[Top] [All Lists]

Re: [CQ-Contest] Inverse Distance Based Scoring Proposal

To: cq-contest@contesting.com
Subject: Re: [CQ-Contest] Inverse Distance Based Scoring Proposal
From: David Gilbert <xdavid@cis-broadband.com>
Date: Sun, 26 Jun 2011 20:07:21 -0700
List-post: <cq-contest@contesting.com">mailto:cq-contest@contesting.com>
This is all pretty weird.  I expressed a short, simple opinion on a 
contesting topic (second post from the bottom) and now I'm being singled 
out (by a non-contester no less) for being a roadblock to its resolution.

Contrary to Ron's assertion, I very much enjoy trying to analyze things 
and Excel has been my friend and companion for at least twenty years.  
Back in 2008 when the sunspot cycle was flat and propagation could 
reasonably be expected to be the same from year to year, I generated a 
large spreadsheet that summed any number of CQWW logs from the same 
general location and created a chart that displayed the total number of 
contacts as a function of zone, band, and time of day.  I made it freely 
available to the public and at least one large contest station in the 
northeast found it useful at the time.

More currently, I am the author and custodian of a very large 
spreadsheet that generates a running 12 month competitive ranking of 
member scores for the Arizona Outlaws Contest Club.  It is a sortable 
online chart based on a WRTC-type scoring mechanism to display club 
member rankings for 16 different band and mode categories.  It is 
viewable by anyone at 
http://www.arizonaoutlaws.net/ladderresults.html.   Ron, if you send me 
the Cabrillo summaries for your contest activities over the past 12 
months I'll make a special run of the spreadsheet to show you where you 

To illustrate my concern about the utility of recrunching existing logs 
to see if the rule changes currently being discussed would affect 
geographic influences, let's look at the published data for the 2010 
ARRL DX CW contest.  All of the top ten finishers in SOAB HP operated 
roughly the same number of hours (I checked the 3830 archives) so there 
is not a huge reason to normalize the results.  K1RX (#2 from NH) and 
N2IC (#3 from NM) had roughly the same scores (4,984K and 4,758K 
respectively), but K1RX had significantly more QSOs (4,474 vs 3,796) and 
N2IC had significantly more multipliers (419 vs 377).   I wouldn't begin 
to try to speak for Steve, who is a far better contester than I would 
ever dream to be, but I'd bet that his operating strategy included 
trying optimize his country count to offset the greater difficulty in 
working tons of Europeans ... which would probably mean spending time on 
bands that didn't necessarily produce high run rates.  Take away the 
country multiplier as part of a switch to a grid based scoring system 
and I'll bet Steve's focus would change.

A similar situation exists between W9RE (#6 from IN) and WC1M (#7 from 
NH).  W9RE won that battle but take away the multiplier contribution and 
the reverse is true.

In each of those comparisons the impact of the multipliers on the 
DIFFERENCE in scores is about 8 to 10%.  That seems to me to be a pretty 
significant error factor if it disappears while someone tries to 
recalculate results based upon a totally different scoring mechanism.  
More importantly, knowing in advance that those multipliers wouldn't 
exist would very likely cause everyone involved to adjust their 
operating strategies to emphasize high volume, long distance bands more 
strongly.  Under the current rules the low bands are very important to 
generate country multipliers, but under a grid based scoring system it 
would probably be more important in general to stay with the higher 
frequency, more populated bands as long as they remained open.  To what 
effect that might alter any geographic influence is indeterminate using 
existing logs, in my opinion.

If I could think of a valid way to use existing logs to gain insight 
into the impact of a scoring change I would do so, but since I can't I 
guess I'll just have to wait until Ron does his analysis and reports 
back here with his results.  It should be interesting.

73 to all,
Dave   AB7E

p.s.   My comments on how any of this would affect the strategies of 
K1RX, N2IC, W9RE, and WC1M are purely conjecture on my part.  I have no 
idea to what degree a rule change would impact their logs or not.   
Which is exactly my point ... I don't think anybody does.

