Hm… Yes.
1 Like
Go is like a dice. All information is seen but no human able to predict it perfectly. Those who predict it better have higher rank.
1 Like
3x3 tic-tac-toe is much easier to predict
It’s it not already pretty meaningless? What is meaningfully forecast?
As @Vsotvep and @gennan say we can apply probabilities to things like dice rolls, coin tosses and card games. It’s just a bit of fun.
However people applying probabilities to things like “will world war 3 start before the end of the year” or “will the stock market be above or below X in Y time” is a bit meaningless.
How about worry a bit less about the probability and a bit more about how to manage in either eventuality or how to prevent it maybe!
Preparing for an event like a collision with an asteroid can cost a lot of effort. For instance we need to calculate the point of impact as early as possible, so we need extra observations and computing power. In addition we need to prepare to evacuate millions of people. The probability calculation tells us if we need to make that effort now, or if this can wait.
5 Likes
In principle yes, but there are practical issues like not knowing initial conditions to infinite precision, and there is also inherent randomness from quantummechanics.
1 Like
There are also millions of asteroids, so narrowing down which ones are more likely than others to hit us is kinda essential.
4 Likes
2010 RF12 is a very small asteroid, classified as near-Earth object
there is a 1-in-10 chance of an Earth impact on 5 September 2095.[4]
1 Like
The movie and the novel called “The Witch of Laplace” was interesting 
So basically, no use worrying about it as a normal human 
No. Not all predictions have a practical use, but forecasting in general is absolutely meaningful. The weather is a pretty good example of a meaningful forecast (be it for agriculture, naval matters or other concerns); the asteroid discussed in this thread is another example.
How about worry a bit less about the probability and a bit more about how to manage in either eventuality or how to prevent it maybe!
It’s not always possible to prepare for both eventualities, sometimes you need to make a choice and to go one route or the other. Sometimes it is possible, but may not be an efficient allocation of resources depending on the likeliness of the event (which is the purpose of forecasting).
6 Likes
I think that calculating the probability of an asteroid hitting the earth is pretty straightforward (you can just run thousands of simulations that fit within the error margins of your measurements, and can say with good statistical certainty how likely it is the simulation is correct, since we have lots of experience with tracking space objects). Similar to predicting tomorrows weather, for that matter.
If the asteroid will hit the earth, let’s say we know it’s going to be central London, then basically our only affordable chance of changing its course, will be when it is going to pass close by the earth in 4 years. Once it’s on its way around the solar system again, it becomes a lot harder to steer it away. But if we can give it a tiny push in the right direction soon after it passes, it’s possible to change its course, as was demonstrated by the DART mission.
In that perspective, getting as much data as we can now, can save millions in the future for not having to send bigger rockets further away in order to save London from its (by that time almost inevitable) demise.
3 Likes
Now I’m starting to wonder if an asteroid can be used as a weapon in a war: steer it towards the enemy’s capital.
1 Like
I think generally it will be more useful if it crosses a certain threshold. Just like in Go, for most players, the win rate doesn’t really matter unless it goes to like 99%. If I walk out tomorrow and the chances of an asteroid hitting me is like 1% or 3% it won’t make much of a difference, but if it suddenly jumps to 50 or 60% I may have to consider staying at home.
A chance of being hit which is 1% is huge. Since my life expectancy at birth is about 30000 days, any risk bigger than 1/30000 is significant. In the case of an asteroid, staying at home won’t improve chances of survival, better go to a country outside the asteroid’s trajectory.
3 Likes
Yes ok I should have been clearer. The application of probability to forecasts is probably meaningless!
The weather example is still a good one. If I claim it will rain tomorrow then that’s one thing. The farmer can wait for the next day to plant the seed. If I claim there is a 75% chance of rain tomorrow, what do the farmer do? 50%? 20%? 1%? If in fact it rains tomorrow then I can claim to have been correct in each case. If on the other hand it doesn’t rain then I can claim to have been correct in all cases except the first. So the only thing that supplying a probably that isn’t 100% or 0% achieves is that I can never be wrong! That feels pretty meaningless to me.
I think this is exactly the point about the asteroid. It’s either going to hit or not not. The consequences of being hit are significant so let’s try and so something about it. But people obsess instead over whether the calculated probability is up a bit or down a bit. Ah it’s down a bit so we can down tools. Oh it’s up a bit, maybe we should do something about it.
1 Like
I can’t disagree more. On the contrary forecasting without probabilities is completely meaningless.
So the only thing that supplying a probably that isn’t 100% or 0% achieves is that I can never be wrong! That feels pretty meaningless to me.
Let’s say that both A and B predict it will rain tomorrow, but A gives it 80% and B gives it 60%. Whether it rains or not tomorrow, in any case you won’t be able to determine if A or B was correct.
However A and B are not just giving that information for tomorrow, they give it everyday. And on this basis, you can certainly assess the validity of their respective models. If you take all the days on which A predicted a 80% chance of rain, how many days did it actually rain on? If it’s close to 80%, it’s a good model.
If I claim there is a 75% chance of rain tomorrow, what do the farmer do? 50%? 20%? 1%?
The point is to provide the farmer with correct information so that they may assess the situation and future behavior; actual consequence depends on the circumstances and associated risks. Of course a decision for a given day may be the wrong one, but overtime they’ll be able to mostly make the right calls.
This information is very valuable and significant investment are made to obtain and refine it. Telling the farmer “well it might or might not rain tomorrow” is not helpful.
5 Likes
Considering that a lot of capitals are near water, it might be simplier to steer it nearby, at the sea. That way you do not need much accuracy.
I remember daydreaming while swimming as a kid about weird oceanic tsunami making weapons… for example, if you were to sneak a submarine near the sea there and explode nukes there (or something really exotic like matter+antimatter 
), totally underwater (thus out of any defense’s power to intercept), would in create a powerful enough tsunami to destroy/damage a city? 
I had no clue 25+ years ago and I still have no clue if that is possible or not. Certainly very hard to accomplish, but it doesn’t seem as certainly improbable. I guess it would depend on the depth and topography of each place.
Indeed, I’m sure the model can be
Until it isn’t.
The example in the Radical Uncertainty book is about financial models which were very good, and had all sorts of refinements and very high validity to allow policy makers (farmers) to assess the situation and future behaviour. Until the real world diverged from the models and had a financial crisis which the models didn’t forecast at all!
A model can certainly cease to be good if the underlying parameters evolve (e.g. an asteroid hitting the Earth and impacting climate, which would require weather models to be re-evaluated).
It doesn’t mean the model was useless to begin with.
You can’t simply decide that, since we can’t forecast everything, we should forecast nothing and just prepare for all eventualities. At the risk of repeating myself:
It’s not always possible to prepare for both eventualities, sometimes you need to make a choice and to go one route or the other. Sometimes it is possible, but may not be an efficient allocation of resources depending on the likeliness of the event (which is the purpose of forecasting).
4 Likes