[CQ-Contest] Kudos to CQ WW for Busting Packet Cheaters!

[David Robbins]

i just thought of another good one.  what would the committee do if a guest
operator at my station using his own callsign worked everything that wasn't a
dupe that was spotted by k1ttt-6 within a minute or two???  and the coincidence
rate of other spots was around the normal so2r level.  

how could this happen?  well, my login script for the instance of wintelnetx
that feeds spots to the ct ethernet in the shack automatically logs in to the
node as k1ttt-6.  the operator is a real s/o but wants to help out the yccc so
sets up ct to be able to send spots but not receive them.  so every time he
dials up the band and works a new qso he does alt-f3 to send it as a spot....
but the spots come out as k1ttt-6 on the spotting network and go into his log at
about the same time +/- node and ct clock errors which could be a couple minutes
either way.  just now i see a spot going though the network that is a full 5
minutes in the future!  now if it was an exact 1:1 correlation that would be
interesting enough, but this operator uses my advice and when he types in a call
and it comes up a dupe, he doesn't just wipe it he alt-f3 and sends it.... so
now it may appear that he is only grabbing the spots that are new qso's or mults
that k1ttt-6 is spotting.

David Robbins wrote:
> 
> "Leigh S. Jones" wrote:
> >
> > Re: Marty NW0L's comments on CQWW Packet Cheaters:
> >
> > It's not a flawless system.  I can imagine a number of situations
> > where a correlation may exist without any real cheating.  For
> > instance, an unassisted single operator station may be located very
> > close to a packet assisted station, the unassisted station may follow
> > the packet assisted station around the band (refine this technique
> > through the use of a spectrum analyzer to help follow the loud signal
> > around the bands) resulting in a strong correlation between the packet
> > spots and his unassisted contacts.
> >
> 
> i could see how a system of correlating spots with qsos could easily lead to
> incorrect disqualifications unless some way to exclude hits from the following
> types of scenarios is devised:
> 
> 1. using a radio with a spectrum analyzer as a second radio it would be easy in
> the evening to be running on 20m and pouncing on the relatively widespread
> pileups (many of which have probably been spotted) on 10m and 15m.  similar
> effects can probably be found on other bands just opening or closing where a s/o
> may be searching for mults this way.  it is amazing to see a new packet pileup
> on a spectrum analyzer, its pretty hard to miss if you are watching the scope.
> 
> 2. in the early morning on 10m and 15m the first loud africans get spotted just
> about the same time a so2r or maybe a plain s/o would be tuning up the band to
> see if it was open.  since the first spots often appear about the same time as a
> savvy single op would be checking the band on a second radio or vfo a close
> correlation is likely.
> 
> 3. on 40m ssb most stations operating split working stateside get spotted often,
> it would be hard to not work something that had been spotted at times.
> 
> 4. on some systems there are users who run up and down the band just to spot
> anything to fill up the band maps of the multi ops.  there are also m/s stations
> with spare receivers that spot lots of stuff to keep their band maps full for
> the next band switch.  with this happening it is very common to be able to dial
> up a band and find a very high percentage of the stations that are cqing already
> in the band map.  a s/o just going to the band and tuning up or down would
> probably work lots of stations that had been spotted recently.
> 
> in general with the spot rates on some nodes running at 400-500 spots/hr it is
> likely that if you dial up or down any given band that lots of the stations you
> hear may have recently been spotted.  and probably a very large percentage of
> rare (contest wise) multipliers have been spotted very recently.  it would be
> good for the contest committee to publish their algorithm and results, while
> they may be good, and their results are final, there is always the chance that
> they missed something. it may also be that some operating methods may leave
> s/o's open for false accusations so they may want to point out those practices
> so they can be taken into account.
> 
> --
> David Robbins K1TTT
> e-mail: mailto://k1ttt@berkshire.net
> web: http://www.k1ttt.net
> AR-Cluster node: 145.69MHz or telnet://k1ttt.net
> 
> --
> CQ-Contest on WWW: http://lists.contesting.com/_cq-contest/
> Administrative requests: cq-contest-REQUEST at contesting.com

-- 
David Robbins K1TTT
e-mail: mailto://k1ttt@berkshire.net
web: http://www.k1ttt.net
AR-Cluster node: 145.69MHz or telnet://k1ttt.net


--
CQ-Contest on WWW: http://lists.contesting.com/_cq-contest/
Administrative requests: cq-contest-REQUEST at contesting.com