mdlee3

Since gambling is such a good example to use, look at it this way. Take a single deck of cards (52 cards) and shuffle them and start dealing blackjack. Everytime you deal a card, the odds change of a certain card coming up next and the odds won't reset until the deck is shuffled again.

Now go play roulette. Unlike blackjack, the last play doesn't determine the odds for the next play as they odds are the same for every spin. This is the same as dying in a plane crash. If the odds are 1 in 10,000,000 then the odds are always 1 in 10,000,000.

mecabq

But there is a difference between 1/10,000 and 1/10,000,000. In each of our lifetimes we should expect to achieve at least one meaningful 1/10,000-probability event (because we likely play a 1/10,000 lottery of one type or another every day), whereas we should expect to achieve zero 1/10,00,000-probability events.

(At least that's how I rationalize asking the beautiful blonde at the bar for a date in whichever city I am in each Saturday night.)

zlc

That's where the flaw of logic is, if you try a lot of 1/10,000 events enough times, you will achieve one some time, this is true, however, for each individual event, your odds are still 1/10,000, and you have no control which event you will be successful. You may score that date with the beautiful blonde, but your odds of scoring one date with any beautiful blonde is much bigger than 1/10,000 since the odds has been aggregated, just like the 1/10,000 of dying from a plane crash during a 20-year periods of weekly flights is already an aggregated odds, though much bigger than the original odds, but still small enough to be a factor of cause.

davef139

I would there is 1. Also by killed in a plane crash are we talking passenger only or bystander, with Sept 11 I would think odds are high for this board.
I will add dying by car is pretty good odds of death in the US.


If you ever look at major political events, generally someone in the top5 line of sucession isn't present in the event of a major disaster.

pinworm

Exactly..the "Gambler's Fallacy"

The odds reset each time you board the plane, the flying you did before has nothing to do with the flying this time. If it's one in ten thousand, it's one in ten thousand EACH time.

However, casinos understand that the longer you play, the more you stand to lose. This is not because the odds reset, it's because they odds are so unlikely that you tend to lose more than win when you throw the dice more, exposing yourself to their advantage.

Still, in theory you could win every time and they could lose every time. Forever

pinniped

Well, with a plane crash, you can only lose once.

Mistuda

Guys..I haven't read this entire thread but come on...people make a living flying planes and are still perfectly fine and kicking...granted there are accidents, if it was so unsafe, no one would fly!

black dawn

Interesting... I always wondered what the chances are of dying in a plane crash.

dabears1020

But a plane crash is not guaranteed death. In fact, more likely than not, you will survive if in a plane crash.

ralfp

Really? People do plenty of stupid risky things, irrespective of the risks involved (I often do).

On the other hand, plenty of people chose not to fly because it's "unsafe". The fact that all alternative methods of getting from point A to point B, where the two are (rolls the BSing dice...) over a few 100km apart, are more dangerous does not seem to matter.

People are morons when it comes to statistics & risk. I consider myself pretty ignorant about these subjects, yet I have met very few people (in person) who have a better understanding of the subject than I.

That probably depends on the definition of "crash".

mecabq

The OP's premise is that it's one in ten million EACH time, but about one in ten thousand that you will be in a plane crash once in your lifetime if you fly 1,000 times. As I said in a previous post, even though each trial is independent, the cumulative probability is relevant if you are looking at the likelihood of being in a plane crash over a lifetime.

Just like the calculation that if there is a lottery whose odds of winning are one in ten million, then someone who buys one ticket per week for 20 years (i.e., 1,000 trials) faces about a one-in-ten-thousand probability of winning the lottery at least once in his life.

I always found it fascinating that the limit of infinite-trials-of-infinitesimally-small-probability-events -- or the limit of the probability of at least one success over N trials, with probability of success in each individual trial of 1/N, as N approaches infinity, is about 63%, or one minus the inverse of the natural exponent e.

zlc

Exactly, these are two different probabilities, one is the probability of a single event, the other is the probability of a process involving multiple events. Yet most people can not properly differentiate the two and start comparing apples with oranges.

This requires college level calculus, and I am certain that at least 99% of people on this board, myself included, don't know/remember how to figure it out.

jbcarioca

Skip the math. The odds of being in an airplane accident, being injured as a bystander of an airplane accident or winning a lottery jackpot are so vanishingly small that there is minimal risk of the event happening. Thus all these events are cheap events to insure against. Having spent a good portion of my life calculating probabilities of improbable events for business purposes I understand, I think, that very unlikely is not good enough if you're designing an airplane, say, or running a casino. You, in those situations, want to be very precise so you can evaluate your risks in a multitude of events. Then mecabq calculations are necessary.

Otherwise why worry? You're far more likely to fall and die in a shower than in a plane. mecabq's calculations, adjusted to consider decreased risk due to the impact of improved experience, can prove that these assertions are true.