ERRORS OF RUSSIAN DX CONTEST 2003 PARTICIPANTS

Errors in logs of the RUSSIAN DX CONTEST 2003 averaged 6.5% of the total number of QSOs. This is about the same as the last three years. However the ratio of the nature of the errors has changed:
 the number of “not in log” QSOs (NIL) has decreased to .5% (.7% in 2002)
 the number of wrong exchange (received) has decreased to 1.2% (1.4% in 2002)
 the number of incorrectly received callsigns has increased to 1.85% (1.6% in 2002)
The other errors are: incorrect time, band or mode  .1% of the total number of QSOs (systematic errors not included).
The UBN files, containing list of errors made, sent out in 2002 contributed to a noticeable decrease in errors in the logs of traditional contest participants.
The most frequent errors of callsigns and exchanges received could be divided in several distinctive groups. Nature of error, number of such errors in logs of all participants and most character examples are listed below.

1. Errors in receiving prefixes
 Errors in the first character of a prefix: RAUA, KNW, GM and similar (for instance, an error in first symbol Of Russian Stations RU and vise versa was registered 658 times). There are 1128 errors of this kind altogether.
Here are few examples:UR 272 UA1ANA  RA1ANA RU 386 RA3RIU  UA3RIU EU 38 EW1GA  UW1GA NK N2OW  K2OW JO JA3YPL  OA3YPL MG M0CSU  G0CSU
 Errors in the second symbol of a prefix: IKISIZ, JAJHJK, OKOM, RARKRZRURXRNRV, SPSQ, UNURUS, UTUY  Totaled 3340 errors. Examples:
ZK 42 RZ9MYL  RK9MYL, IZ8DBJ  IK8DBJ KZ 38 RK4HWW  RZ4HWW AU 90 RA4AO  RU4AO UA 48 RU0LL  RA0LL WA 74 RW9OS  RA9OS NR 34 UN7EED  UR7EED ST US7IGF  UT7IGF UT YU1EQ  YT1EQ
 Similar errors in 2 symbols (for instance, MM0GW0)  212 errors.
 Missed, or, on the contrary, added 2nd symbol of a prefix (For instance, I0IGV instead of IS0IGV) – 576 errors.
 Transposed symbols in "look alike" prefixes RUUR, LYYL, YUUY – 209 total, but part of them might be considered as typos. Examples:
YL 42 YL6W  LY6W LY 19 LY1YK  YL1YK
These occur in prefixes of countries where several prefixes are being used, and "similar" prefixes belonging to different countries. The frequency of such errors is practically same for SSB and CW.
2. Hearing errors, which are different on CW and SSB
On CW the most frequent errors are in receiving symbols which sound "similar" HS (696 total, 527 participants Received "S" instead of "H" and only 169 – "H" instead of "S". Same errors occurred in receiving 2letter Oblasts identifiers, but the frequency of those were lower because of the use of adapted RDXC software.
Examples of CW errors:
HS 527 RA4HT  RA4ST, HM  SM SH 169 ON7SA  ON7HA, ST – HT BD 148 ON4AEB  ON4AED, CB  CD DB 136 UA9CDC  UA9CBC, AD – AB SI 129 RU9SL  RU9IL, SV  IV IS 68 EI4DW  ES4DW, IR – SR UV 112 SU9NC  SV9NC, UD  VD VU 108 RV1AT  RU1AT, IV – IU IE 52 RA4AI  RA4AE, KI  KE EI 28 PI4DEC  PI4DIC, PE – PI XK 40 RX3RZ  RK3RZ, KX 33 OM3EK  OM3EX, KR – XR LR 40 HA0LP  HA0RP, UL  UR RL 25 UA0QBR  UA0QBL MN 29 PA0MIR  PA0NIR, HM  HN NM 25 SN7Q  SM7Q ZG 61 RZ3AZ  RZ3AG AT 46 RK3AWA  RK3AWT, BA  BT WM 38 N2OW  N2OM CK 32 RA1AC  RA1AK, CB  KB DN 31 OH5DX  OH5NX 5H 223 5B4AGC  HB4AGC, S59AA  SH9AA H5 25 UX3HA  UX35A B6 24 BW4/UA3VCS  6W4/UA3VCS
CW errors in numerals reception in call signs and exchange numbers were distributed quite evenly. The majority of errors were with the following numerals:
54 408 UN4L  UN5L, 154  144 45 299 UR5E  UR4E, 014  015
The majority of errors on CW occurred while receiving the last numeral of the number of the QSO.
Excessive number of symbols in call sign reception on CW – 291, at the end of call sign – 251. The majority of errors of this kind – an extra character "T" (55 occurances). Special attention should be paid for CW errors which occur because of sending a callsign and next code combination without an interval.
callsign + "TEST" 55 RK9AY  RK9AYT, EA6UN  EA6UNT callsign + "BK" 54 T92D  T92DB

