Confidence or guesswork?

[alancalverd]

Interesting explanation on today's "More or Less" (Radio 4) of the official IPCC vocabulary.

Apparently "high confidence" doesn't mean 95%, 99%, or 3sigma, like in real science, but 80% probability. Fair enough, you are looking for trends in small changes in a chaotic system, so you'd be pretty brave to go beyond 80%.

But "medium confidence" means 50%. In other words, IPCC considers a completely unfounded guess to be a useful scientific prediction!

So here's my next peer-reviewed publication:

"After studying all sorts of dodgy data, stirring in a few proxies, and applying whatever fudge factors seemed to please our sponsors, we tossed a coin and can now state with medium confidence that the weather is going to get worse.

Allowing a margin for subjective interpretation, we can state with high confidence that the weather will get better if it doesn't get worse."

[principled]

we tossed a coin and can now state with medium confidence that the weather is going to get worse.

Far be it from me to carry out an early peer review of your very interesting, concise paper on climate change probability, but I feel I cannot stay quiet in the matter.
I need to know which side of the coin was facing up when you flipped it and which side you chose to land face up. This is VERY important as depending on which you chose you may have either a 51% or a 49% probability. Now even the IPCC would NEVER accept 49%, would they? So own up, which was it?

www.smithsonianmag.com/science-nature/Gamblers-Take-Note-The-Odds-in-a-Coin-Flip-Arent-Quite-50-50-181022961.html

P

[alancalverd]

IPCC is quite happy to accept complete bunkum as long as it's taxable.

[nickrr]

But "medium confidence" means 50%. In other words, IPCC considers a completely unfounded guess to be a useful scientific prediction!

If someone told you that there was a 50% chance that your house was going to be burgled tomorrow would you take no action on the basis that it was a completely unfounded guess ? Of course you wouldn't because a 50% chance shows that it's quite likely to happen.

Also, you say that it's fair enough to call 80% "high confidence". So where would you define medium confidence, bearing in mind that it needs to be significantly less than 80%?

Added to which in science the important thing is to define what you mean. They have done. What they call it informally is irrelevant.

You really need to think these things through before you post.