On 6/26/2011 9:55 AM, Ron Notarius W3WN wrote:
> While I may have been a touch heavy on the sarcasm, I stand by my comments.
> If we're going to seriously discuss whether or not a change to a distance
> based scoring system is a viable and "fairer" option, we should have some
> hard evidence that this has the potential to do so.  Not guess work.
> IMHO, the simplest and most effective way to do this is to take existing
> logs (with the permission of the operators, of course) and perform some
> statistical research on them.  Specifically, by taking those logs, assigning
> to each station its approximate grid square, and then calculating a new
> "alternative" score for the working contest station in question.
> Then, and only then, will we have a means of seeing how much of a difference
> that distance based scoring could have.  Only an apples-to-apples comparison
> can give us a handle as to how much of an impact -- if any -- it will have.
> Not guess work that "it ought to."
> Of course, any rules change will affect strategies in future contests, from
> determining where to go, how to deploy antennas and operators, what bands to
> work and when, and a host of other factors.  That goes without saying.  But
> to decline out-right to re-examine old results by pre-declaring it
> irrelevant based on what you THINK might happen AFTER the rules are changed?
> That's just silly.
> Are we actually going to overhaul the scoring system without have an actual
> clue as to what it's impact will be?
> And, just in case anyone is actually interested, I am intrigued by this
> idea.  It does appeal to me as making things a little fairer, and by giving
> some otherwise unavailable opportunities to those who have location
> challenges under the existing rules.  That said, I want to know more.
> Intriguing as the idea is, I'm curious as to whether or not it will actually
> have a significant impact.
> In short, I'm open to new ideas, and this one is interesting.  But will it
> actually work?  There's only two ways to find out.  With all due respect,
> AB7E does not seem to want to do any analysis, unless he has a better idea
> he hasn't mentioned yet.
> The other way?  Make the change and hope for the best?  What are we, the
> federal government?
> 73, ron w3wn
> ________________________________________
> From: David Gilbert [mailto:xdavid@cis-broadband.com]
> Sent: Saturday, June 25, 2011 4:46 PM
> To: Ron Notarius W3WN
> Cc: cq-contest@contesting.com
> Subject: Re: [CQ-Contest] Inverse Distance Based Scoring Proposal
> You are WAY over reacting to my comment, but I'll try to make this simple
> enough for you to understand.  Checking existing logs for the potential
> effect of changing the rules as dramatically as is being discussed here is
> like analyzing CQ WPX logs to look for regional imbalances in CQ WW
> results.  You might get an answer of sorts but you've added so many other
> variables that the result would have questionable validity.
> Rant on, Ron.
> 73,
> Dave   AB7E
> On 6/25/2011 11:51 AM, Ron Notarius W3WN wrote:
> Oh.  So instead of establishing a relative benchmark so that we can at least
> try to do an apples-to-apples all-else-being-equal comparison, we're just
> going to keep guessing and make assumptions that feel right.
> Sure.  That makes sense.  Why bother with research?
> Obviously, any significant change in the scoring system will cause
> adjustments in strategy.  I thought that was a "given."  And that's not the
> point.
> I THOUGHT that the question is or was:  Will switching to a grid based (ie
> distance based) scoring system significantly alter the scoring, to make the
> competition in different regions more "fair" or "balanced"?
> Right.  Why bother to find out?
> -----Original Message-----
> From: cq-contest-bounces@contesting.com
> [mailto:cq-contest-bounces@contesting.com] On Behalf Of David Gilbert
> Sent: Saturday, June 25, 2011 12:32 PM
> To: cq-contest@contesting.com
> Subject: Re: [CQ-Contest] Inverse Distance Based Scoring Proposal
> Personally, I don't think that would be a definitive analysis.  I'm
> pretty sure that contest strategies would change dramatically if the
> scoring system changed to a grid based system without country multipliers.
> Dave   AB7E
> On 6/25/2011 6:48 AM, Ron Notarius W3WN wrote:
> Suggestion:  Will distance-based scoring make a difference and alter the
> landscape?  There's one way to find out.  It shouldn't be too hard to
> match
> up calls worked with their Maidenhead grid squares (if not exactly then
> close enough).  Why not ask some of the top operators to "donate" their
> logs, then have a computer program run through and re-calculate their
> scores
> using the grids as multipliers, and see what you come up with?
> It's one thing to guess that it will make a difference.  This way, you can
> actually show what the results would have been, all else being equal, if
> you
> had scored that way.
> 73, ron w3wn
> _______
CQ-Contest mailing list

<Prev in Thread] Current Thread [Next in Thread>