2021-02-22

Busting Calls: CQ WW 2020

Prior posts in this series:
Throughout this post, I apply the procedures developed in the second post above.

For the purpose of this post, only verified QSOs are counted.

Lowest Probability

I begin with an ordered list of the stations with the lowest probabilities of busting a call in 2020 CQ WW SSB.

2020 CQ WW SSB -- weighted mean values of $p_{bust}$ (all)
Position Call weighted mean $Q_v$ $B$
1 LU1FAM 0.0007 1,400 0
2 DQ5T 0.0009 1,005 0
3 DF2RG 0.0010 957 0
4 DK1KC 0.0011 861 0
5 ND9G 0.0012 773 0
6 CR2X 0.0014 2,175 2
7 VE4VT 0.0014 2,174 2
8 UA9BA 0.0014 1,400 1
9 G3VPW 0.0014 671 0
10 W1GD 0.0015 1,336 1

And for 2020 CQ WW CW:

2020 CQ WW CW -- weighted mean values of $p_{bust}$ (all)
Position Call weighted mean $Q_v$ $B$
1 DL4FN 0.0005 1,775 0
2 W2JU 0.0007 1,293 0
3 ES2MC 0.0008 1,166 0
4 LZ5EE 0.0008 1,164 0
5 K3PH 0.0008 2,411 1
6 K1ZZ 0.0008 3,586 2
7 OR2F 0.0009 2,347 1
8 AH6KO 0.0009 1,066 0
9 SQ8N 0.0009 1,061 0
10 WW4XX 0.0009 1,053 0

It is interesting to plot the aggregated probability function for $p_{bust}$, weighted by the number of verified QSOs, $Q_v$, for all stations:

In case it isn't clear, the location of the solid vertical lines represent the weighted means of the probability curves.

We can limit the analysis to calling stations (i.e., not the running station).

2020 CQ WW SSB -- weighted mean values of $p_{bust}$ (no-run)
Position Call weighted mean $Q_v$ $B$
1 DQ5T 0.0009 999 0
2 DF2RG 0.0011 858 0
3 DK1KC 0.0011 856 0
4 K3IE 0.0012 812 0
5 ND9G 0.0013 753 0
6 SP2GMA 0.0015 650 0
7 G3VPW 0.0015 636 0
8 W1GD 0.0016 1,277 1
9 N2RC 0.0016 608 0
10 WY3A 0.0016 1,236 1

2020 CQ WW CW -- weighted mean values of $p_{bust}$ (no-run)
Position Call weighted mean $Q_v$ $B$
1 YU5R 0.0004 2,171 0
2 DL4FN 0.0006 1,526 0
3 SP2LNW 0.0007 1,399 0
4 RU3A 0.0007 1,259 0
5 SE5E 0.0007 1,256 0
6 R5DT 0.0008 1,244 0
7 K1AR 0.0008 1,220 0
8 AA3B 0.0008 1,174 0
9 K3PH 0.0008 1,128 0
10 SQ8N 0.0010 970 0

And similarly for running stations.

2020 CQ WW SSB -- weighted mean values of $p_{bust}$ (run)
Position Call weighted mean $Q_v$ $B$
1 LU1FAM 0.0007 1,353 0
2 CR2X 0.0014 2,169 2
3 DK8AX 0.0018 546 0
4 VE4VT 0.0018 1,687 2
5 YT7AW 0.0019 524 0
6 IK8UND 0.0019 1,025 1
7 N4BP 0.0021 457 0
8 UA9BA 0.0021 936 1
9 9M6NA 0.0022 450 0
10 OR2F 0.0022 903 1

2020 CQ WW CW -- weighted mean values of $p_{bust}$ (run)
Position Call weighted mean $Q_v$ $B$
1 G4PVM 0.0009 1,021 0
2 ZR2A/4 0.0010 951 0
3 IT9SSI 0.0011 904 0
4 K1ZZ 0.0011 2,824 2
5 AH6KO 0.0011 869 0
6 EW8DX 0.0011 867 0
7 VR2CO 0.0012 822 0
8 ES2MC 0.0013 753 0
9 OR2F 0.0013 1,512 1
10 LZ1BP 0.0013 1,486 1

Congratulations to OR2F for appearing in both tables.

We can also look at the changes over the period from 2005 to 2020.

