10 Years of Contests and QSLs

Originally appeared in the Scuttlebutt, newsletter of the Yankee Clipper Contest Club, March 2003.

I moved back to Massachusetts in February 1993. My first log entries from the new station were in the ARRL DX Phone contest of that year. I think I was barefoot to a Hygain multi-band vertical strapped on to a chain link fence in the back yard. Station building commenced in May and the station was fully operational by CQ WW in October.

I decided that the move to the new QTH would also be a great time to convert to full computer logging. I purchased DX4WIN and it became the collection point for all QSOs made in and out of contests.

Of course, we all know that if you make lots of contacts, you will receive lots of QSL cards. I see no need to clutter the QSL bureau system with unwanted cards, so I decided I would only answer QSLs. This enabled me to conduct two long-term experiments: study error rates and study what could be achieved just from incoming cards. The tenth year anniversary seems like a suitable event to present some results.

I love contests. I am not much of a DXer, but I will chase the occasional big expedition and make random DX QSOs around the bands. I do operate lots and lots of contests; the major ones such as CQWW, smaller country contests, and even state QSO Parties. The table below lists the QSO totals by band and mode from March 1, 1993 to March 1, 2003.

I didn't really get into RTTY until 1999, but it offered another fun set of contests to play around in. The other column would be a few PSK31 QSOs and some 6 meter activity with a borrowed radio. I just never got interested in the WARC bands because there were no contests there!

99.9% of my QSL cards come in through the W5 QSL Bureau. At one point near the bottom of the sunspot cycle, bureau manager WF5E said I received more cards than anyone else in the bureau! He made me a separate "letter" and shipped cards to me by UPS on a quarterly basis. This still occurs and is a great way to lower the cost.

When a batch of QSLs was received, I would check them against the computer and mark them for printing a label. If a call could be matched with something in the log, I would note the error in the comments field. This was the main goal of my effort - to determine my call copying error rate. If I couldn't find anything close enough to match, I threw the card away. In retrospect, I wish I had kept statistics on these as well. I suspect many of them were people who confused my call with KZ5D.

The results of the QSL error study are presented below. These include both my operation and those of various guest ops. They correspond extremely well with the typical error range of the contest log checking (UBN or LCR) reports. This clearly demonstrates just how good the computerized log checking process has become.

Note: Since I was only looking at incoming cards where the other station had my call correct, the experiment is only measuring a portion of the possible errors. Errors made in the other station's log are not discovered.

Since I was already in a QSL counting mode, I looked at some other interesting statistics from of this process.

Did you know you could get to 5BDXCC if you work enough people and have enough patience? I am only 3 cards short on 80m. My DXCC totals just from cards received are shown below.

I don't receive many cards from the USA, so 5BWAS is not doing as well.

Another thing I have always wondered about is how long does it take before you have received all the cards you are going to get? The table below shows the QSOs made and corresponding QSL receipts for each calendar year.

The average QSL rate is steady around 14 to 15%. It drops in 2000 and beyond because there are still incoming cards for this period that have not been received yet. Does this constant represent an approximation of the new participants in contesting each year?

About 5000 to 8000 of the contacts each year were made by guest ops and are included in the above. There are 4000 QSOs made with the YCCC club call AJ1I in WPX CW contests that are not included.

So where do these cards come from? What countries are the most prolific QSLers? The answers, at least for my log, are below. These 33 countries represent over 87% of the total QSLs received.

Japan and Germany are expected, but Spain and Belgium were surprises. I think Belgium is helped by the 100% contest QSLing of ON4UN under all of his OT*T calls. Spain and Italy show the QSL rate that a country with lots of new hams can generate.

At the lower end of the list, we see countries that are less affluent or that don't find another USA QSL card that compelling.

Now we can determine the cost benefit ratio of preemptive QSLing. On one hand, you print out the labels and the QSL responsibility is dispatched in one shot. On the other, you waste about 85% of your QSLs (i.e., the other guy didn't want one).

Of course, it takes time to check the incoming cards against the log, but I find that a fun way to appreciate each card and to learn more about the kinds of errors I am making.

The numbers above also give a small indication of what the new ARRL Logbook of the World (LOTW) concept is going to provide. We will have access to a much higher level of instant confirmations, but with 1.5 to 2% (or higher!) error rates in the data.

I can't wait to see what the next 10 years will bring!

Copyright 2000-2008 by Randy Thompson, K5ZD
E-mail to k5zd@k5zd.com