some day the coin gets stuck on edge and heads/tails both lose.

[jonathankaplan]

People die every second. We all are going to do it. A person died on Sunday about whom I'd like to say a few words and tell a story.
His name was Howard Ring. He was brilliant and sensitive, a truly caring man who positively affected almost everyone he met. He had a tough side (he was a top floor trader, after all) but still, I never saw him

Comments

[herooftheage]

If Howard was the one doing the flipping, and it was a fair coin, then the only way I can see for the Flipper to have gotten ahead was the coin-flip analog of wheel-clocking. If he had sharp enough eyesight to see the state of the coin at some useful place in the flip, I suppose he could have made a better than average prediction about how it lands.

[corwyn_ap]

Had a thought on this. Do most people flip the same way each time? If so you can get a good approximation by looking at which side is up before the coin is tossed, and then just add the adjustment.

[herooftheage]

I think that is fairly unlikely. Think about the myriad things you'd like to be able to do the same every time, and require immense practice to manage - swinging a golf club for example. Getting a random physical motion to be uniform doesn't seem to be what humans do very well.

I think even the good pre-computer wheel clockers didn't rely on the motion being the same. What they did (and may still do for all I know) was notice where there wheel was when the ball was released, apply a correction for the speed of the ball, and then try to place a bet in the right quadrant before the croupier announced "no more bets".

[michaelsullivan]

I disagree. I don't know what the odds given were, but if they were small but signfiicant, say 6-5 or so, it's conceivable that you could beat them this way. The adjustment wouldn't have to be perfect, you'd just have to be right more often than the odds require.

If I flip consistently enough that knowing the side up when I start will mean your prediction is correct 65% instead of 50%, and you can see the side I start with, 60% of the time, you will be right 59% of the time, and that's enough to turn those 6-5 odds around and be ahead roughly the equivalent of getting the 6-5 odds yourself.

Is that level of accuracy possible? I'll bet it is. The only problem is that Howard was a *very* astute gambler. He would almost certainly have been aware of the potential, and consciously varying the throw to make this more difficult. That makes me think the flipper must have been getting a picture in the air.

Michael

[corwyn_ap]

It does meet my requirements for betting too long, if you are up, on someone else's proposition, and they want to up the stakes, take the money and get out. Now.

[nokomisjeff]

One of my favorite coin flip proposition bets is to get someone to bet that on 10 tosses you will get exactly 5 heads, 5 tails. After all, it is a 50-50 that you will get heads or tails. I even offer them 3-1 odds. You would be suprised how many math people, traders, and gamblers fall for that one.

Aloha,

Jeff

[michaelsullivan]

At those odds, I think they're only giving you a 1.5% overlay. Still sad, but are you sure you mean 3-1 odds? It looks to me like you'll get 5 heads and 5 tails 24.6% of the time (10-choose-5 / (2^10)) with a fair coin , so 3-1 odds is close to an even bet.

Michael

[nokomisjeff]

Look at it like this need a bit of explanation

[michaelsullivan]

It's not 5-2 in your favor but just barely over 3-1 unless I'm missing something. How do you get 5-2? I get 24.6% which is about 3-1. As you surely know, seeing exactly 5 heads and 5 tails is the most common of your 10 results. The steps you have there should come up a number of times equal to their binomial coefficients, and the total will be 2^10 -- that's how I got my formula.

Michael

[strategery]

As a numismatist, I'm ashamed to admit that I don't know whether US coins show near a 50-50 rate.

I am out of ideas. I probably would have lost a lot more than Howard.

[chrishartman]

Flipping them by hand, you shouldn't have any effects from the details of the coin. (Incidentally, meta-logic could have told you that. Don't you think Howard would have noticed something was up if the Flipper just kept saying "heads"?)

However, there's a nice side bet you can do using pennies standing on edge. Set up 10 or so on a table, on their edges. Now hit the table (straight down, so you aren't biasing the direction) and count the number of heads and tails. Notice anything?

(Pennies are stamped out, the sides aren't precisely perpendicular to the faces.)

Hmmmm. I swear this used to work, and now I just tried it since I couldn't remember whether heads or tails was more likely. Three trials with eight pennies and no statistical significance. Maybe it was nickels?

[strategery]

Notice anything?

Nearly 100% heads in three trials with nine. I faced them differently on the desk on the third trial and still had 8/9. I have entirely too much time on my hands.

[herooftheage]

The flipper should not just say "heads". He should say heads often enough to make his bet a good one, but not so often that Howard would catch on. Now of course, Howard should catch on anyway, but it wouldn't be immediately obvious.

[herooftheage]

There is, of course, another possibility - that there was no trick. It could be the guy was getting the worst of it and still decided to gamble. When I was a serious blackjack player, I kept in my head the sober claim by Wong that you could play a lifetime of great blackjack and still have roughly a 20% chance of being minus at the end of it.

It might be that the flipper had some history with your guy, and was really trying to plant a logic bomb in his head. 6:5 odds on money you decided you could lose might be worth that.

Notice that upping the stakes has two possible interpretations: either he has an edge or he's trying a martingale scheme.

[corwyn_ap]

From the description it sounds like the flipper up the stakes the second time, when he (the flipper) was up.

[herooftheage]

Right. This scenario has him upping his bet when he's down to recoup, but then doing it again when he's up to try to make his killing. Betters at casinos do that sort of thing all the time. One of the reasons why casinos are so lavish - but part of why it is enticing is that you can see it work often enough to be mesmerized.

In fact, it is an interesting question how much it was worth to play with Howard rather than a casino. One of the house's advantages is that it is very rich compared to the player. If the Bellagio let you flip coins as an even bet, it would tend to have the best of it, because you'd ruin long before they did.

[jonathankaplan]

...In fact, it is an interesting question how much it was worth to play with Howard rather than a casino. One of the house's advantages is that it is very rich compared to the player. If the Bellagio let you flip coins as an even bet, it would tend to have the best of it, because you'd ruin long before they did.

Sounds right.
I'd guess this house advantage at pretty small though generally, and even smaller in this specific case. You could flip coins at Howard for 5K a pop for a few days before it would likely affect his wealth greatly, even with some (not overwhelming) edge in your favor.