Better ADL | Crypto Craft

Options

Go to
Permalinks
Print Thread

Similar Threads

A better Editor for mql is available. 100x better than MetaEditor 20 replies

looking for Advance Decline Indicator ADL 13 replies

Broker needs to do better, much better 4 replies

Better momentum and better Sine wave indicator 0 replies

Is Better still better? 1 reply

Attachments: Better ADL
Exit Attachments

Better ADL

  • #1
  • First Post: Edited Feb 27, 2016 11:55pm Sep 12, 2015 7:15pm | Edited Feb 27, 2016 11:55pm
Introduction

I've been thinking about volume and price interaction a lot in the last few months, and playing around with some of the traditional indicators to see if they provide any edge. Indicators such as OBV (On Balance Volume), ADL (Accumulation/Distribution Line), CMF (Chaikin Money Flow), and a few others. In short, the ADL is one of the better ones in my opinion, and divergences between ADL peaks (or troughs) and price peaks (or troughs) can sometimes give clues about forthcoming changes in market direction. However, it's a bit hit and miss and, just like any other indicator, you can't rely on it without reference to other signals for better confirmation.

I think there are two reasons why the ADL isn't very accurate:

1. The possibility that Volume itself might not be a reliable indicator of future price movement.

This is highly debateable and I believe there are good points to be made both for and against this argument. And while I don't might discussing it a bit later in the thread if required, I'd rather leave it aside for now and see if we can obviate the need by producing a stronger version of the indicator. After all, the proof is in the pudding.

2. The ADL doesn't measure the relationship between volume and price very effectively.

While I can't do much about the first reason, I can try to create a better indicator. So in this thread I'll present a new version of the ADL, which I've cheekily called Better ADL (BADL) for now, and explain how it's different from the ADL and how we might be able to use it to our advantage. Please note that I mean no disrespect to the creator of the original ADL indicator, Marc Chaikin, and I value his immense contribution to this field.

As always, please bear with my slow pace. I'll try to post when I can but it won't be every day. Thank you.

Edit 28-Feb-1016: Latest version of the NinjaTrader indicator is here.
You reap what you sow.
  • #2
  • Sep 12, 2015 8:03pm Sep 12, 2015 8:03pm
Original ADL

Before we can start creating a better indicator we have to know exactly what the existing ADL calculates and what we can improve. So, this is the formula for Marc Chaikin's ADL:

ADL = Previous ADL + (((Close - Low) - (High - Close)) / (High - Low) * Volume)

As you can see, it's cumulative - each bar's value is added to (or subtracted from if negative) the last bar's to give a running total. It also proportions the bar's volume by applying a ratio which is based on it's price spread.

We can get a better idea of what it's measuring, and any potential shortcomings, by looking at a few examples:

Attached Image

The ratio applied for the above bar is minus two thirds (negative because the bar is bearish). This looks about right to me. If we assume that the upper and lower wicks represent a balance of buying and selling pressure, then the remaining part (the body of the candle) represents two thirds of selling pressure. So far, so good.

Attached Image

The ratio applied to both bars above is the same: minus one. This suggests that there was huge selling pressure in both bars. But is this true? In the case of (a): Yes. But in the case of (b): No! A movement of 1 pip does not constitute a huge imbalance of power, in my opinion. Yet the same ratio (minus one) is applied to both bars. So if bar (b) had the same volume as bar (a), the ADL would change in exactly the same way.

Attached Image

The ratio applied to the above bar is minus one third. But why? It seems to me that two thirds of that bar was selling pressure, which is 100% more than the ADL ratio! Surely this will produce inaccurate results over many such bars.

I believe that these are just some of the shortcomings of the original ADL indicator. In the next post I'll address how we can make it better
You reap what you sow.
  • #3
  • Sep 12, 2015 10:37pm Sep 12, 2015 10:37pm
Better ADL

I believe there are three ways in which we can make the ADL better:

1. New ratio

BADL ratio = (Close - Open) / (High - Low)

The calculation becomes simply the ratio of the Open-Close spread to the bar's High-Low range. This will produce the same ratios as the original ADL in Examples 1 and 2 above, but a different ratio (minus two thirds) in Example 3. Which is a step in the right direction.

