In Theory

[stavros]

I could "theoretically" win into infinity by winning just 16% of my placed bets if the wins and losses "clustered" in the right way. So, I ask, what exactly constitutes the "house edge" that we are constantly confronted with? Does it simply mean that we can't, over the long term win more bets than we lose on the even chances?

[rjeaton1]

I could "theoretically" win into infinity by winning just 16% of my placed bets if the wins and losses "clustered" in the right way. So, I ask, what exactly constitutes the "house edge" that we are constantly confronted with? Does it simply mean that we can't, over the long term win more bets than we lose on the even chances?

The "house edge" doesn't actually come from the "mystical" number 0 (or double 0).  It comes from the unfairness of the payouts in comparison to what your chances of winning acutally are.

For example:  If you were to place one chip on any number of the layout (lets say the number 30), and leave that chip there and then spin the wheel for 370,000 times, you'd win almost exactly 10,000 of those bets because your odds of winning any particular inside bet are 1 in 37...but the payout for winning a straight up number is 35 to 1.  So even if you win at the expected rate of 1 win per every 37 spins of the wheel, you come out under where you started.

The same goes for the even chances.  It looks like a 50/50 shot to win on Red or Black, but it isn't.  There are 18 Black squares and 18 Red squares...but there is also a green square.

So you can calculate your REAL odds of winning like this if you're betting on an even chance (for this example we'll use red):

Divide all outcomes that would be a win for you (18 squares would be a win as you're betting on red) by ALL POSSIBLE outcomes (37 as there are 37 squares on the board) X 100 = odds of winning.  So, the above looks like this:

18/37 = 0.4864 X 100 = 48.64% chance to win any given bet. Muliply that percentage times 2 (as if you were betting on both red and black at the same time) and the percentage doesn't come out to 100% as it would if it were a "fair" game.  It comes out like this:

48.64% chance to win X 2 = 97.28% - Meaning if you bet both red and black every spin you'd eventually end up losing 2.70% of your money - I.E.- The house edge.

Imagine you bet every single number on the layout (zero included) - You still lose money even if you win (which you would because you were betting every number on the layout), because the payouts are unfair.

What is interesting, is that you can actually INCREASE the house edge in some casinos that are paying less than 35 - 1 on inside numbers.  The casino nearest me only pays out 34 - 1 on an inside number bet.

That means I'm actually BETTER OFF playing anything but inside numbers (outside bets) because playing the inside numbers is actually increasing the house edge because of their reduced payout.  The even chances still pay out even money, so it wouldn't make sense for me to not play them as opposed to inside numbers.

If you're interested, here is the formula for calculating the house edge on any given bet (as quoted from another website):

You can calculate the roulette house edge to play the game online by subtracting the theoretical payment without the house edge from the actual payoff that is publicized in the casino. Multiply this figure with the mathematical probability of gaining your bet. Multiply the sum further by 100 to arrive at the percentage house edge.  If the casino payoff is 1:35, the probability of you winning is 1:36. The house edge is calculated as: [36/1 - 35/1] x 1/37 x 100 = 2.703%

[rjeaton1]

I mentioned above about picking a number and betting on it for 370,000 spins.  I went ahead and did that in RXtreme with numbers downloaded from Random.ORG (I actually placed 369,999 bets) but I won 10,049 of those bets.  Now, I actually won MORE than what my "odds" of winning were supposed to be by 49 bets.  But, I still came out below where I started (because of the unfairness of the payouts):

[attachimg=#]

[Marven]

In my opinion, in the short run randomness and fluctuation is what you're up against. The house edge will only influence your results in the long run.

If you don't have a real/proven edge over the house, then you have nowhere to escape. All your bets will have a slight negative expectation and it will accumulate and manifest if you play consistently enough. One shouldn't quickly convince himself that he has "escaped" the house edge just because he has avoided a short-term cluster of L's, or whatever.

The most important thing you need to do is prove to one's self whether or not he has a real long-term edge, otherwise he is just gambling.

To do this one needs to test his bets over significant sample sizes and then apply standard deviation and chi square tests over the results of his bet selection. These will tell him whether he has a real edge, or 'thinks' he does but is just being lucky (or in more scientific terms: subject to a random positive deviation).

Once one proves this to himself, his play can no longer be considered as gambling.

This is no "theory" but scientific facts.

But as far as roulette is concerned, there are more people who want to 'think' they are winners than there are actual winners.

Some of them will just play 'long enough' and assume they're winners. I'm in no position to say they aren't, just stating the way I, personally, prefer to approach the game.


I highly recommend this: nolinks://vlsroulette.com/reference-area/roulette-probability-made-easier-t2193/
If you haven't already done so, do some research on standard deviation, and chi square tests.
Also, there is a great book written by Taleb Nassim called "Fooled by Randomness: The Hidden Role of Chance in the Markets and in Life"

[kav]

Dear Marven,
If this is the way you approach the game, 99% of the posts here must be BS to you. Not that they aren't, but I wonder why are you still here.

Where I disagree with you is that...
Standard deviation and chi square tests would show something only if you have a relatively big advantage. A 1% advantage canot be validated with standard deviation and chi square tests.

[Tangram]

kav,

A 1% advantage canot be validated with standard deviation and chi square tests.

Why do you say that? If your edge is small it will take a bigger sample to show it, but it can certainly be done.

[kav]

Tangram,

In what kind of sample would a 1% advantage be less than normal fluctuation? In what kind of sample can we be sure of the results with more accuracy than 1%?
I don't believe that such small advantages can actually be proved by standard deviation and chi square tests. It's in the range of statistical error, so to speak.

[kav]

-