Author: Uri Blass
Date: 13:38:26 02/04/02
Go up one level in this thread
On February 04, 2002 at 16:28:37, Miguel A. Ballicora wrote: >On February 04, 2002 at 16:08:19, Uri Blass wrote: > >>On February 04, 2002 at 15:37:38, David Rasmussen wrote: >> >>>On February 04, 2002 at 11:40:07, Miguel A. Ballicora wrote: >>> >>>>On February 04, 2002 at 10:38:04, David Rasmussen wrote: >>>> >>>>> >>>>>Thanks for the link. Great article! I too am a weak chess player, and I have >>>>>also recently had an insight about focusing on tactics instead of positional >>>>>knowledge. I have 40 chess books or so, and of course some of them are about >>>> >>>>I found this kind of things too dogmatic. >>>>The most important part of learning is interest and motivation. If you despise >>>>going through thousands of tactical exercises with nothing in between for >>>>a year most probably your are going to quit after two weeks. This is like >>>>the magic diets where you have to juggle your day around the diet. >>>>As always, improvement is an individual effort and depends very much on the >>>>individual. That's where the importance of the teacher comes, NOT TO EXPLAIN >>>>WHY Ba4 is better than Bxc6 in the Ruy Lopez. >>>>Tactics will be a key for a player, but might not be for another. There >>>>hundreds of details that are important and some of them are not even related >>>>to chess (like attention etc.). In general, tactics are very important >>>>particularly at that level, but it is not wise to separate it from everything >>>>else. >>>>Tactical exercises are good, but it is never good to be 100% of the training. >>>> >>>>Ah! do not forget to play real OTB chess, but not too much. 60-80 slow paced >>>>(anything that last more than 3 hours) games a year, select some and analyze >>>>them to death WITHOUT A COMPUTER, show it to a stronger player or a friend. >>>>Share analysis... Then, use your computer. Keep a notebook with everything... >>>> >>>>Regards, >>>>Miguel >>>> >>> >>>I know what you mean, and I generally agree. I too find the article dogmatic, >>>but that doesn't matter, IMO. Sometimes that's needed to fight another (older) >>>dogma. The articles is even wrong at some points: It is not necesarily better to >>>be able to look 5 moves with "no positional knowledge (not even material?)", >>>than to look 4 moves with Grandmaster level positional knowledge. In chess >>>programming terms: There are a lot of evaluation terms that makes up for search >>>depth: If you have a passed pawn on the 6th rank supported by your king in an >>>endgame, with positional knowledge, you will know with a 0-ply search that this >>>is strong, whereas it takes a 3-ply search with "no knowledge" to see this. >> >> >>I agree that it is not exactly true that 5 plies is always better than 4 plies >>but 5 plies with understanding of only material and pieces square table may be >>practically better than 4 plies of one of most programs. > >No program has an eval of "GM level positional knowledge" so the comparison >is not good. >Better experiment would be a program with no knowledge reaching 5 plies against >a GM in a bullet game with an increment equal to physically make a move. Almost >without thinking. > >Miguel programs evaluation and humans evaluation cannot be compared. No human give exact number to positions in bullet. Humans use something different than static evaluation to evaluate positions. Uri
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.