< Earlier Kibitzing · PAGE 2 OF 2 ·
|Aug-12-07|| ||unixfanatic: I'm astounded Fritz was so weak in this game; even for 1991 this is not very good for a computer. Good thing this guy didn't play Rybka!|
|Feb-08-09|| ||WhiteRook48: shocking play by White|
|Mar-19-11|| ||noah913: Fritz played horribly. Roman played pretty well although he missed 24.Ng5! which totally kills black.|
|Mar-20-11|| ||FSR: Probably Dzindzi was blitzing out his moves at the end, but yes, 24.Ng5! ends it and 27.Qxg7 was mate. Surprising that he would miss that. He has a very quick sight of the board, as those of us who've watched him crush masters at 5-1 time odds can attest.|
|Aug-21-11|| ||Bozantium: 15. Rxe6+!|
|Oct-12-11|| ||DrMAL: <csmath: what you see here is an early version of currently popular software playing on a 80386 processor, about the strength of current pocket PCs.> Thank you for clarifying, I was wondering why Roman would adopt Nimzowitsch system with 4.Nf3 then play 5.Bd3?! sharp but very risky move. 8...O-O? was much more confusing, computers were not nearly as good then but they played at master level ever since 80s, this blunder reflects much lower level. |
I used to play 5.Bd3?! as young boy it got me many wins against French with opponents below, say, ELO 1800. Usual line was different, here is one very old example game: 5...cxd4 6.O-O Bc5 7.Re1 Qb6 8.a3 a5 9.Nbd2 Nge7 10.Nf1 Bd7 11.Ng3 Ng6
Opponent realized 11...O-O? was big mistake 12.Bxh7+ Kxh7 13.Ng5 Kg8 14.Qh5 with mate to follow, I had won that way before. Instead he thought Ng6 protects K and castling was OK rest of game was: 12.Nh5 O-O?! 13.Bh6! gxh6 14.Nf6 Kh8 15.Nxd7 Qa7 16.Nxf8 Rxf8 and, up R for N+P with good position I went on to win.
In analyzing game I learned 14.Qd2! (or 14.Qc1!) was much stronger, game 17 months later against better player went: 14.Qd2 f5 (best move, he saw it coming now) 15.exf6 Kf7 16.Qxh6 Ke8 17.Qxh7 Kd8 18.Bxg6 (Qxg6 was stronger) and two pawns behind yet still on the run he soon resigned.
I much recommend carefully studying this opening for player around ELO 1200-1600 it is very instructive and dangerous weapon at this level. After 5.Bd3, taking with N for swap is at least as good below are computer lines. It is also imporant to know lines for other moves such as 5...Bd7 they get played too.
Houdini_20_x64: 32/75 8:02:29 267,019,628,008
+0.18 5. ... Nxd4 6.Nxd4 cxd4 7.0-0 Ne7 8.Re1 Nc6 9.a3 Qc7 10.Qe2 g6 11.Bg5 Bg7 12.f4 Bd7 13.Nd2 h6 14.Bh4
+0.15 5. ... cxd4 6.0-0 Bc5 7.Nbd2 Nge7 8.Nb3 Bb6 9.a4 a6 10.a5 Ba7 11.Qd2 Ng6 12.Bxg6 hxg6 13.Nbxd4 Nxd4 14.Nxd4
Note black is slightly ahead here, eval is relative to side with move. Computer chooses Q-side play at first to minimize objective error behind 5.Bd3 it assumes opponent is super-GM who knows lines extremely well, this is opposite to K-side theme on which 5.Bd3 was chosen. BTW, I fined tuned Houdini for my needs, preferring hash table at 1024MB it seems best trade-off between evaluating more positions (less hash) better for blitz game versus better choice for depth (more hash with 32/75 instead of 28/84 at small hash) since computation is run for long time. Bigger hash such as 4 GB seemed to decrease number of positions too much but it is also fine to use for long computation.
