., ***ESSENTIAL...: Posts with these titles by Peter
Marko, several times a day, collect recent analysis articles deemed
important, lists of important web sites including links I've used here, and
news and features about this match.
() FEN = "Forsyth Extended Notation", a concise code for positions
used in correspondence chess. White's pieces are uppercase KQRBNP, Black's
are lowercase kqrbnp, each rank starting from *Black's* side is listed
between slashes using numbers for runs of empty squares, and a final +w or
+b tells if White or Black is to move. For example, the position after 50.
h8Q b1Q in our game is 7Q/1p6/3p2K1/6P1/8/8/8/1k1q4+w.
() EGTB = "Exhaustively Generated TableBase"---see item (6)---used
to label positions mathematically proven to be wins or draws. These
positions all have 5 or fewerpieces and can be looked up on-line thanks to
the Huntsville (Alabama) Chess Club at
http://chess.traveller.com/scripts/chess_kt_endings/ --- the URL should be
completed with the "FEN" code of a position. Unfortunately these URLs are
so long that the MSN BBS posting software linebreaks them and makes them
unclickable, so you'll have to cut-and-paste the FEN part. The Club itself
is at http://chess.liveonthenet.com/chess/.
() IMHO marks my own new opinions; KWR marks ones I've given before;
others' opinions are quoted.
() Standard chess annotator's marks: "+" means check, "+-" means
White wins, [] means move is (or looks) forced, ! means move looks
especially good, !? means "worthy of attention" or "speculative", and "?!"
means "dubious".
The above article is copyright (c) Kenneth W. Regan, 9/23/99. Permission
is granted to reproduce it in whole or in part with due attribution on the
MSNBC World Team Strategy BBS and on other relevant Internet sites---and
likewise to include text with attribution in other posts or articles---for
private, non-commercial use. (The "Move Tree" to come within a day will NOT
bear any copyright.)