The ACD Method

I agree with both your points with a caveat. We can't necessarily trust volume because for example in gold the London session is by far the highest volume market. I don't have it handy but I found an academic paper on the gold market that showed that the US market clearly is where the majority of price discovery takes place. I think your point 2 is important. Perhaps volatility is what we need to be looking at more than high/low distribution. I'm coding to do that analysis right now and I look forward to comparing those results with the high low distro in a 24 hour market. I think a 24 hour market is actually easier to analyze than a product with limited trading time. For example, we can see in my 24 hour view of GC that that the official open is still very important, but this could be an inaccurate view in a market that only trades a fixed number of hours a day.

After noticing that 1000 spike on my graph I was looking back at some charts and the last two days in GC is kind of funny with the 0820 and 1000 thing.
 

Attachments

  • 2017-03-18_09-39-48.png
    2017-03-18_09-39-48.png
    95 KB · Views: 27
  • thinkorswim_2017-03-18_09-38-59.png
    thinkorswim_2017-03-18_09-38-59.png
    72.6 KB · Views: 20
baggerlord said:
I agree with both your points with a caveat. We can't necessarily trust volume because for example in gold the London session is by far the highest volume market. I don't have it handy but I found an academic paper on the gold market that showed that the US market clearly is where the majority of price discovery takes place. I think your point 2 is important. Perhaps volatility is what we need to be looking at more than high/low distribution. I'm coding to do that analysis right now and I look forward to comparing those results with the high low distro in a 24 hour market. I think a 24 hour market is actually easier to analyze than a product with limited trading time. For example, we can see in my 24 hour view of GC that that the official open is still very important, but this could be an inaccurate view in a market that only trades a fixed number of hours a day.

After noticing that 1000 spike on my graph I was looking back at some charts and the last two days in GC is kind of funny with the 0820 and 1000 thing.
More...
Can I ask why your research centers around Gold? If you were to look at Wheat for example at 8:30AM CST the wide range off the Open is consistently visible to the naked eye even without backtesting. GC has always seemed like a tougher game to play.
 
I believe gold has the least inputs as far as what effects price, making it easier to analyze and more predictable. I also used to be a bullion and coin dealer so have a decent understanding of the market, and still do some trade of the physical. I'm also learning programming right now so I'm getting real into the weeds on one market. Once my tools are built I can apply them to other markets.
 
Curious for those of you who have used ACD in the context of day trading: it seems like the initial stop is usually a substansial proportion of the daily ATR doesn't this make it hard, over a number of trades, to manage the risk/reward of trades?
 
banco said:
Curious for those of you who have used ACD in the context of day trading: it seems like the initial stop is usually a substansial proportion of the daily ATR doesn't this make it hard, over a number of trades, to manage the risk/reward of trades?
More...
You are absolutely correct, and I believe Fisher has conceded in later videos that the stops advocated don't work anymore. You generally have a limited reward to harvest, if you piss R:R away with wide stops it's a disaster. And it's not just in day trading, a tighter stop is as rewarding in swing trading, which I do, and I work on that.
 
justrading said:
You are absolutely correct, and I believe Fisher has conceded in later videos that the stops advocated don't work anymore. You generally have a limited reward to harvest, if you piss R:R away with wide stops it's a disaster. And it's not just in day trading, a tighter stop is as rewarding in swing trading, which I do, and I work on that.
More...
I agree 100%, I am about to manually backtest multiple small stops/multiple chances instead of just using points B & D with one big stop.
 
baggerlord said:
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.
More...
Yeah, it's the nature of random walks to have their start point end up as the high/low more often than you would expect. I wish I could find it now but there was a blog that talked about this and showed some data similar to yours.
 
I think it has been agreed on for years in this thread that the strategy in the book isn't a functional strategy. I haven't seen anyone on here propose using the stops as described in the book. They might work for trading into a position you intend to hold for weeks or longer. For shorter term stuff you need to create your own strategy based on the info that ACD tools give you.
 
You must log in or register to reply here.
Back
Top