Jun

1

The Chair recently mentioned Point and Figure, and advised that he was reading Tom Dorsey's book on the subject. I'm not sure if that was a recommendation. It won't hurt anyone to read Dorsey. There is another popular author on P&F, whose work I do not recommend.

Others have commented on the plethora of Dorsey followers. I have seen that also, but what has surprised me is that the institutional following of Dorsey is quite large.

There are substantial differences in Point and Figure methodologies that bear noting, particularly to anyone doing statistical/quantitative testing.

Dorsey uses a 3-box reversal. That means that to reverse direction, three boxes are needed, whereas continuing in the same direction requires only one box. That has the quality of making his version of P&F asymmetrical. Before you decide whether or not you like that, consider the following:

If your box reversals are symmetric, then P&F becomes strictly a non-linear filter. The smaller you make the box sizes, the more the P&F filtering of the data will converge to the price. For example, if your box height were the minimum tic, then the P&F filter would be almost identical to the price. (It would be identical unless the market went sideways.) But with the asymmetrical/classic/Dorsey version, reducing the box size changes the chart. But it never converges to the price. You will have to decide if that's the type of filter you want.

Dorsey also has box sizes that adjust to price, but only at certain intervals (like at $20). That creates some problems if the price fluctuates around the appointed level. However, the point is well taken that if the price doubles the box size probably should also, so there should be a better method of adjusting box sizes. John Bollinger had done that by defining Bollinger Boxes as a function of price.

I'm not sure John wants his formula out there, so I won't give it. But it does dynamically adjust box sizes. The values they provide are good ones for most stock prices. The only problem with that formula is that it is poorly defined for values below one.

Now you might question why I am looking at "penny stocks", and the answer is that I am not. But consider that if it makes sense to P&F filter prices, then it might make sense to do that with indicators, oscillators, or whatever. That's where the occasional values below one come into play.


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search