[pepsi_junkie]

Right, about climate change. Which isn’t what the article is about.

[SampleMan]

Climate was the example I used, but my point was that a model that is demonstrated to correlate to past events (assuming those events are truly known) does not by definition have validity based on “its track record”.

The reason being that the model has never actually been tested. It might be valid, it might not be.

Often such models state what appears to be a an impressive time period over-which it has been successfully applied. However, if there are only ten data points that are being matched against the model over a 100,000 million year period, it is still just a model based on ten data points. And one should always raise an eyebrow at a 100,000 million year old data point's accuracy. How does one accurately separate plate tectonics from polar wander? Of the inner core, outer core, mantle, and crust which do we use for our reference? There is no master surveying stake to refer to, when everything is moving.