Yet Another Nats Item: Stunt Circle Allocation

[Howard Rush]

Traditionally, Nats qualification rounds for both Open and Advanced stunt have been spread over four circles, the best five guys from each circle advancing to the next, “top 20” round.  Advanced attendance has been dwindling (see the first figure), and in 2012 we had the somewhat silly situation of only 23 (of 25 entered) Advanced contestants showing up to fly: 5 guys on one circle, 6 on each of the others.  Thus one circle went through the motions for two days for nothing, and the other circles labored for two days to eliminate one contestant each.  Steve Yampolski came up with an improved method, which we employed manually at the last two Nats, causing us to overstay our alotted time at the 180 building.  I have finally automated it and incorporated it into the Nats tabulation program.  Both the number of circles allocated to an event and the number of contestants advancing from the qualifications rounds are functions of the number of contestants in an event.  I fiddled with the inputs to Steve’s formula and came up with the numbers on the second figure.  This works out pretty nicely.  The number of contestants per circle stays between 5 and 8 for most any number of event entries less than 33.  At least half the qualifications-round entries advance to the next round.  Except for the case where there are 9 entries in an event, it avoids the situation where all but one contestant advances from some circles and all but two advance from other circles. 

This method preserves the existing Nats format.  If at least 25 guys show up in an event, that event will be flown on four circles.  If there are at least 33 guys in an event, 20 will advance from the qualifications rounds to the next round.  I hosed these numbers around to guys who have run the Nats and to other PAMPA luminaries and got no negative comments that I remember—I got hardly any comments at all, actually.

The program accommodates any combination of entries for Advanced and Open, allocates them to a number of circles appropriate for the entry level, optimally balances the circles, folds the seeds among the circles used for each event, checks score entries for errors, evaluates judges, and prints all the paperwork including flight schedules, scoresheets, pull test schedules, posters, and reports.  It does so as published in advance with no subjective nor arbitrary inputs.  I think it is cool, but it is an ambitious project for a flaky person such as the undersigned.  If you have Excel, please get a copy of the program and give it a test run to look for stupidity in operation or instructions. 

Everything in the Nats tabulation program is public.  Anybody who wants it can have a copy of the program.  I also made a couple of demonstrators that show how circle allocation, advancement to the next round, and seed folding works.  The files are too big to post here: I’ll send them as email attachments. 

[Derek Barry]

I will fiddle with it if you send it to me. I think I already have at one point because there was a cute little message when I put 50 people in Advanced and the now retired Expert.

Derek

[RC Storick]

here is a program www.random.org its free and easy.

[Howard Rush]

here is a program www.random.org its free and easy.

Thanks.  I've been using Randomize in VBA, followed by Rnd, to generate random numbers I think are actually random.  I've been caught by Excel's repeated "random" numbers before, and I think this gets around it.  I use random numbers in local contests to draw flight orders, but we draw pingpong balls at the Nats because some people don't trust a random number generator they can't see.  The Nats program uses no random numbers except for testing.

[Steve Fitton]

I for one would be okay with doing the drawing by computer.  Then the ping pong balls could go toward Wescott's pilot project.

[Tim Wescott]

I for one would be okay with doing the drawing by computer.  Then the ping pong balls could go toward Wescott's pilot project.

It is comforting to know that somewhere out there, there are people whom I've never met, that still have my back.

[Brett Buck]

Thanks.  I've been using Randomize in VBA, followed by Rnd, to generate random numbers I think are actually random.  I've been caught by Excel's repeated "random" numbers before, and I think this gets around it.  I use random numbers in local contests to draw flight orders, but we draw pingpong balls at the Nats because some people don't trust a random number generator they can't see.  The Nats program uses no random numbers except for testing.

  I think some of the distrust of "random" number is because they used the "scramble" for years, and it wasn't random at all. The rest of it is a problem in psychology, not math!

    Brett

[Howard Rush]

  I think some of the distrust of "random" number is because they used the "scramble" for years, and it wasn't random at all. The rest of it is a problem in psychology, not math!

