Tuesday, April 20, 2010

Formula for predicting rain in L.A.

Today, there is a 60% chance of rain in Los Angeles, which is wrong. I haven't backtested a full statistical analysis of this, but the correct way to predict rain in L.A. is as follows:

% predicted by weather reporters - 30 = true % chance of rainfall.

The elegant thing about this formula is that it transcends all human media factors. No matter which local TV personality you watch or which radio station you listen to, they all predict the same thing. So you don't have to amend the formula if you watch, say, Fritz Coleman over Jackie Johnson. (I feel it is important to note here that Jackie Johnson is hotter than Fritz Coleman.)

Therefore, to take today as an example, there is actually only a 30% chance of rain (60% predicted - 30 = 30% chance). So if you want to make money off some local news junkie friend of yours by placing an even-up bet on the rain, today is your chance. No guarantees, but the odds are in your favor.

Applying the formula elsewhere, we know that most summer days in L.A. feature a 0% chance of rain. This, of course, means that there is actually a negative-30% chance of rain.

Looking at things the other way, when forecasters tell you that there is a 100% chance of rain, they are risking their reputations. Even on such days, there is actually only a 70% chance of rain. The ONLY time there is ever a 100% chance of rain in Los Angeles is when weatherpeeps predict a 130% chance of rain.

I have yet to work out a complementary formula for snowfall predictions in L.A., but I imagine it involves a lot of zeroes.

1 comment:

hotfire said...

You touched on an interesting point early on...the "hotness" factor of newscasters. An interesting survey would be to mine the factor that determines what newscasts men watch...I guarantee a leggy blonde is in the mix (think Fox News.)

To your point about the weather...the inverse of that works correctly in Pittsburgh. Chance of rain + 30%. Given that, a zero percent chance of rain still means there is in reality a one in three chance we will see showers.