First for all QSOs:

Then for calling stations:

And for running stations:
I think it's also interesting to see who appears to have the lowest probability of busting  a call over an extended period. So, for the last ten years:

2011--2020 CQ WW SSB -- weighted mean values of $p_{bust}$ (all)
Position Call weighted mean $Q_v$ $B$
1 JM1NKT 0.0007 3,022 1
2 NW0M 0.0007 2,940 1
3 KV1J 0.0007 4,337 2
4 WA1ZYX 0.0007 1,335 0
5 K4RUM 0.0007 1,267 0
6 ES2MC 0.0008 3,697 2
7 IK4OMU 0.0008 1,136 0
8 NB3C 0.0009 1,066 0
9 K1RV 0.0009 4,340 3
10 LY3CY 0.0009 5,400 4

2011--2020 CQ WW CW -- weighted mean values of $p_{bust}$ (all)
Position Call weighted mean $Q_v$ $B$
1 AD1C 0.0002 4,283 0
2 NW0M 0.0003 3,190 0
3 WB4TDH 0.0004 5,339 1
4 HL1VAU 0.0004 5,225 1
5 K6WSC 0.0004 4,872 1
6 WW4XX 0.0004 2,005 0
7 JA1QOW 0.0005 4,437 1
8 KM3T 0.0005 1,826 0
9 EU4E 0.0005 3,715 1
10 LZ2SX 0.0006 1,606 0

A good argument can be made that a better measure of copying ability is to consider only run QSOs:

2011--2020 CQ WW SSB -- weighted mean values of $p_{bust}$ (run)
Position Call weighted mean $Q_v$ $B$
1 R7MM 0.0008 1,191 0
2 F4FTA 0.0009 1,105 0
3 ES2MC 0.0009 2,250 1
4 EI4DW 0.0012 830 0
5 DL3LAB 0.0012 806 0
6 CF7RR 0.0013 1,557 1
7 EE1A 0.0014 673 0
8 SP9XCN 0.0014 2,764 3
9 WL7BDO 0.0015 660 0
10 VA7RR 0.0016 3,773 5

2011--2020 CQ WW CW -- weighted mean values of $p_{bust}$ (run)
Position Call weighted mean $Q_v$ $B$
1 W3OA 0.0004 2,072 0
2 LY3CY 0.0006 1,654 0
3 RW5CW 0.0006 1,598 0
4 WB4TDH 0.0007 2,884 1
5 VX7SZ 0.0007 1,294 0
6 YU1RA 0.0008 2,558 1
7 KM3T 0.0008 1,186 0
8 DJ1YFK 0.0008 1,153 0
9 EU4E 0.0009 2,306 1
10 SM6FKF 0.0009 1,069 0

Highest Probability

We can also look at the calls associated with the highest probability of busting calls in either the forward or the reverse direction:

2020 SSB -- Most Busts
Position Call QSOs Busts % Busts
1 LZ9W 9,832 174 1.8
2 YT5A 9,413 132 1.4
3 II2S 6,955 131 1.9
4 ZF1A 5,172 126 2.4
5 RA5G 5,273 121 2.3
6 K1TTT 4,799 118 2.5
7 CQ8M 3,608 115 3.2
8 PR4T 4,690 112 2.4
9 EW5A 6,296 111 1.8
10 TM3R 4,684 106 2.3

2020 CW -- Most Busts
Position Call QSOs Busts % Busts
1 EA6FO 6,865 228 3.3
2 UZ2I 7,094 158 2.2
3 OZ5E 3,834 143 3.7
4 E7DX 6,693 132 2.0
5 YT5A 11,260 130 1.2
6 EF1A 4,375 128 2.9
7 UZ4E 2,308 125 5.4
8 PX2A 5,602 124 2.2
9 LN8W 8,994 123 1.4
10 K1TTT 5,558 119 2.1

2020 SSB -- Highest Percentage of Busts (≥100 QSOs)
Position Call QSOs Busts % Busts
1 SP9WZO 105 22 21.0
2 YC9MX 149 28 18.8
3 OM8VL 103 19 18.4
4 R9YBW 276 47 17.0
5 JS1BIB 121 18 14.9
6 R2BAA 143 20 14.0
7 YB2YEN 134 18 13.4
8 IS0AGY 198 26 13.1
9 PY2NRT 101 13 12.9
10 M7MRX 109 14 12.8

2020 CW -- Highest Percentage of Busts (≥100 QSOs)
Position Call QSOs Busts % Busts
1 W1HNZ 298 66 22.1
2 IK0YUO 112 21 18.8
3 N5ER 118 22 18.6
4 HG6E 133 24 18.0
5 DL7CO 128 23 18.0
6 DJ5UZ 119 20 16.8
7 I4LCK 192 31 16.1
8 LZ1BY 248 39 15.7
9 F5OKB 172 27 15.7
10 JM7SKE 154 23 14.9

