Amazing Chess Game: Exchange Sacrifice Power! : Garry Kasparov vs Deep Blue – 1997 – Game 1

Amazing Chess Game: Exchange Sacrifice Power! : Garry Kasparov vs Deep Blue – 1997 – Game 1


Morning all, let’s continue our look at Gary
Kasparov versus the deep blue computer of 1997. So this was the rematch. The first match was in 1996, which Kasparov
won 4-2. Okay so this rematch in which Kasparov nicknames
his computer deeper blue, quite musingly it was just a nickname. He kicked off the first game with the white
pieces with Nf3. Okay so Richard Rhett’s opening, but the
leading hyper modernist around them switches the time. The ones of the hyper moderns and he plays
it in very hyper modern style. He first been chateaus his King’s Bishop. Deep blue ventures Bg4 here and now the Queen’s
Bishop is also prepared to be fianchetto. So we’re going to get a crossfire across two
different diagonals, two different color complexes. So Nd7, Bb2, this isn’t the double fianchetto
had been used against Anatoly Karpov in a key last game which Kasparov had to beat Karpov
to retain his title, very very dramatic. So he has played this kind of double fianchetto
before. So Ngf6, castles c6 and it looks as though
granite is being set up against the g2 bishop. But in the process the dark squares for the
dark square bishop still effective. D3, Bd6. Now Nbd2 and it looks like later e4 could
be prepared and often it’s prepared with Qe1 to avoid a pin and then e4 will be given potential
tempo getting threat of e5 in some cases. After castles h3 taking the bishop and now
actually we see an interesting delayed kind of move. Just e3 here, like a way to move to see what
the computer will do here. So not committing even to Qe1. That kind of curious keeping a lot of options
open. H6 is played and after this h6, now Kasparov
feels it’s right to play Qe1 here. So it looks as though this central frost now
is really on the cards. H6 has induced a subtle weakness really making
this plan a little bit more effective. By going the slower route with h6 there you
could imagine if a five from black f5 would be weakened. A knight on f5, you imagine the knight on
f5. Because of h6, g6 is getting more difficult
to play. So this slight delay of playing Qe1 has reached
this kind of weakness of f5 potentially. Okay so that’s something to bear in mind here. So after Qa5, okay that’s a typical kind of
computer move. It seems fairly pointless into two with the
Queen. But it’s got some tactical threats potentially
like Ba3. So it has got some strategic undertone. If the computer can get rid of this dark squared
Bishop, then the crossfire is going to be reduced. White would like to keep this Bishop ideally. Because it’s pointing also against the king. So he plays actually a3, which stops any Ba3. Bc7 now is played and now again not e4 here. But actually another weakness provoker is
played. Nh4 believe it or not, it’s provoking g5 and
with g5 of course that’s an open invitation later to open lines and if g5 is played, you
know horrendously the diagonal is also going to be emphasized. From a human point of view g5 is an undesirable
move in over 90% of cases really in a game. But you always got to investigate like the
temp sent. Because sometimes g5 is a winning move. But the computer plays g5 here. So we see you know slight emphasis on the
diagonal. Weaknesses created on f6 and h6 and that sets
an undertone later for potentially powerful moves to celebrate this fact. But also you can imagine the potential for
isolated pawns is also increased. If black loses this pawn, then we have some
isolated pawns as well. So g5 from number of angles is quite committal. So what has the computer got in mind with
this? Well the knight retreats here and now e5. Okay and now finally white plays e4 and he’s
really secured that f5 Square potentially. But of course how can white get a Knight to
f5? G5 at least has ruled out Nh4. But it is a pretty crude way of ruling that
out with some consequences on this diagonal potentially. But there’s another route to get to f5, a
slow route again. You can imagine if the Queen moves somewhere,
the rook moves will have this Knight maneuver to f5. A slow Knight maneuver to exploit that f5
square. So these little seeds of destruction starting
