Vaughan Pratt:
I fully agree. What you’re pointing out (very reasonably) here would seem to strongly support the idea that tossing a coin 14 times and getting a perfect heads-tail alternation every time ought to be highly improbable.
All that excerpt said is you were (seemingly) wrong on one specific point. Specifically, I said a certain trait gives us no predictive information. This was the exact opposite of what you said about that trait. I can’t see anything in the quoted portion of comment that would “strongly support” that idea.
Why do you believe otherwise?
If I were married, I’d tell you if I had stopped beating my wife yet. Since I’m not, I’ll just point out I never said anything about how probable I think that pattern is. I’ve said your 1/8192 value is wrong because it was calculated in an inappropriate manner, but that’s it.
That would be a fair criticism if your method of analysis bore the slightest resemblance to mine. Since it doesn’t, all that does is to call your approach into question. You made no attempt whatsoever to suppress the considerable high frequency noise in HadCRUT3.
What in the world are you talking about? You posted a link to a couple graphs that showed non-overlapping point estimates from a smooth. I posted a graph showing all point estimates for the smooth. Not only does my “method of analysis” bear “the slightest resemblence” to yours, it is identical.
The fact you’ve used a different method elsewhere doesn’t magically make the method you used here stop being what it is.
You are trying to make far too much out of a simple caricature.
You’ve said several different things about probabilities relating to the “simple caricature” that seem to be baseless, if not false. You shouldn’t be surprised I disputed them. You certainly should think I am “trying to make far too much” of anything when all I’ve done is try to get you to either explain or retract statements that seem wrong.