2020 SSB -- Most Reverse Busts
Position Call QSOs Reverse Busts % Reverse Busts
1 DF0HQ 9,402 263 2.8
2 OM2Y 3,290 240 7.3
3 IR4X 2,157 195 9.0
4 TM3R 4,684 195 4.2
5 EA8RM 6,867 192 2.8
6 JH8YOH 1,568 185 11.8
7 EW5A 6,296 142 2.3
8 NN3W 3,086 139 4.5
9 IQ4FA 4,376 137 3.1
10 PX2A 6,022 130 2.2

2020 CW -- Most Reverse Busts
Position Call QSOs Reverse Busts % Reverse Busts
1 DR1A 7,885 308 3.9
2 UA4S 5,962 296 5.0
3 KP3DX 5,541 283 5.1
4 TA1C/2 1,513 274 18.1
5 SQ3PM 253 253 100.0
6 ES9C 8,107 251 3.1
7 YT5A 11,260 225 2.0
8 OK1MV 2,180 202 9.3
9 CR6K 8,650 192 2.2
10 9A3XV 3,316 190 5.7

2011--2020 SSB -- Most Busts
Position Call QSOs Busts % Busts
1 LZ9W 88,289 1,534 1.7
2 CN3A 80,599 1,393 1.7
3 A73A 61,796 1,222 2.0
4 PJ2T 66,003 1,214 1.8
5 OT5A 63,373 1,201 1.9
6 V26B 55,555 920 1.7
7 EF8R 54,369 875 1.6
8 HG7T 60,587 845 1.4
9 RT6A 46,042 792 1.7
10 ED1R 53,533 787 1.5

2011--2020 CW -- Most Busts
Position Call QSOs Busts % Busts
1 LZ9W 102,022 1,230 1.2
2 PJ2T 87,250 1,204 1.4
3 PV8ADI 7,851 1,118 14.2
4 PI4CC 47,401 908 1.9
5 ZW8T 12,253 881 7.2
6 D4C 58,867 801 1.4
7 RW0A 47,680 790 1.7
8 G3V 32,634 764 2.3
9 PJ4A 66,880 759 1.1
10 HG7T 66,589 739 1.1

2011--2020 SSB -- Highest Percentage of Busts (≥500 QSOs)
Position Call QSOs Busts % Busts
1 PV8ADI 796 127 16.0
2 K2JMY 1,802 287 15.9
3 EA7JQT 572 71 12.4
4 EA1HTF 1,389 153 11.0
5 OH1TS 561 57 10.2
6 EA4GWL 931 91 9.8
7 K8TS 615 59 9.6
8 LU4DJB 672 64 9.5
9 YB9KA 706 67 9.5
10 PU2TRX 803 76 9.5

2011--2020 CW -- Highest Percentage of Busts (≥500 QSOs)
Position Call QSOs Busts % Busts
1 4Z5FW 527 108 20.5
2 W2UDT 837 169 20.2
3 BD3MV 1,099 220 20.0
4 DJ5UZ 885 155 17.5
5 JA3AHY 554 92 16.6
6 WP3Y 570 93 16.3
7 LZ1BY 708 115 16.2
8 AE3D 1,016 161 15.8
9 YO7LYM 1,558 244 15.7
10 IK0YUO 670 104 15.5

2011--2020 SSB -- Most Reverse Busts
Position Call QSOs Reverse Busts % Reverse Busts
1 DF0HQ 81,900 2,403 2.9
2 JA3YBK 36,974 1,126 3.0
3 K3LR 66,312 1,054 1.6
4 TM3R 19,555 891 4.6
5 CN2R 51,062 877 1.7
6 CN3A 80,599 862 1.1
7 IK2YCW 26,491 857 3.2
8 W3LPL 51,773 801 1.5
9 HK1NA 40,363 788 2.0
10 HG1S 34,848 762 2.2

2011--2020 CW -- Most Reverse Busts
Position Call QSOs Reverse Busts % Reverse Busts
1 JS3CTQ 24,551 3,055 12.4
2 DF0HQ 79,932 2,817 3.5
3 ES9C 68,266 2,058 3.0
4 DR1A 55,033 1,588 2.9
5 W2FU 55,290 1,434 2.6
6 K3LR 70,578 1,369 1.9
7 HG7T 66,589 1,367 2.1
8 RM9A 54,157 1,194 2.2
9 EF8R 58,148 1,145 2.0
10 NR4M 50,909 1,143 2.2