As I understand the Betty Adamisin Scramble, the first round is drawn randomly, by pingpong balls if you wish, then the other rounds are mapped from that by a process that can be published in advance for all to see.  I thought that was a cool idea, but Paul nixed it because it still looks mysterious to some.

[BillLee]

Howard, the sun rising in the east and setting in the west is mysterious to some. 

[Dennis Vander Kuur]

"when I put 50 people in Advanced and the now retired Expert".

I knew there was some talk about this but didn't know if it was official. Has the "Expert" class in stunt been dropped from the NATs?
DennisV

[Brett Buck]

Howard, the sun rising in the east and setting in the west is mysterious to some. 

   That is an outstanding point - although I would use "claimed to be mysterious" rather than simply mysterious. If the results to not come out to the right people's preference, they will claim that the process was mysterious, obscure, etc, and thus leave the implication that it was rigged. You could count the decays of a lump of radium (a way to generate a *genuine* random number) and there would likely be someone willing to claim something was crooked about it.

   Paul and Howard have done a great job coming up with reasonable algorithms to perform most of the functions that were previously done by judgement. There wasn't anything wrong with it before, but no you can't complain about about Shareen and Warren any more.  However, that only solves a math problem, not the underlying psychological problem. As predicted, complaints have shifted from who is doing the job, to the perceived flaws in the algorithm, but haven't significantly diminished, particularly with the seeding and judge selection algorithm. Howard posts the results, sends the algorithms out to anyone who asks, and yet it is still subject to the same silly arguments and "I am offended that my hero didn't come out on top" posts every single year.

  I think we have made a lot of good decisions in stunt over the years, and a few whopping mistakes. One of the few whopping mistakes is being *far too concerned* over a few people with idle, innocent-but-baseless, and malicious complaints. I think we have *grossly* overreacted in some cases. Its magnified by the internet and the passion of a few, but many of these complaints are simply not relevant, and we shouldn't bend over backwards to try to accommodate them, at the cost of lots of extra work and pain.

   Brett

[Brett Buck]

"when I put 50 people in Advanced and the now retired Expert".

I knew there was some talk about this but didn't know if it was official. Has the "Expert" class in stunt been dropped from the NATs?

  Yes. It didn't accomplish the goal of attracting more entrants, it just seemed to split Advanced in half, at the cost of much additional work and a bit more cost. There are still events to enter for everyone who wants to come, so it didn't seem like there was any advantage. In the words of one notable personage, it was a "failed experiment". Worth a try, maybe, but it didn't work.

    Brett

[Tim Wescott]

  Yes. It didn't accomplish the goal of attracting more entrants, it just seemed to split Advanced in half, at the cost of much additional work and a bit more cost. There are still events to enter for everyone who wants to come, so it didn't seem like there was any advantage. In the words of one notable personage, it was a "failed experiment". Worth a try, maybe, but it didn't work.

Offhand, I can't think of anyone in the Pacific Northwest who consistently shows up for Expert with a not-pilot-built plane, and that's the only reason that I can think of for someone who flies Expert locally to want to fly Expert at the Nats.

Unless you want to drive half way across the country for a trophy in "not good enough to win Open", which boggles my mind because I'm averaging around 500 points in contests locally and if I could afford to go to the Nats it would be to enter Open, with the expectation of not making it into the top 20.

[john e. holliday]

Have you ever been to a NATS?   I miss going to the NATS, mainly for the commoradery.   What events I did fly was just for fun.

[Tim Wescott]

Have you ever been to a NATS?   I miss going to the NATS, mainly for the commoradery.   What events I did fly was just for fun.

No.  I got serious about competition in 2008, roughly at the same time that the @#$% hit the fan.  My accountant tells me that I did quite well compared to most small businesses since then, but that mostly means that I still own my house and no one here has starved to death.

Someday I may be able to afford it.

[Howard Rush]

You could afford it now if you hadn't spent all that TUT money on the Lear Jet, but how you allocate your income is not for me to say.