[Paco]

If computer models are so accurate, they should be able to take data and readings from the last 100 years, run the models in reverse and show what the climate was like in the past 100-500 years and more. Wouldn't that be logical? Has that been tested anywhere?

[nevergore]

Nah....if you spin the record backward.....

It just says “Paul is dead”

[rjsimmon]

It has, the process is called linear regression. There are a few sites that cover it but fact of the matter is, temperature readings from the past are not that accurate. Visual recordings (people reading a thermometer) are subjective and localized. Core temperature readings (trees, ice cores, etc.) are also subjective. So there is no way to accurately read temperature from the past but we do have to give more credence to visual recordings.

The most damning thing to AGW is the predictions made and how none have come to pass. In 1977, predictions were that in 50 years the Earth would have an increase of 6 degrees Celsius if trends continue. Well, CO2 emissions have increased and the average temperature has increased, but only by roughly 0.5 decrees C. There are numerous possibilities and variables that go into climate, the first and foremost is the Sun. Second is weather patterns such as el Nino. In 1977 there was a blizzard of epic proportion, AGW types did not take that into consideration, else the average temperature would have shown virtually no increase.

A simple Google search will show lots of diagrams and slick charts, but the source is all the same and the majority are flawed and have been proven to have been altered. The below chart comes from satellite recordings (which are the most accurate) and is for the past 40 years:

It shows an increase, but it is 1/4 of 1/10th of a degree. Imagine trying to read that on a thermometer hanging outside your office...

Climate averages

[Basket_of_Deplorables]

In the nomenclature, that’s called, “retrospectively validating the model,” and yes, they tried it. And yes, they fail.

[neverevergiveup]

"they should be able to take data and readings from the last 100 years, run the models in reverse and show what the climate was like in the past 100-500 years and more. "

That's an excellent point, but my concern is that they write their code to get the answer they want, and would either alter the input data or the code in this instance as well. The code they use was paid for by government grants, and should therefore be made public - where it can be suitably examined, and tested in the manner you suggest.

[Freedom_Is_Not_Free]

Sure it has been tested and failed miserably. Just look at the predicted models from 1980 to now vs. actually observed temperatures. The models have been completely wrong.

[DuncanWaring]

They’ve tried that, and failed.

If they had succeeded, they would be shouting it from the rooftops.

[SgtHooper]

The employed pseudo code:

If question is “what is temp 100 years ago”?
Then, output 20 degrees C.

Takes care of that nonsense. :-)