Re: [chrony-dev] [GIT] chrony/chrony.git branch, master, updated. 1.25-pre1-18-g20a4340

[Bill Unruh <unruh@xxxxxxxxxxxxxx>]

On Fri, 15 Apr 2011, Miroslav Lichvar wrote:

> On Fri, Apr 15, 2011 at 01:45:08AM -0700, Bill Unruh wrote:
> > How the hell can it diverge? Is it possible to get the values of w[i]
> > for this divergent example?
>
> I think it's the distribution of the weights what makes it unstable
> and it has to stay that way long enough to reach the inf (e.g. the
> exponentially decreasing dispersion in refclock after a frequency
> step), otherwise it will stabilize. Absolute values of the weights
> don't seem to matter.
>
> The columns are number of the sample, weight, peer_distance.
>
> 0 1.331934e+300 3.021604e-04
> 1 7.203082e+299 2.907540e-04
> 2 3.078620e+299 2.797782e-04
> 3 7.403268e+298 2.692167e-04
> 4 1.000000e+00 2.590539e-04

But then W= sum1/w[i] is 1,  not very large, and dividing by 1 should not do
anything strange. Which says that both P and V are very very small.
Which I guess makes sense. It also means that the results are pretty
unreliable.

The problem with using the unweighted variance is that it sort of obviates the
use of the weights. It certainly is an overestimate of the variance, but
probably not too bad a one (far better than the underestimate of the current
procedure) and it almost certainly will not be unstable.

> 0 3.577760e+302 2.907540e-04
> 1 1.934848e+302 2.797782e-04
> 2 8.269600e+301 2.692167e-04
> 3 1.988620e+301 2.590539e-04
> 4 1.000000e+00 2.492748e-04
>
> 0 1.011771e+305 2.797782e-04
> 1 5.471643e+304 2.692167e-04
> 2 2.338597e+304 2.590539e-04
> 3 5.623706e+303 2.492748e-04
> 4 1.000000e+00 2.398648e-04
>
> 0 1.676595e+308 2.797782e-04
> 1 1.031369e+308 2.692167e-04
> 2 5.577631e+307 2.590539e-04
> 3 2.383896e+307 2.492748e-04
> 4 5.732643e+306 2.398648e-04
> 5 1.000000e+00 2.308100e-04
>
> 0 inf 2.590539e-04
> 1 inf 2.492748e-04
> 2 inf 2.398648e-04
> 3 inf 2.308100e-04
> 4 1.000000e+00 2.220971e-04

--
William G. Unruh   |  Canadian Institute for|     Tel: +1(604)822-3273
Physics&Astronomy  |     Advanced Research  |     Fax: +1(604)822-5324
UBC, Vancouver,BC  |   Program in Cosmology |     unruh@xxxxxxxxxxxxxx
Canada V6T 1Z1     |      and Gravity       |  www.theory.physics.ubc.ca/

---
To unsubscribe email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject.
For help email chrony-dev-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.