Kasparov-Deep Thought 1997 (second game)

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

User avatar
Leto
Posts: 2071
Joined: Thu May 04, 2006 3:40 am
Location: Dune

Re: Kasparov-Deep Thought 1997 (second game)

Post by Leto »

Lyudmil Tsvetkov wrote:
Leto wrote:
Can you post Kasparov's analysis of this position (45.Ra6) from his new book?
after I buy it, definitely, but still have not done so.

maybe someone knows more.
I now have the book, here are some excerpts from the book about this position:

-----
Much to my dismay, Deep Blue again refused to play like a machine. Instead of grabbing the pawns, it played a bishop move that felt like the last nail in my coffin. After nearly four hours of torment in the sort of passive position I despised, it had somehow become even worse. A fatalistic depression set in and I could barely make the next few moves as white's queen and rook invaded on the a-file. My only hope was to establish some sort of blockade, but I couldn't see any way to achieve it. I gave a last check with my queen, almost out of spite, and barely even noticed that Deep Blue moved its king out of check by moving into the center instead of the more natural retreat toward the corner.


On move forty-five, it attacked my queen with its rook and it was all over. My queen couldn't escape without abandoning my bishop. I could sacrifice the bishop to get in a few desperation checks with my queen against white's open king, but that also looked hopeless. If computers are good at anything, it's seeing long sequences of checks, the most forcing move in the game. After such a powerful performance, it was inconceivable that Deep Blue would allow its king to be chased around for a draw when it had passed up simple ways to secure it.

The entire game had been a demoralizing experience and I just wanted to get as far from the board as possible. My mind was already racing, wondering how in the hell the time-wasting computer from game one had achieved this positional masterpiece in game two. That I was already thinking about anything other than the game was a typical human frailty that we just cannot avoid. It felt physically painful to keep looking at what I was sure was a totally lost position. I wanted to resign with at least a little dignity left and to save some energy for the next game instead of continuing in a hopeless cause.

I resigned and stormed away from the board, replacing my disgust with anger as quickly as I could. I was in no mood to face the audience or the commentators or anyone else. My mother and I left the building with no delay, leaving the Deep Blue team to their moment of glory.
-----

Later on about the day after losing game 2:

Little did I know that night that if it was already going to be very difficult to recover after such a loss, it was about to become impossible. My team Yuri, Frederic, Michael, and Owen and I were walking to lunch down Fifth Avenue the next day when Yuri approached me with the face of a man about to tell someone that a close family member has just passed away. "The final position of yesterday's game was a draw"￾ he told me in Russian. "Perpetual check. Queen to e3. Draw."

I stopped dead still on the sidewalk with my hands on my head for a moment. I looked at each of them since they had all obviously known this and had been debating if, when, and how to break the news to me. They could barely meet my gaze, knowing how horrified I was by the news. I had lost one of the worst games in my life in front of the entire world and now I was finding out that I had resigned in a drawn position for the first time in my life. I was in disbelief, a feeling that I was becoming all too familiar with in this match. A draw?!

The Kabler-Ross model, better known as the five stages of grief, is a set of emotions that terminally ill patients and others experience after they receive terrible news: denial, anger, bargaining, depression, and acceptance. I spent the rest of lunch in a form of incredulous denial, staring at the walls for a few minutes running the variations through my head before I began hammering my poor team with questions. How could Deep Blue miss something so simple? It played so well, it played Be4, it played like God, how could it miss a simple repetition draw?

To psychoanalyze just this once, with twenty years to cycle through the stages, this was also me saying to myself, my god, how could I miss something so simple? When you are the world champion, the world number one, any defeat can be viewed as self-inflicted. This is not exactly fair to my opponents, many of whom would count their victories over me as the pinnacle of their careers, but after such an incredible revelation I wasn't in the mood to be fair to anyone.

The discovery had been made thanks to the power of the Internet to connect people around the world. Even before I resigned game two, the millions of chess players who were following the match got to work analyzing it and sharing their results. By morning, these armchair analysts, also armed with strong engines, had demonstrated that Deep Blue would not have been able to win the final position had I played the best moves instead of resigning. This unbelievable news was verified by my team that morning before they broke the news to me. The queen infiltration I discarded as desperate and pointless was, in fact, a saving resource. The white king would not be able to escape the checks by my queen, eventually resulting in what we call a three-time repetition of position draw. Several of Deep Blue's final moves had in fact been blunders that would have let its brilliant victory slip away had I only been alert to the opportunity.