2. Weight the spread

We still have the problem discussed in Example 2, which is the same ratio applied to bars that have very different Open-Close spreads. An obvious solution to this is to apply a weighting based on the bar's spread, so narrow-spread bars produce a smaller ratio than wide-spread bars.

There are several ways to do this, but the one I've chosen is to average the spread of the last n bars, where n is configurable, and divide the current bar's spread by this average. So the new ratio becomes:

BADL ratio = ((Close - Open) / (High - Low)) * (|(Close - Open)| / Average Spread)

where Average Spread is the mean of the absolute values of (Close - Open) for the previous n bars. Note that the absolute value is used for the current bar's spread too, so the weighting is always positive.

This method isn't perfect of course, one reason being that spreads tend to vary depending on the time of day. Also spreads usually narrow before important economic data releases. Hence the first wide bar after a series of narrow ones might get an unfairly high weighting applied to it, and subsequent bars might get an unfairly low weighting. But such is life. And hopefully the next modification will help to alleviate the problem.

3. Fractionalise the data

There's probably a better word than "fractionalise" but I can't think of it right now. What I mean is "divide the data into smaller parts". For instance, if you're running the ADL on a daily chart it will calulate it's values based on the daily bars. Fine maybe, but not very sensitive to the huge number of buying and selling battles that occur within each day. The daily bar only gives the final result of the day, not the details.

You could of course simply load a lower-timeframe chart to see the details, but then you've lost the wider-angle perspective. Wouldn't it be better to calculate the ADL on very low timeframe bars (1 minute or less), but show the result on higher timeframe bars? This is what I've decided to do. What's more, the problems mentioned above in weighting the spread might be alleviated if we break the data down into smaller pieces.

So to summarise, this is the formula of Better ADL:

BADL = Previous BADL + (((Close - Open) / (High - Low)) * (|(Close - Open)| / (Mean of |(Close - Open)| for previous n bars)) * Volume)

You can choose period n as a parameter, and another parameter allows you to choose the bar period (e.g. M1) that you want to apply it to, regardless of the period of the bars loaded on the chart.

I'll post the actual indicator next, and present some charts to see how it compares to the original ADL. But that will be later next week.
You reap what you sow.
  • #4
  • Sep 12, 2015 11:14pm Sep 12, 2015 11:14pm
nice presentation NorthTrader, this will be very intresting. volume trading is always intresting. is this something simular to gadi_obv presented here a fiew year back?
  • #5
  • Sep 12, 2015 11:41pm Sep 12, 2015 11:41pm
Quoting NorthTrader
Introduction I've been thinking about volume and price interaction a lot in the last few months, and playing around with some of the traditional indicators to see if they provide any edge. Indicators such as OBV (On Balance Volume), ADL (Accumulation/Distribution Line), CMF (Chaikin Money Flow), and a few others. In short, the ADL is one of the better ones in my opinion, and divergences between ADL peaks (or troughs) and price peaks (or troughs) can sometimes give clues about forthcoming changes in market direction. However, it's a bit hit and miss...
North, the issue is that ADL is predicated on volume as you know. There is no way to fix this without a centralized exchange. Volume analysis works, but not in FX I believe. What is even worse is that broker volume by bar is based on the opening of the broker. So my volume here in NYC will be different than if you live in Tokyo. I don't know how you get around this. The thing is that your formula has some variables in it that cannot be made a constant. That means your formula cannot be axiomatic as you are trying to hit a moving target.

