[CQ-Contest] Summarizing the Skimmer Accuracy Thread

[Marijan Miletic, S56A]

HS0ZCW wrote:  Just so you know, I am breaking my face with a grin at all
the bad news re Skimmer...

Let me combine these two comments from my 1991 ARRL CW experience with over
200 QSO worked by robot.  I was #5 at Dayton and twice EU CW simulated
pileup champ with about 60% correct calls.  IBM AT with 12 MHz clock using
TI DSP board at 30 MHz beat me badly with 80 %.  I got humble and let it run
on the air.  No problem apart from the band opening with the ringing
signals.  I was better then knowing to EXPECT KC1XX, W3LPL, K3LR and all the
other big guns I worked million times.  It was not my ears and CW brain
behind it but PRIOR KNOWLEDGE.  My repeated CW robot efforts recently
managed Morse Runner in HST mode at 50 wpm and 3K points.  My score was 1K7.


Now I can use master.dta for additional verification on my GHz TB PC.  EK3LR
is not there but I might miss TO5X, HA30S! 

Beware that we catch Skimmer faults on the callsigns we mainly know but he
digs a lot of weak ones we might skip in time/space collision.  Skimmer does
it 24/7 with accuracy better than 1 %.  Digital modes entered 21st century
while hams are in the dark with SSB speech recognition despite limited
spelling vocabulary.  

Thanks to Tor, N4OGW for great work on SDR contesting!  ARRL & IARU:
Hamradio is the best scientific hobby of 21st century.

LP MMM S56A

P.S.  Bob, EK1LZ was spoted only 20 times!  I forgot to check OC Krassy.  I
am more on numbers than Latin - hi.

-----Original Message-----
From: w5ov at w5ov.com [mailto:w5ov at w5ov.com] 
Sent: Wednesday, February 20, 2013 10:22 PM
To: Marijan Miletic, S56A
Cc: cq-contest at contesting.com
Subject: Re: [CQ-Contest] Summarizing the Skimmer Accuracy Thread

Mario,

Your analysis is not correct. It has nothing to do with EK3LR spots as I
would not have seen them. Locally, this is blocked from the K3LR system and
we don't see them or send them out.

It is the many, many other errors that persisted throughout the contest of
calls being spotted wrong, over and over and over.

Some examples:
C6AGP was spotted as many other callsigns HA30S (H A Three Zero S) was
spotted as T30S and others N2NT was spotted as T2NT K1LZ was spotted as
EK1LZ et al ad infinitum

Yes, there were many others and they varied.

It became exhausting to clear the bandmap of all of these repeated errors
over and over and over again.

It got to the point that there seemed to be more busted spots than good
ones. At that point, the value of this over the old-fashioned(?) human
packet network really becomes a concern.

Skimmer is a good thing, but propagating these repeated errors all over the
globe makes little sense to me.

W5OV