The ACD Method

redbaron1981 said:
I used the criteria that you defined for a trend day so please do not make false accusations.



There is absolutely no point in debating this as I agree with you about when lows and highs are formed and have carried out the research to verify this.



If I can be bothered then I will do this when I have the time. My OHLC data was purchased from a very reputable source and was not cheap. I am also very hesitant to share it online because I am unsure if doing so will breach the data vendors policies. To be honest spending time on here arguing with people like yourself is not as financially rewarding as other ACD projects that I am working on.

I have no idea why you skipped of one subject onto another without finishing first and I am quite tired with your smart ass attitude. For the first time in years I am considering hitting the un-subscribe button to save myself having to read your posts anymore.
More...
I don't trust the accuracy of your study. I don't know if your ATR calculations are forward looking/data snooping or lag one day. I don't know how you calculated the ATR itself. I am just asking for you to output your data and values. Let's forget looking at 30 different ETFs for 1.5 years and instead take an extensive look at one etf (like SPY) over many years.

We are not just debating the data output, we are debating the accuracy of how the data is derived. I am able to share years of data without any problem in a format that is accessible to all. I am just asking you to do the same.

Let's consider the following format:
DATE,5 DAY ATR VALUE,OR% VALUE,TREND DAY (YES/NO),
 
So I got some code working and have some interesting results.
This is GC top is time highs are found, bottom is time lows are found. Hours are 0820 to 1600. Looks pretty good.

But next is 1100-1600. Hmm.

Finally 0000-2359. I think this shows that we need to take the nature of random walk from a given start point into account, and also make sure we understand the configuration of data we are working with.
0820-1600.png
1100-1600.png
 

Attachments

  • 0000-2359.png
    0000-2359.png
    48.9 KB · Views: 15
So here are some things I am considering now.

1. Clearly the 0820 open is important.
2. Is big money also flowing on a midnight based cycle, and an 1800 based cycle, or are these spikes based on random walk and the cycle that my data uses to declare a daily high and low?
3. If the first data point on any time period is significant due to random walk, does that mean that I can use ACD to trade random data?

I don't have answers this is just what I'm looking to explore next.
 
Gold is tougher to prove. Probably not a good example to win someone over who is new to ACD. When Mark built the ACD method pit trading in every contract was still around and it was easy to see a spike not only in volume but range on the open. Some people who struggle to find a meaningful opening range end up widening the range time.

Remember this is called the "OPENING RANGE" keyword being "OPEN". You can't just pick anytime of day to do a study. That does not align with what Mark Fisher originally did. The best thing you can do is to read The Logical Trader a few times yourself.

You say 0820 is meaningful, is your data Eastern or Central Time Zone? 0820 Eastern is the old pit open I believe. Look at Wheat and S&Ps to really see what I am talking about.

I looked at the Gold ETF (GLD) which opens at 9:30 EST like all of the other stocks. I know this is an etf and not GC contract data but Fisher's 20% Rule still holds.

Here is 8 Years of Data from 2008-06-02 to 2016-04-25:

Out of 1989 Trading days:
The High of the day was in the first 30 mintues 606 Times, 30.46%
The Low of the day was in the first 30 minutes 571 Times, 28.70%

You would expect to see a random/evenly distributed data around 7.69%.

As always I will attach the data I used to back up these claims so that all may have equal access to it.
 

Attachments

Last edited:
I'm not trying to prove or disprove ACD. I'm a believer already. I'm showing that a distribution doesn't necessarily mean what we think it does. My output that only includes data from 1100 shows this. For example, try cutting your gld data to start at noon EST and see how it looks. I can do that when I get back on the computer later if you want.
 
Also note the increase at 1000. I don't daytrade so don't follow news releases much. This makes me wonder if there are 10 am releases that are important to gold. If so, and we filtered only for these days, would we find a different opening range time to be suitable for these days only?
 
I will chop the data to see what I get like you did.

I have one other interesting thing, Fisher did not come up with this to my knowledge. If you divide the trading day for stocks into 13 Half Hour buckets the first 30 Minutes has the greatest High-Low Range. So if you measure each 30 Min Bar High-Low Range the 1st Bar has the largest range on the Open 25% or more of the time.

GLD is attached.
 

Attachments

Now that I'm on a computer I'll give you a better answer. I hate typing on a smart phone.

I have read the ACD book several times. His reasoning for opening ranges is that they are significant because they hold the highs and lows more than they should. He sells his data and for each product, he uses different time periods, because each product is different. I am just trying to develop my own code to generate this information, because I don't want to pay for it, I want to truly understand where it comes from and what it means, and also because I want to have the ability to do my own research and test my own ideas. Gold is traded 24 hours and has opening times in London, the US, and Australia/Asia. Which one should I use for the opening range? The London market has tons of volume but the US has more volatility. Asia is slow but seems to be pretty trendy vs the volatility. If I'm going to trade gold I need to have the tools to profile each time period. How do you generate a number line for this market? Should I use all of them, a 24hr one, or some sort of weighted derivative?

I guess what I am getting at is that I'm not trying to "prove" ACD or convince anyone of anything. ACD is a set of tools and they need to be understood and applied correctly. A commodity like gold is far too complex to just slap an OR around the pit open and call it good IMO.

Please don't take me wrong either. I'm a blunt person and I think you are too. :) This is good discussion and I'm just throwing out some snippets of what I am discovering as I learn to work with data.

Edit: I missed your earlier question. Yes I'm using EST time period and 0820 is the official open.
 
I chopped the data/moved forward my opening times like Baggerlord suggested and it looks like you get the same exact results no matter what Opening Time you select. Move your Open forward or backwards, the results are the same. It looks like Fisher's 20% Rule does not have the original intended significance that I thought it had. I originally thought that you could get this 20% number only by associating it with the "true Open".

Now because of Baggerlord's insights it appears that you can make anytime the "Open" and still get the 20% no matter what time is selected.

Better to know the truth than to put blind faith into a lie (not that Fisher is a liar but you get what I mean). Thank you Baggerlord. However,I still believe the Open carries significance for these 2 reasons. These are not Fisher's reasons, just my own.

#1. The highest amount of Volume is on the Open. I know this is not Market Profile but Volume distribution appears to have a bias on the Open.

#2. Let's say you divide the trading day into 30-Minute Bars/Candlesticks. The High-Low Range (High minus Low) Range for each individual Bar is greatest in the first 30 Minutes on the Open 25% of the time. You will not be able to change this by chopping data/moving Opening times.

Thank you again Baggerlord.
 
Last edited:
You must log in or register to reply here.
Back
Top