Posts

Once in 37 Spins Is Not Once Every 37 Spins: a 36.29% Chance of No Zero, and Why Probabilities Don't Add

The misconception A European roulette wheel has 37 pockets, numbered 0 to 36. Zero comes up with probability 1/37 on any spin. From there, a very common line of reasoning runs: one in 37 means it should appear once in 37 spins, so after 37 spins without a zero, it is close to certain on the next. Both halves are wrong. Over 37 spins there is a 36.29% chance that zero never appears at all. And the 38th spin carries the same 1/37 as the first. The error comes from adding probabilities that should be multiplied. Why adding fails Adding 1/37 thirty seven times gives exactly 1, which would mean certainty. The problem is visible with a smaller example that can be checked by hand. Take a six sided die and ask for the chance of at least one six in two rolls. Adding gives 1/6 + 1/6 = 12/36, or 33.3%. Now count directly. Two rolls produce 36 equally likely pairs. The pairs with no six use only the faces 1 to 5 on both rolls, so there are 5 x 5 = 25 of them. That leaves 36 − 25 = 11 pairs ...
Read more

Same 2.70% Edge, Nearly Six Times the Spread: Standard Deviation on Red Versus a Single Number

A £1 bet on red and a £1 bet on a single number at European roulette carry exactly the same house edge, 2.70%. The common conclusion is that they are therefore the same bet in different clothes. They are not. They share an average and almost nothing else, and the difference shows up in a number most players never calculate: the standard deviation. The misconception, stated plainly "If two bets have the same house edge, it makes no difference which one I play." The first half of that sentence is about expected value, and on expected value it is correct. Over a very large number of spins, both bets lose 1/37 of every pound staked. The second half is wrong, because a session is not a very large number of spins. A session is 50, 100 or 300 bets, and over that range the shape of the outcomes matters at least as much as their centre. Expected value tells you where the results cluster in the long run. Variance tells you how far from that centre any one session is likely to land....
Read more

The Martingale Doubling System, Worked Through: a 99% Chance of £5 Against a 1% Chance of Losing £635

The Martingale system rests on one claim: double the stake after every loss on an even money bet, and the first win recovers everything lost plus the original stake. The arithmetic of that single sentence is correct. The conclusion drawn from it, that the system cannot lose, is not, and the expected value calculation below shows exactly where the loss hides. The Misconception, Stated Precisely On European roulette, a £5 bet on red that loses is followed by £10, then £20, then £40, and so on. Whenever red finally lands, the payout on the current stake covers every previous loss in the sequence and leaves £5 profit. Because red "has to come eventually", the belief goes, a patient player always finishes each sequence £5 up. Two facts break this. Stakes cannot double forever, because bankrolls and table maximums are finite. And the house edge on every individual bet is unchanged by the size or order of the bets around it. The Maths, Using a Single Zero Wheel A European whee...
Read more

What a 2.5% House Edge Actually Costs You Per Hour, Not Per Session

A house edge of 2.5% or 0.5% sounds small enough to round to nothing, and that reading is where the misconception lives. The number is small because it is a percentage of every pound wagered, not a percentage of the money sitting in front of a player, and the missing variable that turns a small percentage into a meaningful hourly cost is simply how many bets get placed. Two games with identical edges can cost wildly different amounts per hour, and the same game can cost ten times more or less depending on nothing except pace. What "2.5% House Edge" Actually Means House edge is defined as the casino's expected return as a proportion of total amount wagered, sometimes called handle, across many bets. A single-zero roulette wheel carries a 2.70% house edge because, averaged over the 37 possible outcomes, the payout structure returns 97.30% of every pound bet on it in the long run. That figure says nothing on its own about how much a specific player loses in a specific hou...
Read more

The Gambler's Fallacy: Why a String of Losses Doesn't Make a Win "Due"

Six reds in a row on a roulette wheel does not make black more likely on the seventh spin. This misconception, that a streak makes the opposite outcome "due," is the gambler's fallacy, and it survives because it feels like reasoning even though the underlying process it describes has no memory at all. The misconception, stated plainly The gambler's fallacy is the belief that if an outcome hasn't occurred for a while, it becomes more likely to occur soon, to "balance out" a run in the other direction. It shows up as "black is due," "the machine hasn't paid in ages so it's about to," and "three heads in a row, next one's definitely tails." All three statements assume the process is tracking its own history and adjusting future outcomes to compensate. Independent random processes do not do this. The maths: why independence rules this out A process is independent when the outcome of one trial has no effect on th...
Read more