Heikin-Ashi: another candle
motivated by e-mail from Ken G.
We've talked about candlesticks before.
Ths "usual" candlestick generates a neat chart, using only today's Open, High, Low and Close ... but ignores any previous data.

>Previous? Like last week's High, Low ... ?
Actually, I want to talk about the Heikin-Ashi candles which use one piece of info from yesterday's stock numbers: the Close.
It also uses some of its own numbers from yesterday.

The H-A candles generate a different set of High, Low, Open and Close ... like so:

If O(n), H(n), L(n), C(n) are today's stock Open, High, Low and Close
and
xO(n), xH(n), xL(n), xC(n) are today's Heikin-Ashi Open, High, Low and Close
then
    xC(n) = ( O(n) +H(n) +L(n) +C(n) )/4    the average of today's stock numbers
    xO(n) = ( xO(n-1) + C(n-1) ) / 2    the average of yesterday's H-A Open and yesterday's stock Close
    xH(n) = Max[ H(n), xO(n), xC(n) ]    the Maximum of today's stock High and today's H-A Open and Close
    xL(n) = Min[ L(n), xO(n), xC(n) ]    the Minimum of today's stock Low and today's H-A Open and Close
>Huh? That's confusing! I mean ...
Okay, here's a picture which illustrates the standard candles and the Heikin-Ashi candles:  

>That's still confusing!
No matter. The interesting thing is what Heikin-Ashi does to the slick candlestick patterns, like these:

 
For example, in the Doji, the Open and Close are equal (or very nearly so).
You do the Heikin-Ashi ... and it's gone.

Indeed, in order to see a Heikin-Ashi Doji, we'd need xO(n) = xC(n):
( xO(n-1) + C(n-1) ) / 2 = ( O(n) +H(n) +L(n) +C(n) )/4
which is unlikely. In fact, I played with several stocks and found no Dojis so I figure that ...

>You played ... with what?
A spreadsheet. It's just a modification of an earlier spreadsheet, but you can pick reg'lar or Heikin-Ashi and ...

>Can you just show the spreadsheet?

 

>It's guaranteed to work?
Nope.