It was a crushing blow, as if I had lost the game twice. Resigning in a drawn position, unthinkable! I would never have given up so pathetically against any human player in the same position, of that I was sure. I had been so impressed by Deep Blue's play, so demoralized by the way the game had gone, so annoyed at myself for letting it happen, and so sure that a machine would never commit such a simple mistake.
-------------

Later on on chapter 10 he says this about the match and game 2:

After saying all of that, we come to my own confession. On what mattered most, on what really destroyed my composure, I was wrong and owe the Deep Blue team an apology. The moves in game two that left me with a lost position and crushed morale were unique only for the time. Within five years, commercial engines running on standard Intel servers could reproduce all of Deep Blue's best moves, even improving on some of the humanlike￾ moves that so impressed me and everyone else at the time. The engine on my laptop today slightly favors the shockingly humanlike￾ move 37.Be4 from game two in less than ten seconds, although it rates it nearly equal to the queen sortie I had expected because 37.Be4 turns out not to have been as superior as we all believed at the time. Had I played better defense instead of collapsing and resigning, game two would have been considered a very impressive game for a machine but nothing more, no matter the eventual result.

This also highlights why it was so critical that I never saw a single game of Deep Blue's before the match. Had I seen it make a single move demonstrating the uncomputerlike positional approach of game two's Be4, for example, or the surprising h5 pawn push from game five, my play and my reactions would have been completely different. Keeping Deep Blue completely hidden was the strongest move of the match, but it was made by IBM, not by either of the players.

In turn, by understanding now that Deep Blue was very strong but still making plenty of inaccuracies, the fact that it missed the perpetual check draw at the end of game two becomes more comprehensible as well. Still strange, considering its powers of calculation, but no longer inconceivable. If I had had any way of knowing this during the match, perhaps the story would have turned out differently, but I'm not sure. My premature resignation in game two and the intense shame and frustration it produced in me were what made it nearly impossible to play.

------------------

In conclusion unless I'm missing something the original poster of this thread, Franco Kanizsa, is incorrect, Kasparov does not claim the game was lost after 45.Ra6. Perhaps he came to that incorrect conclusion after reading this sentence from the book: "On move forty-five, it attacked my queen with its rook and it was all over." As you can see in the excerpts I posted above Kasparov gives his reasoning at the time for resigning but he does not claim that the position is lost for black after 45.Ra6, on the contrary he believes he missed a draw.
Rodolfo Leoni
Posts: 545
Joined: Tue Jun 06, 2017 4:49 pm
Location: Italy

Re: Kasparov-Deep Thought 1997 (second game)

Post by Rodolfo Leoni »

Then, we must at least acknowledge him for being honest.

I'm rather impressed by the accurate description of his emotional state while in game, and after. No doubt that he had an an impact, and it was stronger than I expected.
F.S.I. Chess Teacher
MikeGL
Posts: 1010
Joined: Thu Sep 01, 2011 2:49 pm

Re: Kasparov-Deep Thought 1997 (second game)

Post by MikeGL »

Leto wrote:
Lyudmil Tsvetkov wrote:
Leto wrote:
Can you post Kasparov's analysis of this position (45.Ra6) from his new book?
after I buy it, definitely, but still have not done so.

maybe someone knows more.
------------------

In conclusion unless I'm missing something the original poster of this thread, Franco Kanizsa, is incorrect, Kasparov does not claim the game was lost after 45.Ra6. Perhaps he came to that incorrect conclusion after reading this sentence from the book: "On move forty-five, it attacked my queen with its rook and it was all over." As you can see in the excerpts I posted above Kasparov gives his reasoning at the time for resigning but he does not claim that the position is lost for black after 45.Ra6, on the contrary he believes he missed a draw.
You mean Kasparov have 4 strong analysts and all of them believed it was
a draw? This can't be, Lyudmil felt there's a win hidden in the position.

But anyway, this statements from ex world champ took my attention:
My only hope was to establish some sort of blockade, but I couldn't see any way to achieve it. I gave a last check with my queen, almost out of spite, and barely even noticed that Deep Blue moved its king out of check by moving into the center instead of the more natural retreat toward the corner.


