Very compelling visuals in the Lindzen video, Anteros, thank you for posting that.
However one has to wonder whether there was anyone in that audience both competent in statistics and willing to challenge Lindzen on the following omission from his presentation.
If each point in the right slide is obtained as the average of 100 more or less normally distributed points in the left slide, the errors bars shrink by a factor of sqrt(100) = 10. Lindzen did not mention this.
I took Lindzen to be implying, both by this omission and his subsequent remarks, that in fact they don’t shrink, and that it is therefore misleading to zoom in on the right by a factor of sqrt(n) (n the number of points on the left producing one point on the right) without also increasing the length of the error bars in proportion.
Now imagine that McIntyre was in the audience. Would he have raised this point with Lindzen at question time, or would he have passed over it in silence?
Now further imagine that the speaker had been Mann instead of Lindzen, with the exact same talk, slides, and emphases, and ask again what would McIntyre have done.
It would be a very interesting poll to see who believes McIntyre would be just as likely to have raised this point with Lindzen as with Mann, and who believes otherwise. Especially if McIntyre himself were among those polled.
As the one who posted this video, Anteros, what do you think?