with White’s e3 provoking h6 and now it’s g5 has created a very committed pawn structure. Probably above the horizon depth of calculation
for deep blue at the time. Modern engines maybe have better evaluation
functions and wouldn’t like commit to this, unless there’s a really tangible gain. Okay Rfe8 and we see now Nh2 and it looks
as though potentially f4 now is on the cards. Okay and this Queen hasn’t clearly moved. So the idea of Qe2 is pinning the knight. So this this looks as though Queen, the idea
of rewriting this is not on the cards at the moment, f5. So after Qb6, now we see a very interesting
move, Qc1. Just waiting, what is black doing here? What is the purpose of this move? Is it potentially going to pounce in some
way on this we can pull chain over here? Black plays now a5, now Re1 finally making
way potentially for a knight, either Knight to come and maybe reroute to f5. Bd6 and that’s actually the D Knight is chosen,
not the H Knight. That has some advantages. For
Example the queen is now seeing g5 and h6. So maybe even later that would be useful. Okay now the computer plays its first pawn
exchange. The first pawn exchange of the game is now
played. Otherwise it was a like a closed position,
which usually is a good recipe for beating our computers or it was to keep the positions
closed. Because in closed positions, the brute-force
approach, you know the if then move type analysis is less effectual. Because there’s no hand-to-hand battle, it’s
all behind the scenes. Does this really open up the position? Dxe4 though. It’s only one pawn exchange and it’s the symmetrical
pawn structure is of course kept intact here. Bc5, the first frets on a soft spot as well. Something to be alarmed about or not, Ne3. Rad8 and it looks as though if e5 wasn’t a
problem, then maybe black can also reroute a knight to say d4. That might be quite unpleasant potentially. Nhf1 and now another weakening kind of move
is played here from the computer, g4. Okay so what is going on here? Well Kasparov takes on g4. After the Nxg4, he has the option to play
f3, which he takes here. The knight on e3 is supported. So even though there’s a battery here which
looks kind of dangerous, it’s okay Nx3, Nxe3 and we know here we should know isolated pawns
and structural damage generally. Be7, it looks as though the computer is interested
in pinning the night against the Queen here. Which after Kh1 it does. Now in this position Re2 is played. Which might have the idea of things like Qg1
and then maybe putting the rook here and proceeding from there for example. Okay alternatively maybe just well you wouldn’t
want to use the d file here. So maybe Qe1, anyway in this position the
computer plays a4, softening up a bit maybe the Queen site. But after b4 it’s not that soft and now we
see a very interesting move. If the Knight can be unpinned as mentioned
that maybe for example Qe1, Qg1, there is a basic purpose of that
for Nf5. Because this pin is holding up Nf5. If Nf5 is played, then there’s more pressure
around blacks King and maybe this Bishop can be liberated further with later f4. Assuming it’s not losing too much material. But in this position the computer plays a
very dynamic move trying to exploit this pin here, f5. So clearly the knight can’t take. So Kasparov takes to the pawn and then we
see the piercing, seemingly piercing e4. So it looks as though the rook, the bishops are
all pointing at sensitive squares in White’s position. But what about this diagonal? And what about this pawn sack? Does it have consequence for black? Well here if what can white play without crumbling
basically. If he plays g4 it looks as though this is
going to start crumbling, tumbling down this position after he takes for example and maybe
Bishop takes looks kind of painful attacking the Queen and something like this and black
will be a piece up. So okay White’s got to be careful. Will engine check after? I don’t want to spoil the flow here by going
into detailed operations or checking with the engine here at this point. In the second part of the game, we will look
more frenzied than this. But after e4 there’s an interesting option. Which Kasparov chooses, which is actually
a very powerful option as well. So how does white solve all this pressure? All these problems of the pressure of these
bishops raking down these two diagonals. If I give you ten seconds or you may want
