Author: blass uri
Date: 04:27:00 07/15/98
Go up one level in this thread
On July 15, 1998 at 02:54:18, fca wrote: >On July 14, 1998 at 18:27:26, blass uri wrote: > >> >>On July 14, 1998 at 17:36:42, fca wrote: >> >>>The reference is: http://www.rebel.nl/rebel10a.htm#ANTI-GM >>> >>>First may I say I am excited about anti-GM, and very positive about the four >>>follow-up examples given of its operation. It looks like a real advance! I >>>will certainly buy R10, but that is not so strange as I bought R8 and R9 (and >>>many other programs) too... >>> >>>But the original / basal example troubles me a lot: I have told Ed of this, but >>>as the example stays on the site I must consider that I may have missed >>>something. >>> >>>In the example, the move is 1. Rxe6 (selected quickly by anti-GM). If 1. Rxe6 >>>is _not_ played, IMO the position is -= . Black has a pawn, but White has some >>>compensation. Probable outcome: draw. >>> >>>Now it is not difficult IMO to show that now 1... fxe6 wins for Black. There >>>are some mates to avoid, but otherwise _natural_ play by Black seems to result >>>in a comfortable win. Not hard to find the winning play (I did). But it is >>>likely that a GM sees more dangers than I. >>> >>>Now what I have already written does not in itself mean that anti-GM has >>>selected a bad move, for if a GM etc. would not select 1... fxe6 (say time was >>>quite limited, and GM got scared) then anti-GM would have worked! >>> >>>I believe Amir's opinion (please excuse if I've been misadvised) was that "they" >>>may not select 1... fxe6 in such circumstances (short time). >>> >>>I disagree. I think there are no sensible alternatives to 1... fxe6, so the >>>sacrifice must be accepted. Here is some manual analysis (interesting to see if >>>programs agree) to support my view. The reason I did the analysis by hand is >>>simply that I do not have a cc here: but I suppose it is relevant because we are >>>looking at what a human Black would play. (A computer black obvioulsy grabs the >>>Rook, and wins). >> >>It is not truth >>The programs I have do not grab the rook if I give them 3 minutes. >>Rxe6 is a good anti computer move. >> >>Uri > >After 1. Rxe6, please Uri what is the suggested move by your programs? Genius3 suggested Rc7 I do not remeber what Fritz5 or Junior4.6 wanted to play but I think Rc7 or 0-0 I gave Gunius3 the position after Rxe6 0-0 Bh6 exf6 Nxe6 and it evaluate the move Qa5 as a clear advantage for black because after Qa5 Bxd7 Rf7 Bxc8 Bxc8 black lose a knight maybe white can play this line better(I did not check the position for more than some minutes with the computer) Uri >>> >>>1. Rx6 0-0 2. Bh6 exf6 3. Nxe6 forking Q & R leaves White at least += - but >>>there may well be something _even_ better for White (as per Alekhine, the >>>makings of a combination are here) >>> >>>1. Rxe6 0-0 2. Bh6 Re8? (if the Rook-capture fear still persists) 3. Bxd7 fxe6 >>>(3... Qxd7 4. Rxg6+ 1-0 ) 4. Qxe6+ Bf8 (4... Rg8 5. Nxe6 +- ) and after the >>>(big) simplification White stays a pawn up, maybe with a better position +- >>> >>>1. Rxe6 Rxc1+?! 2. Rxc1 exf6 (best, why now if not earlier? 2...Nxf4? 3. Rxe7+ >>>+- ) Nxe6 Qb8 Nxc7+ +- as the knight is poisoned, and a King escape to f-land >>>will leave White well ahead in material >>> >>>So... what is left? >>> >>>1. Rxe6 Rc7 but surely not? How boring! At best, Black has lost all >>>of his advantage (I think White is *tactically* winning here, but the problem is >>>too hard for me!). Would a GM actually choose ..Rc7 ? White's threat is still >>>in place... he has not removed anything. Pins... soon. more pins, Rook doubling >>>on e-, etc. Would a human GM care to have Black here? > >Kind regards > >fca
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.