Statistics 101

Guys, if you want to teach Crash some conditional probability, then try to do it with some examples other than the limiting cases of P(A) and P(A/A). No wonder he thinks it's nonsense And please please remember Chiroptera's essential point early on - probability only applies over repeatable trials. Consequently, trying to imply that Mod's points imply some kind of telepathy is sheer nonsense. Funny how related this is to Randman's assertion that entanglement leads to acausal communication...

As both a teacher of statistics and a UK national examiner for statistics, I can assure it's not Hmmm, perhaps you need to come to one of my lessons

Sorry Crash, Mod's understanding is fine. As is PaulK's. This is all about conditional probability, and these are indeed time dependent as knowledge is time dependent. Chiroptera does a great job of explaining this, hitting the all important point of making you put your hypothetical money where your mouth is.I'll try a shorter example:I privately roll 3D6 - probability of 3 6's is 1/216. I tell you at least one of the dice is a 6 - the only probability you care about at this point is the conditional probability of P(three sixes given at least one six) = 1/36. Does this mean that there is a 1/36 chance of having three sixes? I'm not even sure what that means, the dice have been rolled and I know what they are. BUT, if we perform this exercise a million times, then of those times that I tell you there is at least one six, 1 in 36 of those times will have all sixes. This is what the probability means. It has no meaning when applied to a single trial.

I see. So RR's rather poor understanding of stats is the proof that Chriop and me, professional mathematicians no less, are talking nonsense. Glad we cleared that up

I'm advancing a proper understanding of conditional probability, a subject you seem adamant to avoid. Nothing inconsistent at all.Chirop once again nails the crucial point:

quote:

---

What is goofy is the ambiguity of so-called "plain English".

---

You cannot muck about with the event in question in probability. It has to be precisely defined else you are talking bollocks.Probability of getting three sixes on three dice given that you already have a six on one of the dice = 1/36. Precisely defined.Probability of getting three sixes on three dice given no other relevant information = 1/216. Precisely defined.Probability of getting three sixes on three dice given that you have an even number on one of the dice = 1/108. Precisely defined.If your books disagree with the above, your books are wrong. Send me the details and I will take it up with the respective publishers.

Bugger, what system is this???? I remmeber it but i can't for the life of me remember which games use it. I was asked this same question then, and I seem to remeber drawing up some rather nice charts. But this was 10-12 years ago...

If you mean for one draw - i.e buying 146 million tickets in one week, and picking every possible number, then the probability is theoretically 1 (practically slightly less for all the usual caveats, acts of God, etc) i.e. play one ticket with the same number for 146 million weeks?Eaxactly the same as if you randomly picked your numbers each week, or counted through the numbers sequentially, or any other method whatsoever. And the answer is 0.632 (3sf)

The only reason to not play that combination is because so many people already play those numbers and you would end up splitting the prize-money, should you win, with thousands of others.If you think there is any other reason not to play those numbers, you do not understand probability - neither theoretically nor practically.