Computer Chess Club Archives




Subject: Strength of the engine in chess programs

Author: Rolf Tueschen

Date: 07:15:44 05/20/02

Excuse me if I ask a question already answered a hundred times or more in this
forum. Could someone explain, perhaps in understandable numbers of percentages,
how important the strength of the engine is in chess programs, are there
differences between commercial and amateur programs?

Let me demonstrate a little thought experiment. If I would gauge (in 2002) the
actually most known chess programs against say 1000 human chess players (first
step) to get some insight into the Elo numbers, I would expect that the top
programs would at best get Elo performances of 2200 - 2350, if I let the engines
play without books and implemented book-like tricks. Now, if I'd do some comp vs
comp over a period of a decade or such, I'd expect the leading engines to reach
astonishing Elo of 2600, maybe 2700! So, what we had found were two things:
without books engines would be outplayed by better human chessplayers but
through imbreeding processes the Elo of the engines could still reach Super-GM
Elos. Now, at that moment I'd organize show matches between the engines and
former or actual champions, with a guaranteed sum of say 1 million of USD for
the champion no matter if he'd looses or wins. What would be the next step?
World champion the engine XY on 1 million GHz?

A fair copy of this:

Enough interest=money provided naked engines of chess programs would be
dispersed by human players from the quality above expert status. Humans will
learn to pay attention to the difficulty of tactical play resulting from the
overall depth of 6 to 10 moves at maximum. The rest of the time will be used to
discover typical exploitations of horizon. Humans will adapt to a completely
different chess. New patterns/ algorithms will be developped for early
spottings. Depending of the specific engine 'early' could well be a whole book
with chapters about "Winning from move 1 on against FRITZ 25" or "How to survive
in a fortress against JUNIOR 12b" etc.

Still, we had the programs with books.

Now, for these programs we need only players from a level of Elo 2500 upwards.
Eidetical talents are absolutely required! Then we can repeat the whole
procedures from above.

Still, we had the Elo numbers due to our imbreeding technology.

In pure comp vs comp matches we could still fabricate magic Elos (We let older
programs play on older hardware vs new programs on new and stronger hardware!).
Then we make some show events with tricky programs, with newest books doctored
until the morning of the first game. The results affirmate by far our Elo
numbers by imbreeding.

But back to the question, what is the real strength of the chess engine? How
would you measure it? When will the engine itself begin to reflect its 'chess'?
How many years from now it will take to develop a real chessplaying robot who
could participate in human tournaments completely on his own? Buying new books
he reads, asking collegues for some information about this or that,
differentiating between truth, lies and irony.   ;-)

Rolf Tueschen

This page took 0.3 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.