[FOM] Chess as science?

[Harvey Friedman]

On 2/5/04 2:23 PM, "Timothy Y. Chow" <tchow at alum.mit.edu> wrote:

> If all that is required for a subject to be "ripe for foundational
> exposition" is that there exist powerful intellectual intuitions, then
> sure, chess is ripe.  Chess was ripe four hundred years ago, for that
> matter; people had powerful intellectual intuitions back then.

The question is whether there is any prospect for analyzing what is behind
these powerful intuitions in any appropriately deep way.

There are plenty of foundational enterprises of clear importance for which
ripeness is unclear. E.g.,

*what is an important mathematical theorem?*

I believe that there is a reasonable chance that one can focus on some
suitably interesting aspects of the above question, and do something really
substantive, foundationally. But this is not at all clear.

What is clear is that I, and many people on the FOM email list, are in a
position to at least try seriously to deal with the above question, and try
seriously to put it into some manageable form. Whether we succeed or not, is
less clear.

But it also appears that, perhaps for many reasons, very few people on the
FOM email list, and very few people outside the FOM email list, are in a
position to do anything serious about the foundations of chess.

But how do we even look at chess as a subject?

If we are going to look at chess as a subject, then we could view

*over the board chess games*

as a special form of experimental testing of ideas, in a rigid debating
format. Thus there is an analogy between

*professional over the board chess*

and 

*real time debate*

say as in Conferences or Meetings or Symposia, where there is generally
carefully timed interaction that is publicly documented.

We could also view

**correspondence chess**

as another form of experimental testing of ideas, perhaps of higher quality
but obviously of lower quantity.
 
> I personally think that "powerful intellectual intuitions" by themselves
> aren't enough.  It is my ("mere") opinion that, at minimum, one needs to
> be able to *evaluate the correctness* of some of those intuitions in a
> relatively objective and definitive manner.  Otherwise, all we'll get is
> a proliferation of mutually inconsistent "foundational expositions" with
> no way to adjudicate between them.

But there are at least reasonably looking ways to evaluate the correctness
of some of the professional work in chess, as you call for.

1. There could well be a reasonably coherent theory that is able to single
out certain chess positions - constructed or naturally occurring - and
asserts that they are advantageous. And this may be highly nonobvious, maybe
paradoxical. And also at odds with what computers recommend. The computers
are obviously tactically vastly superior in certain ways. As numerous games
are played from such positions - positions where theory says that the
position is advantageous - the play comes out as a win, even against other
humans plus computer. Such theory cannot be expected to assert anything
about most positions - just certain kinds of positions. This would lead to
an expansion of theory - e.g., how do we maneuver to get into such subtly
advantageous positions?

2. One should also bear in mind that, even in physics, every single
interesting theory that has ever been proposed, and probably that ever will
be proposed, is false and probably refutable. However, we think and hope
that they steadily march toward the TRUTH - just as in chess.

> Chess is not at that stage.  I don't know what confuses you about Silman's
> review of Berliner; is it that Silman and Berliner disagree?  But this is
> the norm in chess.  One grandmaster confidently dismisses "f3?" as a dud;
> another grandmaster wins a high-profile game with it, and it becomes
> "f3!". 

The dispute between Silman and probably almost all GM's, versus Berliner, it
at a much deeper and much more interesting level than that.

> For a long time grandmasters regarded K+Q versus K+R as trivial;
> then Belle came along with a complete analysis and bamboozled the
> grandmasters.  

The steady march to the TRUTH can be slow.

>Andy Soltis, in a column reprinted in the anthology "Karl
> Marx Plays Chess," gives plenty of examples of opening lines that were
> definitively rejected by top players only to be revived by the next
> generation.  This tendency shows no signs of abating.

But you are just taking one superficial aspects of the march towards the
TRUTH - the particular tentative conclusions concerning some specific cases.
I would appear that that is the wrong way to look at chess progress.

> The only method of definitively settling these debates is exhaustive
> tactical analysis, which in spite of advances in computer technology
> still leaves most of chess untouched.

As indicated above, this can be done and is being done with selective types
of positions. That is how real progress towards the TRUTH can be made.
 
> Watson leans towards 2.  I can't summarize the entire book in a few
> sentences, but part of the reason for thinking that the super GM's are
> better athletes is that post-mortem analysis of their winning games almost
> always reveals that the win was due to something mundane---superior
> opening preparation, lower rate of calculational errors, greater mental
> toughness, and so on.

The over the board games are just timed debates, not the real thing. They
don't reflect chess as science.

>I can't think offhand of a practical example that
> is like the solved endgames---where both sides maneuver *inexplicably* in
> an apparently level, quiet position for dozens of moves with no
> discernible mistakes but gradually the position becomes hopeless for the
> losing side.  

Are you sure about this? My impression was just the opposite. I have seen
quotes from great players that they find the machinations of other great
players as inexplicable.

>Typically the grandmasters will rely on a rather simple
> statement about what the winning chances consist of, if there are any
> winning chances, and it's a matter of concrete variations whether the
> winning chances can be converted into an actual win.

Super GM's are not going to give away great insights - assuming they have
them - for free. They won't even play serious chess for free (smile).

> Watson might be wrong of course; maybe 1 is true.  If super-GM's were
> able, for example, to play the 6-piece endgame that I posted previously
> nearly perfectly, without memorizing gigabytes of variations and without
> being able to articulate carefully what strategic principles they were
> following, then I would consider this strong evidence for 1.  (Of course,
> they can't.)  Or if the super-GM's did not bother with extensive opening
> preparation or studying thousands of current games, that would also be
> some evidence towards 1.  Indeed, in the past, Capablanca got away with
> doing very little opening preparation, but that was because none of his
> opponents did that much of it either (by modern standards).  That approach
> just doesn't work any more today.

Lasker expressed complete disdain for opening analysis, if I remember
correctly. 

CONJECTURE: sometime this Century someone will have such great insights into
the TRUTH that they will crush their opponents (in matches with or without
computers allowed) without spending any serious time on opening analysis.

Harvey Friedman