[mistermann]

Surely there was one model that happened to come up with the right prediction, I would assume there were multiple. But my questions is whether there was one model with consensus support (which seems to be the claim), which would be necessary to rule out survivorship bias.

[phaemon]

> Surely there was one model that happened to come up with the right prediction

What does that mean? This "model" is someone sitting down and, using what we know of physics, calculating how something will warm up when you add some extra energy (in this case, the earth). You surely don't think that physics works by people randomly guessing answers and then one of them "happened" to be correct?

You don't, do you?!

[jazzyjackson]

When a thousand hedge funds hire mathematicians and economists to create models of the global economy so that they can predict whether individual stocks will move up or down over different timescales, their task is similar to predicting the weather. Some will beat the market, many won't. Are they guessing?

Maybe that's a bad analogy, but my point is that climate is a chaotic system, and detailed models might be right and might be wrong. It's not that the scientists are guessing, of course the science is methodical - but the performance of the model might not be better than guessing, hence the speculation of survivorship bias. There were lots of climate models over the last century, was there consensus around a paper that is now proven correct? Or were there a hundred models that were off and one that, yes, happened to be correct. I don't know the answer.

[monknomo]

I think an individual stock in this analogy would be "Will it rain in Duluth". Climate predictions, to use your economic model, would be more along the lines of "Will stocks continue to rise on the whole over the next 100 years?"

[mistermann]

> What does that mean?

It means that if you are dealing with a complex multivariate system where the behavior of individual variables is not perfectly understood, let alone the behavior in an interactive system, and you are building multiple systems to model the environment and output a prediction between 1 and 10, undoubtedly some of the models will be correct (since you are predicting within such a small range), but there is no good way of knowing whether the correct prediction is due to proper modeling, to chance, or to some combination of the two.

> You surely don't think that physics works by people randomly guessing answers...

No, I do not think this.

> ...and then one of them "happened" to be correct?

Yes, in that I am suggesting that this is a possibility.

> You don't, do you?!

Rereading my comment, I see no actual content that suggests I believe this.

I wonder, if you were to reread my comment while keeping in mind the theory of System1/System2 thinking [1], do you still reach the same conclusion?

And for clarity, while this may seem (again, see [1]) that I am antagonizing you, I am actually and sincerely trying to make what I consider to be an incredibly important and overlooked point: the manner in which human beings think, on particular (culture war, identity related) topics, is a core problem that is preventing forward progress on not only climate change, but a wide variety of issues.

I am completely open to the idea that this theory is incorrect, but it seems almost no one who is even willing to acknowledge the possibility that it may be correct - an observable phenomenon which I proclaim further supports the theory.

[1] https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow#Summar...

[breakyerself]

Even zero and one dimensional ultra simplified energy balance models come pretty close to approximating how much warming we've had based on the amount of co2 we've released.

Global average temperatures are governed simply by the need for incoming energy to be balanced by outgoing energy.

You're confusing the complexity of weather with the complexity of climate. A lot of the value of increasingly large scale, complex, supercomputer based climate models is the spacial resolution to be able to tell people of particular regions how their local climate might be affected. For instance a lot of work is going into understanding how the monsoon season will be affected.

[rbavocadotree]

From the 1975 US National Academy of Sciences/National Research Council Report

...we do not have a good quantitative understanding of our climate machine and what determines its course. Without the fundamental understanding, it does not seem possible to predict climate... [0]

The point is that there were a lot of climate models done in the 70's, and only a few of them were accurate. No one was randomly guessing, but the model you gave was not the consensus.

[0] https://skepticalscience.com/ice-age-predictions-in-1970s-in...