There has been a very nice multi-blog discussion, started here by Random Roger, and advanced by Michael Taylor of Taylor Tree, as well as Anumati, about the Tom Dorsey idea that once a stock gets to 90, its a buy because it will continue to 100. (Arguably, Ugly foreshadowed this idea here.)
To recap, Roger put the idea on the table, quoting Dorsey. Michael looked at the numbers, and said, while it's a little difficult to get comprehensive data because of splits, etc, it appears to happen "only" 55-60% of the time. Anumati then said this makes it an example of confirmation bias. All true, but let me put the Jaloti spin on it.
If this were a simple 50/50 bet, i.e. goes to 100/doesn't go to 100, and the odds were 55-45, you'd take that bet all day long--right? The casinos make billions on tighter odds than that. What is really be asked with this one, is how long does it take to get from 90 to 100, AND what are the chances it goes to 0 first? The way to make money off of this one is to ask, once a stock hits 90, how often does it get to 100, before it gets to 80 (or 85, or whatever margin of safety you want, to pervert the value investors' catchphrase). If, for instance, there was good data over time that after 90, 55% go to 100 before they go to 85, I'd take that one in a heartbeat.
Cut your losers short, and let your winners run, right?
Subscribe to:
Post Comments (Atom)
3 comments:
As far as the timeframe, from my test results it took anywhere from 15 to 19 weeks on average to hit the $100 target and giving the stock a time limit of 52 weeks to achieve its goal.
You're asking the essential question...how much are you making versus what you are risking. That's the beginning of a system you can trade. Also, when you add a margin of safety to the test you will indeed limit your risk but typically at the expense of profits.
Filters can improve your profit margin while limiting your risk...but it limits the number of data points in your test. Thus, drawing into question the reliability of your results.
It's fun isn't it! :-)
One more thing to think about. The easiest way to limit your profit growth is to set price limits in your system. So, if I was to continue on with this 90 to 100 idea...I would change it to 90 to ????. ???? being whenever I get stopped out by some form of a trailing stop. This allows my profit margin to grow exponentially...while keeping my risks the same as with the profit cap.
In fact, it would be very interesting to see the results of that type of system tested on the stocks that did indeed split in the past and are thus not included in the current test sample.
Pulled all together nicely. Good stuff.
Post a Comment