The Flow plus ALR

Post your EA's using ALR here
Dewey McG
Trader
Posts: 435
Joined: Sat Nov 26, 2011 4:20 pm
Location: Tampa FL

The Flow plus ALR

Post by Dewey McG »

TambaTrader » Mon May 18, 2015 2:19 am wrote:Hi Dewey,

I've only just noticed how this thread has developed. As you know I've been using ALR2. I have been doing some backtesting and have found that over the long term (5 years) it is better to allow a large number of ALR turns. So your max of 15 is a good idea. There are many settings with ALR though that lose massively. You get good runs for ages e g. a year and then it gets lots of failed alr sequences in a row. From memory when I tested it with 12 turns as the max I had only 13 losses over 5 years and that one doubled it's money in 5 years. It did however have a year and a half where it was only breaking even as there were slightly more failed alr sequences than normal. I have only been backtesting eurusd and gbpusd so far.

If you tell me the sequence of trades e.g. trade 1 SL 30, Tp 90, lot size 1.2 etc. I will put it in my EA and backtest how often the alr sequence would have failed. (I can only go up to 12 turns in mine though - I need to get that changed).

I also find reducing the take profits at the end of the sequence helps. I have a lot more work to do to find the optimum ALR settings.

By the way I have a really simple EA where you manually type in the individual trades. I had it coded for backtesting purposes. I could share that if anyone is interested (the person who coded it for me said he didn't mind). Am on phone at moment so I'd have to do it later.

Shona
Thanks--please do.

I had planned on doing a bunch of back testing myself over the weekend but for some reason when I downloaded 5 years worth of tick data from Tickstory it simply didn't work. I was showing 10% quality which was worthless. I did a little bi over the last few months just to make sure it was working but will have to start all over.

Before anyone jumps in and points out all the pitfalls of back testing--I am well aware. I never use back testing to determine if a strategy will be successful. It can however be useful for seeing if an EA works, comparing settings and observing characteristics of specific pairs.
taipan
Trader
Posts: 355
Joined: Sat Feb 16, 2013 11:27 am

The Flow plus ALR

Post by taipan »

Dewey McG » Mon May 18, 2015 8:58 pm wrote:
Thanks--please do.

I had planned on doing a bunch of back testing myself over the weekend but for some reason when I downloaded 5 years worth of tick data from Tickstory it simply didn't work. I was showing 10% quality which was worthless. I did a little bi over the last few months just to make sure it was working but will have to start all over.

Before anyone jumps in and points out all the pitfalls of back testing--I am well aware. I never use back testing to determine if a strategy will be successful. It can however be useful for seeing if an EA works, comparing settings and observing characteristics of specific pairs.
Dewey,

You can download Dukascopy tickdata from:

http://www.strategyquant.com/tickdatadownloader/

and it is free.
Dewey McG
Trader
Posts: 435
Joined: Sat Nov 26, 2011 4:20 pm
Location: Tampa FL

The Flow plus ALR

Post by Dewey McG »

Since the EA version is working well I am going to discontinue the manual test. With a little over 2 months data I think we can agree this is a viable strategy. Total gain is 24.6% at 10.41% a month with only a 5.03% draw down.

Before anyone bets the farm on this I still have more testing to do before I am comfortable going live. Gareth has been helping me with a lot of the testing and so for every pair we have tested looks fine---except one. The GBPNZD is not a good pair for this strategy. Regardless of settings it keeps hitting SL's even with as many as 20 turns allowed. I was actually surprised because this is considered a more volatile pair.

As to why this one does so poorly when all the others seem to do OK, here is what I think: The GBPNZD is more random than other pairs. It will go through periods of wild volatility and then long periods of little to no movement (which is a killer for ALR).

There will be losses over time. What ALR does is reduce these to a minimum to the point of being rare and allow us to choose settings that are not account killers the way a martingale would be. If we can make 10% a month and keep full SL's down to 1-2 per year (total, not per pair) then we have a real winner. I want to test the other pairs to make sure we don't have any other consistent losers before going live. Once we do go live there is no guarantee that ever changes in the market won't cause another pair to lose more frequently but I think we would have a robust enough strategy to withstand it.
You do not have the required permissions to view the files attached to this post.
Raj4x

The Flow plus ALR

Post by Raj4x »

Dewey McG » Thu Apr 16, 2015 5:42 am wrote:Here is a tease of something I am working on to add to this. Since we are trading daily charts we can look at what the big boys are doing to know when the wind is at our backs. One of the tools for this is the COT report which shows us what the asset managers, banks, big speculators, leveraged funds etc. are doing.

Now this isn't perfect since the numbers are always a little behind, but also remember these guys don't make complete reversals on a dime. It does help to get an idea of sentiment for longer term trades. Here is a weekly chart Showing the EURUSD and what the big guys have been doing. You will notice how the black line (EURUSD line chart) tends to follow the red line (Big guys).
I really like this. Very nice and valuable addition. :yahoo:

Cheers :good:
Post Reply

Return to “Automated Loss Recovery”