Friday, February 18, 2011

Monte Carlo Method

Yahoo Finance:
T. Rowe Price examines this question in a new study that uses Monte Carlo probability analysis to look at likely outcomes of different retiree responses to a bear market
FLG always chuckles when somebody says Monte Carlo probability analysis or Monte Carlo method.  It sounds so sophisticated.  Some black box that only the high priests of statistics and finance have access.

What it means, in practical terms, is to throw a lot of random shit at the wall and see what sticks.  Maybe the kitchen sink does.  Who knows?  They don't.  Otherwise, they wouldn't be doing a Monte Carlo probability analysis.

Don't get FLG wrong.  It's useful so far as it goes.  It's just a fancy pants name for throwing random numbers at the fucking model and saying what happens.


Andrew Stevens said...

I love the name. I think Von Neumann was a genius at picking names as well as the many other things he was a genius at.

If you think about what Monte Carlo is famous for, you would immediately deduce that it relies on random numbers and I certainly don't think the name is "fancy pants."

LibertyAtStake said...

Assumption: The model itself is a reasonable representation of reality.

Constraint: The modelers who think the best available method is throwing random shit at the wall to see what sticks are probably the same modelers who built the model.

"Because the Only Good Progressive is a Failed Progressive"

FLG said...


I'm not saying that the name "Monte Carlo Method" is literally, in and of itself, fancy pants. It's the way in which people use the term.

Blah, blah, we used a Monte Carlo simulation to analyze this model, blah blah.

That sounds all fancy, when in truth it's just a bunch of random numbers thrown at it.

Creative Commons License
This work is licensed under a Creative Commons Attribution-No Derivative Works 3.0 United States License.