Hi Bob.
If you are willing to do the analysis, I can provide you the mdb, or plain
text file with all the callsings spotted and counts for each of them.
I don't think anyone is supposed to defend the RBN. It's already being
defended by all those who use it.
It would be a whole lot better to see how it can be improved.
You are right when you say the same error doesn't get repeated over and
over with human generated spots.
I suggested a web service to get feedback from users and ban bad calls on a
per-contest basis. But no one replied. So dunno to what point the
developers want to perfect it.
Anyways. I can live with the current error rate. I listen with my ears all
the time, and I will log a station only if I know who other station is.
When I know I've been wrongly spotted, I simply don't log the callers. You
realize that, because dupes start calling you in a row. It is their job to
make sure who I am and I won't dupe them since I will get NILs if I do.
73.
Martin, LU5DX
On Thu, Feb 21, 2013 at 10:36 AM, Bob Naumann <W5OV@w5ov.com> wrote:
> First of all, I suspect there are far more than 400 errors out of 1 million
> spots. The % is definitely far higher than .04%.
>
> I did my time with this error-prone RBN by clearing the bad spots from the
> bandmap again and again this past weekend.
>
> I have no desire to invest more time in it to count them up now. (I
> downloaded the files, and they're too big ~34meg total for both days).
>
> Second, why is everyone so defensive about this RBN?
>
> It makes many of the same errors over, and over, and over, and over! Why
> defend it and pretend that it does not?
>
> It makes many repeated errors - this is a fact. I lived it. I know.
>
> The packet spotting network quality was better and it had less errors when
> the packet system was driven by human input only. I lived that too and I
> know.
>
> Sure, there were spot errors, but they did not happen over and over and
> over
> for the whole contest!
>
> Were things missed that the skimmer/RBN system does not?
>
> Of course - and that's not the point.
>
> The benefit of the skimmer/RBN is obvious.
>
> The problem is that there's way too much bad stuff that comes along with
> the
> good.
>
> Put some energy into fixing it instead of minimizing the problems!
>
> Please!
>
> W5OV
>
> -----Original Message-----
> From: CQ-Contest [mailto:cq-contest-bounces@contesting.com] On Behalf Of
> Guy
> Olinger K2AV
> Sent: Wednesday, February 20, 2013 9:32 PM
> To: Michael Adams
> Cc: cq-contest@contesting.com
> Subject: Re: [CQ-Contest] Summarizing the Skimmer Accuracy Thread
>
> Hmm. A little attention to the math. Looking at the spot counts in their
> analysis app, there's something in the magnitude of a million spots over
> the weekend. 400 errors through the process out of 1,000,000 spots is
> 0.04%. Your city drinking water should be so good.
>
> And the RBN folks ARE actually working on ways to improve that 0.04%
>
> Leave 'em alone. They're doing great. And giving away the spots for free,
> a gift.
>
> 73,
> Guy, K2AV
>
> On Wed, Feb 20, 2013 at 9:09 PM, Michael Adams <mda@n1en.org> wrote:
>
> > On Wed, Feb 20, 2013 at 5:13 PM, <w5ov@w5ov.com> wrote:
> >
> > >
> > > If the 20 EK1LZ spots were the only errors, that would not be so bad -
> I
> > > guess... but that's not the case.
> > >
> >
> > It's probably worth noting that in all 20 cases, the EK1LZ spots appear
> to
> > have been unique, coming from only one skimmer. (Different skimmers at
> > different times, but only one at a time.)
> >
> > --
> > *Michael D. Adams* (N1EN)
> > Poquonock, Connecticut | mda@n1en.org
> > _______________________________________________
> > 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
>
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|