Courtesy of NOAA
It's from this page. Kind of frustrating, isn't it?
liberals are devoid of common sense and feel their way through life !
*****
Recognizing this is the first step to engaging a lib in something resembling a conversation. Logic is perceived as a threat to them, so all questions must be phrased in a "how do you feel when..." or "how do you feel about..." style.
Ex: How do you feel about China's increasing prosperity?
[Leading in to China's huge use of high sulfur coal-fired factories]
The moment you mention statistical probability, their eyes will glaze over.
It's from this page. Kind of frustrating, isn't it?
Then you must have noticed that the author of this article doesn't really understand statistics at all.
He states that if we multiply 100 variables, each "known to an accuracy of 99%", that the result has only 36.6% accuracy.
He confuses probability with the (somewhat vague) concept of "accuracy". He attempts to determine the accuracy of a computation involving the multiplication of 100 variables. He does not seem to have a clear definition for "accuracy" and the computation he performs is effectively meaningless.
It is true that if each of the 100 measurements was 1% too small, the computed result would be 36.6% of the true value. But this number has nothing to do with a 1/3 probability as Mr. Lewis states.
Suppose we have 100 independent random variables, with an underlying distribution that allows errors of at most 1% in each variable. How do these errors combine when we multiply the variables together? In general a lot of the errors will be in opposite directions and "cancel out" so that the error in the total product will only grow approximately as the square root of the number of terms. In other words with 100 variables with random 1% errors, we would expect an error of about 10% in the product, which the author of this piece would define as "90% accuracy".
Of course in a real world example we can't guarantee that all the variables are independent. Also the result of the model would be some complicated function of the variables rather than a simple product as Mr. Lewis assumes. Nevertheless it is clear that Mr. Lewis has violated the first rule of scientific criticism, that the critic should know something about the subject he is talking about.
I read a science fiction book called "Fallen Angel". It takes place in a near future world where Canada and the Northern US are being overrun by glaciers. The cause was that we were so aggressive in combatting global warming that we ended up lighting the fuse on an ice age.
Point is, sometimes doing NOTHING is the best course.
The article ignores the fact that climate scientists can analyze the effects of many variables and determine which are important and which aren't (as well as the time-scales on which various factors are significant). The article is anti-scientific in the sense that it doesn't allow for the ability of skillful science to actually figure some things out!