Re: I'll Bet

1

Play at the nontransitive dice table, and make sure the other player picks first.

horizontal rule
2

How does this work?

horizontal rule
3

Well, you're guaranteed to at least break even if you double your bet after every loss (or something- I forget if it varies with odds). Of course, you need an infinite bankroll for it to really work. So, maybe not so much. I once cleaned up at roulette, using a similar strategy, but we might have been using lame, non-standard odds- the only prizes were for the top few winners of the night with no conversion back to cash, so the hosts could afford to have games paying out more than they took in (basically, everyone got some seed money at the start).

Barring that- go for games where there's less pure chance and more strategy and bluff (poker- the less shared cards, the better, I say- goddamn Texas hold'em- harder to bluff).

horizontal rule
4

After all, unlike going to Vegas, we're not looking maximize our absolute winnings, just our winnings relative to the other participants. Any gamblers/gamers/math geeks with ideas?

What are the rules? Do you need to be one of the top X winners to get prizes? In that case you should know that, unlike Vegas, you need to be more worried about maximizing your upside than minimizing your downside risk. Assume that through random luck some people in the building are going to do reasonably well (for whatever definition of reasonably there is). That means that if you're competing with them there's no difference between being up 25% or down 100% at the end of the evening -- in either case you won't be in the top X.

My suggestion would be to find something that offers a 1/3 chance of trippling your money and bet all your money on it. If you win then having 3 times the money you started with should be good for a prize, if you lose, you lose, but you have a 1/3 chance of winning.

The law of large numbers is not your friend in this case.

The problem with this strategy is that it's no fun.

horizontal rule
5

The only way to beat the casino without cheating is to count cards at blackjack. There are some systems that are really simple to use provided the house isn't reshuffling all the time, using a huge number of shoes, or looking to throw out counters. I can't remember the exact details, but it's something along the lines of adding a point to a running count for every low card that comes up, subtracting a point for every 10 or higher, and then increasing your bets when the count reaches a certain threshold. A little internet research could probably get you a nice, simple system.

You could also do team play, if you wanted. It's probably totally unnecessary in this context (it's just to avoid getting kicked out), but it might be fun to collaborate with your friend, and it would probably make you feel more like spies. Ben Mezrich's Bringing Down the House is a terrible book, but covers this stuff pretty well and is a predictably quick read.

horizontal rule
6

It sounds like everybody gets a fixed set of X dollars at the beginning of the night and then people with money left over at a certain time can use that money in an auction for the Fabulous Prizes. So, it sounds like you must be in the top N% of winners to get a Fabulous Prize or else be able to beat up a number of people outside of the top N% and take away their money.

Sounds like we should go for the big payoff games instead of a slow and steady game like blackjack. What's the best big payoff game?

horizontal rule
7

Ben Mezrich's Bringing Down the House is a terrible book, but covers this stuff pretty well and is a predictably quick read.

I liked the Wired article discussing the same events, which I believe is by the same author, but was never tempted to read a book length account.

horizontal rule
8

7: yeah, I tore through it on a vacation where I'd finished the book I brought unexpectedly early. It was interesting enough, I guess, but also fairly ridiculous.

horizontal rule
9

Bringing Down the House is a terrible book if being judged as one would judge a normal book. If being judged as overnight international flight reading material, it is more than adequate.

horizontal rule
10

You should also ask around to see what the cutoff appears to be right before the auction. If you are missing the apparent cut-off, put all your money on "red" in roulette until you make the cutoff or lose all your money.

horizontal rule
11

What's the best big payoff game?

Totally slots.

horizontal rule
12

The best part was when, in trying to convey the sheer astounding amazingness of the facts he was about to relate, Mezrich used some variation on "despite having the fevered and verdant imagination of a fiction author, I could never have prepared myself for the following revelation..." Also, the other time when he did it, fifty or so pages later.

But we're getting off-topic! How best to defraud what I assume is a charity event... Hmm. Do you know what the chips they'll be using look like?

horizontal rule
13

Even better -- it's a "See! Our company is so much fun our employees are forced to want to hang out with each other on a Friday night!" event, not a charity event, so I feel I can totally check my scruples at the door. Defraud away!

horizontal rule
14

Can you pool your winnings to make it past the cutoff? (I assume you mayn't, but it might be possible to anyway.)

horizontal rule
15

For optimisation problems of this sort where terminal wealth is all that matters and there is no statistical edge, the optimal rule is:

1. If you are the richest person in the room, don't bet

2. If you are not the richest person in the room, then where X is the difference between your wealth and the wealth of the richest person in the room, bet Y=(X+1)/2 on black at roulette.

3. If Y is greater than your wealth, bet Y on black at roulette.

(for "roulette", substitute whichever game gives you the best odds in the room; this might be blackjack if you can be bothered playing basic strategy really well with a favourable set of rules, but now we are getting into the deepest tedium).

The idea is that you want to end up with $Z+1, where Z is the second place gambler. Any more, and you risked too much.

horizontal rule
16

btw, there is a lot of information in the Mezrich book which is purely and simply crap. Ed Thorp's "Beat the Dealer" has the massive advantage of being reconcilable to the published mathematical literature, and being a full set of instructions, without any black-box "and then they did something secret" bits.

horizontal rule
17

We might be able to pool our winnings. I'm not sure. Thanks, dsquared!

horizontal rule
18

Just keep doubling your bet until you win. The more times you lose, the more you'll end up winning. But you have to keep playing until you win.

horizontal rule
19

Just keep doubling your bet until you win. The more times you lose, the more you'll end up winning.

Actually this isn't true. Your total is the same (when you eventually win) no matter how many times you lose.

I guess the main difference between dsquared's advice and mine is that his would suggest waiting and placing no bets until near the end of the event and spending the intial time trying to get a sense of how much money other people have.

horizontal rule
20

Depending on the size of the crowd, you might get somewhere by trying to talk other gamblers (using whatever means seem necessary or appropriate) into gambling incautiously.

horizontal rule
21

The idea is that you want to end up with $Z+1, where Z is the second place gambler. Any more, and you risked too much.

Actually I'm not sure that I agree with this. I think it places too much emphasis on knowing how much money other people have.

Imagine that you have $80, the top person in the room has $140. You have a choice between betting $61 (dsquared) on black or $80 on black (my advice). If you bet $61 and lose you have $19 which probably isn't enough to get back into the running (you'd have to double your money 3 times to take the lead) vs $0 in my scenario but if you win, in my scenario you have some cushion in case you miss-counted and the leading person actually has $145.

Thinking about it, I might be inclined to the following: if Y is the value in dsquared's calculation (Y=(X+1)/2) and R is your total wealth, I might be inclined to bet Y + (R-Y)/2) on the game. That way if you lose, you will require 1 more doubling to get back into the lead (4 successfull bets rather than 3 in my example), but you hedge for imperfect information.

horizontal rule
22

Even if you can't pool your money, you can put it all down on "red" and she can put it all down on "black".

horizontal rule
23

That's actually the best thing I've heard yet. Do that twice, and you have a 50/50 chance of quadrupling your starting stake, which seems pretty likely to put you in the money at the end of the night.

horizontal rule
24

Many excellent strategies abound!

horizontal rule
25

I haven't seen any conclusive proof that they're excellent.

horizontal rule
26

She didn't say they abound here.

horizontal rule
27

Update: Thanks everyone! I won one of the Fabulous Prizes!

horizontal rule