I've been following this thread and don't understand the problems. We have
been doing this with the IOTA Contest for the past few years.
If you correlate the spots with logs within, say, 5 minutes of a spot
appearing (our software allows us to vary the threshold) you get not a
normal distribution but two very clear and distinct peaks. And, guess what,
most of the entrants appearing in one of those peaks have claimed
"Assisted". But a few haven't! So they get an email suggesting that maybe
they clicked the wrong button when sending their log ..... And we take it
from there. Seems to work fine!
IOTA Contest Manager
On 18/12/2007, Denis K7GK <firstname.lastname@example.org> wrote:
> All this talk about packet cheating, public logs and christmas wishes made
> me wonder how hard would it be to write a small program that correlates (now
> public CQWW) logs and DX Cluster spots. Here is what I came up with in a few
> scarcely available hours.
> The archive contains a command line executable (no fancy UI = time
> savings), readme.txt file with some help, sample cabrillo log and a
> capture of cluster spots found on JG1VGX's web site for CQWW CW 2006. Rin's
> site (http://www.jg1vgx.net/en/) also has the cluster spots for CQWW SSB
> for the same year.
> The program calculates the numbers and percentages of frequency jumps
> found in the log where QSOs are made a short time (5/10/20 min) after the
> same call appeared on the DX Cluster. To get an idea whether someone might
> have used a cluster I compare an output produced with two control logs - a
> known non-assisted entry, a known assisted entry and with the log in
> question, all from the same geographic area. The output is very much open to
> interpretation and in my experience does not provide hard evidence of
> cheating. But sometimes I find the information produced in this way quite
> 73, Denis - K7GK
> Get the power of Windows + Web with the new Windows Live.
> CQ-Contest mailing list
CQ-Contest mailing list