Removing the personalities, this seems to me to be the central problem in a peak oil prediction. Calling any peak, in any domain, is harder than naming the general trend. The example that springs to mind now is the housing bubble. I know people who have been calling a top for 20 years, but were wrong for 19 of those years.

It is always the best bet to say that next year will be a lot like this one. That is the statistically correct prediction. It will be right for say 19 years, until it is wrong.

Calling the 20th year, or peak oil, is tricky business. I think we have to hold ourselves in check by saying that we have indications rather than any sorts of proofs.

Good points.

The main mistakes with the past peak predictions are the same ones being made repeatedly on this board. Some have been calling for a peak for nearly two decades now - eventually they will be correct.

Well, I told my brother back in '04 that he should sell his appt. in Washington DC because the market was overheating. He didn't take the bait. Even now, why sell?

At the same time I argued that China would continue many years into the future with 10% growth because there is a dynamic behind it that can't (at least not with any normal measures) be shut off. For these reasons it only makes sense to continue to do as the EI what have yous do and assume that the past will continue into the future. That is also the meaning of time elapsed growth shift seen in Gilles' figure 1. There is no real "thought" behind it. The trend just remains your friend, right? Until it switches, of course.