[<a href='13687.html'>Next Message in Time</a>] | 
[<a href='13685.html'>Previous Message in Time</a>] | 
[<a href='13687.html'>Next Message in Topic</a>] | 
[<a href='13685.html'>Previous Message in Topic</a>] <br/><br/>
<b>Message ID:</b> 13686 <br/>
<b>Date:</b> Tue Jan 25 21:46:16 GMT 2000 <br/>
<b>Author:</b> Ryan Honeyman <br/>
<b>Subject:</b> RE: Deepwater Gobs <br/>
<br/><br/>
<div id="ygrps-yiv-412397010"><blockquote><span title="ireply">&gt; Just a brief word on statistics.  If you want to get a reasonably good<br/>
&gt; approximation of what % something drops, you need a number of tests one<br/>
&gt; order of magnitude higher than the range.  So if its 5% you need at least<br/>
&gt; 200 tests for statistical accuracy.  This isn&#39;t necessarily true for all<br/>
&gt; randoms, but for psuedo-random numbers such as computers produce, it is<br/>
&gt; generally the applicable rule.<br/>
<br/>
 </span></blockquote>  I&#39;ll gladly defer this to you and John(tm).  I&#39;m weak at<br/>
  statistics ;-)  All I know is I&#39;ve been really damn lucky<br/>
  and really damn unlucky hehe.<br/>
<br/>
  Harm.</div>