Reverse NILs in CQ WW: 2005

Using the augmented public logs of the CQ WW contests, one of the many analyses that can easily be performed is to see which entrants cause the largest number of NIL (Not-In-Log) strikes against stations who have logged them.

I should make it clear from the outset that NILs can and do occur naturally in contests. A typical instance would be when station A is CQing and is called by stations B and C simultaneously. Station A goes back to station B, but station C, for whatever reason, thinks that station A has come back to him. The common end result is that station A logs station B, but both stations B and C log station A. This will cause a NIL strike against station C -- this is a legitimate NIL strike, because station C never heard his call transmitted by station A, there was no QSO between those two stations, and station C should never have logged the non-existent QSO.

NILs, though, have occasionally been used as a weapon when a station deliberately does not log a perfectly valid QSO: this has no cost to the station that omits the QSO, but can cause a penalty (and perhaps loss of a multiplier or QSO points) against the station that he did not log.

Because of the way that scoring works in CQ WW, this behaviour is particularly pernicious in the case of intra-W QSOs (i.e., QSOs between two US stations). Intra-W QSOs are worth no points, but they do provide multipliers. As QSOs aren't worth any points, it is frowned on for an American station to make more than a handful of QSOs with other Ws, to obtain the necessary multipliers, as the calling station is generally wasting the time of the called station while he obtains the needed mult. The result of this is that if a W station does not log all W QSOs, he is likely depriving the other stations of necessary mults, as they will often not take out insurance QSOs to be sure of the mult, not wanting to waste the time of another station. Failure to log the QSO gives the station that does so a competitive advantage, as it therefore deprives other stations of a legitimately-worked mult, and causes penalties to be applied to the innocent parties -- at no cost to himself.

So I thought that it might be instructive to look at the public CQ WW logs to see which stations cause the largest numbers of NILs to be applied against stations who have claimed a QSO with them. (I call a NIL that is applied to the other station's log a reverse NIL, or rNIL.)

I made a couple of small adjustments to the basic idea outlined above:

1. I required a minimum of 50 appearances in other stations' logs for a station to be included in the analysis (25 appearances in the case of purely intra-W QSOs);
2. I removed all stations for which the analysis showed a 100% rNIL rate: this rate almost certainly occurred because of some basic error in the station's log, such as submitting a log that used a different callsign from the one actually used in the contest.

The code to perform the analysis is available here (feel free to let me know of bugs :-)). Below are the results of running this code against the logs for 2005.

2005 SSB:

Callsign	Total rQSOs	Total rNILs
HK3JJH	1820	1790
M7Z	1242	1166
EA1BVP	1957	972
IR8P	934	927
UU7J	4518	837
SP8IMG	1103	834
OR5N	678	668
JA6GCE	770	660
IZ8DPL	678	643
RU6MM	613	609

Callsign	Total rQSOs	Total rNILs	% rNILs
IT9RBW	383	382	99.7
RU6MM	613	609	99.3
IR8P	934	927	99.3
DL5MK	132	131	99.2
W7QDM	114	113	99.1
TA0U	393	389	99.0
YB0AI	563	555	98.6
OR5N	678	668	98.5
FR1HZ	581	572	98.5
9A4RV	190	187	98.4

Callsign	Total rQSOs with Ws	rNILs against Ws
KC1XX	139	22
N2IC	91	17
N3RS	67	15
NQ4I	247	14
W4MYA	84	12
W0AIH	136	11
W3LPL	183	11
WB9Z	38	11
K3LR	226	10
W7WA	30	10

Callsign	Total rQSOs with Ws	Total rNILs against Ws	% rNILs against Ws
W7WA	30	10	33.3
K5NZ	25	8	32.0
WB9Z	38	11	28.9
NE3F	25	6	24.0
N3RS	67	15	22.4
N2IC	91	17	18.7
KC1XX	139	22	15.8
W9RE	32	5	15.6
W6YI	26	4	15.4
W4MYA	84	12	14.3

2005 CW:

Callsign	Total rQSOs	Total rNILs
J88DR	1734	1706
SM7YEA	1496	1488
OJ0B	2126	824
ZY7C	2316	715
UN6T	674	672
RK9CR	666	631
W3BGN	1648	611
Z37M	3983	605
KT1V	2557	550
S54A	973	545

Callsign	Total rQSOs	Total rNILs	% rNILs
UN6T	674	672	99.7
W4ZW	329	328	99.7
SM7YEA	1496	1488	99.5
PV8AZ	82	81	98.8
S52P	389	384	98.7
PY8MGB	457	451	98.7
SP9UOP	204	201	98.5
J88DR	1734	1706	98.4
W1WFZ	176	172	97.7
K1OZ	66	64	97.0

Callsign	Total rQSOs with Ws	rNILs against Ws
W5UN	34	27
K9NS	181	25
NQ4I	229	12
K3LR	170	11
K1TTT	131	7
K1RX	125	7
KT1V	37	7
K5GO	171	6
K1AR	55	6
W5KFT	66	6

Callsign	Total rQSOs with Ws	Total rNILs against Ws	% rNILs against Ws
W5UN	34	27	79.4
K2QMF	25	5	20.0
KT1V	37	7	18.9
W3BGN	25	4	16.0
K9NS	181	25	13.8
K0EU	26	3	11.5
W2RE	27	3	11.1
K1AR	55	6	10.9
K2LE	47	5	10.6
K4XS	40	4	10.0

A word about these old logs: they contain many formatting and other errors compared to modern logs (for example, the "zones" field might contain serial numbers, even though the correct zone was actually transmitted), so many of the results should not be taken at face value. For example, the second and sixth tables above indicate an unrealistic percentage of rNILs from all the stations in the table. On the other hand, the large difference between W5UN and all the other lines in the final table suggests that one might want to investigate the cause, if one were interested. Generally speaking, and especially in these old, low-quality logs, it's a good idea to focus on what appear to be anomalous stations, rather than attaching too much credence to the raw numbers in the tables.

(If one suspected a particular station of purposefully failing to log QSOs, several statistical approaches suggest themselves. One presumes that nowadays the contest committee would quickly resolve the issue by requesting the required audio recording of the contest from the suspected station.)