Mmmm. Point.
Well, if you are the attacker, there are a plenty of ways you could rig the analog raffle. I don't see how the device the picks the winning ticket makes a difference.
- - - Updated - - -
Interesting analysis!
BTW, yes, I am generally having fun with this. I realize one definition of insanity is repeating a behavior and expecting a different result. I propose that a second definition could be, "enjoying raffle preparations".
Which brings us to your point: how ARE we going to draw the winners? That is to say, the winning tickets that aren't tragically stuck in some seam of the drum. :-(
Let me begin by saying that Jessica has just spent an hour analyzing the rotation of the drum and its behavior with respect to ticket-snagging and ticket-losing. Of the two, it is by far the ticket-losing behavior that I care about most. We just plain can't have the occasional ticket drifting through the hatch seam and fluttering down to the ground. In Jessica's tests, every few minutes a ticket or two would waft out of the drum's hatch. It's bad enough under controlled conditions in the laboratory that is our kitchen, where we can see the tickets fly out and grab them and stuff them back in. Imagine the same thing happening in a crowd of 50-100 people, plus add alcohol, plus add a touch of hysteria. This is how middle-aged middle-class people wind up meeting the police late at night. Worse... how are we to KNOW that we caught ALL tickets that fell out?!
Also... after the drum is loaded and zip-tied shut, we are going to let people come spin it pretty much as they wish. My 11-year old boy, John, is fascinated by the whole thing and I can guarantee that there will be a LOT of spinning and at varying speeds, too. He's studying centrifugal force. A practicum, if you will.
Jessica figured out how to tape the edges of the hatch so that the hatch would seal shut. Subsequent QA has demonstrated no leakage. Whew.
As for ticket-snagging, first we see a reduced incidence of ticket snagging due to the hatch tape upgrade. Second, we have developed a highly-customized ticket de-snagger. It's available to the military through a standard procurement contract and costs only $695. It's also available in a civilian version -- obviously less robust and unwise to use in a life threatening situation -- for about $3. It's civilian name is "that multi-tool screwdriver thingie that John won at Dave & Buster's". A little poking through the sides of the drum and any snagged tickets are freed. I think when we're doing the drawing we might also open the hatch and, under the camera's watchful gaze, allow a highly-trained Professional Ticket De-Snagger sweep the inside with his/her hand to jar loose any remaining snagged tickets. Plus a small drum rotation to move things around again.
Those of you looking for the point I promised four paragraphs ago may now wish to congratulate themselves for their patience. Here's how I think we'll do the drawing:
1. Pick the Grand Prize winning ticket. Probability 1/N where N=number of actual valid tickets put into the drum. EVERYONE has a shot at the Grand Prize. Ticket goes into an envelope. Envelope is held by an ex-Babson College Women's Basketball Team Captain, who is simultaneously gorgeous and frighteningly tough.
2. Pick the Second Prize winning ticket. Probability 1/(N-1). EVERYONE who DIDN'T win the Grand Prize has a shot at Second Prize. Ticket => Envelope. Envelope => Amazon.
3. Repeat for Third Prize.
4. Repeat for Fourth Prize.
5. Repeat for Fifth Prize.
6. Repeat for Sixth Prize.
7. Repeat to see which people are so drunk at the party that they've forgotten there are only 6 prizes.
Now, we ask our awesome Amazon to give us the Sixth Prize envelope.
We open the envelope, announce the winner. If the winner isn't in the room, we get on the speakerphone and call the winner. 50-100 semi-inebriated people compete with each other to congratulate the winner. We may or may not pretend to be calling from the Nobel Committee.
Repeat for Fifth, Fourth, Third, Second.
For Grand Prize, I simply rip up the envelope and award it to Bonnie.
All good? I'm definitely interested in proposed improvements.
Alan
Good question - it might "feel" unfair, but it wouldn't be unfair - it would just be the breaks. I'm not a stats person (can still hear my German stats professor in college spitting out the words "stochastic processes"), but it seems to me that rationally (and yeah I get it, we're not rational, and it's only wise-human to recognize that that's okay - that's why it's fun that Alan is taking this so seriously, and again, I just hope he's having fun with it - good deeds shouldn't be punished) these two situations are probabilistically equivalent and technically fair:
2000 tickets:
"feels fair"
1 in 2000 chance of getting Tesla Model S or
1 in 1999 chance of getting Prize 2
1 in 1998 chance of getting Prize 3
1 in 1997 chance of getting Prize 4
1 in 1996 chance of getting Prize 5
1995 in 1996 chance of getting nothing but the satisfaction of supporting an awesome cause, and all the fun of anticipating winning
"still fair, but feels (perhaps) unfair"
1 in 2000 chance of getting Tesla Model S or
1 in 1999 chance of getting Prize 2
1 in 1998 chance of getting Prize 3
1 in 1997 chance of getting Prize 4
1 in 1996 chance of getting Prize 5
1975 in 1996 chance of getting nothing but the satisfaction of supporting an awesome cause, and all the fun of anticipating winning
20 in 1996 chance of having your ticket stuck in the seams so you get nothing but the satisfaction of supporting an awesome cause, and all the fun of anticipating winning
Of course, if you break it down and consider all the permutations of what if they clear tickets out of seams between drawings or not, I still think the aggregated odds will work out as above. We will all be just as likely to be more or less selectable in any given drawing.
A more interesting question (just thinking through the emotional journey) to me is: which one will be drawn first? Personally, as much as we'd all be happy to win anything, would it be better to win something and then not be eligible for grand prize (unless we've bought multiple tickets, which is another darned good reason to do so)? or to find out the grand prize winner first and then after being vicariously happy for grand prize winner, keep up hopes for runner-up prizes?
@Pollux, I can't believe you know that eBay has an old ping pong raffle machine - I can't imagine auditing (and painting numbers on) slippery, bouncy balls - that would totally give me nightmares, so if you go that route - may divine forces of good bless you with many assistants!!
- - - Updated - - -
See, this is what I like: rather than throwing in the towel on Plan A, figure out how the heck to make it work! We've got our thought experiment: clearly, in a cement mixer, we could use the original stubs (in their envelopes), and get satisfying mixing and randomness. Now all we need to do is come up with a smaller cement mixer. I like it!
See previous post; The Wife solved the containment problem. We will be able to maintain fusion conditions indefinitely.
Duct tape. Inside seams and door. Test it. Should work. Keep the odds equal.
Tesla has to be drawn first for all tickets to have equal odds.
Plan A might work if you rent a large cement mixer. Probably not welcome in the house though.
- - - Updated - - -
I'm not bright enough for stats. I'm just determined that the tickets stay in the drum, we unstick 'em from the sides so that they're in the big pile in the middle, and we stick in a child's grubby, filthy paw to poke around at random and draw out a ticket or maybe a toad.
Simple question: How do the odds *not* change if your ticket is eliminated from the pool to begin with?
Why not just throw out half the tickets to start with -- according to you, the odds of not winning don't change. It makes no logical sense.
It's not being "advantaged" or "disadvantaged".. it's being completely eliminated from the pool to begin with!
- - - Updated - - -
So here are the actual odds:
Assume 20 tickets get "stuck".
Your chances of not being one of the stuck tickets is: 1980/2000: 99%
Your chances of being the winning ticket: 1/1980: 0.0505051%
So the chances of both of those happening: 0.05%
If zero tickets get stuck, your odds are 1/2000: 0.05%
Nevermind.