ERRORS OF RUSSIAN DX CONTEST 2004 PARTICIPANTS

The RDXC 2004 results are finished now. As usual, we are trying to summarize and compare these year changes.

We have checked 1 503 211 QSO altogether, 1 446 326 QSO (96.2%) of them were verified, i.e. they do not contain errors.

non-confirmed (NIL) QSO number reduced - 0.4% (0.5% in 2003);
incorrect control errors increased - 1.6% (1.2% in 2003);
incorrect callsign errors is the same - 1.6% (1.6% in 2003).

The rest of the errors, such as incorrect time, band or mode, are 0.22% from the total QSO number (systematic errors excluded).

It is necessary to note that systematic errors were quite many this year. About 50 logs contained local times.

The second prevailing error was incorrect time. It could differ from correct time up to 5-20 minutes over the whole log.

Some 20 logs were incomplete, but single-band only.

Those hams, who participate in the RDXC many years, have certainly decreased their log errors level. Actually, after studying their UBN, their log errors level goes down to 20-80%.

It is interesting; - many club stations have actually the same number of penalty points every year. Most likely, this is caused by large number of the team operators and by their different operating skills. The number of NIL QSOs of club stations in S&P mode with run op blocking facility has increased.

It is significant that Russian calls errors (bad call) decreased among Russian participants. By the way, penalties to Russian and foreign participants in summary tables are different. This is caused by different scoring principles. Russian participants get maximum 5 points for QSO and 15 points penalty as a result. On the contrary, foreign participants get maximum 10 points for QSO and 30 points penalty accordingly. This proves that foreign stations have to pay rapt attention to the correct Russian callsigns logging.

One can determine a few typical groups for the most often callsign errors and control number errors. For all errors, error type, number of such errors in all logs, including SWL logs and typical examples are listed.

1. Prefix reception errors.
This occurs often with countries that have similar prefixes and with similar prefixes for different countries. You can see such errors equally often both in SSB and CW.

The first prefix symbol errors: RA-UA, K-N-W, G-M and the similar. Total 624 errors (1128 errors in 2003). Examples:
```
U-R	111	UA9ACJ - RA9ACJ
R-U	154	RA1ACJ - UA1ACJ
N-K	27	N4AF - K4AF
M-G	17	MU3GSY - GU3GSY
```
The second prefix symbol errors: IK-IS-IZ, JA-JE-JH-JK, OK-OM, RA-RK-RZ, RA-RN-RU-RV-RW-RX-RZ, SP-SQ, UN-UR-US, UT-UY � total 1616 errors. Examples:
```
Z-K	69	RZ9MYL - RK9MYL
K-Z	53	RK4HWW - RZ4HWW
A-U	48	RA4AO - RU4AO
U-A	47	RU0LL - RA0LL
W-A	47	RW9OS - RA9OS
N-R	23	UN2E - UR2E
U-T	23	YU1EQ - YT1EQ
J-E	19	JA7NVF - JE7NVF
N-T	15	UN1O - UT1O
S-T	12	UR4III - US4III
T-U	11	YT7WA - YU7WA
```
The reason of these errors, in most cases, is operator�s reception and logging carelessness.
Similar errors in two symbols at once (e.g. EA1AES -YU1AES) - 161 errors.
Missed or added second symbol (e.g. I3OWC instead of IV3OWC) � 470 errors.
Mixed symbols in similar prefixes RU-UR, LY-YL, YU-UY - 277 errors, though, some of these errors occur due to �keyboard� errors. Examples:
```
LY-YL		51    LY6A - YL6A
RU-UR	39    RU3QR - UR3QR
YL-LY		34    YL1ZS - LY1ZS
```
It is necessary to say that LY and YL stations suffer most of all from such errors every year.

Auditory errors that have different nature in CW and SSB.

Similarly sounding symbols reception was the most often error: H-S (651, 449 logged "S" instead of "H", and only 202 logged "H" instead of "S"). Same errors occurred in 2-symbol oblast ID reception, but their rate was lesser due to special software adapted for RDXC. Some CW auditory errors examples:

H-S    357    HA0HH - HA0SS (22 times), ZL4HH - ZL4SH, HM - SM
S-H    182    HA9HU - HA9SU, RW9OS - RW9OH (12 times), ST - HT

B-D    207    RK3BA - RK3DA (12 times), CB - CD
D-B    188    9A6DM - 9A6BM, UA9CDV - UA9CBV (9 times), AD - AB

U-V    166    RU3AA - RV3AA (10 times), UD - VD
V-U    153    EA5FV - EA5FU (10 times), NV - NU