2011--2020 SSB -- Highest Percentage of Reverse Busts (≥500 QSOs)
Position Call QSOs % Reverse Busts
1 CW90A 1,370 30.9
2 CE6VMO 593 21.2
3 BA8AG 752 16.4
4 OG60F 4,258 11.4
5 ZP6DYA 1,226 11.3
6 LU9DDJ 800 10.6
7 V84SCQ 806 10.5
8 BV55D 919 10.2
9 YB2IQ 1,052 10.2
10 JG3SVP 1,270 10.1

2011--2020 CW -- Highest Percentage of Reverse Busts (≥500 QSOs)
Position Call QSOs % Reverse Busts
1 G3RWF 1,459 64.8
2 YT65A 1,149 37.2
3 5K0A 1,853 32.1
4 PE75W 1,408 29.3
5 OG55W 3,265 28.6
6 TA1C/2 1,513 18.1
7 DP65HSC 516 16.9
8 5J1E 1,523 15.6
9 SB0A 1,102 15.5
10 LZ1195IR 576 14.6

In tables of reverse busts, one sometimes finds what seems like an unreasonable number of reverse busts (as, in the last table, for RZ3VO and G3RWF). This is generally caused by a discrepancy between the call actually sent by the listed station and the one recorded as being sent in at least some QSOs in the station's log.

2021-02-16

Continent-Based Analyses from 2020 CQ WW SSB and CQ WW CW logs

In addition to zone-based analyses, we can perform similar analyses based on continent rather than zone using the various public CQ WW logs (cq-ww-2005--2020-augmented.xz; see here for details of the augmented format) for the period from 2005 to 2020.

Continent Pairs

We start by looking at the number of QSOs for pairs of continents from the contests for 2020.

The procedure is simple. We consider only QSOs that meet the following criteria:
1. marked as "two-way" QSOs (i.e., both parties submitted a log containing the QSO);
2. no callsign or zone is bust by either party.

A counter is maintained for every possible pair of continents and the pertinent counter is incremented once for each distinct QSO between stations in those continents.

Separate figures are provided below for each band, led by a figure integrating QSOs on all bands. The figures are constructed in such a way as to show the results for both the SSB and CW contests on a single figure. (Any pair of continents with no QSOs that meet the above criteria appears in black on the figures.)

Continents and Distance

Below is a series of figures showing the distribution of distance for QSOs as a function of continent.

Each plot shows a colour-coded distribution of the distance of QSOs for each continent, with the data for SSB appearing above the data for CW within each continent.

For every half-QSO in a given continent, the distance of the QSO is calculated; in this way, the total  number of half-QSOs in bins of width 500 km is accumulated. Once all the QSOs for a particular mode have been binned in this manner, the distribution for each continent is normalised to total 100% and the result coded by colour and plotted. The mean distance for each continent and mode is denoted by a small white rectangle added to the underlying distance distribution. The 99% confidence range of the value of mean is marked by a small blue rectangle (typically entirely subsumed by the white rectangle). The median is marked with a vertical brown rectangle.

As usual, only QSOs for which logs have been provided by both parties, and which show no bust of either callsign or zone number are included. Bins coloured black are those for which no QSOs are present at the relevant distance.

The resulting plots are reproduced below.

Half-QSOs Per Continent, 2005 to 2020

A simple way to display the activity in the CQ WW contests is to count the number of half-QSOs in each continent (a single QSO contains two half-QSOs, so a single QSO may contain two different continents or the same continent twice). We count half QSOs, making sure to include each valid QSO only once (that is, if the same QSO appears in two submitted logs, it is counted only once).

If we do this for the entire contest without taking the individual bands into account, we obtain this figure:

The plot shows data for both SSB and CW contests over the period from 2005 to 2020. I include only QSOs for which both parties submitted a log and neither party bust either the zone or the call of the other party. The black triangles represent contests in which no half-QSOs were made from (or to) a particular continent. We can, of course, generate equivalent plots on a band-by-band basis:

As in prior years, the activity from EU so overwhelms these figures that in order to get a feel for the activity elsewhere, we need to move to a logarithmic scale:

Intra-Continental QSOs

We can also easily look at the percentage of QSOs that are between two stations on the same continent, and in particular between two EU stations:

So, for example, in CQ WW CW in 2020, somewhat more than 30% of all QSOs were within the same continent; about 27% of QSOs -- more than a quarter -- were between two European stations. Yet, judging from the name, which is "CQ World Wide DX Contest", this is supposed to be a world-wide DX contest, not a European QSO party.

