2 ms·
I remember someone took one of those global models and noticed all the equations were time reversible. So they asked "what does the model 'predict' the past lo
by pfdietz 5y ago
I remember someone took one of those global models and noticed all the equations were time reversible. So they asked "what does the model 'predict' the past looked like?"
And it turned out the model predicted the world population was infinite sometime in the 1800s. :)
I have very little faith in any of these models.
- titzer 5y agoThe beauty of open mathematical models is you can show your work. So show your work. Link? Study? Paper?
- pfdietz 5y agoThis was decades ago, I don't have the reference, and it would be on paper not the net anyway. I THINK it may have been in this book: https://www.google.com/books/edition/Models_of_Doom/neaRvwEACAAJ?hl=en https://www.google.com/books/edition/Models_of_Doom/neaRvwEA...
- anyfoo 5y agoAny sources for that? Right now it's just "I have very little faith in these models because I remember someone trying to reverse them and coming up with nonsensical results", probably coupled with "and that is so much more comforting".
- wcoenen 5y agoIf many different initial conditions can evolve to almost the same state, then reversing time must necessarily blow up small errors/perturbations. It's the same reason why driving with a trailer backwards is more difficult than forward, even though Newtonian mechanics are time reversible.
- pfdietz 5y agoI understand they original researchers didn't even validate their models by running them forward from, say, 1600 or 1800. They didn't even try it.
- brandmeyer 5y agoPlenty of ODEs are stable in forward time, but unstable in reverse time. Just because you can run some equations in reverse doesn't mean that the result is relevant.
- vharuck 5y agoAll models are wrong. But some are also useful. Don't be surprised when a screwdriver isn't useful as a hammer.