A Super Mega Bowl guide
Probability is famous for producing answers that feel wrong. A handful of the same mistakes trip up almost everyone, from casino floors to courtrooms. Here are the big ones, each with the mistake, an example, and the fix.
The mistake is thinking that a run of one outcome makes the other outcome due. After five reds at roulette, black feels overdue. It is not. Independent events have no memory, so the odds reset every time. The fix is to remember that the past cannot change the probability of the next independent trial. This one has its own full guide: the gambler's fallacy.
The mirror image of the gambler's fallacy is assuming a streak will continue, that a slot machine or a player is hot. With truly independent events, a streak carries no predictive power either way. The fix is to treat a run as a normal product of chance, not a trend, unless you have real evidence that outcomes are actually linked.
This is the costliest mistake in the list. Suppose a test for a rare disease is 99 percent accurate, and you test positive. Your chance of actually having the disease can still be low, because the disease is rare to begin with. If only 1 in 1,000 people have it, most positive results come from the 1 percent false positives in the huge healthy majority, not the tiny sick group. The fix is to always start from the base rate, how common the thing is, before trusting a test result.
People often rate a specific, detailed story as more likely than a general one. In the classic example, told that Linda is outspoken and studied philosophy, many judge "Linda is a bank teller and an activist" as more probable than "Linda is a bank teller." That cannot be true. Adding a condition can only make an event less likely or equal, never more. The fix is that more detail means fewer, not more, of the possible worlds where it is true.
A small sample swings wildly, and people read too much into it. A hospital that delivers a few babies a day will have far more days that are 70 percent boys than a hospital delivering hundreds, simply because small samples vary more. The fix is to trust large samples and stay skeptical of dramatic results from tiny ones, exactly what the law of large numbers is about.
The chance of a positive test given the disease is not the same as the chance of the disease given a positive test. Confusing these two, sometimes called the prosecutor's fallacy, has sent innocent people to prison. "The chance of this DNA match by coincidence is one in a million" is not "the chance the defendant is innocent is one in a million." The fix is to keep straight which condition you are given and which you are solving for.
We judge odds from the winners we can see and forget the losers who vanished. Stories of college dropouts who became billionaires ignore the vast majority of dropouts who did not, because they never made the news. The fix is to ask what is missing from the sample before drawing a conclusion from it.
Train your intuition → Flip a long run of coins and watch the streaks. Seeing chance behave normally is the fastest cure for most of these mistakes.Most probability errors come from a few reliable traps: treating independent events as linked, ignoring how common something is, adding detail that feels likelier, over-reading small samples, flipping conditional probabilities, and only counting the survivors. Knowing their names is half the battle. The other half is why our brains fall for them in the first place.