Avoid Parlays!

People have very poor intuitions about probability, and some current events give good examples. For example, let's consider "The Islamist Threat", which makes Robert Bidonetto say: "National security is the only issue that truly matters these days; all else pales by comparison". In order for it to truly be a problem worth governmental action now, we need the following confluence of events:

1) Radical Islam truly is a growing and evil movement, not just one that happens to be in the news these days because of Iraq.
2) Radical Islam continues to grow, while retaining its evilness
3) We can do something now to nip it in the bud (ie action now is worth much more than action later)
4) Our selfish, bureaucratic, incompetent government finds that something
5) That same government correctly implements the something

This is what's known in the gambling world as a "Parlay" - meaning that a sequence of things must all be true. The final probability is P(1) * P(2 given 1) * P(3 given 1 and 2) etc. And as a skeptical libertarian, I am awfully suspicious of 4 and 5. Even if you convinced me that 1, 2, and 3 had an 80% chance each of being true, I would have a tough time seeing 4 and 5 as any higher than 50% each. Which leads to a final probability of 87.2% that government action now is a waste of money.

The same is true of global warming. Here the first couple steps are different:

1) Global warming is a significant net harm, ie the loss of coastal areas outweighs longer growing seasons and cheaper heating bills.
2) Global warming will continue to be bad, ie private action will not invent Mr. Fusion or discover cheap nano-based solar panels

But steps 3-5 remain the same, although of course they will have different probabilities for different events. Again, even if each step is somewhat likely, the chance that they will coincide into a perfect storm is pretty small.

Step (3) is a crucial and often-ignored one. This whole argument is an argument from doubt - doubt that a problem is real, doubt that it will get worse, doubt that the gummint will be able to fix it. And one great strategy when you are in doubt is to wait for more evidence. For this strategy to be bad, the problem must be fixable now but not later. For global warming, for example, that doesn't seem true at all - with the explosion of the private space industry nowadays, it seems to me that it will only get cheaper over time to reduce the Earth's insolation by putting lots of opaque discs into space (or whatever). And for the Radical Islamofascist Threat, well, we managed to beat Hitler. It's not like the bad guys automatically win if they manage to control a few countries. Yeah, it gets more expensive to beat them - but you've got all the money you saved from not going after every scare du jour.

So of course, in the end it all comes down to the relative benefit of money spent on a threat vs. the chances that the threat is real. I'm just saying that you gotta seriously discount those chances because of the tricky parlay involved.

Share this