You fitted a graph using a variant of your dependent variable as an independent variable. No wonder you got a good fit.

For this reason, correlation does not equal ________. It rhymes with “oscillation,” but has fewer syllables.

What was actually discovered by the research is that the temperature decline from 1941 to 1973 was 57% less than it would have been without the high sunspot time-integral during that period and the temperature increase from 1973 to 2005 was 47% more than it would have been without the high sunspot time-integral. There is no ‘fudging’. It is a straight forward application of the first law of thermodynamics (I called it ‘energy balance’) and numerical integration.

If the graph on the page you refer to — one of sunspot oscillation — were used to predict global temperature across the 20th century, it would do a very poor job. This is because sunset activity is one and only one of many factors influencing the climate. In the pdf linked to from that article, you note that you have to use an “offset” — meaning that you move your curve around — in order to get your predictors to match reality.

Also, you ought to note that you (since you wrote that “model”) are predicting temperature variation with temperature variation. You call it “PDO” (for Pacific Decadal Oscillation), but to cut through the verbiage what that measures is the temperature of the water in the Pacific Ocean. LO and BEHOLD!, the temperature of the Pacific Ocean tracks very closely to global temperature. You may achieve a close match, with some additional fudging, but it’s a trivial one.

The model, with an eye-opening graph, is presented in the October 16 pdf at http://climaterealists.com/index.php?tid=145&linkbox=true.

This model predicted the ongoing temperature decline trend. None of the 20 or so models that the IPCC uses do.