Short answer: In a WPX-style contest each new prefix is a multiplier, so the number of prefixes grows with every hour of operation. The curve is steep at the start and flattens later, because the easy prefixes come first. By looking at the growth by hour and by band in your statistics, you can see when new prefixes stopped coming, where you spent time without gaining multipliers and where to look for more next time. Check the current rules for how multipliers are counted in your contest.
Why prefixes behave like a growth curve
In CQ WPX-style scoring, a prefix is the part of a callsign that identifies a country or a territory in a particular way, such as the letters and the first digit. The rules explain how prefixes are formed, including how portable stations are treated, as described in CQ WPX prefixes explained. For a station, every prefix that has not been worked before counts as a multiplier, and in many versions of the contest a prefix is counted once for the whole contest, whatever the band. Check the current rules to see how your contest does it.
This produces a typical curve. In the first hours, almost every QSO is a new prefix, because the log is empty. As the contest goes on, more calls repeat prefixes that you already have, and the number of new prefixes per hour falls. After a day, you need to work rare prefixes or new areas to keep the curve moving. The shape of your curve tells you how the contest went.
How to get the curve from your log
You need the callsigns and the times of all QSOs. For each QSO, find the prefix, and for each prefix, note the first time that it appears. Then count the number of new prefixes per hour, and add them up to get the cumulative number. A spreadsheet can do this in a few minutes, and a statistics tool does it automatically. Contest Analyzer shows WPX prefixes among its full statistics, together with rates, QSOs by hour and band, countries, continents and CQ and ITU zones. You can see how the count of prefixes grew and where it stopped.
Look at the curve as a whole, then at the details. A smooth curve suggests steady work on the whole. A curve with steps means that the new prefixes came in groups, for example when you moved to a new band, or when a pile-up of stations from a rare area appeared. The article on multipliers per hour shows the general method, and this article applies it to prefixes.
The early hours
In the first hours, the curve is steep. Almost every QSO is a new prefix, and the multiplier count grows quickly. This is the time when the rate is often highest, and the temptation is to run and pay little attention to prefixes. Because the early prefixes are common ones, they come naturally. A good habit is to check, after the first few hours, how many prefixes you have and which common ones are missing. A quick look at the bandmap for stations that you have not worked is often enough to fill several gaps.
If your curve rises more slowly than in past years, check the rate and the bands. Perhaps you spent the first hours on a band where activity was low, or perhaps a technical problem cost an hour. The article on the first hour of a contest and the one on reading your statistics after the first 24 hours help to read the early part.
The flattening point
At some point the curve flattens. New prefixes still come, but only a few per hour. This is normal, and the interesting question is when it happened. If it flattened early, in the first third of the contest, you may have been working the same areas again and again. If it flattened late, you used the time well. Find the hour where the number of new prefixes per hour falls below a small number, and look at what you were doing: which band, run or search and pounce, which areas were active.
The flattening point also shows the value of search and pounce. In the tail of the contest, a run frequency gives you QSOs but few new prefixes, and a search of the bands gives you fewer QSOs but more multipliers. The article on run versus search and pounce describes the trade-off. Use your numbers to see how much you gained by switching, and when.
Prefixes by band
Depending on the rules, a prefix may count once per contest or once per band. If it counts once per band, the picture is different: each band has its own curve, and the sum of them is your multiplier count. In that case, look at each band separately. Which band added the most new prefixes? Which band gave nothing? A band that gave many QSOs but few new prefixes may be a good place to run but not a good place to look for multipliers. See band ratio in contest statistics for another way to look at the balance of bands.
If the prefix counts once for the whole contest, the band matters less, and the key question is which band gave you new prefixes in the late hours. Often the answer is a band with fewer stations, where each call is more likely to be from a new area. Check your own log for this pattern, and use it next year.
Rare and special prefixes
Some of the new prefixes in your log come from stations that use special calls, such as event stations, club calls or portable stations. These stations may create new prefixes that are not on any normal list, and they can be a good source of multipliers. Be careful with how you copy them. A special prefix is easy to bust, because you may not know that it exists. The articles on /P, /M, /MM and /AM in contests and callsign structure: prefix, digit and suffix help to read them.
After the contest, check the prefixes that appear only once in your log. These are the most likely to be errors. Contest Analyzer checks every callsign and every zone against official license databases and several independent reference sources and shows only the few QSOs worth a second look. It also lets you edit calls and download a corrected Cabrillo, so a busted prefix can be fixed before the log goes to the organisers. It works with Cabrillo from N1MM Logger+, DXLog and Win-Test, and the contests page shows the supported contests.
Comparing years
The most useful comparison is with your own earlier logs. Take the cumulative prefix curve for this year and for the previous year and put them side by side. Do they start in the same way? Where do they separate? A difference in the first hours may show a difference in the rate. A difference in the tail may show a difference in how you looked for multipliers. Year-on-year comparison with your own logs in Contest Analyzer gives you this view without extra work, and the article on year-on-year comparison of the same contest explains how to avoid unfair comparisons, for example when propagation was different.
Do not read too much into small differences. A contest has many variables, and a change of ten prefixes in a total of several hundred may be luck. Look for large and repeated patterns, and use them as ideas for your plan.
Planning from the curve
Once you know the shape of your curve, you can plan. Set a target for the number of prefixes after the first day, based on your past results and the conditions. Plan some time in the late hours for a systematic search, and decide in advance which bands you will use. Keep a list of the prefixes that you tend to miss, and look for them in the bandmap. And decide how much rate you are ready to give up for multipliers. The article on setting realistic contest goals from last year’s statistics shows how to turn numbers into goals.
Using the numbers for the next contest
Write down three numbers after each contest: the number of prefixes at the end of the first day, the number at the end of the contest and the hour in which the growth fell below a few new prefixes per hour. Over several years these numbers become a personal benchmark. When you see in the next contest that you are behind the benchmark after the first day, you know that you need to spend more time on search and pounce or on bands that you have neglected. When you are ahead, you can afford more running.
The benchmark must be adjusted for the conditions. A year with poor propagation will give lower numbers than a year with good propagation, and a weekend with less operating time will give lower numbers than a full one. Note the conditions together with the numbers, and compare like with like. The point of the benchmark is not to give you a score to beat, but to give you an early warning during the contest and a clear picture afterwards.
It also helps to look at the prefixes that you worked on only one band or only once. These are your most fragile multipliers. If any of them was a busted call, you lose the multiplier as well as the QSO, so these lines deserve a second look before the log is sent.
Related reading
The bottom line
Read your prefix curve by hour and band, find where it flattened and plan your search time and band choice for the next contest. Check the current rules for how prefixes are counted.
FAQ
Does a prefix count once per band or once per contest?
It depends on the contest and the category. Check the current rules.
Why does the curve flatten?
Because the common prefixes are worked first, and new ones are rarer later in the contest.
Is a prefix that appears only once an error?
Not always, but it is worth a second look, since busted calls often create unique prefixes.
Can I compare my curve with others?
Compare it with your own earlier logs first. Other stations have different conditions.
Does the statistics view show WPX prefixes?
Yes. WPX prefixes are among the statistics that Contest Analyzer provides.