S-I    117    RN4SS - RN4SI, RN4IS (10 times), SV - IV
I-S     54    IV3OWC - SV3OWC (4 times), IR - SR

U-A     95    S50U - S50A, UD - AD
A-U     62    YO6ADW - YO6UHW, MA - MU

W-A     65    YU1WC - YU1AC
A-W     50    W8CAR - W8CWR, KA - KW

I-E     61    YZ1ZI - YZ1ZE, KI - KE
E-I     31    YU1EL - YU1IB, PE - PI

Z-G     60    Z35M - G3HM
G-Z     42    WX4G - WX4Z, VG - VZ

L-R     57    YU1EL - YU1ER, UL - UR
R-L     33    YU1RA - YU1LA, RA - LA

X-K     55    UX1UX - UK1UX,
K-X     42    RX6LDK - RX6LDX, KO - XO

A-D     50    W8CAR - W8CWR, MA - DA
C-K     45    YU1ACR - YU1AKR, CU - KU
M-N     45    Z35M - Z35N, MU - NU
F-R     44    UT7QF - UT7QR
A-T     42    WB2AA - WB2TA, SA - ST
N-M     42    YU1BN - YU1BM, NN - NM
R-A     42    UT1IR - UT1IA, KR - KA
W-M     42    YV5OHW - YV5OHM
D-N     41    W7DRA/7 - W7NRA/7, UD - UN
K-A     41    YL2HK - YL2HA, KA - AA
F-U     35    YU1DF - YU1DU
R-L     33    YU1RA - YU1LA, RA - LA
R-W     33    WA9IRV - WA9IWV, VR - VW
G-M     32    VE4MG - VE4MM, VG - VM
A-I     29    YL2PA - YL2PI, MA - MI
O-M     29    YU1YO - YU1YM, OM - OO
J-O     28    YO7VJ - YO7VO,
P-W     27    YT1NP - YT1NW, PE - WE
M-O     24    YU1LM - YU1LO, MO - OO
P-J     24    YL2PJ - YL2JJ, PM - JM
W-J     24    WW2Y - WJ2Y
X-B     22    UY0IX - UY0IB
Y-C     22    WW2Y - WW2C, YA - CA
A-R     20    VE2XAA - VE2XRA, AR - AA

Those stations that have double numbers in the call, especially if they are located together, suffered most of all from �number � letter� type errors. �Extra Risk� group are A61AR, T96Q, Z35M, and stations with stroke in the call (5B/RA6LUX, BW4/UA3VCS). However, this did not prevent them from reaching high scores. Examples:
```
5-H     50    5B/RA6LUX, Z35M - Z3HM
6-B     40    A61AR - AB1AR, W6VUN - WB4UN,
4-V     22    P40A - PV0A
```

CW error in logging numbers and control numbers distributed quite evenly. Certainly, most often �neighboring� number was mixed up. The biggest error value in number reception:

2-1    459    YO2RR - YO1RR, 299 - 199
1-2    318    ZL1TM - ZL2TM, 041 - 042

3-4    408    YP3A - YP4A, 138 - 148
4-3    317    UZ4E - UZ3E, 004 - 003

2-3    397    YL2PQ - YL3PQ, 238 - 338
3-2    318    YP3A - YP2A, 301 - 201

7-8    313    YU7FU - YU8FU, 367 - 368
8-7    414    YO8MI - YO7MI, 084 - 074

8-9    293    YL8M - YL9M, 080 - 090
9-8    304    YI9ZF - YI8ZF, 119 - 118

5-4    281    YO5CBX - YO4CBX, 715 - 714
4-5    308    ZL4HH - ZL5HH, 304 - 305

0-1    258    YU0OTC - YU1OTC, 370 - 371
1-0    433    UX1VT - UX0VT, 321 - 320

7-6    232    YU7OP - YU6OP, 700 - 600
6-7    239    YL6W - YL7Y, 065 - 075

5-6    230    YO5CBX - YO6CBX, 540 - 640
6-5    227    UY6IM - UY5IM, 006 - 005

0-9    226    YL0A - YL9A, 1204 - 1294
9-0    224    WA9IRV - WA0IRV, 091 - 001

In CW, the biggest error value was during the last symbol of the QSO number reception. Redundant symbols in CW � 571, including in the end of the callsign - 396. The biggest value of such error falls on logging letter �T� in the end of the callsign (78 times). Special attention should be paid to CW errors caused by joint callsign and the following code transmission:
```
Callsign + "TEST"   78     RK9AY - RK9AYT (6 times),
                           RA0AY - RA0AYT (5 times),
                           HA0LP - HA0LPT (3 times)
Callsign + "R"      29     YT1BB - YT1BBR
Callsign + "BK"     25     LY7A - LY7AB
Callsign + "K"      23     UA9QM - UA9QMK (3 times)
```
Missed symbols in CW (excluding calls with stroke):
Missed symbols in the middle of the call - 227.
- �� "E" - 58 times,
- �� "T" - 51 times,
- �� "I" - 31 times.
- �� "J" - 2 times.
Missed symbols in the end of the call - 594.
- �� "T" - 53 times,
- �� "K" - 28 times (it should be noted that very few calls are ended with "K").
- �� "F" - 2 times.
Missed figures in control number - 311. This, in the most cases, is not the last figure reception in 4-digit numbers. Every year these errors decrease. This means, permanent RDXC participants get used to control numbers over 1000 in this contest.

3. SSB errors associated both with operator�s personal symbol interpretation of English phonetics and their specific pronunciation, and with recording devices quality.

Again, like in the last year, letters V - W became �error heroes�. Main error sources are Russian callsigns that include letter V! Letter combination G-J became active again, last year these errors were twice lesser.

  V-W   321  	ES5TV "suffered" the most from this error (ES5TW - 26 times)
  W-V   150  	most "suffered" IV3OWC (IV3OVC - 32 times)
  G-J   148    	most "suffered" JL8GFB (JL8JFB - 22 times)
  X-Z    96    	here, our sympathies go to RX3ZX, whose call was 15 times logged as RZ3ZX
  A-D    56   	OZ1ADL - OZ1DDL (33 times)
  C-K    51   	YO3CZW - YO3KZW (16 times)
  O-A    50   	UA6HON - UA6HAN (10 times)
  U-A    49   	YU1EQ - YA1EQ
  J-G    47   	UN7JX - UN7GX (3 times)
  Y-J    45   	YT1BB - JT1BB (6 times)
  V-Z    42   	RV9AZ - RZ9AZ (6 times)
  X-E    32   	RK2FXB - RK2FEB (3 times)

Oblast ID errors were much lesser due to special RDXC software. The reason is also, that �difficult� symbols are not used in oblast IDs. There are no letters "F", "Q", "W", 'X", "Z", starting 2004 there is no letter "J".

Besides that, alternative IDs helped to prevent errors (JA = YR, JN = YN, YA = SH, etc) and were credited as well.

It was ascertained that no test software databases contain SSB error calls made in the contest. Therefore, SSB errors were not provoked by the software �prompts�. If different operators from different countries receive any callsign many times similarly incorrect, the participant should find the reason in his operation style. Mistake may be caused by both specific symbol incorrect pronunciation, and equipment improper operation. This should be remembered that new callsigns that never participated in the RDXC before might cause errors more often especially if they include �problem� letters.

Below are the most often errors in SSB number reception:

  2-3     70	"record" went to PS2F, who was logged as PS3T 14 times.
  4-3     47   	G4FKA - G3FKA (4 times)
  8-9     45   	EX8AA - EX9AA (5 times)
  3-2     42   	YP3A - YP2A (3 times)
  1-9     39   	HA1AG - HA9AG (8 times)
  3-4     38   	RA3RHA - RA4RHA (3 times)
  6-3     36   	GX6UT - GX3UT (5 times)
  4-2     33   	G4FKA - G3FKA (4 times)
  9-0     33   	RZ9OQ - RZ0OQ (5 times)
  4-5     32   	IK4QJM - IK5QJM (4 times)

Most probably, they are �keyboard� errors, as numbers are differently pronounced but located close to each other on the keyboard. Besides that, errors "8-9" are pronounced different, but Russian operators mix them most often, and just like this: "8" instead of "9", but not vice versa.

There are not too many excess symbols in the middle and in the end of the call in SSB callsign errors. There are many missed symbol errors in the middle of the call (157), and in the end of the call (290). This year we have extensive growth of such errors: there were 97 last compared to 210 in 2004. Most often letter �A� disappeared in the end of the call � 23 times, least often � letters "H" and "S" � 2 times each:
```
A     30        RZ3TZA - RZ3TZ (4 ����)
F     17        UN7LFF - UN7LF (4 ����)
M     15        UA9CN - UA9C (5 ���)
```
The first figure in control number is suffered most of all in SSB error rate.

Symbols interchange - 675 errors (579 in 2003).

