Jump to content


Photo

Classifier HHFs about to get real


  • Please log in to reply
287 replies to this topic

#281 GuanoLoco

GuanoLoco

    Minister of Culture

  • Members
  • PipPipPip
  • 4,245 posts
  • LocationMy Safe Space

Posted 02 May 2018 - 09:19 PM

Software, it’s a thing now, or so I hear.
  • slemmo, ralloway, TCB and 1 other like this
Are you now, or have you ever been a member of the Doodie Project?

You'll never find a more wretched hive of scum and villainy.

In Doodie, Veritas.

"You might be a little thin-skinned for this forum". -Scott.

#282 GuanoLoco

GuanoLoco

    Minister of Culture

  • Members
  • PipPipPip
  • 4,245 posts
  • LocationMy Safe Space

Posted 03 May 2018 - 03:19 PM

I'm not being at all dismissive of the will or effort, but IMHO this should be mostly handled in software, even on the initial pass.
An 'exceptions' filter could be developed and used to identify (and optionally remove) anomalous scores in the database. The same filter could be used to reject such scores in the future BEFORE they are put in the database.
Once you have plausibly clean data to work with, programmatic identification of the 100% HHF based on the average of the highest 10 scores/classifier (or whatever criteria) isn't hard.
I would love to see the recorded scores for a Classifier graphed as a distribution curve. By comparing distribution graphs you would quickly identify which classifiers have abnormal distributions relative to the others. A real stats guy could probably assess/automate this far better than I (yea I took a 400-level stats class but it wasnt my thing).
Anomalous Classifiers could be further assessed and deprecated/replaced as desired.


I might have mentioned the need for automation about 10 months back...but what do I know.
  • Sweet T and ralloway like this
Are you now, or have you ever been a member of the Doodie Project?

You'll never find a more wretched hive of scum and villainy.

In Doodie, Veritas.

"You might be a little thin-skinned for this forum". -Scott.

#283 snark

snark

    Aberrant Freak

  • Members
  • PipPipPip
  • 1,097 posts
  • LocationNull Island

Posted 05 May 2018 - 08:16 AM

That is the stupidest bullshit I have ever seen written, and untrue.  I've never awarded a gun for any class win, just order of finish.  I don't care who you are, when you walk, what you get, just that we can explain how we do it.  I have a few hundred witnesses on this one.  

 

You should probably go back more than one page before getting triggered by something posted on Doodie although that would take the fun out of it.

 

For you new whippersnappers, minimizing sandbagging was the entire reason behind the 'no more than one lower' classification thing.  It was added to the system when it became too big a problem to ignore (ie: somebody sandbagged the wrong people).  Then a couple years later they had to start saving classifications for expired shooters so they couldn't drop membership for a year and get a reset (another sandbagger won the C-limited gun at Nats)

 

That shit was at least a dozen years ago.  "Only one class lower" was a hot discussion topic ... in 2001.


Use the rules. Don't DQ someone who doesn't do anything DQable. -- Da Beard.

 


#284 Mike Foley

Mike Foley

    Advanced Member

  • Members
  • PipPipPip
  • 907 posts

Posted 05 May 2018 - 02:36 PM

😂

#285 Cajun cracker

Cajun cracker

    Advanced Member

  • Members
  • PipPipPip
  • 296 posts
  • LocationLa.

Posted 03 June 2018 - 04:04 AM

Logic tells me that even those guys can't nail the exact same hit factor every single time on demand. Those guys are good, but they're not that good. I've shot with most of them and seen them make mistakes.



#286 Cajun cracker

Cajun cracker

    Advanced Member

  • Members
  • PipPipPip
  • 296 posts
  • LocationLa.

Posted 03 June 2018 - 04:05 AM

Easy, one of em might get wind of this blasphemy you speak of.... 😝
Maybe just take the shooters on all top 20 list and bump em every year? That would ruffle some feathers. If they worked on their shooting like the do the sandbagging. Smh seems like a lot of work to stay .2% from the bump year after year. That would be some funny shit. You email would BLOW the hell up...

#287 jakemartens

jakemartens

    Advanced Member

  • Members
  • PipPipPip
  • 2,351 posts
  • LocationIndianapolis, IN

Posted 03 June 2018 - 09:13 AM

 Jake Martens hates this method, because it is a lot of work and takes some time after the match, but we continue to do it.  Everyone gets something, appropriate to their finish.  

Actually what I said was "I hate prize tables", doesn't matter the method


  • Ben likes this

#288 foofighter

foofighter

    Advanced Member

  • Members
  • PipPipPip
  • 1,388 posts

Posted 04 June 2018 - 12:40 AM

The outsourced match deal is dead.  Multi-gun is the last one to have this arrangement, and I'm not doing it outsourced again if I can help it.

 

oh shit. MG nats is going to start sucking ass again?






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users