3 ms·

I have general comment on what is discussed in several individual threads here: Such "simulations" should be seen as a tool for thought, not as truly accurate

I have general comment on what is discussed in several individual threads here: Such "simulations" should be seen as a tool for thought, not as truly accurate scenario of a most probable future. It is rather like assigning some numbers to a qualitative argument in a debate. To keep the complexity in check, you leave out everything that you think is not relevant to the overall outcome of the simulation. To challenge such a simulation, a critic must present his or her own simulation showing that the omitted factor may well be relevant instead. In the Princton simulation the participants use only a bit more than 10 percent of their arsenal. Assuming that most warheads get through, simulating an intercept rate would be redundant if one also assumes that the margin of error on how many of their arsenals participants would use is much larger. Likewise, focusing on Russia and the USA already covers almost 90 percent of all nuclear warheads. Simulating the other nuclear powers would probably not change much in the overall picture. However, in view of the complexity of the influencing factors and the multitude of possible scenarios, the fundamental question arises as to whether such simulations can at all provide deeper general insights into the topic than what we already vaguely suspect.

Nomentatus 4y ago
Especially since finding participants here whose psychology and job experience parallel Putin's quite closely will not be easy. One the one hand: Putin bluffs about whether he's bluffing about bluffing. His word is chaff. If he says he won't be invading, it's a pretty good indication he will. Etc. On the other hand: When one person is vital to the course of events and can't be removed, prediction generally becomes a fool's game, either way; because human brains are, needless to say, highly complex. Concentrating power means anything can happen (but not that anything will, of course.) Pick your poison. With Putin being the other player, no course of action, including pacifism, is anything but risky. Gird your loins (as none of the kids say.)
data_maan 4y ago
This comment is wrong and I'm negatively surprised it made it to the top, since I would have believed HN had a more discriminating approach. The comment mirrors a very similar argument made against climate change: Since we can't simulate it accurately, we say the entire simulation is useless (" whether such simulations can at all provide deeper general insights into the topic than what we already vaguely suspect"). The key benefit of such simulations is not always scientific insight, but making the problem and its urgency palpable for the non-expert. You "vaguely" suspect the problem, but that is not enough to make you go out and do something against it. You need to fully understand it to get the motivation to act in some way (write to your local politician, set up a fundraiser against nuclear proliferation etc.).
- godelski 4y ago
  That's not how I personally read the comment. But let's also recognize that the accuracy of war games is very different than the accuracy of climate simulations. War games like these typically are one offs. Climate simulations have substantially larger teams, use ensembles of models, and have a long history of verification. The above comment reads to me that they are saying the simulation should be treated like a Fermi estimate. That people can come back and say "well this is wrong and that is wrong" but the point is to get a reasonable estimate. I mean let's be real, the accuracy of this simulation is pretty low. It doesn't even consider the effects of fallout. But that's outside the scope of the simulation and that is perfectly okay. It does exactly what it is supposed to, put a pin as a starting point for conversations. We can add complexity on top of it and calculate ensembles of simulations for the different potential scenarios. But one can argue that there is a lot more randomness involved in a simulation like this than compared to climate because it depends on something that is almost impossible to model: decisions of singular persons. There's several very critical points in these simulations that depend on such decisions. > The key benefit of such simulations is not always scientific insight, but making the problem and its urgency palpable for the non-expert. But I'm a bit confused, because this line makes me think you two are actually in agreement.

Nomentatus 4y ago

Especially since finding participants here whose psychology and job experience parallel Putin's quite closely will not be easy. One the one hand: Putin bluffs about whether he's bluffing about bluffing. His word is chaff. If he says he won't be invading, it's a pretty good indication he will. Etc. On the other hand: When one person is vital to the course of events and can't be removed, prediction generally becomes a fool's game, either way; because human brains are, needless to say, highly complex. Concentrating power means anything can happen (but not that anything will, of course.) Pick your poison. With Putin being the other player, no course of action, including pacifism, is anything but risky. Gird your loins (as none of the kids say.)

data_maan 4y ago

This comment is wrong and I'm negatively surprised it made it to the top, since I would have believed HN had a more discriminating approach. The comment mirrors a very similar argument made against climate change: Since we can't simulate it accurately, we say the entire simulation is useless (" whether such simulations can at all provide deeper general insights into the topic than what we already vaguely suspect"). The key benefit of such simulations is not always scientific insight, but making the problem and its urgency palpable for the non-expert. You "vaguely" suspect the problem, but that is not enough to make you go out and do something against it. You need to fully understand it to get the motivation to act in some way (write to your local politician, set up a fundraiser against nuclear proliferation etc.).

godelski 4y ago
That's not how I personally read the comment. But let's also recognize that the accuracy of war games is very different than the accuracy of climate simulations. War games like these typically are one offs. Climate simulations have substantially larger teams, use ensembles of models, and have a long history of verification. The above comment reads to me that they are saying the simulation should be treated like a Fermi estimate. That people can come back and say "well this is wrong and that is wrong" but the point is to get a reasonable estimate. I mean let's be real, the accuracy of this simulation is pretty low. It doesn't even consider the effects of fallout. But that's outside the scope of the simulation and that is perfectly okay. It does exactly what it is supposed to, put a pin as a starting point for conversations. We can add complexity on top of it and calculate ensembles of simulations for the different potential scenarios. But one can argue that there is a lot more randomness involved in a simulation like this than compared to climate because it depends on something that is almost impossible to model: decisions of singular persons. There's several very critical points in these simulations that depend on such decisions. > The key benefit of such simulations is not always scientific insight, but making the problem and its urgency palpable for the non-expert. But I'm a bit confused, because this line makes me think you two are actually in agreement.