|Feb-28-12|| ||peristilo: Can I get a Fried Fritz, please? LOL! Great moves by Roman. This GM and his videos on youtube have given a great contribution to chess all over the world!|
|Nov-15-12|| ||carapov: Wow. GM Houdini spots mate in 4, beginning with 15. Rxe6+.|
(15. Rxe6+ Bxe6 16. Ne5+ Kf6 17. Ne4+ fxe4 18. Qg5#)
|Nov-15-12|| ||shivasuri4: <carapov>, indeed, that's elegant. In that line, if 17...dxe4, 18.Bg5# is lovely too!|
|Feb-18-13|| ||FSR: Roman Holiday.|
|Apr-23-13|| ||sofouuk: <carapov: Wow. GM Houdini spots mate in 4, beginning with 15. Rxe6+>|
click for larger view
15. Rxe6+ Bxe6 16. Ne5+ Kf6 17. Nh7+ Qxh7 18. Bg5
click for larger view
<very> pretty :)
|Dec-04-13|| ||cyborg077: After rook moves at 26 move then is mate one move before, because bishop is pinned:)|
|Dec-23-14|| ||al wazir: 14. Nxe6+ was worth considering.|
|Dec-23-14|| ||Cheapo by the Dozen: He was toying with the poor machine.|
|Dec-23-14|| ||perfidious: <Cheapo> These days, it would be the other way round.|
|Dec-23-14|| ||morfishine: Tip #3 under 'Tips for Writing Good Puns': "Obvious and tired puns like "Tall" for "Tal" are rarely funny"
Pun Submission Page
This polite suggestion by <CG> refers to such players as Sax, Krush, Dzindzichashvili (Roman), etc. And please divorce yourself from any idea of submitting anymore puns for Don Wedding. The honeymoon is over for that name
<al wazir> 14.Nxe6+ followed by 15.Nxd8 wins instantly, not much to consider
If forced to wring out anymore worn out play on words for Roman, I'd venture "Roman knows best" (play on 'Roman nose')
|Dec-23-14|| ||Once: To put this into perspective, this was the very first version of Fritz, probably running on an MS-DOS PC. The whole program probably amounted to no more than 400 Kb - say about the same size as 1½ angry birds games on an ipad.|
Within a couple of years, Fritz2 and 3 were beating the top grandmasters at speed chess.
Yes, we can look back at these early games and snigger about how poor the early computers were compared to today. But at the time we were pretty amazed by how good they were and how quickly they were improving.
|Dec-23-14|| ||varishnakov: Back when computers could be beaten tactically|
|Dec-23-14|| ||AylerKupp: <Once> Although some are large, chess programs are usually not very large. For example, the executable version of Gull 2.8b, ranked #4 in the latest (Dec-20-2014) CCRL tournament rating list is only 311 Kb; Critter 1.6a, ranked #6, is 546 Kb; IvanHoe 946f, ranked #9, is 602 Kb; BlackMamba 2.0, ranked #9 is 235(!) Kb; Hannibal 1.4, ranked #13, is 330 Kb; and Bouquet 1.8, ranked #8, is 447 Kb. These numbers are for the 64-bit versions; the 32-bit versions are usually a little bit smaller. Then again, they don't have a GUI. And, needless to say, the current versions play a lot better.|
So 400 Kb for a 1991 version of Fritz would not be unusual since the GUI, if included in the program, would likely have been quite rudimentary. Does anyone know how large the current version of Fritz 14 is?
|Dec-23-14|| ||1 2 3 4: <perfidious> <These days, it would be the other way round.> No, they don't toy with us. They destroy us, but that's how it is. Computers can't toy with us, they're programmed to destroy us.|
|Dec-23-14|| ||BOSTER: < morfishine: The honeymoon is over>.
Thanks for the link < Pun Submission Page>.
I've never read it.|
|Dec-23-14|| ||kevin86: Fritz looked like it was on the fritz. lol|
|Dec-23-14|| ||chesssalamander: Isn't 27.Qxg7 checkmate?|
|Dec-23-14|| ||RookFile: Sure. A few have pointed it out in this thread.|
|Dec-23-14|| ||Whitehat1963: Crushing! I love to see someone seemingly leave pieces en prise while they continue to force mate.|
< Earlier Kibitzing · PAGE 2 OF 2 ·