
Even worse maths
The appeal to probability against the likelihood of the changes to the graph seems rather poorly thought out, and even more poorly argued.
The odds of flipping a coin 70 times and getting 55 heads would actually be (unless probability has changed in the many years since I did GCSE maths):
Combinatorial (55, 70) / 2^70
or about 1 in 1.6 million, so about 6 orders of magnitude more likely than claimed. Not sure where the 2^40 figure came from. However, this all relies on each result being independent. The keyword appears in the article itself, 'systematic', and therefore the changes are not independent and the probability cannot be calculated in this way. In fact we can draw no conclusion at all from this argument without fully understanding what changes were made to the graph and why.
The probability argument is therefore moot. This sort of poor, superficially attractive, but completely worthless argument is unfortunately common on both sides of the 'debate'. A more interesting question is why the different data sets do not agree, but the writer already knows the answer to that without needing to consider it: NASA bias of course!
Come on Reg, we expect better.