The Motley Fool Discussion Boards

Previous Page

Investing Strategies / Shorting Stocks


Subject:  Re: hussman Date:  11/24/2002  7:36 PM
Author:  bdixon619 Number:  27607 of 44633

Hey Warren,

This really warms the heart -- "Generally speaking, we consider an indicator to be useful if it is effective in three samples of data - one sample used to actually derive and optimize the indicator, and two separate samples (preferably very different in their overall characteristics) used to verify the measure "out of sample" - that is, outside of the sample of data used to create it. Alternatively, if the approach is based on purely theoretical grounds (so there is no first sample used to derive and optimize it), we generally require it to show good characteristics when the data is split into two samples."

I've done internal figures on both the Naz and S&P, now. And I have not only 1-day probabilities but 2-day probabilities as well. Remembering, those times you got the figures for me when I couldn't, thanks. I'm still working with that forecasting tool and got an e-mail from the developer this morning I haven't had time to work through. I'll post the probabilities for both markets before tomorrow. As in all forecasts, if you're right they're golden if you're wrong you're the goat. I think it might be a help to some or even a majority to have a framework of probabilities to bet with or against. It seems not only more general (no specific price targets) but more precise (the market will likely go up tomorrow) than just trying to go with the flow. Especially, for those of us apt to make mistakes it helps to have sort of an estimate or overview of what the day might do independent from what the market may or may not be saying currently.


Copyright 1996-2022 trademark and the "Fool" logo is a trademark of The Motley Fool, Inc. Contact Us