to pause the video, what would you play here as white starting from now? Okay, okay what are we seeing lot on this
channel? Positional sacrifices and exchange sack in
particular with f4 are now invited. F4, so there’s a couple of options you might
think definitely need to be investigated. Bxe2 of course and Bishop takes f4 maybe as
well to investigate. Let’s check out maybe Bf4 after in the second
pass through the game. Here Bxe2, so Kasparov takes on g5. So an exchange sacrifice and it looks like
a very logically justified one. Because Black has not got a dark squared Bishop,
this dark square Bishop is running riot down the whole diagonal and it’s got potential
coordination with Queen and Knight and also these pawns can act as attacking resources
as well as Kasparov notes many times that pawns around the opponent’s king could often
be like an extra piece. So Ne5 here and a pawn is pushed to g6, you
might wonder well hold on, hold on, why didn’t black take on g5? It is a good question and one we should really
investigate maybe in second pass through the game. There are a number of options here it seems. So okay so in the game Ne5, g6 holding on
and creating two connected pass pawns around the King. So there’s an unchallenged Bishop. Okay, but still Bf3, how’s Kasparov get
rid of this horrible pressure he’s under? Bc3 secures the second rank a little bit. Qb5, looks as though there’s an invasion or
Qe2 planned and against this Kasparov offers the exchange of Queens and of course often
with simplification pass pawns gain in strength. That’s a fundamental kind of human pattern
we deduce that’s often towards the endgame. King safety often becomes less rather than
the king needs be more active in endgames and pass pawns gain strength. So are these two connected passed pawns gaining
strength here? Qxf1, Rxf1. Now after h5 is played. Kasparov unpins here, his bishop with Kg1. Kf8, okay and it looks as though you know
potentially there’s an option of Be5 here. But maybe as well as Rc5, there’s Bg2 to consider. So Kasparov for the moment he plays actually
Ba3. He is keeping both bishops on the board. B5 and that brings his king to f2. Kg7 which goes straight into a pin. You might think that’s a remarkably strange
move to play Kg7, why would black want to set up a pin here? Okay well it looks as though in this position
these two pass pawns there might be an option at some point if there wasn’t Rd2++ at some
point of f6 being on the cards. Okay so let’s bear that in mind anyway. Kg7, going straight into the pin which Kasparov
exploits that pin by playing g4. Which he can afford to do and on brief engine
inspection earlier, this look, this next move looks like a howler. But it’s a bad position for black. Black is going quite downhill now. Because g5 will force through potentially,
will make this pawn a stronger at least for potential f6 at the right time possibly. So Kh6 is played. But Rg1 is again renewed as a kind of possibility
and now the deep blue takes on g4. So more simplification now. Bxg4, but again with this simplification the
power of the pass pawns is increasing, and this is actually simplifying now into a completely
hopeless position for the computer. After Rxg4, these pawns are really really
powerful. Black is going to have to give up material
soon. Okay so we see now Rd5 and f6 is okay, supported
by the bishop. In this position now on move 44 the computer
play Rd1 and I was reading some notes, there’s some notes on wiki about this match. Apparently, a psychologist has attributed
this to giving Kasparov some anxiety while the computer would play this move. So it has some higher intelligence behind
it. It turned out that this may have been a bug. Maybe the deep-blue software did have a few
bugs. It’s not like it’s got the exposure of the
market exposure a modern engine, which you know many thousands of chess enthusiasts have
got to get rid of bugs and you get new version releases. This was specialist, a specialist system. Maybe this was a bug. But it seems I hope this position anyway and
in fact after g7, deep blue, the team resigns here. It looks completely hopeless, because f7 looks
unstoppable. Now let’s try and do some clinical analysis
