[CQ-Contest] Uniques, Busted, Bads in CQWW

[kr2q at optimum.net]

Interesting discussion on the survey part II.  Some confusion is obvious.  Try this:

1.  All uniques are NOT busted (or bad) callsigns, though statistically, many are.

2.  Not all busted (or bad) calls are uniques.

3.  Some busted calls are uniques

4.  Some uniques are busted 

A unique in your log which is successfully cross-checked to another log (using established
parameters), will be identified in the LC as a busted / bad call and it will be treated as such
in terms of scoring.

Currently, uniques which are in your log which cannot be cross-checked, may or may not be
identified in LC as a busted / bad call, depending on the quality of the call (determined by LC
s/w).  EG: if you log F79XA23, it is clearly a unique (and busted) call.  There would be zero
chance of "cross-checking" that to a good call, but it still fails as a valid calls and is treated
as such.

If a Unique cannot be categorized as as "busted" call (we did not find in the other guy's log, same
date, ~time, same band, etc), you currently get to keep it.  We do not assume it is bad/busted, 
although statistically, it might be.

The bigger question, at least for me, is DOES IT MATTER?

Almost everybody who is in the top 10 or 20 of their (competitive) category works all
categories of calls (good, bad, unique, etc.).  And believe it or not, the percentage of 
calls that remain as uniques, per log is very, very small (usually under 1%).

The exception to rule includes guys who are "rare DX" (or zones).  If a lot of guys need
it for a new band/mode or, as they say in DXing parlance, an ATNO, then guys will turn on
their radio and work JUST HIM (her).  It is not impossible and DOES happen.  But those are 
outlier cases and easy to "figure out."

How likely is it for a guy in NA or EU to work a genuine unique - not very.  Is it possible - yes.

If a contest sponsor removes all uniques, ala WRTC at least one year....maybe more, then the
scores might be cleaner for the majority of entrants (certainly most of NA and EU, but not all).

This raises at least 2 issues:
1. Historically, this has never been done.
2. According to the survey, some guys, especially in North America, are going to feel cheated.

And an offshoot of that is that there is a 100% (certainty) chance that SOMEBODY is going
to produce a QSL card for a QSO that was removed as a Unique.  Very bad PR for the sponsor.

When I did an actual study of this in CQWW many years ago (pre-Randy Surveys), I found almost 
100% "no change" in order of finish.  It other words, it just didn't matter if Uniques were left in or 
taken out.  The one change I recall too place at something position of finish #156.  So the guy who 
finished 156 was now in position 157.  Yawn.

So, faced with a choice of removing uniques and creating "issues" with the entrants versus doing 
nothing and having essentially zero impact on order of finish, what would you, as the contest 
sponsor, do?  <grin>

The last piece not yet discussed are "excessive" uniques in a log.  That is either the sign of
a poor operator (new guy? language issue on SSB?) or intentional cheating (log padding).  
As we found out last year, there are other ways to pad your log without resorting to creating 
unique calls.  Excessive rate of uniques is definitely a flag for further scrutiny to the log.  
With SDR, it is easy to go back and figure out what happened and then take appropriate action.

Hope this helps,
de Doug KR2Q