Okay,
a number of issues here then
Log scale - this guarantees that the same percentage move occupies the same number of boxes onscreen.
Dorsey/Zieg - Dorsey doesn't mention it, Zieg actually says that a 2% move (p33) is preferred - for some obscure reason they then go on to invent a new fixed scale rather than drawing the obvious inference, log 2% does exactly what they suggest.
("Hello foot, like my new gun? BANG!, oops, how did that happen? "
)
Don't get too stuck in Dorsey, he wrote during a long bull market when even Tarot cards worked. Weber/Zieg is better, although I would suggest this is a beginner's book primarily - they raise some interesting ideas, but there is no substitute for carrying out your own testing.
I've spent over a year now trying to improve the underlying 'win rate' of P&F by applying filters of one sort or another, and in all that time I've managed to gain about 10% over trading simple double tops. You must be extremely careful not to read too much into these texts - very few will quote an x% win rate for tactic A compared to y% using the plain double top or whatever. I've read a few, and you have to be aware that they're aiming for a style that looks 'clever than I am' with a few nuggets that look new - I've yet to discover the Holy Grail in any of them.
As an example, Marconi tumbled a few years back from £10 to about, oooh, a quid. No fixed price scale will handle that - if you wanted to trade the dead cat bounce (assuming there was one), for example, a log scale would cope - the box size on a 2% log chart at £10 would be 20p, after the crash the box size would be 2p, and you'd see the whole move on one chart, from start of fall to end of bounce. A fixed price scale like say 10p boxes would show the ups and downs prior to the fall, but be far too coarse to show any movement at the £1 end,
Price scalings like the traditional scaling (seen in Dorsey, on Stockcharts, and so on) or the variation in Weber and Zieg ensure the box size scales with the stock price, but in fairly 'grainy'/coarse steps - eg you might see a 10p box at £10, a 5p box at £5, but you also get a 5p box at 9.99 - the problem, with this style of scaling, is that at/around the cutoff values a share slightly under can have a radically different box size to a share slightly above - log scaling is smooth, the scale changes by an almost infinitessimally small amount every box, so the scale constantly adjusts to suit the share price.
THAT is why I favour it - I can see no advantage to rescaling every £X when you can do it just as easily every Xp.
The log scaling is appropriate in any circumstance, in my view, re Jeremy's bit on the Updata school log is particularly appropriate at those times, it would be (in my view) a serious error to assume that is was not so at other times also.
MS9 allows you to set box and reversal, the problem is that you only get to set one of each, when any P&F user knows full well that the optimum is to set each share's box and reversal individually - to suggest a blanket 1x3 (as the default setting) will be okay for the whole SP500 is laughable, Equis avoid this problem by introducing the filter that - instead of changing the box to a more appropriate size - simply rejects any share not within the chosen price band.
It would be all to easy to go OTT here, after all I'm not arguing with Equis and I like Metastock - I just consider the P&F toolbox is a long way from what their site describes - it's 'depth' (as in their site statement <<< No other analysis package gives you the depth of Point and Figure analysis that MetaStock does.>>> is, in my view, 'shallow'.
I suspect this will improve in time, and Metastock has many, many good points - it is far from my first choice for P&F however, and this toolbox reinforces my view that they do not consider this a significant area of TA.
Dave
