|
< Earlier Kibitzing · PAGE 3 OF 3 ·
Later Kibitzing > |
Dec-27-07
 |
| nimh: It is easy to criticize computer, but try to beat it on equal basis, or find a better move by yourself. |
 |
| Dec-27-07 |
| square dance: its not the computer that i was criticizing. |
|
Dec-27-07
 |
| Eyal: Btw – if we actually try to look at the position, we can see that the 0.83 difference of evaluation (in my own Fritz 10 it's even bigger) between 8.N1c3 and 8.Nd6+, as played by Morphy, is very reasonable. 8.Nd6+ followed by 9.Nxc8+ doesn't contribute to White's development, helps Black to develop and relinquishes control of d6. On the other hand, 8.N1c3 prepares 9.Qd6 with a very strong grip on Black's position (an immediate 8.Qd6 is also possible, but perhaps a bit premature after 8...Qxd6 9.Nxd6+ Ke7 10.e5 Ng4) – the only way to prevent that is by 8...Ke7, which in this case is very awkward and much weaker than in the game. |
|
Dec-27-07
|
| nimh: The whole case was started by ibm's pointless nitpicking over the naming of errors ang distrust towards computer's opinion.
It would be interesting to see what alternative he would propose instead of Morphy's or Rybka's eighth move. |
|
| Dec-27-07 |
| D4n: Pushing pawns wins games... |
|
Dec-28-07
 |
| InspiredByMorphy: <nimh> <The whole case was started by ibm's pointless nitpicking> We learn by analyzing not only the game but each others analysis. We then express our opinions regarding it. This is called constructive criticism. My opinion is that given by a human, to only give the computers analysis expressing no opinion of your own is only partially beneficial. You are human with a brain which makes you unique unlike the box your analysis derives from. You are negatively defensive when anyone questions the silicon master. I would recommend working on this to get better interaction with those on this site. Of course if you don't care about interaction and only want to filibuster then by all means continue. Just remember, it is the combination of human and computer analysis that will yield the greatest results. Not one without the other. |
|
| Dec-28-07 |
| beatgiant: <nimh>
I don't have the engine, so to me the computer checks are informative. But except for positions where a complete search is possible, computers have their limits too. In particular, the current incarnation of Rybka seems to be weak in the endgame, as shown in McDonnell vs La Bourdonnais, 1834 where its evaluation swung wildly and flagged "blunders" in a near dead drawn position. I haven't studied the current game enough to comment on this opening, but in the current state of the technology, programmers usually rely on a preset "book" since computer evaluation functions aren't deep enough to rely on their own strategic ability in the opening phase. So, maybe you should increase the "blunder" threshold when evaluating opening and endgame positions. |
|
Dec-28-07
|
| nimh: <You are negatively defensive when anyone questions the silicon master.>
The attacking side should always bear in mind that defender will do his job as good as he can. <This is called constructive criticism>
Fine, show you are capable of it by pointing out flaws in Rybka's ideas, or describing why Morphy's choice 150 yrs ago was better than Rybka's present opinion. <You are human with a brain which makes you unique unlike the box your analysis derives from>
<<laugh>>I'm not even 2000 ELO. Analysis of silicon brain + analysis of my brain = analysis of silicon brain. <Rybka seems to be weak in the endgame>
Yes, I know it, but I'm sorry, I cannot do anything about it.
I think not so best analysis is better than no analysis at all. :)
And also I'd want to note that having threshold should prevent a comp's preference to be called merely 'a matter of taste'. <So, maybe you should increase the "blunder" threshold when evaluating opening and endgame positions.>
I have done it already. I used to check wiht the threshold of 0.25, but thanks to someone's advice I decided to raise it a bit.
If I used too high limiter, we would lose something in accuracy.