Flogging a dead horse, on 160m nearly 70% of QSOs in this "world wide DX" contest were between two European entrants, even in the more DX-friendly mode. On SSB, 90%(!) of similar QSOs were between two European entrants.

2021-02-11

Zone-Based Analyses from 2020 CQ WW SSB and CQ WW CW logs

A huge number of analyses can be performed with the various public CQ WW logs (cq-ww-2005--2020-augmented.xz; see here for details of the augmented format) for the period from 2005 to 2020.

As usual, there follow a few analyses that interest me. There is, of course, plenty of scope to use the files for further analyses.

Below are some simple zone-based analyses from the logs.

Zones and Distance

As in prior years, we can examine the distribution of distance for QSOs as a function of zone.

Below is a series of figures showing this distribution integrated over all bands and, separately, band by band for the CQ WW SSB and CQ WW CW contests for 2020.

Each plot shows a colour-coded distribution of the distance of QSOs for each zone, with the data for SSB appearing above the data for CW within each zone.

For every half-QSO in a given zone, the distance of the QSO is calculated; in ths way, the total  number of half-QSOs in bins of width 500 km is accumulated. Once all the QSOs for a particular contest have been binned in this manner, the distribution for each zone is normalised to total 100% and the result coded by colour and plotted. The mean distance for each zone and mode is denoted by a small white rectangle added to the underlying distance distribution.

Only QSOs for which logs have been provided by both parties, and which show no bust of either callsign or zone number are included. Bins coloured black are those for which no QSOs are present at the relevant distance.

The resulting plots are reproduced below. I find that they display in a compact format a wealth of data that is informative and often unexpected.

Zone Pairs

As in prior years, We can examine the number of QSOs for pairs of zones from the 2020 contests using the augmented file.

The procedure is simple. We consider only QSOs that meet the following criteria:
1. marked as "two-way" QSOs (i.e., both parties submitted a log containing the QSO);
2. no callsign or zone is bust by either party.

A counter is maintained for every pair of zones (i.e., 1-1, 1-2, 1-3 ... 40-39, 40-40) and the pertinent counter is incremented once for each distinct QSO between stations in those zones.

Separate figures are provided for each band, led by a figure integrating QSOs on all bands. The figures are constructed in such a way as to show the results for both the SSB and CW contests on a single figure. (Any zone pair with no QSOs that meet the above criteria appears in black on the figures.)

It is clear from these figures, as from those for earlier years, that CQ WW is principally a contest for intra-EU QSOs, and secondarily one for QSOs between EU and the East Coast of North America. This format is undoubtedly popular, as CQ WW, in both its SSB and CW incarnations, would seem by any reasonable measure to be the most popular contest of the year. But one does wonder whether there isn't some other format that would strongly encourage participation from other parts of the world, instead of concentrating on these limited areas.

Non-Zero Zone Pairs

The activity between pairs of zones in the CW and SSB CQ WW contests over the period from 2005 to 2020 may be usefully summarised in a single figure:

There are 820 possible zone pairs: (z1, z1), (z1, z2) ... (z1, z40), (z2, z2), (z2, z3) ... (z39, z39), (z39, z40), (z40, z40). The above figure shows the number of different zone pairs actually present in the public logs, for each mode and for each year for which data are available, separated on a band-by-band basis and presented in the form of percentages of the maximum possible count (i.e., 820).

The top two lines require some additional explication: the line marked "MEAN" is the arithmetic mean of the results for the six separate bands for the relevant year and mode. The line marked "ANY" is also constructed from the data for the individual bands, but such that any give zone pair need be present on any one (or more, of course) of the individual bands to be included on the "ANY" line.

Half-QSOs Per Zone for CQ WW CW and SSB, 2005 to 2019

A simple way to display the activity in the CQ WW contests is to count the number of half-QSOs in each zone. Each valid QSO requires the exchange of two zones, so we simply count the total number of times that each zone appears, making sure to include each valid QSO only once.

If we do this for the entire contest without taking the individual bands into account, we obtain this figure:

The plot shows data for both SSB and CW contests over the period from 2005 to 2020. As in earlier posts, I include only QSOs for which both parties submitted a log and neither party bust either the zone or the call of the other party. The black triangles represent contests in which no half-QSOs were made from (or to) a particular zone. By far the most striking feature of this plot is the way in which activity in EU overwhelms that in the rest of the world.

We can, of course, generate equivalent plots on a band-by-band basis:

The activity from zones 14, 15 and 16 so overwhelms these figures that in order to get a feel for the activity elsewhere, we need to move to a logarithmic scale:

The figures speak for themselves.