Everybody has models/assumptions that are variously flawed? But the evidence for climate change with an element of man-made causal factors remains the most persuasive!

Albeit coal industry stooges will hotly contest with pseudoscience and bought and paid for academics?.

But no matter how they wriggle and obfuscate, what they cannot whitewash with highly credentialled academics and their rented conclusions is. What cannot simultaneously occur is record temperatures/heatwaves/fire seasons, for weeks, months, years or decades. During a natural waning phase of the sun, which we have been in since the mid-seventies! (NASA)

El Nino cannot be trotted out as the culprit, given 2017, the second hottest year on record was also La Nina!

So what's is left for a 3 trillion dollar fossil fuel industry?

Well, they can emulate the tobacco and asbestos industries, which shamefully had their own scientists telling their employers the cold hard facts and mountains of, shovelled by the shipload, BS for Joe average?

Modelling can be made to imply everything is hunky-dory, and just to protect the mountains of cash the most powerful folk on planet earth "earn"!?

All one needs are a few (I'm alright Jack) bought and paid for highly credentialled academics, faked science, a few false premises, a few courageous assumptions and abracadabra doubts replace decisive affirmative action!

And all these (warm and comfortable) academics need do is, Don, constantly question the mountains of credible (frog in bowl) evidence and this multi-trillion train rolls on and on!? TBC.
Alan B.

The article really needs to construct a model or models in some detail around a current issue - with numbers-statistics. eg. sea level rises from global warming.

The article then needs to stick to that issue and in so doing illustrate problems with models.

eg.

A. how many mm has the sea level risen on north coast NSW over the last 30 years?

and

B. What is the likely reactions of homeowners who live by north coast beaches on each perceived mm of sea level rise?

Just talking models without an example (WITH NUMBERS) is too waffly - and I predict few comments on this OLO article.

Being stuck in model-land bings amazing outcomes such as Boris,Uncle Donald and Scott Morrison; all by sirprise!
But maybe for some, such surprises are unwelcome.

Dan

I think Don Aitkin is alluding to the fact that, despite his qualifications and background, even he can't always follow the logic of the models, but he can recognise that there are often logical falacies arising in the analysis of statistics. Many people that develop models do not have a strong grounding in statistics (see 6Sigma qualifications for example), and even when they do Mathematicians accept the output without doing sanity checks. Physicists are trained to do sanity checks on output- which is perhaps why in this article these Statisticians have called themselves Physicists- though I'm sure they are both. Analysis of the sanity of models is often a very subtle science- while it can sometimes appear as Expert-ism- my recent reading of Baysian vs Dutch Book attests to it's validity.

For the uninitiated- I agree with Alan B that you don't want to fall into the trap of Expert-ism and assume that everything you don't understand is right- argumentum ad baculum- even though the so called Left often use the same strategy- however you also don't want to assume everything you don't understand is wrong. So what to do- listen to both sides of the argument and build up a picture factoring in the interests and resources of the participants- not all arguments are similarly impenetrable- try to understand a little statistics- there are standard paradoxes in Mathematics and Science to be familiar with- in the end you may not always choose the correct side- but sometimes it's more important to be decisive- you can always change sides...

Often as in the Baysian vs Dutch Book discussion both arguments are perhaps partially correct- so strictly both are somewhat correct- but what does this mean in the real world?

My forte was economic policy rather than modelling, but I first used large-scale modelling in 1966 (with Dick Lipsey et al), I've been involved in many modelling exercises since, and on one occasion I was the only non-modeller invited to a top-modellers' workshop on a new model being developed by the Productivity Commission. So I have some background.

The main lesson I have learned is that you can not depend on models. They are heavily dependent on the modeller's assumptions, which often reflect a bias, political or otherwise, of the modeller. I recall an exercise where three leading modellers (well, some thought the third one was tops, I didn't) were asked to model the same thing for the Hawke government, they came up with wildly different outcomes, which went back to their unproven assumptions. So when I used models, I regarded the output as indicative, and never ever looked more than ten years ahead - the unknowns are too great, and discount rates render the longer term fairly irrelevant to current decisions.

A major factor is that climate is an extremely complex chaotic system, which no one - no one - understands. So it can not be well modelled, it is far, far harder than economic modelling, which very often proves unreliable. And the modellers are very often climate scientists - I use the latter term loosely - rather than expert modellers.

So I have argued for over fifteen years that costly emissions reduction programmes are not the way to go: the future is always uncertain, we don't know what it holds, and to focus policy on one highly uncertain area on which our greatest efforts will have little effect is surely madness. Better to focus on policies which foster resilience, enterprise, innovation etc which will stand us in good stead whatever happens, and which - contrary to the warming scare policies - involves less government, less regulation, less assumptions that a few mortals know far better than the rest of us.