"Mike Murray" wrote in message
...
On 17 Mar 2007 16:42:08 -0700, "help bot"
wrote:

I see a few players whose Playchess ratings seem pretty stable over
time, but many who experience up and down swings of several hundred
points. I think this is relatively rare OTB.

I'd be interested in other players' perspectives on this.

Many times I have seen people talk about how they
sometimes (or even often) allow others to play under
their handle at rated chess online, and I have also
witnessed several cases of this first hand. I have also
seen players who wanted to "teach someone a lesson"
who fired up a chess playing program, to deal a blow
to some fellow's ego -- someone who had ticked them
off. Obviously, it is much harder to do this sort of
substitution OTB, so the online ratings are far less
meaningful than an OTB rating.

Good points, although until proven otherwise, I'd think this
relatively rare.

Not only is it rare, but the way an elo chess rating works is that even if
you throw some games, for whatever reason, and take a rating dive, it will,
at best, be very temporary. Your rating will quickly get back to a normal
level. What I think is the biggest factor in this whole issue is honesty and
objectivity. I believe that the majority of chess players, especially those
at the lower skill levels, do not have either the ability, or the desire, to
assess their skill level in an honest and unbiased way. They will make
excuses and say nonsensical things such as "blitz chess isn't chess" and
"online chess isn't real", etc. to rationalize their poor performances.
You'll find that even when you beat these people at slower time controls and
OTB they'll come up with some new excuses to try to discredit you. In the
end it's the result that tell the tale.


This also would explain (in part) why the online
ratings are more volatile, even apart from any added
playing activity as described far above.

But added to this is another aspect of online play:
often it can be hard to find an opponent unless you
are willing to play fast time controls, such as blitz.
Because of this, even a player who much prefers
slower chess can be forced into playing blitz or
even bullet chess, where he fails miserably --
relative to his OTB skill level, that is.

Playchess maintains three ratings: slow, blitz and bullet. And there
are many players with a w..i..d..e difference between their ratings at
these three modes.

Lastly, there is the small matter of mouse,
typing, or touch-pad speed differences among the
machines themselves to factor in. I have known
players who claim that typing the (blitz or bullet)
moves is significantly faster than using a mouse,
or who even use a third party to do this for them
while they call out the moves verbally! Obviously,
this makes any closely-contested games depend
all too heavily on secondary issues, rather than
what we normally consider to be real chess skill.
It also seems to give the younger players a big
edge over their older, slower-reflexed rivals.

For this reason, players with a bullet rating over 100 points higher
than their blitz rating go on my "ignore" list. It's also why I have
pretty much settled on 3.2 (three minute game with a 2 second per move
increment). It's fast enough to discourage alternate computer
cheating and slow enough to make mouse skill less decisive.

There's really no such thing as "mouse-skill". I'm assuming you're using a
decent, wired mouse thats comfortable, doesn't skip or have a mind of it's
own (logitech are known for this) and that you have it set to an
appropriate speed. Beyond that, it's about the judicious use of premove and
in making good decisions quickly.


As an experiment, I played quite a few 3.0 games and found the rate of
lag cheating unacceptably high.

Lag cheating does occur. After you iggy all the usual suspects who do this
you won't notice it that often, and the vast majority of the cheaters are
under 2000, so if you set your formula above that you won't have to deal
with them.

JMR