Re: The Future of Chess and Getting It "Unstuck" (potential heresy

On Apr 2, 9:43 am, richardhut...@[EMAIL PROTECTED]
 wrote:
>
> I believe the framework of chess can be addressed now so that we never
> turn chess into a solved game.

When you say "turn Chess into a solved game" it sounds like something
that will ruin the game.  When Chess is eventually solved that will be
testimony to the state of the art of computer science and nothing
more.  It won't have any impact whatsoever on Chess gameplay among
humans.  Checkers was solved.  Does that mean that you or I can now
sit down and force a draw against the world's leading Checkers
player?  No.  Does Checkers being solved have any impact whatsoever on
its gameplay among beginners, intermediate, or even expert players?
No.  Not unless you happen to have a supercomputer handy that can play
perfectly.

More and more Chess opening plays get cataloged but that doesn't move
Chess any closer to being solved.  It just creates more opening plays
that aspiring Chess masters have to learn.  Eventually (or perhaps
already) there will be a balance between the number of new openings
that are sufficiently unique and thus worth memorizing and the number
of openings that can reasonably be memorized.

> ... (stuck here means set on a path
> to being "solved", without a way to adjust before it does).
>

All games are on a "path to being solved".  That's a consequence of
ever increasing computational power.  By the way, Backgammon is no
further from being solved than a lot of other games without dice.  Yes
the dice rolls between moves increase the branching factor, but
Backgammon is played on a one dimensional board with simple, uniform
playing pieces which results in a small branching factor (for a given
dice roll) compared to something like Grand Chess.  Solving a game has
absolutely zero consequences for gameplay among humans.  It's only an
issue for robots.

Making these occasional changes to Chess that you suggest doesn't
really make sense.  Not everyone is going to agree on what the changes
should be, how often they should occur, etc.  The notion of keeping
Chess "off balance" like this, to prevent anyone from ever becoming
truly expert at the game is ill fated.  People like to become expert
at things and they're not going to take kindly to having the rug
pulled out from under them just when they're approaching grandmaster
status.

One thing you seem to be getting at here is game tree size.  You could
play Grand Chess on a 16x16 board but that has four times the number
of squares an 8x8 board has, which has a monumental impact on the size
of the game tree.  A relatively huge game tree can create tedious,
lengthy, difficult games for beginners.  Beginners love Chess.  They
don't have a problem with it.  It takes a half hour to 45 minutes to
play and produces a winner almost every time.  But there's a very
broad range of skill levels in an established game like Chess.
Experts are so much better than beginners that they wring the game
out.  Sure you could fix it for experts but at the cost of ruining it
for beginners.  You have Chess960, etc. but this creates problems for
beginners too.  You can't get started learning a few basic opening
plays.  Now you have this big, complex range of opening possibilities,
as well as the problem of randomness (a topic in its own right).

The only games which are immune to these tradeoffs are scalable games
like Othello or Rush.  For scalable games you just increase the board
size a tad for increasing skill levels, gradually varying from 9x9 to
11x11 to 13x13, etc.  It's exactly the same game - just a little
bigger.  Chess is not a scalable game.  You don't have that option
with Chess.