[Skimmertalk] Bad Skimmer Spots

[Pete Smith]

There is a third possibility, but the problem is a little trickier than 
I had thought.  Here's an excerpt I just sent to someone on another list:

"As for the theoretical error rate, let's take an example.  Say you have 
N6NC CQing for an hour and being heard by 10 RBN Skimmers.  Each Skimmer 
will spot you correctly 5-6 times (every 10-12 minutes).  But if some 
local event - transmitter splatter, QRN, QRM (especially things like 
clicks from a loud adjacent station) causes a mis-spot, then whichever 
Skimmer that mis-spotted you will immediately send out a bad spot.

Say just one of the Skimmers listening to you sends one bad spot of you 
in that hour. That's 1 bad spot.  Meanwhile, the good spots are being 
filtered out as dupes by client software, so that clients are only 
receiving 5-6 good spots of you.  Holy cow, that's a 16-20 percent 
apparent error rate.

It's the aggregation, which is why the RBN has to come up with a 
solution.  This is a fairly tricky problem, because the system has a 
built-in preference for new spots, whether good or bad.  The bad spots 
can come at any time when you are CQing.  Meanwhile, the good spots will 
tend to cluster, because most people who spotted you correctly will have 
done so within a few seconds of your first CQ, and then at 10-12 minute 
intervals thereafter.  I don't know just how this one will be solved, 
but I think it can be.  The devil, as usual...

73, Pete N4ZR

The World Contest Station Database, updated daily at www.conteststations.com
The Reverse Beacon Network at http://reversebeacon.net, blog at reversebeacon.blogspot.com,
spots at telnet.reversebeacon.net, port 7000



On 2/3/2011 11:38 PM, Igor Sokolov wrote:
> Lee, I think it would be excellent if VE7CC client software could have
> filtering and validation facilities. There were so many broken spots during
> CQWW 160 that filled the entire band map and their percentage was reaching
> 80% on the last night of the contest.
> I see two methods of validating spots:
>   1) pass the spot if the call has been mentioned by 2 or 3 other skimmers
> 2) Pass the spot if it is in the master. dat file.
> Combination of these two would probably cut the number of broken spots to
> acceptable figures.
> I think telnet client software is the best place for filtering (better then
> filter in the logger or at the skimmer)
>
> 73, Igor UA9CDC
> ----- Original Message -----
> From: "Lee Sawkins"<ve7cc at shaw.ca>
> To:<Skimmertalk at contesting.com>
> Sent: Friday, February 04, 2011 3:12 AM
> Subject: [Skimmertalk] Bad Skimmer Spots
>
>
>> Here is a small subset of bad skimmer spots. I am looking at a full feed
>> from the RBN.
>>
>> These are only spots for USA calls that have not been issued.  Many times
>> the first letter of a call is missing.  Many of these are actually Italian
>> calls with the "I" missing.
>>
>> http://www.bcdxc.org/ve7cc/notissue.txt
>>
>> I will update the list from time to time.
>>
>> CC Clusters drop these bad calls.
>>
>> There are many other non USA calls that are wrong which also have the
>> first
>> letter missing. These are harder to find.
>>
>> Other bad calls occur when the last letter is missing or truncated or is
>> added
>> +in error.
>>
>> "E" and "T" seem to be the most often added bad letters.
>>
>> Lee
>> _______________________________________________
>> Skimmertalk mailing list
>> Skimmertalk at contesting.com
>> http://dayton.contesting.com/mailman/listinfo/skimmertalk
> _______________________________________________
> Skimmertalk mailing list
> Skimmertalk at contesting.com
> http://dayton.contesting.com/mailman/listinfo/skimmertalk
>
>