[dangus]

A few problems with your charitable naivete:

1. The proponents of the hockey stick still publsh it, cite it, promote it, defend it, and attack as holocaust deniers anyone who questions it.

2. It was a formula not a “bug” or logic flow that is the issue here, and it’s a work of considerable statistical expertise. Saying that it’s an accident is like saying you’ve accidentally assembled a nuclear reactor.

[ModelBreaker]

A few problems with your charitable naivete: 1. The proponents of the hockey stick still publsh it, cite it, promote it, defend it, and attack as holocaust deniers anyone who questions it. 2. It was a formula not a “bug” or logic flow that is the issue here, and it’s a work of considerable statistical expertise. Saying that it’s an accident is like saying you’ve accidentally assembled a nuclear reactor.

You have obviously looked at this more closely than have I. I know the problem was with calculating principal components. That is a well known algorithm; so I assumed the problem was a bug in the author's implementation. Did they change the PCA algorithm and, if so, what was their justification--was it ever peer reviewed and published or did they keep it secret?