Missed symbols on CW (not counting call signs, having “/” as a part of callsign):
 missed symbols in the middle of a call sign – 227.
 most frequently missed characters are: "E"  44, "I"  24.
 there are no missing letters "O" and "W" at all.
 missed symbols at the end of a call sign – 672.
 most frequent: "T"  66, "K"  32, "C"  31. It should be taken into consideration that only few callsigns end with "K". Most rarely missed symbols – "F" 3, "Y" – 1.
Missed numerals in exchange numbers – 232. Mostly it’s a last numeral in a 4 digit exchange. Back in 2002 there were 1.5 times more errors. Looks like the participants became used to the fact, that exchanges often exceed 3 digit numbers.
3. Typical errors on SSB are linked to accustomed interpretation of English phonetics by different participants, particularity the pronunciation of their callsigns by different participants, DVP performance. For instance, in many countries “Victor” is associated with letter “W”, “Canada” – with “K”, “Easy” – with “I”.
Participants should pay attention to these symbols. There are significantly less errors in reception of oblasts identifiers because of the contest software. Besides the alternate identifiers contributed to avoid errors (JA=YA, JN=YN, etc).
VW 342 RV0AU  RW0AU, HA8JV  HA8JW, VL  WL WV 108 RK6HWR  RK6HVR, EW7KR  EV7KR XZ 85 RZ1CXO  RZ1CZO, LX1NO  LZ1NO GJ 75 RG4A  RJ4A, 9A2GA  9A2JA JG 45 RA9MJ  RA9MG, HA8JV  HA8GV YJ 56 UA9AYA  UA9AJA, LY1YK  LJ1JK JY 25 RK9JWV  RK9YWV , IZ8DBJ  IZ8DBY CK 51 RA9CO  RA9KO, YL7C  YL7K KC 13 RA1QKB  RA1QCB, LZ1KSC  LZ1CSC OA 40 RK9CZO  RK9CZA, H2O  H2A AO 21 RU6FA  RU6FO, LA6HJA  LA6HJO MN 35 UA9AM  UA9AN, LY3UM  LY3UN NM 13 UA9HN  UA9HM, YL1XN  YL1XM EI 38 RA3EA  RA3IA, PI4DEC  PI4DIC 0Z 39 I0ZQB  IZ0QB
Here are the most frequent errors in reception of numerals in call signs and exchange numbers n SSB. According to the frequency of errors on SSB the first numeral in the exchange is most vulnerable.
23 135 RN2FB  RN3FB, DL2JRM  DL3JRM, 231 331 32 124 RA3AJ  RA2AJ, LY3BH  LY2BH, 063 – 062 42 110 9A4KA  9A2KA, UU4JWF  UU2JWF, 947  927 89 102 IZ8DBJ  IZ9DBJ, US8IM  US9IM, 681  691 43 87 RG4A  RG3A, YO4KBJ  YO3KBJ, 405  305 98 59 HG9M  HG8M, RA9FLW  RA8FLW, 1349  1348 19 52 YT1BB  YT9BB, 019  090 90 46 DH9SB  DH0SB, RZ9UGN  RZ0UGN
Excessive symbols on SSB: in the middle of callsign  29, in the end of callsign – 119. None of the symbols prevail. Missed symbols in callsigns on SSB: in the middle of callsign97, at the end – 210. The letter “A” was the one to disappear 23 times. The least frequent to disappear – "H" and "S" – twice each only.
4. "Keypad" Errors
The next group of errors are the typos and graphic errors of those who had handwritten logs or those who entered the log into a computer after the contest was over. Unfortunately it’s extremely difficult to qualify the nature of these errors, such as transposition of symbols in a call sign, errors in absolutely different Symbols.
For instance, symbols "T" and "Y" could be transposed because of similarity in handwritten logs or because they are adjacent on a keyboard. Very often errors in reception of numerals could be explained because the appropriate keys are adjacent on a keyboard.
Transposed symbols – 579. The occurrence is as more often as the key is closer to the center of a keyboard.
H 41 IK2HKT  IK2KHT, SP7LHX  SP7HLX, HM  MH J 34 SP9LJD  SP9JLD, UA9MJW  UA9MWJ K 34 SP8KEA  SP8EKA, I4MKW  I4KMW, KN  NK G 31 F5HGT  F5GHT F 30 UA9AFM  UA9AMF T 27 IK2HKT  IK2HTK L 24 UA9LCY  UA9CLY, UA9CLB  UA9CBL, KL  LK D 23 SP9KDA  SP9DKA, DL9HD  DL9DH
In received exchanges the adjacent numerals were transposed most often:
56 51 756  765 67 39 076  067 78 37 1871  1781
The number of typos and graphic errors increased at night time (local for each participant). It should be taken into consideration that handwritten logs, as a rule, are being written in evenings or nights.