Moreover, how is it possible to set a different threshold for different game phases? |
|
| Dec-28-07 |
| beatgiant: <If I used too high limiter, we would lose something in accuracy.> Right, which is why I suggest it for those parts where it is already inaccurate. <how is it possible to set a different threshold for different game phases?> Easy. Run it twice - once at threshold 0.33, say, and once at 0.66. Apply first results to the middlegame and second to the other phases. You can make reasonable definitions like opening = first 10 moves, endgame = position with less than 4 non-pawn pieces on each side. |
|
Mar-25-08
 |
| The Chess Express: <InspiredByMorphy: My opinion is that given by a human, to only give the computers analysis expressing no opinion of your own is only partially beneficial.> I find long lines of analysis with no comments quite boring actually. Maybe that's because I'm not a good enough player to understand it all. |
|
| Mar-25-08 |
| mrbiggs: I actually (gasp) kind of like nimh's game analysis because it gives me a place to start when I'm informally analyzing a game. I don't have his computer power, so it's nice to see an idea of when a good computer thinks the greats went wrong, even if it's not necessarily a 100% certainty. |
|
Mar-25-08
 |
| keypusher: What surprises me most about rybka's analysis is that it apparently approves of Morphy's Nb5 sacrifice. My Shredder agrees that 6....Bxe3 7. fxe3 Qxe3+ 8. Be2 is good for White but is ambivalent about 7....Nc6, which looks pretty good for Black to me. In a subsequent game Morphy preferred 6. Nc3 and Paulsen went horribly wrong with 6....Qxb2 7. Ndb5 Bxe3 8. Rb1. Morphy vs Paulsen, 1857 |
|
Apr-07-08
|
| The Chess Express: Yea, after 6. Nb5 Bxe3 7. fxe3 Nc6 8. Nd6+ Ke7 9. Nc4 Qc5 it looks like black is becoming pretty active. |
|
| Apr-08-08 |
| RookFile: Given the way the game went, I think black had every reason to be satisfied with his position out of the opening. This game is quite impressive, I think both players were ahead of their time. |
|
Apr-08-08
|
| The Chess Express: Yep, I agree 100%. |
|
Apr-15-08
 |
| whiteshark: In the position after <19.Qd2> 
click for larger view19…Qd4+ 20.Qf2 Qxf2+ 21.Rxf2 d6 or 19…d6 20.Rfd1 Qd4+ or 19…Nh5 20.Rad1 d6 would have secured a small black advantage. |
|
| Apr-15-08 |
| Riverbeast: Beautiful game, I love the purity and clarity of Morphy's play |
|
May-18-08
 |
| heuristic: This is game #5 of the final round of the 1st American Chess Congress |
|
May-18-08
|
| heuristic: <In the position after <19.Qd2>
19…Qd4+ 20.Qf2 Qxf2+ 21.Rxf2 d6 or
19…d6 20.Rfd1 Qd4+ or
19…Nh5 20.Rad1 d6
would have secured a small black advantage>however, these lines seem to equalize :
19...Qd4+ 20.Kh1 Nh5 21.Rfd1 Ng3+ 22.Kh2
19...d6 20.Rfd1 d5 21.Qf2 dxe4 22.Bxe4
19...Nh5 20.Rad1 Nf4 21.Rf2 Rgd8 22.c4
|
|
May-18-08
|
| heuristic: <Rybka 2.3.1 mp, AMD X2 2.01GHz, 10 min per move>
i don't want to regurgitate the previous "man vs engine" thread.... i'm curious why rybka rated 61.Rxd4+ higher than 61.Ka5.
Personally i like Rxd4+, but my exposure to engines lead me
to believe that Ka5 would be evaluated higher.
61.Rxd4+ Kxd4 62.c7 Ra1+ 63.Kb3 Rb1+ 64.Ka3 Rc1 65.b6 Rc3+
61.Ka5 Rh5 62.c7 Rxb5+ 63.Ka6 Kxb4 64.Rxd4+ Kc5 65.c8Q+ Kxd4
66.Kxb5 #11
note that i used 62...Ra1+ in the first line to remain compliant
with Rybka's analysis.(Ra1+ = 6.45 versus Rh7 = 25.95) note that the second line leads to a forced mate. (but it
requires a depth of 22 ply)
With engines, it's a delicate balance between branch depth and
tree width; i'm wondering if the 10min/move parameter influences
width more than depth.
and i'm not critizing nimh's work. I find his work & analysis
with "complexity of positions" to be excellent.
|
|
May-20-08
|
| heuristic: <In the position after 19.Qd2>
i misunderstood <whiteshark>'s comment, even if the lines equalize; BLK's position is still better than with the game move 19...Rg7. even 19...d5 20.Rad1 Rgd8 21.Qf2 dxe4 22.Bxe4 is better than an engine-best line of 19...Rg7 20.Rad1 d6 21.Rf2 Nh5 22.a4. it sure looks like 19...Rg7 is not a good line, in fact it might be considered a mistake :-) |
|
| May-24-08 |
| addiction to chess: Morphy was truly Great. |
|
Sep-22-08
|
| The Chess Express: He’s my favorite early chess player. |
|
| Dec-13-08 |
| Bear With Me: It is easy to overlook that Morphy had a good understanding of endgames, as well as being noted for his tactical play. In this respect I think he is similar to Tal, who as we all know was a tactical genius, but has many fine endgame wins to his credit where he shows excellent technique, but is never really given proper recognition, in same way as it wass for Morphy: |
|
Jun-02-09
 |
| David2009: There is a very interesting variation proposed by <Calli> in Morphy-Paulsen 5th match game 1857. After the moves 47...Re5! (in place of 47...Ke5) 48.Ra6 Re3+! 49.Kf2! Rc3 50.Rxb6+ Ke5 51.a5 Rxc2+ 52.Kf3 Rc3+ 53.Ke2 d4, one reaches with White to play
click for larger view
This gives Black better practical chances than in the game but still seems to be a win for White.
I can see no defence to 54 a6, e.g. 54.. Ke4 55 Rb5 Rc2+ 56 Kd1 Kd3 57 a7 and Black has insufficient counter-play. Comments welcome. |
|
 |
|
< Earlier Kibitzing · PAGE 3 OF 3 ·
Later Kibitzing > |
sac: 61.Rxd4+