Understand, I know what you can do, and I am not challenging you as much as being the devil's advocate.
If you are good at something, never do it for free--Joker
  • #6
  • Sep 13, 2015 2:51am Sep 13, 2015 2:51am
An idea inspired by similarity stuff: Mid-bar transient zones denote bars where significant buying and selling pressure has occured. One can use for example h=1 transient zone to minimize the lag, and use a low tf (1 min).
ADL_transient[i] =ADL_transient[i+1] +/-1 (if the bar i+1 is transient and up/down, respectively).
One could also instead of the binary +-1 add the height of the zone etc.
Basically this measures how much the price is keen to go up or down. Divergences in the indicator vs. price could be a good measure of what happens next...
Cheers,
k
  • #7
  • Sep 13, 2015 4:59am Sep 13, 2015 4:59am
â € <â € <
  • #8
  • Sep 13, 2015 5:58am Sep 13, 2015 5:58am
Quoting slicken
nice presentation NorthTrader, this will be very intresting. volume trading is always intresting. is this something simular to gadi_obv presented here a fiew year back?
Thanks slicken. I think it's important for people to know exactly what an indicator does before they use it, which is why I'm trying to explain it carefully before posting the indi. As for the gadi_obv indi, I'm sorry but I'm not aware of that one. I think it was introduced before I joined FF. I've had a quick search but couldn't find the formula for it, so not really sure if it's similar to mine or not.
You reap what you sow.
  • #9
  • Sep 13, 2015 6:05am Sep 13, 2015 6:05am
Quoting jmn5611
Understand, I know what you can do, and I am not challenging you as much as being the devil's advocate.
Hey Jim, I always appreciate constructive criticism so thanks very much for bringing it up. And you're absolutely right - this indicator will only work with decent volume data, which I realise is hard to come by in the FX market.

Personally I use the CME futures contracts when analysing FX volume. I also trade the futures usually, but there's no reason why you couldn't trade spot based on futures volume, which apparently is a good representation of the entire market. If you can't access futures data, then try running it against your broker's volume data. It might work surprisingly well
You reap what you sow.
  • #10
  • Sep 13, 2015 6:13am Sep 13, 2015 6:13am
Quoting kprsa
An idea inspired by similarity stuff: Mid-bar transient zones denote bars where significant buying and selling pressure has occured.
Thanks k, that's a nice idea! And would eliminate the need for volume data. Have you tried coding this, or is it just an idea at this stage? I still believe there might be something to Mid TZs, and they're still on my list of things to analyse. Actually I'm being a bit lazy because I know someone who's doing a lot of work on Mid TZs offlist so I'm kind of waiting to see what he comes up with before doing the work myself. In the meantime, I'm focusing on volume and price. Cheers!
You reap what you sow.
  • #11
  • Sep 13, 2015 6:16am Sep 13, 2015 6:16am
Quoting mbertaggia
< <
Sorry mate, I can't read your post at all for some reason. It's all squiggles. But thanks anyway!

Right... back to my Sunday evening of excessive eating and drinking, so that's all from me for today. Have a good one
You reap what you sow.
  • #12
  • Sep 13, 2015 7:22am Sep 13, 2015 7:22am
Quoting NorthTrader
{quote} Thanks k, that's a nice idea! And would eliminate the need for volume data. Have you tried coding this, or is it just an idea at this stage? I still believe there might be something to Mid TZs, and they're still on my list of things to analyse. Actually I'm being a bit lazy because I know someone who's doing a lot of work on Mid TZs offlist so I'm kind of waiting to see what he comes up with before doing the work myself. In the meantime, I'm focusing on volume and price. Cheers!
The first one I have coded (with binary additions of the ADL) before, now I made also the non-binary one (adding the height of the zone). It looks to me that the first one is kinda better for trading -> by divergences, like standard ADL. The second one does not distinguish itself from the price that much...
Anyways, here they are.

