Roger Coppock wrote:
On Dec 13, 11:25 pm, Whata Fool wrote:
(Peder B. Pels) wrote:

Whata Fool wrote:
You goof, there isn't any physical quantity that can
be expressed to 10 decimal places, let alone 18.
Since when has a confidence factor been a physical quantity?
An "estimated" confidence factor, computed from data
cherry picked.

I did not cherry pick.

No, they are magic numbers of unknown lineage and parentage.

And Roger has 36 digits of confidence in a linear analysis
of weather, that's the scary part.

But let's see where it takes us.

Rxy 0.85554 Rxy2 0.73195
TEMP = 13.614826 + (0.005547 * (YEAR-1879))
Degrees of Freedom = 125 F = 341.335554
Confidence of nonzero correlation = approximately
0.999999999999999999999999999999999999 (36 nines), which is darn close
to 100%!

TEMP = 13.614826 + (0.005547 * (YEAR-1879))

Year Temperature
========== ===========
-1,000,000 -5543.81
-500,000 -2770.31
-100,000 -551.51
-10,000 -52.28
-5,000 -24.54
-3,000 -13.45
-2,000 -7.9
-1,000 -2.35
-500 0.42
0 3.19
500 5.97
1,000 8.74
2,000 14.29
3,000 19.83
5,000 30.93
10,000 58.66
100,000 557.89
500,000 2776.69
1,000,000 5550.19

Yes folks, Roger's analytic techniques show that 1 million
years ago the temperature was about -5500 degrees. And we're
in big big trouble in the distant future. It seems obvious
what killed the dinosaurs, no?

Damn that AGW.

Roger also states that "We know that a 'running average'
destroys the value of correlation data." Roger uses a 30
year rolling average himself and states a "Confidence of
nonzero correlation = approximately
0.999999999999999999999999999999999999 (36 nines)".
So from whence does his confidence come?

It's pretty well agreed that weather is chaotic, and a
simple linear model seems totally inappropriate.

Cheers,

Rich