5. Errors in receiving fractal callsigns:
 absence of a fractal part (RU3DX/3  RU3DX, W2/UR5DEM  UR5DEM)  219 errors
 incorrectly received fractal part (MM0DFV/P  MM0DFV/M, RV6LGY/9  RV6LGY/1)  106 errors
 missing symbol in fractal part (BW4/UA3VCS  BW3/UA3VCs, YU8/9X0A  YU/9X0A)  79 errors
 transposition of a fractal part (ZA/UT7DW  UT7DW/ZA, EA6/DL6KAC  DL6KAC/EA6)  41 errors
 more complex errors (HL1/WX8C  HL1GWX/8C, RN4ABJ/6  RN6/RN4ABJ)  52 errors
Considering that in addition to above mentioned errors, traditional errors can be added, but those participants can be considered “High Risk Group”. Nevertheless the constant number of different DXpeditions participating in RDXC is very welcome.
6. Errors in receiving of the Russian Oblasts Identifiers
Errors in receiving of the Russian Oblasts Identifiers by participants who were not familiar with the rules of the contest, or international system of callsigns, should form a separate group. In this case an Oblast Identifier was interpreted by them as number of QSO, for instance:
VO  040 NV  094 TV  004 TB  006 NN  099 TA  001 SV  054 TN  009 BA  061 HA  051
These errors are very difficult to systemize, and the total quantity of such is 71.
7. Intime error distribution during the Contest
In 2003 intime error distribution smoothed out noticeably and did not have any splashes neither at the beginning nor at the end of the 24hour period, as it happened in previous years.
Within the first hour of the test the number of errors were averaging 11%, regardless of a rate within the same period. Mostly those were errors in callsigns reception and “not in log”. Exchange numbers were being received correctly. If the participant began working not from the beginning of the test he did not have such an error growth.
By the 4th hour of the test the average percentage of errors decreased down to 4% and remained at the same level up to 02 UTC. The ratio among all kinds of errors was pretty much equal to average during the entire contest. A second splash of errors could be seen in logs of those who took lengthy timeoff. This continued up to 1 hour after resuming operation. The errors reached up to 14%, those were wrong callsign and exchange reception. If a participant operated without timeoff, the percentage of errors grew noticeably at 05 hours participant’s local time and often exceeded 1015%. This “splash” was the longest one and continued from 1 to 3 hours. The same was true for MOST, although the operators could change. At this period for some reason only exchange numbers were incorrectly received, everything was OK with reception of call signs. During the last hour of the test the percentage of errors had increased again. Incorrect reception of exchange numbers, including identifiers of Russian Oblasts prevailed all. The Number of NIL’s also increased.
8. Average percentage of errors in logs of participants of RDXC 2003
Number of errors on CW % of errors on CW 
1.6% 
1.1% 
0.47% 
Number of errors on SSB % of errors on SSB 
2.1% 
1.3% 
0.54% 
Total % of errors 
1.85% 
1.2% 
0.5% 
We would like participants to pay attention to the fact that the total number of CW errors seemed to be larger than that of SSB. The fact should be considered that in RDXC2003 more than 59.7% of QSOs were made on CW, the remaining 40.3%  on SSB. The following is the ratio of errors to all checked QSOs percentage wise for each mode and total.
Mismatches of time of QSO, Mode (only for MIXED Group) and Band (only in multiband Groups) constitute total of .12% for all logs.

9. Errors band distribution
 1.8 MHz  26% (majority – errors in call signs  17% and NILs  7%)
 3.5 MHz  19% (nature of errors is even)
 7 MHz  15% (nature of errors is even)
 14 MHz  12% (nature of errors is even)
 21 MHz  7% (majority NILs only  5% from 7%)
 28 MHz  21% (majority – incorrectly received exchange numbers  8% and NILs  11%. Looks like quick propagation change on this band was the contributing factor.)
Regardless of Mode and nature of errors they are distributed as follows:
The organizers of the Russian DX contest hope that these statistics will help the participants to improve their results in this Contest. We wish everybody success and good luck in RDXC2004! 73!
The analysis was prepared by RW1AC and UA1AAF