Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Benchmarking chess algorithms

Author: KarinsDad

Date: 12:31:21 07/20/99

Go up one level in this thread


On July 20, 1999 at 14:38:49, Dann Corbit wrote:

[snip]
>I don't think I could possibly disagree more strongly.  I think the notion that
>you invent your own stuff and assume it is better is a truly horrible idea if
>you don't know exactly how good the alternatives are.

1) I do not invent my own stuff in a vacuum. I investigated what is out there
(and am still researching via CCC and other sources) and am trying to improve
upon it.

2) I do not assume my stuff is better.

3) How many non-commercial chess programs out there have a design/functional
specification like I do? For that matter, how many of the commercial programs
have one? Your message here implies that everyone who doesn't use your idea of
scientific methodology is just tinkering (as per what you said below).

4) How many people on this forum besides myself have read Plaat's 180 page
thesis? 5? 10? 50? Dave and Don have probably read it. I do not understand
everything that I read, but at least I read what I can and try to understand as
much as I can.

5) It is arrogance on your part to think that your scientific method of testing
is better than my scientific method of research.

6) Let's take the example of Vincent's claim that MTD(f) is inferior. I'll bet a
handful of Charms lollipops that he didn't read Plaat's thesis and he didn't
spend a year trying to improve it within his program, but he makes claims on
it's validity or lack thereof like an expert. Horse hockey.

  Your notion of checking
>to see what has already been done is an excellent one that I strongly agree
>with.  However, having a controlled framework for testing ideas is an idea whose
>time has come.

7) Not to insult you Dann, but you seem to have testing on the brain. Research
can also resolve questions like this as can just simply asking questions of the
developers.

>
>If you don't know how good the competition is, you have no idea how good your
>new ideas are.

8) People do extensive tests of one program versus another both on the servers
and privately all of the time. People can write unit test code to test a variety
of conditions. Are you saying that people cannot make valid judgements on what
works and what doesn't without knowing exactly how their "competition" is doing
it? Again, horse hockey.

>
>Take the example of good programs.
>Perhaps one program does one thing better than anyone else, and that is why it
>is in the top ten.  Perhaps a different program does a different thing better
>than anyone else and that is why it is in the top ten.  Yet another program may
>do a bunch of things fairly well and that is why it succeeds.  Without knowing
>the true behavior of the algorithms involved, it is not science.

9) With the large number of elements in a chess program, the moment you change
more than one of them to do a comparison is the moment the validity of your data
starts to become unstable. Yes, you can get a fair idea of which algorithms have
which specific behavior, but unless you can enumerate ALL of the reasons (based
on the algorithm or otherwise) for a program, you are just making guestimations.
It just sounds to me that what you are proposing will result in very little gain
for a lot of work which can just be done via a little research. Who cares if
there are people arguing on this forum about one technique or another? That is
how ideas get exchanged.

  It is not art
>either.  It's just monkeying around.  Tinkering can be fun, but someone who
>understands fully all the processes involved and the principles behind them will
>kick the butt of a tinkerer every time.

10) We'll see.

>
>Having a carefully prepared list that describes in detail the behavior of every
>algorithm at your disposal is *essential* in order to even know if your idea is
>better!

11) Having a carefully prepared list will stifle innovation. People who WANT to
tinker will be told "Well, the list says that the algorithm XYZZY sucks in
practice, so don't bother to use it.". Essential is hardly the word I would use
to describe such a list.

I am not trying to be antagonistic here with you Dann (regardless of the tone
above), but I think your idea is lukewarm at best. You are one of those people
who appear to like to prove things and to categorize things. Let's prove that
one algorithm is faster or better than another and write it down for everyone to
see. I am one of those people who like to investigate things just for the sake
of investigating. Let's see if this idea is good or not. I do not need to know
to the nth degree just how many percentage points in speed one algorithm gains
over another and why that might be, just that one worked better for me.

Different strokes for different folks I guess.

KarinsDad :)



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.