here with our engine friend Houdini 2.0. So here in this final position, it looks hopeless. I mean if check, we can play here. If the rook moves back, okay we just play
f7 and we’re queening a pawn here or winning a rook. That’s going to be okay just to win a rook. End of game. So it is pretty hopeless. There’s no way really of stopping these pawns. Okay let’s go with Rd5 as an example. Is there a way to go wrong? There is a way to go wrong with f7 here. Because of the check winning the pawn there. But instead of that just Ke2, so renewing
the threat of f7. Alright let’s imagine Rf5 then as a try. Check, now we’ve got here in this position
Rh8, pass pawn have got lost to expand as Fischer would say. So here how can this be saved? It can’t. If Kf7 we just take and then with Queen. So it’s very hopeless for the connected pass
pawns that triumphing here, triumphing in the variations. So Rd1 though gave Kasparov anxiety for the
next game. Which is video annotated on the channel. But let’s go back to this. The Roy Lopez in the next game caused a lot
of controversy basically in this match. But let’s check out this game again for some
critical moments. So the double fin château hyper moderns,
the hyper moderns would be proud that their ideology has being used to defeat the strongest
computer in 1997. So e3 deliberately slow path to plain at e4
reaps a huge benefit of provoking h6. But mind you maybe you know the computer is
kind of wary about the bishop not wanting to lose the bishop, you know giving an option
of going back to h7. Maybe it was factored in. You know if you have things like Nh4 chasing
the bishop after. So h6 did give also like a retreat square
for the bishop for g6. So h7, by Qe1 so we’ve got this idea of playing
for e4. But again it was played very very slowly and
cautiously. Only in fact when black can play e5. So when black deprived itself of the e5 square
for a piece basically and we definitely had a locked pawn situation. So this Nh4, in the style of the great Tigran Petrosian, just provoking
more weaknesses, setting the scene for a later powerful positional sacrifice theoretically
based in the weakness of the long diagonal and other weaknesses. So there’s a sound justification positioning
for the positional sacrifice coming up. So now that black can’t use e5 because it’s
occupied now by a pawn, can’t use it by a piece. Now the pawn chain is locked against each
other, jammed in. Nh2, okay we saw now Qc1. It’s a quite position. If you look at these variations, there’s nothing,
it shouldn’t be anything really radically going on here. Because it’s a very locked closed position. So dxe4, okay a pawn exchange, but it’s still
a symmetrical pawn structure. A soft spot issue isn’t much of an issue. It’s funny Houdini is giving after g4, now
it’s changing the valuation. Up to here was like liking black tiny bit
on this very brief death. But it seems g4 maybe was a step too far with
pawn moves. It seems white now technically has the advantage
according tiny tiny advantage, we wouldn’t really speak this as an advantage. Just more than equality basically now for
white after g4. But okay and now after Nxd4, in fact Bxg4
was preferred. Possibly the preference is to do with Ng4
not being, I don’t think Ng4 is an option. Because the pressure on f2, that’s pretty
nasty. That’s why I’m getting the pawn structure
sliced. But okay so in the game Nxg4 and Nxg4 is an
option here. Because there is a huge pressure on f2. So Ne3 now fends off f2 by the way. So, ok Kasparov though didn’t choose Ng4,
he played the strongest move it seems on this brief analysis here or playing f3. So not minding a Knight exchange and whites
advantage seems a bit more pronounced there, approaching almost, almost half pawn. Even though there’s a couple of pins set up,
which seemed a bit of a menace. This pin here and this and potentially well
this pin here is about to be set up now. After this move, it is going to about equal. I don’t know really but why. Because the Queen has to do something surely
to get this rook into the game. How those white get the pieces into the game? A4, b4, that seems to be a slight subtle I
don’t know something which isn’t great. But f5 is very very committal. If black didn’t play f5, alright let’s go
with Qa6. Let’s do something with the Queen, Qe1. Same like b6. Bc1, yes, the challenge that Bishop. Nf5 in here you would imagine visually that
