[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] Re: Strange FIBS ratings

From: Douglas Zare
Subject: Re: [Bug-gnubg] Re: Strange FIBS ratings
Date: Tue, 9 Sep 2003 15:34:15 -0400
User-agent: Internet Messaging Program (IMP) 3.2.2

Quoting "Christopher D. Yep" <address@hidden>:

> > > > > On Mon, Sep 08, 2003 at 11:36:33AM +0200, Jim Segrave wrote
> > > > >
> > > > > Kees' experiments show that cube decisions errors don't weigh as 
> > much as
> > > > > chequer play errors.

They have been mispresented multiple times. I think there is still confusion
about what he showed: In a limited context, the conversion from EMG to
mwc/ratings is different for cube errors versus checker errors. 

This is not the same thing as saying that humans or bots give up more equity (by
either measure) in checker play than with the cube. It suggests that if you
make errors like gnu, then a 2 mppm EMG checker play error rate is worse than a
2 mppm EMG cube error rate when converted to mwc. If your cube error rate is 10
mppm and your checker play error rate is 1 mppm, then you would still give up
more equity through cube errors. The experiments suggest that when stated in
mwc, your cube errors would not be 10 times as large, but perhaps 3-5 times as

 I can't offer any explanation for this, other
> than
> > > > > gnubg's chequerplay is much better than the cube play???

No, this is not correct and is irrelevant.

> I think this phenomenon has been known for many years now.  Kees' 
> experiments and Douglas Zare's research on Gammonvillage are just the 
> latest examples supporting this conclusion.

Some have known it, others have not. I have been arguing that 

1) Humans give up more equity through checker play. 
2) Using EMG overstates the amount given up through cube play.

Many people have not been convinced (mainly weaker players), and I hope that my
column will convince them.

The second point is closer to what KvdD's experiments show. My data says that
human cube errors happen when less mwc is at stake, on average, than checker
play errors. His says that when gnu is told to play stupidly, its cube errors
happen when less mwc is at stake.

As I described elsewhere, I think KvdD's study was quite flawed. The conclusions
were also much more limited than mine. I do not want the two conflated.

Douglas Zare

reply via email to

[Prev in Thread] Current Thread [Next in Thread]