Climate model chart
Scott Adams, author of the famous Dilbert Cartoon, has challenged readers to find a qualified scientist who thinks climate models do a good job of predicting the future.

The Climate Science Challenge
I keep hearing people say that 97% of climate scientists are on the same side of the issue. Critics point out that the number is inflated, but we don't know by how much. Persuasion-wise, the "first offer" of 97% is so close to 100% that our minds assume the real number is very high even if not exactly 97%.

That's good persuasion. Trump uses this method all the time. The 97% anchor is so strong that it is hard to hear anything else after that. Even the people who think the number is bogus probably think the real figure is north of 90%.

But is it? I have no idea.

So today's challenge is to find a working scientist or PhD in some climate-related field who will agree with the idea that the climate science models do a good job of predicting the future.

Notice I am avoiding the question of the measurements. That's a separate question. For this challenge, don't let your scientist conflate the measurements or the basic science of CO2 with the projections. Just ask the scientist to offer an opinion on the credibility of the models only.

Remind your scientist that as far as you know there has never been a multi-year, multi-variable, complicated model of any type that predicted anything with useful accuracy. Case in point: The experts and their models said Trump had no realistic chance of winning.

Your scientist will fight like a cornered animal to conflate the credibility of the measurements and the basic science of CO2 with the credibility of the projection models. Don't let that happen. Make your scientist tell you that complicated multi-variable projections models that span years are credible. Or not.

Then report back to me in the comments here or on Twitter at @ScottAdamsSays.

This question is a subset of the more interesting question of how non-scientists can judge the credibility of scientists or their critics. My best guess is that professional scientists will say that complicated prediction models with lots of variables are not credible. Ever. So my prediction is that the number of scientists who ***fully*** buy into climate science predictions is closer to zero than 97%.

But I'm willing to be proved wrong. I kind of like it when that happens. So prove me wrong.
He also tweeted this:


Climate models are the core of the climate scare, but even the scientists who produce them know their predictive powers are weak. The scientists bundle model output up into an ensemble on the assumption that this will help cancel individual errors, but in doing so they make a very shaky assumption that errors in individual models are independent from each other, and that an averaging process will therefore tend to cancel them out. If the models all share underlying systemic errors, such as shared mistakes in their basic assumptions, bundling the models into an ensemble will do nothing to improve accuracy.

The following presentation by Pat Frank details some of the devastating predictive weaknesses of climate models, especially their poor statistical management of uncertainty.


Will any scientist rise to the Scott Adams challenge?