this is very attractive achievement and yes it should reap some benefit technically one
would expect. Qc3, there was pressure on black. But again it’s not really that clear-cut. What made this game more clear-cut was this
f5 now for white having the advantage going in materialistically mainly to exploit the
pin here. Which looks very very dangerous. So f4 only move basically. G4 as I mentioned I think this just crumbles
white. Exf3 and its disaster. I think just Bxe3 is a big advantage or not. Qe1 is a resource okay. But actually there’s something stronger than
in this position. Just Bishop f7 and how does white defend e3,
e3 is dropping here. So let’s talk the action, okay. So exf3, Bxf3, yeah Bf7. So in the game Kasparov plays f4 positional
sacrifice. So it’s got the undertone position that the
diagonals weakened, there’s weak squares around the king and also these pawns are very useful
coming up. They gain in strength. Now I was wondering about Bxf4 earlier. So here okay let’s have a look. Gxf4, it does open the g file, Bishop takes. Now not Qg1, apparently that’s not the strongest
move here by far. Okay Qg1 is strong. Qd2, that I think the idea is to get the battery
going on the diagonal to exploit those weak and dark squares not minding the shock of
the Knight exposure to the rook. So say Nf6, well that gives up the bishop. So if the bishop moves, hang on what’s going
on here? I guess Qc3, Qc3 is strong or even Ng4 is
strong and then Qc3 we have totally wining, sky’s the limit. So that would be, that’d be very unpleasant
basically for blacks King safety, so black plays Bxe2 keeping it about level at the moment. After fg, we’re wondering, I was wondering
also about hg. So in this position okay a knight’s act just
to get Qg5 in either on c4 or d5 will do. So if Nd5 for example takes Qg5 is a mate
in three. So Kf7, Qg7 always putting a queen in the
way of course computer move. So that would be very crushing just with Qg5
Black having to give up the Queen to stave off mate. So that’s not such a hot idea. Okay so basically Ne5 allows Kasparov to keep
two connected pass pawns near the opponents King. Bf3, look at the evaluation. That’s not too bad it seems. Qb5 and now after Qf1, our Houdini here starts
to still favor white slightly. But I think the advantages get further pronounced
soon with the next set of exchanges. Well g4, here it seems h4 would be according
to Houdini more determined resistance in this position. If g5, well you might think that is there
a threat of f6, because of Kg6. Well Kf8, let’s see what’s going on or why
Kf8? Why not say with Rd7f6, the rook kind of can’t
do anything here. If Re7 f6 where do, we move the rooks? Let’s just do a token rook move. Okay Ng2 would be powerful here. Just going to take on h4 or something. Okay so maybe h4 doesn’t really save black. So whites’ advantage is getting absolutely
out of control now after Rg1 but also h5 is apparently strong. But Rg1, White’s advantage is still growing
hugely now after hg. So I don’t think it really matters after h4
again. If this Ng2 resource is really that powerful
or here Bishop g2, okay. So anyway the simplification now and whites’
advantage is phenomenal. It’s recognized immediately by Houdini as
a winning, nearly winning position. So f6, this isn’t mentioned Rd1. So it does look like some sort of bug occurs. It isn’t mentioned as a strong move for maximum
resistance, Rd1 at all. It is not fourth, it’s not faith. It’s nowhere to be found. So maybe it was just you know a bug, Rd1. Okay after g7 finally okay the team resigns,
the deep blue team resigns. So it was a kind of hyper modern game played
by Kasparov waiting and crating weaknesses through provocation in the style of Tigran Petrosian, kind
of sitting on the base line like Petrosian would do. Waiting for the opponent to do over stretching
or creating weaknesses and then pouncing later in a protein style really to hijack diagonals
exploit weaknesses with this powerful positional exchange sacrifice. So it seems thoroughly based in the positional
advantages, concretely based in positional advantages and the endgame transitions just
amplified and amplified the connected pass pawns, the power of those connected pass pawns. I hope you got something from it. Comments or questions on YouTube. Thanks very much.

