If they are using time between events as their only
measure (surely not), then given the sample size of
stations in the contest joined with the fact that a station
is on the air to be spotted, the folks at the contest
committees had better go have a long hard look at the classic
probability first year section canard called "the birthday
problem".
This is the one where the professor bets his class that
two will have the same birthday. Here we are not dealing
with things that are as uniformly distributed as birthdays
so our likelihood of having an intersecting event is EVEN
HIGHER than the birthday problem would imply. Suppose I
just end CQ'ing and start S&P. SOMEONE has to find the station
on the air or no one would be around to "hear the tree fall
in the forest" and to spot the station. If I am unlucky
enough to do this a couple of times, I can be accused of
using spotting when nothing of the sort occurred. These
kinds of filters are tremendously difficult to get right.
Hopefully, if such is in use, they really go out of their
way to look for extreme cases of blatant cheating.
I suppose the contest committee could in fact gather all
spots into a database and then match them against logs
for coincidental timings. If you get too many coincidences,
then it is likely to not be a coincidence. You can get
a good estimate of the "rate of coincidences" from using
ALL logs where unassisted was claimed and use logs where
assist was claimed to try and estimate the sigmage between
the two distributions. This will also enable you to
get a theoretical answer (maybe based on the BS assumptions
you just made) as to how MANY times the coincidence has
to occur before you no longer believe it is a coincidence.
Hopefully this confusing diatribe by a Ph.D. in mathematics
(probability theory and stochastic processes) convinces you
this is not a thing for the hot and angry to institute as a
test but should be done by someone with professional experience.
You will need to do this on each and every contest as the
distributions will change from contest to contest.
Bob
N4HY
-----Original Message-----
From: cq-contest-bounces@contesting.com
[mailto:cq-contest-bounces@contesting.com]On Behalf Of K4SB
Sent: Thursday, July 31, 2003 14:39
To: cq-contest@contesting.com
Subject: Re: [CQ-Contest] CQ WW Packet Use Statement
Robert McGwier wrote:
> You have got to be kidding right?
> Why would they tell you how to defeat
> their filter by telling you exactly
> how it works?
Bob is of course, absolutely correct. However, the only constant I can
think of is time.
Don't know or care, but they're probably using some period of time
from the packet post to the logged time. Maybe 3 mins or so, perhaps
even less.
73
Ed
---------------------------------------------------------------
The world's top contesters battle it out in Finland!
THE OFFICIAL FILM of WRTC 2002 now on professional DVD and VHS!
http://home1.pacific.net.sg/~jamesb/
---------------------------------------------------------------
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|