Author: Robert Hyatt
Date: 19:33:33 10/04/99
Go up one level in this thread
On October 04, 1999 at 21:08:03, Ricardo Gibert wrote: >On October 04, 1999 at 18:22:29, Robert Hyatt wrote: > >>On October 04, 1999 at 13:54:55, Ricardo Gibert wrote: >> >>>On October 04, 1999 at 13:07:51, Robert Hyatt wrote: >>> >>>>On October 04, 1999 at 11:43:47, Ricardo Gibert wrote: >>>> >>>>> >>>>>But what about the example of a drawing sac? Why leave that one out? >>>>>Technically, mate (mate is really misnomer) is a "forced concession of the >>>>>battle", you don't actually kill the king. That's why we have a complicated >>>>>mating rule. It would be simpler to just capture the K, but the inventers wanted >>>>>to keep the game "polite" and leave the really _fun_ part out. Perhaps, because >>>>>the game was actually anticipated to be played by real kings, they decided it >>>>>was "wiser" to leave the last grisly detail out. Killing a king is not really >>>>>fun when a _real_ king can really kill you in return. But this not _my_ reason >>>>>for calling it a sac. It is a sac when I give something up and I don't collect >>>>>the material back, either because of mate or a forced draw or whatever (=your >>>>>version?). Being "prevented" from getting the material back, because an opponent >>>>>resigns or draw is agreed, does not count of course. >>>>> >>>> >>>>I didn't leave the draw option out. It is _identical_ to the mate situation. >>>>A program plays a move, and discovers that it can checkmate you no matter what >>>>you can do. Is that a sac? >>> >>>It is if material was given up to do so. >>> >>> >>> A program is behind (losing to you) and discovers >>>>that if it makes a move it can force a repetition no matter what you can do. Is >>>>that move a sac? >>> >>>It is if material was given up to do so. >>> >>> >>>Or is it suddenly unimportant since it starts a forced >>>>combination that ends the game forcefully, whether winning or drawing? >>>> >>>>That is the ambiguity I don't like in the general use of the word "sac". >>>>"sacrifice" means to give up something. I am not giving up something if I >>>>mate you in the process... I am getting back more than I gave up many times >>>>over. >>> >>>Of course you are getting back. That's the point of the sac. Why sac if you >>>don't get a net gain? >>> >>> >>>Ditto for drawing if I am in a losing position... >>>> >>> >>> >>>What is given up is _material_. What you don't get back is the _material_. What >>>you get back is something else. That's a sac to put it simply. When get nothing >>>or too little back, it is not a sac, it is a blunder. >>> >>>> >> >> >>that definitely isn't true. I have seen _many_ gm players make what they call >>a positional sacrifice... they allow their pawns to get ripped for what they >>hope is a compensating long-term weakness in the opponent's position that they >>can exploit _first_. So you can sacrifice material or positional considerations, >>and you can find combinations that win material or positional compensation for >>your side... >> >>the term is _still_ way too fuzzy, IMHO. Because sacrifice means to give up >>something you consider valuable. Not trade it for something even _more_ >>valuable... IE the "ultimate sacrifice" where soldiers give up their lives >>for no gain that they will ever realize... they do it for reasons that they >>consider more important than living themselves... > >The issue is material, not what is more valuable. Sheding material is a sac. > _IF_ you are 'shedding it'. If you are playing a combination that wins material, it isn't a sacrifice, just because the first capture loses material, as in my Rxb BxR NxB example... RxB is not a sacrifice by any definition I know of. And you can sacrifice a positional concession just as well. IE I have, on occasion, played gxf6 after white played Bxf6, choosing to wreck my pawn shelter for the use of the g-file. Without seeing how it will turn out further along. That is usually considered either a positional sacrifice or a positional concession (same thing)... >> >> >> >>>> >>>>> >>>>>By my way of looking at things, a mating combination may or may not include a >>>>>sac. I prefer to keep the distinction, rather than meld them together as you do. >>>>>People enjoy the sac version of a combo more. Reason enough to keep the >>>>>distinction IMHO. >>>> >>>>Doesn't bother me... if we agree that the definition of a 'sacrifice' is to >>>>give up material at some point, whether you get it back or not... The only >>>>problem with that definition is that in computer games (particularly those >>>>against humans) it happens so often that it becomes a bland happening. I see >>>>such 'sacrifices' (combinations) in most every game against a human played at >>>>blitz time controls... >>>> >>>> >>>>>Definitions are fairly arbitrary. I choose an objective >>>>>definition that can be applied by anyone. You choose one that is subjective. >>>>>Taking the Rebel game as an example. If a 1200 played that "sac" it would be a >>>>>true sac, since he cannot reasonably be expected to calculate all the >>>>>ramifications. If a GM plays the sac, maybe it is a sac or not, depending on the >>>>>player, time situation, etc. A GM could calculate it to the point where he gets >>>>>the material back or play it based on intuition. I don't care for this >>>>>ambiguity. >>>>> >>>> >>>> >>>>Not a problem for me either way. If the word must stay totally ambiguous, >>>>I can live with it... But a sac from a GM and a sac from a program are two >>>>totally different concepts. Because a program will _never_ do it 'just because >>>>it feels right' while a GM will do it regularly for that reason, without being >>>>able to calculate all the repercussions... >>> >>>A program will do it as a gamble. You clipped an important part of my post that >>>addressed this. What gives? >> >> >>Programs do _not_ gamble. They search a finite tree. They choose the path >>that leads to the best score within that tree. No way can that be considered >>'gambling'... > >Huh? When you make decisions based on a incomplete information, it's gambling. >Case closed. Then the entire game of chess is based on a coin toss? Because the computer certainly can't see mate when it makes most captures. I don't buy that definition at all. As far as the _computer program_ is concerned, it is _not_ making a decision based on incomplete information. It makes a decision based on the exact tree that it searched, and to a program, that tree is the _only_ reality that exists, until it can go deeper... > >> >>If you want to say that some evaluation terms are 'speculative' I would agree. >>But even then the program isn't gambling at all. The programmer is by having >>such a term...
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.