On move forty-five, it attacked my queen with its rook and it was all over. My queen couldn't escape without abandoning my bishop. I could sacrifice the bishop to get in a few desperation checks with my queen against white's open king, but that also looked hopeless. If computers are good at anything, it's seeing long sequences of checks, the most forcing move in the game. After such a powerful performance, it was inconceivable that Deep Blue would allow its king to be chased around for a draw when it had passed up simple ways to secure it.
By the looks of it, Kasparov resigned without doing concrete analytical work with Qe3! and instead he just used his logic and doing shortcuts. Thought process ran like this: Since Deep BlueII didn't play Kh1 or Kh2 securing its K at the corner, there surely isn't perptual check on this position, because come on, this is Deep Blue II and it played Kf1 without seeing any danger.

But kudos to the legendary ex World Champ for his honest confession on what really happened, including his thought process during the game.
The comment of most GM's were correct, Kasparov lost before the match started.
Rodolfo Leoni
Posts: 545
Joined: Tue Jun 06, 2017 4:49 pm
Location: Italy

Re: Kasparov-Deep Thought 1997 (second game)

Post by Rodolfo Leoni »

MikeGL wrote:.................
By the looks of it, Kasparov resigned without doing concrete analytical work with Qe3! and instead he just used his logic and doing shortcuts. Thought process ran like this: Since Deep BlueII didn't play Kh1 or Kh2 securing its K at the corner, there surely isn't perptual check on this position, because come on, this is Deep Blue II and it played Kf1 without seeing any danger.

But kudos to the legendary ex World Champ for his honest confession on what really happened, including his thought process during the game.
The comment of most GM's were correct, Kasparov lost before the match started.
And that's the most interesting aspect of all that stuff about 45.Ra6. How emotional sphere can interfere with a normal human chess analysis. One of engines advantages is they aren't distracted by human factors. And they haven't any consideration like how opponent precise could have been, or so. The max "distraction" from an objective evaluation is a contempt factor, if any.

I wish to thank Leto for his effort of making us partecipe of real Kasparov thoughts. I hope that was a simple copy/paste....
:)
F.S.I. Chess Teacher
first25plus5
Posts: 11
Joined: Sat Jul 22, 2017 2:50 am
Location: New Zealand

Re: Kasparov-Deep Thought 1997 (second game)

Post by first25plus5 »

Regarding this analysis line at final iteration depth 20/52:

Instead of 13....Qxh4+ ,Black should not take the pawn here and instead go for the move ....Qd2+. This is seen when letting a good engine analyse at this point.
first25plus5
Posts: 11
Joined: Sat Jul 22, 2017 2:50 am
Location: New Zealand

Re: Kasparov-Deep Blue 1997 Game two, Final position summary

Post by first25plus5 »

I've got a file collating analysis from various places including the excellent postings by Louis Zulli in this whole topic further in this topic.
Here is a good summary: - all has been analysed before:

45....Qe3 46.Qxd6 Re8 47.Qd7+ Re7 48.Qc6 Qxe4 49.d6 Qd3+ 50.Kg1 Re8 51.Ra1 h5

From here all the lines are leading to draws. Again many thanks to Louis Zulli for posting more drawing lines. I've spent some time checking these myself and agree with the conclusion, as well as going back over previous moves and lines. Of course I also went back over the initial moves as I've listed.
I'm believing it's now a draw after never being fully convinced.
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Kasparov-Deep Blue 1997 Game two, Final position summary

Post by Lyudmil Tsvetkov »

first25plus5 wrote:I've got a file collating analysis from various places including the excellent postings by Louis Zulli in this whole topic further in this topic.
Here is a good summary: - all has been analysed before:

45....Qe3 46.Qxd6 Re8 47.Qd7+ Re7 48.Qc6 Qxe4 49.d6 Qd3+ 50.Kg1 Re8 51.Ra1 h5

From here all the lines are leading to draws. Again many thanks to Louis Zulli for posting more drawing lines. I've spent some time checking these myself and agree with the conclusion, as well as going back over previous moves and lines. Of course I also went back over the initial moves as I've listed.
I'm believing it's now a draw after never being fully convinced.
I had already buried this thread, as seemingly nothing more to be said, and I got tired of it.

I guess this will be my last post, even if someone posts again, as the very long thread already gets unreadable.

draw is only in your head, Caleb, maybe the result of too frequent earthquakes on the island. :)

sorry, no time to check your h5 move and where it belongs, but you certainly might be posting here for as long as you wish.

if the topic is posted in another cleaner thread in say, couple of months' time, it might make sense to go over it again on a fresher head.