Tuesday, December 6, 2011

Reducing the number of "false positive" CQ spots

Have you seen multiple stations spotted on the same frequency by the RBN, gone there, and discovered that only one of them is really there?  If so, you've encountered what may be the last major problem with CW Skimmer, the tendency to spot calling stations as if they were the one CQing.

There was a bug in versions of Skimmer prior to 1.71, and in Skimmer Server prior to version 1.30, which caused stations that had been calling CQ to continue to be spotted as such even if they had left their run frequency and begun to search and pounce.This was fixed in the current versions of the programs, and I think I have written to everyone still using an earlier version to request that they get current.

So, what else can you do to reduce your individual contribution to this problem?  The biggest thing is never to use "minimal" validation (set on the Telnet tab of either program.  Instead, use at least "normal" or, even better, "aggressive".  Each of these levels requires progressively more iterations of a callsign before it "believes" it, so if we use the higher levels the number of "false positive" CQ spots should be much reduced.

How about giving it a try this weekend?

73, Pete N4ZR 


  1. When I go to Aggressive during the week when most of the operating is casual, I get about 1/3 of the spots that I get on Normal. Consequently, I run Normal during the week and then switch to Aggressive during contest weekends. During contesting I think most stations call enough that the loss is modest and I don't want my call associated with a lot of errors.

    I use an Auto Hotkey script to switch.

    Wes, WZ7I

  2. Aggressive is good but not good enough for contests since it is quite common for the calling stations callsign to be sent 3 times (twice by the calling station and once by the CQ'ing station). CW Skimmer needs another validation level (contest mode?) between aggressive and paranoid that would require the callsign to be heard 4 times if in the master list, and 5 times if not in the list.

    One other improvement needed in CW Skimmer is the ability to copy varying speeds during a CQ. One station during CQWW, G1A, was calling CQ by sending 'G1A G1A TEST'. G1A was sent at about 20 WPM, but the word TEST was maybe 35 WPM. Skimmer could not copy the word TEST correctly and always mangled 'TEST' with the last G1A to get something like G1AEET. G1A was spotted 18 times on Dx Summit, but was only copied by 3 skimmers, K3LR, K8ND, & PA5KT.

    Mark, KB9AMG

  3. Both comments noted, and thanks!

    I know that Aggressive validation isn't a panacea, but it should be better than Normal in this regard. I am a bit concerned that the stepped up validation causes such a large drop in the raw number of spots, but Wes knows what he is talking about. Alex had told me to expect about the same number of spots, just delayed.

    I have just switched back to Normal until the weekend, and I have already noticed a big increase in the number of spots. Thanks, Wes!