Interchange errors happen the more often the closer the button is to the center of the keyboard:

 O-0   	44     	N8OO - N800,

 Y-U   	36     	SP5VYI - SP5VUI, UA9AYA - UA9AUA, YN - UH,
 U-Y   	27     	LU5FC - LY5FC, CU - CY,

 K-L   	29     	DL1EKO - DL1ELO, KR - LR,
 L-K   	25     	IK2LTR - IK2KTR, LO - KO,

 F-D   	30     	UA3QDF - UA3QDD,
 Y-T   	24     	LY2TA - LY2YA, YR - YT

We�d like to draw your attention to the last additional symbol errors (over 100) caused by accidental pressing on the neighbor button. These errors are annoying, it is very simple to correct them, - just check your log through before mailing it. Example: RU4WA/A - RU54WA/A

5. Callsigns with stroke
Callsigns with stroke errors may be allocated separately.

No stroke copied (EA5BWR/M -EA5BWR, RW1QM/3 - RW1QM) - 240 errors
Incorrect copied fraction part of the call (BW4/UA3VCS - BW3/UA3VCS, I1/IT9LNH - I2/IT9LNH) - 75 errors
Errors and missed symbols in the fraction part (BW4/UA3VCS - BW/UA3VCS, BW3/UA3VCS) - 81 errors
Fraction part shift (HL1/WX8C - WX8C/HL1) - 36 errors
More complex errors (RZ3FA/3 - RZ3FAX/3, UA9CE/6 - UA9CEB, BW4/UA3VCS - BW4S) - 32 erros

It�s pity, these errors did not decrease. Stations with stroke in the call should try to get their calls copied correctly.

6. Russian oblasts ID errors decreased, there were 54 errors (71 in 2003). Number instead of letter errors remained mainly on top bands. Typical errors:

VO - 040
NV - 094
NN - 099
SH - 055

7. Time error distribution during the contest.
Time error distribution in 2004 does not differ much from the last year. 28 MHz propagation was poor, last years this band was a source of many errors and NILs in the beginning and in the end of the contest.

During the first contest hour average errors rate was about 9% (11% in 2003). There were mainly non-confirmed QSOs (NILs). We are pleased that the error percentage in callsign reception in the beginning of the contest went down to 1.5 times. Numbers were copied correctly. Average error rate decreased down to 4-5%, and was permanent until 00 UTC. Error types ratio corresponded with the average Contest error rate.

As usual, incorrect callsign logging error rate of stations operated all 24 hours in the Contest, increased up to 7-8% in the second part of the night. This applies to the club stations also. Error rate of those participants having rest break increased after this break. Usually this happens in the morning local time and lasts 1-3 hours. Error rate is up to 14%, - mostly calls and control numbers were logged wrong.

As in 2003, the last contest hour error rate increased. This enables us to recommend all participants do not try to increase QSO rate in the beginning of the contest until all participants get used to operate in the contest. After break, please remember of possible errors and to operate the first hours more concentrated. Pay attention to the control numbers logging during the last contest hour.

8. Average error rate in RDXC 2004 logs

	Incorrect callsign	Incorrect serial number	Not in LOG
CW Errors % CW	8869 1.5%	8425 1.7%	3240 0.4%
SSB Errors % SSB	6769 1.8%	4858 1.4%	2027 0.4%
Total errors Total %	15638 1.6%	13283 1.6%	5267 0.4%

Ratio percentage of callsign errors and NIL errors decreased both in CW and in SSB. Control number errors were 4000 errors more than in 2003.

CW errors increased again in RDXC-2004, they were more than 61.9%, and 38.1% were SSB errors (CW/59,7 % and SSB/40 ,3 % in 2003).

QSO time, mode (for MIX entry) mismatch and band (for multiband entries) mismatch were totally 0,22% in all logs.

9. Band error distribution.

1.8 �� - 26% (prevail callsign errors - 17% and NILs - 7%)
3.5 �� - 20% (even error type)
7 �� - 22% (even error type)
14 �� - 16% (even error type)
21 �� - 12% (prevail callsign errors - 6% and NILs - 4%, control number errors - 2%)
28 �� - 4% (callsign errors and NILs � even error ratio)
Total - 100%

We understand, there is no way to avoid errors in the Contest. The one makes no mistakes who does not operate on the air.

With letters H-S in CW, and with letters V-W in SSB.
Pay your attention to the callsigns with strokes.
Do not be in a hurry at the first and the last contest hours, do not try to make fast QSOs.
Correct control numbers logging.
When operating top bands, especially 1.8 MHz, keep in mind that many newcomers are there.

The RUSSIAN DX CONTEST sponsors hope that our statistics analysis will help all participants to improve their scores in our Contest.

We wish all of you good luck in RDXC-2005! 73!

by RW1AC and UA1AAF