Date: Fri, 2 Jun 2006 13:56:11 -0700
From: David Fetter <david@fetter.org>
To: Michael Dean <mdean@sourceview.com>
Cc: Greg Stark <gsstark@mit.edu>, josh@agliodbs.com,
	Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@postgresql.org
Subject: Re: More thoughts about planner's cost estimates
Message-ID: <20060602205611.GA23684@fetter.org>
References: <14821.1149104747@sss.pgh.pa.us>
	<20060602010412.GI10906@fetter.org>
	<87ac8vfptu.fsf@stark.xeocode.com>
	<200606021223.35168.josh@agliodbs.com>
	<87pshre28d.fsf@stark.xeocode.com>
	<4480A204.3070407@sourceview.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4480A204.3070407@sourceview.com>
User-Agent: Mutt/1.4.2.1i

On Fri, Jun 02, 2006 at 01:39:32PM -0700, Michael Dean wrote:

> I'm sorry to interrupt your esoteric (to me) discussion, but I have
> a very simple question:  would you define a "good unbiased sample"?
> My statistics professor Dan Price (rest his soul) would tell me
> there are only random samples of some sort, and "other", which are
> always biased, and never good.

What's at issue here is a biased estimator, not a biased sample.  If
you know an unbiased estimator of multiplicity based on a random
sample, that would be a great :)

Cheers,
D
-- 
David Fetter <david@fetter.org> http://fetter.org/
phone: +1 415 235 3778        AIM: dfetter666
                              Skype: davidfetter

Remember to vote!