37 thoughts on “Amazing Chess Game: Exchange Sacrifice Power! : Garry Kasparov vs Deep Blue – 1997 – Game 1

  1. Yes I am experimenting with this new 2-phase approach – just go over the game first without any engine use – and in the 2nd pass through use the engine. Is it okay to do this you think ?!

  2. yeah, works for me, As you said it saves disturbing the flow of the game ,So an over view and summery approach I would say is the way to go, But what's ever easier for you KC also the madding crowd may prefer the old way of doing things or something in-between one thing is for sure you'll never keep us all happy 🙂

  3. I concur. Just a quick move through the whole game gives the viewer a more organic perspective and a better vantage point for analysis. I would prefer the first pass to be a bit quicker.

  4. Yes, I think in the first pass I will try and emphasise evaluation of positions and general ideas – because my technical variations will likely be totally wrong in sharp positions in any case. The 2nd pass where forensic investigation can take place, can examine the very technical variations and questions in more detail.

  5. Please add this #chess video of Kasparov vs IBM Deep Blue to your favourites if you like it. It gave mankind hope against the machines, if not for just one game 🙂

  6. I like this approach/format much better. !st pass just show the game, added with commentary and personal analysis, as to not interupt the flow. Then 2nd pass through, break out the engine and get more in depth. It's alot better imo.

    It's quite interesting how many more times Kasparov played houdini 1st or 2nd candidate moves compared to Deep Blue. It seems modern engines are able to play much more human like, not so materialistic more positional chess. Quite amazing.

  7. I would use the engine first time just to keep an eye on things, otherwise the two passes is fine and Be careful to avoid repeated cliches such as "on the cards". DB's designer reported that it was still having bug problems during the match. It was ok until the losing Kg7 – it can stop the pawns by returning material – and its evaluation function was probably already compromised at that point as witnessed by the nonsensical Rd1. That would also explain the Ra6 letoff in game 6 when GK resigned.

  8. It was a very instructive game how to play closed positions against agressive style players.
    Thanks for sharing

  9. Deep blue is definitely not as strong as modern engines Kasparov said the deep junior he played in 2003 was stronger than deep blue.

  10. Considering your rate of output I don't really think you can go much wrong with experimenting a bit with order of analysis and such. At worst you'll get a small decrease in video quality over a couple of tens of videos and at best you'll improve quality over thousands of videos. So anything you think might be a good idea is probably at least worth trying.

    Also, thanks for all the videos. I really appreciate the work you put in. 🙂

  11. I liked the video but not the idea that people use engines to boost theyr stiles, it means, if the power runs out, theyr outwitted against those oponents who "play in the desert" and dont need engines, or just cant afford one, and play with intuition. Paul Morphy was such a player. He resigned from chess simply because he couldnt find worthy oponents in his time. In the end he became abit senile too, before his death. :s

  12. LOL at the thumbnail photo. The dude Kasparov is playing is clearly cheating. He has a computer right in front of him.

  13. Kingscrusher, if you have not watched yet i recomend "Game Over-Kasparov And The Machine (2003)" is a good documentary with some conspiration theories.

  14. What about 36…Ng4, aiming at exchanging the knights and cementing the f-line and the g-line?  Maybe 37.f6 is the strongest continuation, willing to sacrify the white knight (37….Nxe3? f7+), but after 37…Re6 instead, I can't see how white can push his pawns further forward to win.  

  15. I would like it if you put ''analysis'' somewhere in your titles because I keep hitting your videos thinking that it shows the actual game live with that thumbnail you put

  16. Hi Kingscrusher, very compliments for your videos, I'm watching all them. Fantastic match by Kasparov here, only the best player of all times could beat a super computer like Deep Blue. Anyway, it seems very strange that the strength of Deep Blue increased so much between the first and the second match against Kasparov in 1997. What do you think about? Did Deep Blue receive an "external assistance" in your opinion?

  17. in the press conference after this game kasparov made some joke about none of his pieces going past the fourth rank in the game and still winning (except for the two passed pawns). it is quite interesting to me that he played sort of passively though dynamically and still won

  18. Replayable game link: http://www.chessworld.net/chessclubs/ltpgnviewer32/ltpgnboard.asp?GameID=91591&v=ILfrKOPdRdM –
    Join me or other Youtubers for a game: http://www.chessworld.net/chessclubs/asplogin.asp?from=1053 – Cheers, K

  19. Hi guys can any1 tell me wer can i watch the actual game not just the bloody documentary??? Jeeezz😡😡 i wana watch the actual All game with Kasparov vs the Machine!! But i cant find it here in youtube its all just Bloody documentary!!😩😩😩😩😭😭😭😡😡😡😡

Leave a Reply

Your email address will not be published. Required fields are marked *