Yup. It is pretty easy to get rid of busted calls in the RBN.
The RBN should allow for feedback from humans, so that you could send the
information about a busted call, RBN would propagate it to the skimmer
servers connected and it would not get spotted again during that contest,
that should be all.
But you have to consider there should also be a way to "undo" that blocking
of a callsign. What if LU5DTU is spotted in place of LU5DX, it gets
reported as a busted call, later during that contest the real LU5DTU gets
on and no skimmer is spotting him...
This should be done in a "per contest basis", since what's a busted call
during say ARRL CW can be a good one during WPX CW and so on.
Vy 73.
Martin
On Wed, Feb 20, 2013 at 11:17 AM, Bob Naumann <W5OV@w5ov.com> wrote:
> K3TN said: "The accuracy of the spots coming from the RBN are generally
> higher than those coming from human beings."
>
> I don't know if that summary is correct and my gut feel is that it's not.
>
> What's the difference?
>
> Well, when a person makes a stupid spot error, someone else usually tells
> them so and they stop. They don't continually make the same mistake
> throughout the contest like the skimmer/RBN does.
>
> During a contest, the RBN continually and repeatedly, ad infinitum, makes
> the same errors over and over and over which fills the bandmaps with
> useless
> stuff.
>
> This volume of useless clutter never happened with human-driven spots.
>
> And, as I reported earlier, the habit of the RBN reporting the stations
> calling into W3LPL, W1MK, K1LZ and others is most annoying and again, it
> rarely happens with human spotters - unless someone is a new spotter and
> doesn't know better, but that doesn't last very long because someone tells
> that guy to stop it.
>
> In contrast with this skimmer/RBN thing, who do you tell to knock it off?
> Is there some way to fix this real time?
>
> The result is continual culling of the crap from bandmaps everywhere.
>
> I think this idea of having skimmers from all over the world linked like
> this is a big mistake.
>
> A local skimmer + the human packet network might be a better and less
> frustrating solution until the problems are worked out.
>
> de W5OV
>
> -----Original Message-----
> From: CQ-Contest [mailto:cq-contest-bounces@contesting.com] On Behalf Of
> jpescatore@aol.com
> Sent: Wednesday, February 20, 2013 4:42 AM
> To: CQ-Contest@CONTESTING.COM
> Subject: [CQ-Contest] Summarizing the Skimmer Accuracy Thread
>
>
> Skimmer accuracy had gotten much better, but it is certainly not perfect
> yet. The accuracy of the spots coming from the RBN are generally higher
> than
> those coming from human beings.
>
> There are techniques for maximizing the odds of getting spotted by a
> skimmer, and for minimizing the skimmer busting your call:
>
> Send CQ TEST and YOURCALL at the same speed
> If you send your call twice in a CQ, leave an extra half space between the
> 1st and 2nd instance of your call
> If you have a call that is often busted by humans, (ie, TN copied as G or
> C)
> use extra spacing to lower skimmer busts
>
> There are techniques for minimizing busted skimmer (or human) spots from
> showing up on your bandmap/logging software:
>
> Use software like VE7CC;s AR User to limit skimmers to those that are
> geographically meaningfully to you, and to spots that are meaningful to the
> contest.
> Use the settings in N1MM or other contest software (blacklisting/available
> mult/Q window filtering, etc) to minimize display of meaningless spots.
>
> When jumping on spots, throw in a little XIT dither of 50 -200 hz to avoid
> the zero beat lemmings leaping on spots problem.
> _______________________________________________
> CQ-Contest mailing list
> CQ-Contest@contesting.com
> http://lists.contesting.com/mailman/listinfo/cq-contest
>
> _______________________________________________
> CQ-Contest mailing list
> CQ-Contest@contesting.com
> http://lists.contesting.com/mailman/listinfo/cq-contest
>
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|