Attached File(s)
File Type: mq4 ADL_transient.mq4   2 KB | 1,343 downloads

Attached File(s)
File Type: mq4 ADL_transient2.mq4   2 KB | 1,249 downloads
  • #13
  • Sep 13, 2015 7:58am Sep 13, 2015 7:58am
The idea looks promising. I am waiting to see your work...Regards
Simplicity is the key to brilliance.
  • #14
  • Sep 13, 2015 7:08pm Sep 13, 2015 7:08pm
Quoting kprsa
Anyways, here they are.
Thanks k. That was quick! And very generous of you to share them here. I'll have to convert them to NT7 to see them in action, which I'll do after I've finished presenting my BADL indi. Should be interesing comparing them all. Cheers
You reap what you sow.
  • #15
  • Sep 14, 2015 1:00am Sep 14, 2015 1:00am
Quoting NorthTrader
{quote} Thanks slicken. I think it's important for people to know exactly what an indicator does before they use it, which is why I'm trying to explain it carefully before posting the indi. As for the gadi_obv indi, I'm sorry but I'm not aware of that one. I think it was introduced before I joined FF. I've had a quick search but couldn't find the formula for it, so not really sure if it's similar to mine or not.
You can download it in the killingzone thread here is the link http://www.forexfactory.com/showthread.php?t=381696. This is what I am using for trading accully. If you want source then ask me in pm. You can find it in the thread as well but as you can see its a huge thread.
  • #16
  • Sep 14, 2015 1:16am Sep 14, 2015 1:16am
Quoting NorthTrader
{quote} Hey Jim, I always appreciate constructive criticism so thanks very much for bringing it up. And you're absolutely right - this indicator will only work with decent volume data, which I realise is hard to come by in the FX market. Personally I use the CME futures contracts when analysing FX volume. I also trade the futures usually, but there's no reason why you couldn't trade spot based on futures volume, which apparently is a good representation of the entire market. If you can't access futures data, then try running it against your broker's...
I know what you are about North. That's why I follow your posts. For all the drama that the fellow from Ghana caused, the best traders on this board came together to discuss concepts. I will definitely give the broker volume a go and compare it with the CME.
If you are good at something, never do it for free--Joker
  • #17
  • Sep 14, 2015 1:17am Sep 14, 2015 1:17am
Quoting kprsa
{quote} The first one I have coded (with binary additions of the ADL) before, now I made also the non-binary one (adding the height of the zone). It looks to me that the first one is kinda better for trading -> by divergences, like standard ADL. The second one does not distinguish itself from the price that much... Anyways, here they are. {file} {file}
If you are good at something, never do it for free--Joker
  • #18
  • Sep 14, 2015 4:29am Sep 14, 2015 4:29am
Quoting NorthTrader
{quote} Thanks k. That was quick! And very generous of you to share them here. I'll have to convert them to NT7 to see them in action, which I'll do after I've finished presenting my BADL indi. Should be interesing comparing them all. Cheers
Hey NT,
Thanks for starting the thread and sharing your ideas in the first place!
If you don't mind, I would invest a bit of time to code your ideas for ADL for MT4 (as far as I can understand them) and share the indicators here?
Cheers,
k
  • #19
  • Sep 14, 2015 7:24am Sep 14, 2015 7:24am
Quoting slicken
{quote} You can download it in the killingzone thread here is the link http://www.forexfactory.com/showthread.php?t=381696. This is what I am using for trading accully. If you want source then ask me in pm. You can find it in the thread as well but as you can see its a huge thread.
Thanks, it must be a good indi if you've been using it all this time. I read on that thread you linked to that the indi uses no price information whatsover, only volume. Wow! Mine uses price and volume, and kprsa's uses only price, so it's going to be great comparing them all on the same chart one day to see which one gives the best signals

I'll PM you when I need the source thanks, but it won't be just yet.
You reap what you sow.
  • #20
  • Sep 14, 2015 7:27am Sep 14, 2015 7:27am
Quoting jmn5611
{quote} I know what you are about North. That's why I follow your posts. For all the drama that the fellow from Ghana caused, the best traders on this board came together to discuss concepts. I will definitely give the broker volume a go and compare it with the CME.
That's a great compliment Jim, thank you very much However, I can assure you that I'm not one of the best traders on this board . Not yet, anyway.
You reap what you sow.
0 traders viewing now