Log in

No account? Create an account

Super Bowl Indicator

I didn't watch the Super Bowl. I am not a football fan. Sometimes I watch it for the commercials, but often not even that.

An interesting bit of folklore associated with the game is the superbowl indicator. This is the belief that you can determine how the stock market is going to do by which league's team wins the Super Bowl. (Up for NFC, down for AFC). According to investopedia, it is correct about 85% of the time.  (Although curiously, I have seen others say it's more like 80%).

Since the NFC team (the Bears) lost, the indicator says we are going to have a bad year in the stock market. It's quite possible we will, because we have been in a bull market for a while now and are due for a downturn. So this could well be a year that the indicator is correct.

That said, most folks look at the Superbowl indicator as superstition, and with good reasons. For one, causation is not correlation. If the stock market goes down this year, it will not be because the Bears lost. This is just so much common sense. There is another, more subtle reason not to put too much faith in the indicator. That is that an 85% reliability for an indicator is not as good as it looks.

Let's consider the following hypothetical case. Over 8 years, the following results occur:

4 AFC victories with the stock market down
2 NFC victories with the stock market up
2 NFC victories with the stock market down

The indicator is right 75% of the time. That looks pretty good. Now let's say that the NFC wins. You think that since the indicator is reliable 75% of the time, you have a 75% chance that the stock market will be up. But not so fast. Looking at the above statistics, you can see that the indicator only works half the time when there's an NFC victory. This scenario shows that the reliability  of the indicator can differ for the NFC and the AFC.

In general, "outcome of A determines whether B is true or false x% of the time is not " is not the same as "outcome 1 of A determines B is true x% of the time".

But just to muddy the waters even more, I don't think this happens with the actual superbowl indicator. I used to think it might be that the NFC won more often, and since the stock market is up more often than it is down. But according to this table of outcomes since the first superbowl, the NFC has won 21 games and the AFC 20. So the two leagues are equally likely to win.

At any rate, it's probably best to treat the SBI as an urban legend, albeit an interesting one. And of course you can also say that a loss by the Chicago Bears could just as easily mean that bears will not prevail in the market.