3 ms·

Other people in this thread make other very good points, but I think this one needs to be said. If I have a single unknown variable in an equation, and I have

Other people in this thread make other very good points, but I think this one needs to be said. If I have a single unknown variable in an equation, and I have no way of estimating it, or knowing its variability, or anything, then anything I put down will most likely be wrong and will likely be off by several orders of magnitude. Let me give an example. You are playing a first person shooter and you are told that you are going to pick how much ammo you'll need for the entire game at the beginning of the game and that if you pick too much "bad things will happen". How do you possibly decide? You can estimate based on other games you played, but if you look at all of them, you'll realize that they all need very different amounts! The number you pick when you start the game will be completely arbitrary and will have VERY different effects on the gameplay. Let me give another example, you are playing the stock market on a day trading scheme. You want to know if you'll profit, so you look on google finance and see a chart that is trending very slightly up but with a lot of variability. You say, well it looks like it is going up RIGHT NOW! Indeed you execute the trade and then you realize this: You have no idea when the end of the spike is. You have no idea when to pull out, and indeed the time you do will have a heavy financial impact.To paraphrase from the show House MD, there is a right answer. You may have no way of knowing what the right answer is, but you'll still be wrong if you're wrong. This is the point that the guy was trying to make. When you know nothing, the only thing you can do is manipulate the equation to suit your perception. In the end you will only deceive yourself.

dejb 18y ago
> If I have a single unknown variable in an equation, and I have no way of estimating it, or knowing its variability, or anything, then anything I put down will most likely be wrong and will likely be off by several orders of magnitude. Actually if you know nothing about the variable then you will almost certainly be off by an infinite number of orders of magnitude (log(infinity) = infinity). Saying that you might be wrong by several orders of magnitude implies that you do know quite a bit about the range of the variable. By definition most judgements are based on incomplete information. Otherwise no judgement is needed. So the argument that an action should or should not be taken because the evidence is not 100% complete is spurious.
- wlievens 18y ago
  Most of the Drake Equation factors are fractions in the range [0, 1]. So you do know a lot about the range of the variable.
  - mindslight 18y ago
    But that's just self-deception facilitated by using a linear instead of a log scale. It's easy to choose a conservative "really small" number of 1% or .01%, but what's to say that you're still not off by many orders of magnitude?
    - yummyfajitas 18y ago
      In simple terms, this bound tells us that the probability that aliens exist lies somewhere in [0,1].
      - dejb 18y ago
        By definition all probabilities lie in this range. It doesn't change the fact that within this range of R (Real Numbers) the number of logarithms are uncountable. In simple terms... for every real number >1 (let's call that number x) the number 1/x also exists. Mind blowing huh! .. Also I was primarily talking about the TTAPS equation which seemed to be the main focus of the parent's threads.
jcl 18y ago
Yes, I agree that a single, pure unknown will render a formula useless. But are there any pure unknowns in the TTAPS formula? It's hard to say, looking at the article, because Crichton trails off with "etc.". But are the unknowns as bad as Crichton suggests? Almost certainly not. The very first value is "number of warheads". This number can certainly be estimated... There are probably people in the U.S. whose entire job is estimating this number. The numbers given in non-proliferation treaties are likely not more than an order of magnitude off. Likewise, we can probably estimate the "size of warhead" and "height of detonation". I don't know how well the others can be estimated; they appear to be physically based, so perhaps small-scale experimentation would yield reasonable values and probabilities. But Crichton insists none of these numbers can be known at all -- that how much smoke a city gives off when it burns is every bit as unknowable as the probability that a planet will generate intelligent life. I think he may be exaggerating a little.
- SapphireSun 18y ago
  The problem isn't necessarily that the estimates could be off by an infinite amount in this case. Weather is a chaotic system, so any models that don't get it (almost?) exactly right are going to be WAY off. That's just how it goes :(

dejb 18y ago

> If I have a single unknown variable in an equation, and I have no way of estimating it, or knowing its variability, or anything, then anything I put down will most likely be wrong and will likely be off by several orders of magnitude. Actually if you know nothing about the variable then you will almost certainly be off by an infinite number of orders of magnitude (log(infinity) = infinity). Saying that you might be wrong by several orders of magnitude implies that you do know quite a bit about the range of the variable. By definition most judgements are based on incomplete information. Otherwise no judgement is needed. So the argument that an action should or should not be taken because the evidence is not 100% complete is spurious.

wlievens 18y ago
Most of the Drake Equation factors are fractions in the range [0, 1]. So you do know a lot about the range of the variable.
- mindslight 18y ago
  But that's just self-deception facilitated by using a linear instead of a log scale. It's easy to choose a conservative "really small" number of 1% or .01%, but what's to say that you're still not off by many orders of magnitude?
  - yummyfajitas 18y ago
    In simple terms, this bound tells us that the probability that aliens exist lies somewhere in [0,1].
    - dejb 18y ago
      By definition all probabilities lie in this range. It doesn't change the fact that within this range of R (Real Numbers) the number of logarithms are uncountable. In simple terms... for every real number >1 (let's call that number x) the number 1/x also exists. Mind blowing huh! .. Also I was primarily talking about the TTAPS equation which seemed to be the main focus of the parent's threads.

jcl 18y ago

Yes, I agree that a single, pure unknown will render a formula useless. But are there any pure unknowns in the TTAPS formula? It's hard to say, looking at the article, because Crichton trails off with "etc.". But are the unknowns as bad as Crichton suggests? Almost certainly not. The very first value is "number of warheads". This number can certainly be estimated... There are probably people in the U.S. whose entire job is estimating this number. The numbers given in non-proliferation treaties are likely not more than an order of magnitude off. Likewise, we can probably estimate the "size of warhead" and "height of detonation". I don't know how well the others can be estimated; they appear to be physically based, so perhaps small-scale experimentation would yield reasonable values and probabilities. But Crichton insists none of these numbers can be known at all -- that how much smoke a city gives off when it burns is every bit as unknowable as the probability that a planet will generate intelligent life. I think he may be exaggerating a little.

SapphireSun 18y ago
The problem isn't necessarily that the estimates could be off by an infinite amount in this case. Weather is a chaotic system, so any models that don't get it (almost?) exactly right are going to be WAY off. That's just how it goes :(