[borisp]

[ QUOTE ]
[ QUOTE ]
Ok, I got over my disgust at the first sentence. On to serious question(s) for David:

What do you assess is the probability of there existing some event that occurs exactly once throughout the duration of the universe? Or is it even possible to assess this quantity?

Is it sensible to talk in the language of physics about these events, if they exist?

[/ QUOTE ]

You can't mean what you are asking. Someone getting eight royals in a row might be favored to happen exactly once.

As for my disbelief in miracles, it does not stem from the fact that I think their probability is so low that it is favored that they will never happen. It is because I believe that there is a high probability that their probability is zero. The answers come out the same but their is a difference. Sort of like the difference between my saying a baseball team is 60% to win and a sprinter is 60% to win. In the second case I am actually saying that there is a 60% chance that the sprinter is (virtually) 100%.

[/ QUOTE ]

I understand the difference between those two notions, far better than you give me credit for, I believe. But let me try to repeat what you have said in a very different way.

You are saying that you have a planned a way to measure a property of a sprinter, and likewise of a baseball team. When you take this measurement the answer will take one of two values, say "win" and "lose." There is a physical process that happens in between, and it is called a "game," and each has its own game (= "event") description.

There is a key difference between these two events. Your knowledge of the sprinter that you have now is, in principle, all of the knowledge that is needed to completely determine the outcome of the sprinter's game. This is because the sprinter's state will not change in a relevant way between now and the time of measurement; i.e., the game is very nonrandom and has virtually no extreme occurrences.

But...your present knowledge of the baseball team is easily corruptible by completely random events that happen during the baseball game, and it is easily known that the game itself can take on very extreme configurations (no hitters, 10 run innings, etc.) Hence, you will admit that in principle you CANNOT determine the outcome now.

Or, to quote the greatest author ever on how to make 100k a year gambling, "baseball is the only sport in which a college team could conceivably beat a pro team."

Anyway, when you are announcing the probability of an event, namely the measurement taking a particular value, you are announcing the numeric value of an integral; you are "adding up" the probabilities of all the paths that lead to "win." This amounts to an integral of integrals, and that is where tricks of language can confuse even the most careful reader of 2+2.

The sprinter has two paths that lead up to the measurement, "going to win," and "not going to win," since his state will not change between now and the measurement. (There are a few "quantum" paths where he (or his opponent) pulls a hamstring or tests positive for steroids and he suddenly jumps from one path to the other.) Your assessment of his probability to win is your assessment that he is on one of these two paths. You perform two integrals, and you are done.

The baseball team has many paths leading up to the measurement, the totality of which is impossible to describe. Now there is randomness in your knowledge of the state, and you cannot escape it. So you integrate over a much larger state space, and you "average out" a lot of extreme events that cancel. For every walk off grand slam, there are so many no hitters, for every blown save, there are so many 9th inning ending double plays, etc. Your knowledge of these cancelling factors, largely a series of estimates, determines your overall success.

At the end, you announce the likelihood that the baseball team is on one of the infinitude of paths that happen to land on "win."

Now to the point:

You don't say that no paths lead to miracles.

You don't say that the integral over the space of all paths to all miracles produces zero.

You do say that any meaningful integral over those paths that do lead to a SPECIFIC miracle must produce zero.

Correct?

Then Bayes theorem allows two conclusions. The probability that there are miracles, given the physics that we know, is zero. And the probability that the physics we know will ever be a complete understanding of the universe, given a miracle, is also zero.

In other words, Bayes theorem has allowed the mutual exclusion of these two states, "miracle" and "modern physics." You choose the non miracle side, because in your view, modern physics is more useful to you than belief in miracles.

Realize that this is a choice, and mathematics was only useful in determining that it was a choice. Physics does not "prove" this choice, it is what shaped your gut feeling about how you made the choice.

Mathematics has only made it obvious that you are either willing to discuss miracles with an open mind, or you are not. You should admit to yourself that you have chosen the latter and move on.

Or, you can admit that there actually may exist miracles, and that the methods of probability are obviously useless for predicting these miracles, by definition. It is still reasonable to go about recognizing when they happen, or even asking others about their efforts to do the same. Try church, I guess, I dunno.

Personally, looking for miracles seems stupid and boring, and I much prefer trying to predict and asses things that the tools of mathematics allow me to predict and assess.