[Jack Simth]

Originally posted by: Krsqk
From what I understand of radioisotope dating, it rarely comes out with the correct date on samples of known age. We assume it works on samples of unknown age. There are dozens of examples of horrendously inaccurate dates published in scientific journals. Even parts from the Mt. St. Helens ordeal have been dated in the low millions.

Krsqk -
You are exactly right - a date of 1 million for a Mt. St. Helens rock historically known to have been formed two decades ago would still be "correct" if the listed marig of error was 1 million years or more, as that would include the proper age of a little over two decades - but they don't come back that way. Instead, the results often come back more along the lines of one milion years +/- one hundred thousand years, a decidedly false result.

Originally Posted by: Imperator FyronThe thing about that is a few million years of innaccuracy don't matter for dating things that are hundreds of millions of years old.

Radioisotope dating is not supposed to be relatively accurate for "newer" objects, only for "older" ones. It is never meant as an absolute indicator. People that use it as such are indeed using it incorrectly.

Fyron -
Krsqk was referring to correct, you were referring to accurate, which are two very different things. Correct would refer to the entire range of values - one million years +/- one million years is correct if the real age is 20 years, but it isn't very accurate. However, like I told Krsk, that isn't the returned result in the majority of cases. For objects of known age that get tested, the testing method is usually demonstrated false. Yet you seem to hold that the method holds for large ages? Fine - based on what evidence? If it doesn't work on objects of known age, clearly the method hasn't been properly calibrated. If it cannot work on objects of known age, clearly the method cannot be checked at all for accuracy. How then can you hold to it?