ModusPonens wrote:
ruveyn wrote:
ModusPonens wrote:
Nope. The probability of a person winning the lottery, in that cenario, is 100%. The probability of a specific person winning the lottery is 0.1%
Then you are doing conditional probability. The conditional probability of ticket 1000 winning give that 1 thru 999 did not win is 1.
ruveyn
I don't get it.
# of people who can win the lottery: 1000
# of possible lottery numbers: 1000
Therefore, the probability would be 1.
Probability is sometimes very straightforward and sometimes very tricky. Am I missing something?
If we are looking for the probability that one particular person among the ticket holders will win, then it is 0.1%, as stated above. If you're looking for the probability that there will be a winner among the 1,000 holders of the 1,000 tickets, then it is 1, which is the sum of the probabilities (i.e.: 0.1%) of all the individual holders. This is just a restatement of what ModusPonens wrote.
If you're looking at conditional probability, then you simply need to know who among the 1,000 ticket holders won, as Ruveyn explained.
One issue here is whether or not conditional probability applies. As stated in the opening of the thread, there's no need for Bayesian statistics; you're not conditioning on anything.
The other is the nature of probability. It is a measure of likelihood over a given set. The total likelihood is 1, meaning that someone wins. Given the size of the set, the likelihood of any particular individual winning is, of course, not high. So the probability of any one individual is 0.1, and that gives a total probability of someone winning as 1, as desired.