Hi Pedro,
Thanks for replying and clarifying.
I agree now.
That is in fact how it happens. They know what to do and how to do it, in
order to avoid being detected.
Fortunatelly, not all the checking is automated and human eyes can still
see what the numbers don't tell. But still, it requires a tremendous effort
by the organizers.
73,
Martin, LU5DX
On Wed, Dec 17, 2014 at 1:51 PM, Pedro Colla <pedro_colla@hotmail.com>
wrote:
>
>
> >> Martin, LU5DX Commented:
> I have attempted to comment in your blog post with no luck.Your analysis
> is not useful to prove who has cheated and who has not.Statistics here do
> not solve the problem, at least in the way you outlinedyour analysis.
> =================================================================Martín,
> Probably my point didn't get thru.
> My attempt is not to try to proove this point using this approach, if you
> understandthat then I'll completely goofed into explaining myself.
> My point was to follow the claim from an earlier poster that the analysis
> of QSO vs.Mult migh throw some light into the subject; and thus validate
> that it will not becauseof the analysis in strict statistical terms just
> lack power to reach that conclusion becauseof the dispersion of the
> samples. In other words proove that this technique isn't worthto fulfill
> it's intended purpose (say identify cheaters).
> Also that what can be seen and validated is that the majority of the
> population in that sample (and also in few others I tried) exhibits a
> significant signature of beingskewed towards more multipliers (assisted) or
> more QSO (non-assisted); this can besaid with some robust likelihood.
> And if any "cheater" hides his assisted nature it falls into the pattern
> of an assistedstation (dominion of Mult over QSO) so his/her advantage
> isn't that obvious.
> The use of outliers, monster level skilled operators and special cases can
> not be used neither as examples or counter-examples; the average population
> should be inspected for that.
> You have far more experience than me in contests, but I do have far more
> experiencethan you in statistics; the signature of normal and abnormal
> behaviour does appears, it'snot conclusive but it's there. If you run with
> a proficient rate you have not much roomto make a perfect conceal of your
> traces, sooner or later the signature will appear.
> And there are a number of techniques in that direction, from rate profile,
> to correlationbetween spots and work, to correlation of UBN with bad spots,
> with the analysisof your relative performance against the peers in your
> category with some field levelingin terms of propagation. Techniques do
> exists, I know them as everybody else does.
> Why the contest organizers choosed not to use it to some extent at least
> to cast somedoubts and require further inspection is something that should
> be asked to them.
> 73 de Pedro LU7HZ
>
> Dr. Pedro E. Colla
> Va.Belgrano-Ciudad de Cordoba
> Cordoba- Argentina
> "Que el hombre sepa que el hombre puede.".A.Barragán, Expedición Atlantis..
> _______________________________________________
> CQ-Contest mailing list
> CQ-Contest@contesting.com
> http://lists.contesting.com/mailman/listinfo/cq-contest
>
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|