Better ADL

That's it from me for now, unless there are any questions or comments. I hope you'll find these new versions of the ADL useful. I for one will certainly be using them from now to help with my trading decisions.

Good luck everyone

This is fascinating NT. Price action alone used as a volume proxy...
Good outside the zone thinking kprsa.
I was thinking of something similar as a directional indicator but not at h=1. Nice

I have been thinking more about this and wondered if the TZADL indi could be used as a currency strength meter...
We could tally the results from various pairs and create a single TZADL for each currency.
Both the binary and total pips version could work, but the total pips version would need to be normalised in some way as not all pips are equal.

What do you guys think?

Sorry NT dont mean to hijack your thread. Just ideas.

Hi Innate,

Thanks for the nice idea and don't worry, you aren't hijacking the thread! One of the reasons I make my ideas and code public rather than keeping them to myself is to help generate new ideas and code from smart guys like yourself

Regarding the currency strength meter, personally I've never used one and don't know the strengths and weaknesses of such an approach. Maybe that's because I only trade one or two pairs at a time and can fairly easily gauge the relative strength and weakness of those currencies against each other, both long and short-term, by looking at the price and volume action on the chart. So while it is a good idea, and could well produce a highly-accurate currency strength meter, I'll pass on it for now and let someone else work on it if they want to.

Personally I'm going to see if BADL and/or TZADL can be applied to the best-performing First Opposite TZ system that came out of the Statistical Analysis thread, to help improve the results. Should be interesting.

Hey guys,
sorry I was away from ff a bit.
currency strength meter makes sense with the indicator in oscillator mode, not in the aggregated mode. I have this version and would upload it here soon. then it can be in principle applied to indices etc.
k

The oscillator mode of the ADL_transient:


The positive and negative contributions separated (ADX-like):


Cheers,
k

Thank you very much for this! I will check it out in the weekend when I have some time. I really appreciate the source code so I can read and understand what is going on.
Nice work k.

I've been battling with BADL Version 2 for the last few days. It's been there or thereabouts for a while, but not quite perfect. The reason is that it's very difficult to perfectly synchronise multi-timeframe bars in Ninja both in historical and realtime modes, and especially at the moment the processing switches from historical to realtime (MTF Ninja coders will know what I'm whinging about here!).

But having it working perfectly in both historical and realtime is important to me, so that's what I'm trying to achieve. Unfortunately I can't do it at the weekend when I have more time, because the markets are closed.

It also features a new, simpler and more effective formula for calculating the BADL; an optional EMA; and Debug mode.

I'll post when it's ready. Almost there...

I tried MarkoInflux but make a lot of "array out of range". I saw the code and it's not checked array ranges at start of every loop. I made fast fix but I'm not sure if calculations will affected

At last, here is version 2 of my BADL indicator for NinjaTrader. After a lot of fiddling and testing it seems to be working well in a variety of conditions (historical, live, disconnected, connected, main only, main and sub), and the two data series (main and sub) are always in synch.

Here's what's new in version 2:

1. New and simpler formula to calculate the BADL:

BADL = Previous BADL + ((Spread / (High - Low)) * (|Spread| / TickSize) * Volume)

where Spread is:

a) (Close - Open) if not considering gaps, or
b) (Close - Previous Bar's Close) if considering gaps

and TickSize is the size of one tick / pip

2. Calculate on Bar Close is set to false by default, which means it will update the current bar's BADL in real time, and show the last (partial) bar's BADL when disconnected from the data feed. However, to save CPU load note that:

a) If using a sub data series, it will only update the BADL each time a sub bar closes, and not on every tick change. So if the sub data series is set to 1 minute (recommended), it will update the BADL for the last chart bar every minute.

b) If not using a sub data series, it will only update the BADL for the last bar when the bar closes.

3. New EMA feature. This saves you having to apply an EMA to the indicator manually, and speeds up processing because the EMA is calculated and plotted at the same time as the BADL.

4. New Debug feature. I recommend leaving this set to false, however, unless you want to check that the bars are in synch or suspect there is a problem.

Here it is! The other parameters are the same as version 1.




Some finer points (OK to skip):

1. If your main (chart) bars are Daily, and your sub (calculation) bars are Minute, then the NinjaTrader session you are using for that particular instrument might affect how well the main and sub bars are in synch. If you can't find a suitable session to get them in synch (I can't find a suitable one for N225M, for instance), then you can resort to using 1440-minute bars instead of daily. This always works no matter which instrument or session you are using.

2. Although perfectly-synchronised bars aren't absolutely necessary, if you're obsessed with it (as I am!) then you can check how synchronised they are by using the Debug mode and studying the output in the Output window. If the times and volumes of each main bar and its sub bars are the same, then your data is perfectly in synch.

3. At the start of a new main bar the BADL might be a little distorted. This is because the main bar's BAD value is derived by dividing the accumulated sub bars' BAD values by the number of sub bars. In other words, it's an average of all the sub bars within the main bar. So if you start a new main bar with a large-spread high volume sub bar, then the BAD value of the new main bar will seem excessive because it's being divided by 1 at this point, and not by the eventual number of bars within the main bar. I've thought about always dividing the sub total BAD by the total number of sub bars, e.g. 60 for hourly main bars, but I'm not sure if this is the right approach. Also, some hourly bars won't have 60 sub bars within them on some instruments (e.g. N225M day session finishes at 15:15, so the 15:00 to 16:00 bar will only contain 15 sub bars).

4. The code has become quite complicated because of the need to keep the two data series in synch for both historical and real time processing, and especially the switch between them. I've tried to make it readable though, with plenty of comments, so if you want to dive in and take a look, convert it to MetaTrader, or whatever, then please do... and good luck! Any questions, please ask.