[akd]

A lot of reasons can sound totally justified in retrospect. "The Colts will win the Super Bowl because they have superior talent." "The Saints will win the Super Bowl because the Colts' ego will go to their head." A lot of people take actions and have reasons for them -- when those actions turn out to be right, the reasons are often given undue credibility.

[fauigerzigerk]

Right. It's the classical data mining problem. What should be done is to take the reasons given and use them as a hypothesis to predict other events. Reinhart and Rogoff have done just that in their book. After reading it I find it very difficult to accept the claim that people predicting this crash were just lucky or that all their reasons are bogus.

And I think you're making a mistake to equate ball games with debt markets. The fortunes in ball games can quickly and unpredictably turn. Debt doesn't work that way. If many people on low income take out mortgages on a teaser rate that goes up two years later to levels they cannot afford unless house prices keep rising at historically exceptional rates so they can remortgage, you know you're going to be in trouble eventually.

[zaphar]

If the Colts did indeed have superior talent and then win the super bowl then obviously the reason was justified. If the Colts' ego is obviously making them overconfident and does indeed go to their head and cost them the game then again the reason is justified.

The point is that the reason isn't justified because the person correctly predicted an outcome. The reason is justified because the reason was based on true fact and is demonstrably capable of causing the effect.

Now if the Colts win the super bowl despite having worse talent because the opposing Saints team members all got drunk the night before and were playing with hangovers, then you would be justified in saying the predictor was lucky. It's not enough to say someone could have been lucky and then dismiss them. You should show they were lucky or accept they were correct in their reasoning.

[jerf]

"If the Colts did indeed have superior talent and then win the super bowl then obviously the reason was justified."

It should be pointed out that with Aristotelian logic, that is false. "Previously I said X will happen because of Y, and X happened." is basically "Y implies X; X is true; therefore Y".

Nevertheless, it is true that if a person makes a claim that something will happen for a given reason, gives plausible logical reasons, that we have priors that also indicate the causative connection is sound, and the prediction ultimately comes true, we should not just ignore it. We may not be justified in using Aristotelian logic to conclude with 100% confidence the person's reasoning was correct, but we are justified in taking it into account and concluding that there is some reason to believe this person, who made a prediction contrary to many people's other beliefs, may in fact have something.

After all, if you're not going to listen to people who make accurate predictions with plausible reasons used for their predictions, you can just give up on science right now, because that's all it is.