I am trying to determine the best heuristic rules to know if an outside bar (engulfing bar) has broken the previous bar high or low first on past historical series. Can you help?

This:


or that:

Just spit-balling, but would the bar's direction work(open-close relationship)?

I did a complicated zz indicator a while back for NT7, and that was the logic I used, I believe I had several relationships for dealing w/outside bars.

You mentioned heuristic, so I tend to think that you expect it to only be an approximation, as you'd never REALLY have a way of knowing unless you saw the bars that made up that bar?

But could logic such as...

 

Code

if(Close > Open && outside bar) //then Low came first

forrestang 

Just spit-balling, but would the bar's direction work(open-close relationship)? I did a complicated zz indicator a while back for NT7, and that was the logic I used, I believe I had several relationships for dealing w/outside bars. You mentioned heuristic, so I tend to think that you expect it to only be an approximation, as you'd never REALLY have a way of knowing unless you saw the bars that made up that bar? But could logic such as...   Code if(Close > Open && outside bar) //then Low came first

I think this is one good way to do it on past historical bars. I also thought that if the bar that preceeds the outside bar has its close > 50% of its range then i might assume the next bar would continue higher first, i.e, the outside bar high would come first even if the outside closes up.

Also, if there is the case where the outside bar close and open are equal?

trendisyourfriend 

I think this is one good way to do it on past historical bars. I also thought that if the bar that preceeds the outside bar has its close > 50% of its range then i might assume the next bar would continue higher first, i.e, the outside bar high would come first even if the outside closes down. Also, if there is the case where the outside bar close and open are equal?

That's actually a good idea.

I never thought of that with the ZZ I made, but I did use the relationship of the bars prior to the outside bar. I.e., weather price was heading up or down.

But I think using a % of the close is very logical for this type of thing.

In this case I would just add a smaller data series underneath.

@trendisyourfriend

As per an earlier request...

I actually had a catastrophic HDD failure where I keep all my trading files, so I dunno if I can find my older indicators just yet? I'm still sorting that out.

However, I did upload a version of this zz I built for NT7 in the downloads section. Find one of my indies, click the question mark next to my name to show all MY indies, and It is likely the only ZZ I put on this site.

If I remember it is the ZZ that starts over every session, and doesn't plot continuously. But I'm sure the logic is the same. I haven't looked at it in a while, but if you do grab it, and need help figuring it out, lemme know... it's actually quite bloated if I remember? I built it so it would render quickly.

Also, with this thread, are you wanting to build a zz, or do something else? In looking for ideas, it might be easier for us to all brainstorm up ideas if we knew what you are trying to accomplish.