Automatic Loss Recovery System (ALR) - read this first

Post your EA's using ALR here
Post Reply
Lagrange
Trader
Posts: 35
Joined: Fri Aug 01, 2014 1:21 pm

Automatic Loss Recovery System (ALR) - read this first

Post by Lagrange »

woodlands » Wed Sep 10, 2014 7:18 pm wrote:I am using Spiders ALR2 with manual entry from Bold Trader's FXD01 (dashboard by Elixe). The loss recovery is working fine with very occasional irrational positions but as they are stops can be adjusted. So far it seems to be a comfortable combination for my taste in risk management.
http://www.myfxbook.com/members/woodlan ... vps/999214
Good to hear Woodlands,

my experience is most positions can be closed at BE or slightly better.
But, I had only one bad trade/position killing all profits.
It hit the ALR border at ECB news after a few previous 'hits'.
So: a large position-after a few ALR band hits- with wide spread during news can be nasty.
Backtests will not show this.

and again: stoporders used in these EA's can give LP's carte blanche.

best L.
User avatar
woodlands
Trader
Posts: 226
Joined: Sun Mar 23, 2014 6:16 pm

Automatic Loss Recovery System (ALR) - read this first

Post by woodlands »

my experience is most positions can be closed at BE or slightly better.
But, I had only one bad trade/position killing all profits.
It hit the ALR border at ECB news after a few previous 'hits'.
So: a large position-after a few ALR band hits- with wide spread during news can be nasty.
Backtests will not show this.

and again: stoporders used in these EA's can give LP's carte blanche.
This has been my experience also in that, as with any EA, it can only 'manage' under manual direction and is really a labour saving device at best, your comments about stop hunting are also very true, enough has been written already about the MM's methods to profit from the markets but then they do create the volatility to benefit from if on the right side of a profit run ( trying to think positive here :roll: ).
I did bail out when the ECB market news drove it against me and have since recovered a (small) part of those losses. It was 'nasty' as you say. :shock: and a difficult decision to cut or wait for recovery, I still have USD/JPY and CAD/JPY in a hole.
Steve says as standard in his post 1:
"•This is not a set-and-forget ea; there is no such thing and anyone who tries to claim there is, is either stupid or lying. This ea requires frequent manual intervention.
•At best, a trading robot is only 90% as good as the manual strategy it trades. At best. At worst, it can be much less effective."
(90% is perhaps optimistic )
http://www.myfxbook.com/members/woodlan ... vps/999214
Thanks for your reply,
Tony
TambaTrader
Trader
Posts: 51
Joined: Sat Jun 28, 2014 5:27 am

Automatic Loss Recovery System (ALR) - read this first

Post by TambaTrader »

Hello everyone,

I'm a bit late joining the ALR party, there are so many EA versions and calculators which look awesome. I haven't had time to look at them all yet though I am definitely going to. Apologies for jumping in without reading them all but I am keen to show people my work and get feedback before I go too far with it.

I have been looking at the EA which you can use with manual trading and want to backtest how often ALR would have failed. The EA is called "ALR with spread v8.3" from wealthmaster.

The problem that needs solving
As I see it the excel calculators make assumptions about how often the SL or TP would have hit based on randomly moving prices. E.g. with a TP to SL ratio of 3:1, the probability of TP getting hit is 25%. This is the best you can do without backdata but it ignores important elements such as the volatility of the pair. You can't backtest an EA such as wealthmaster's in Empty4. You therefore need a better analysis tool. I can use SAS (a statistical analysis system) to analyse what really would have happened if the ALR had kicked in at a particular time.

The question I am asking multiple times in my analysis is "If the first ALR trade started a specific time how many turns would it have taken to hit the TP, given that the take profit level is X and the stop loss is Y". Then I am using backdata and SAS to ask this question for every hour over a given time period. I have asked the question about each hour but I used minute level data to do the analysis for increased accuracy.

The idea is that you can then use this to make decisions about what are the best SL and TP levels for each currency pair.

I have attached an example of what my output data is looking like at the moment, you can see for every hour how many turns would have happened before the TP would have kicked in. (These are not real results as they only include four days - you need a much longer time period). I used EURUSD data by the way.

Points to note
- I am only trying to look at how often the ALR would have failed in this analysis. I am ignoring lot sizes for now. The excel calculators seem like they would work well for calculating what the loss would be if you finally get one.
- I only have four days in this example but the intention would be to do a whole year, (to include all seasonality) and to use the most recent data possible for each currency pair analysed. I probably won't do every hour as it took a long time to run. Perhaps every few hours.
- If I am especially interested in a currency pair I may do many more years but reduce the frequency even further, ask the question at one time per day for the last x years.
- My SAS code doesn't deal with what happens on the last turn yet. I was allowing 15 turns so I never got to it but I might improve it later by coding up that part properly.

I would like people's thoughts on if this is a worthwhile thing to look at, if you can see any problems with the methodology or anyway of improving what I am doing. The sas code I have written is very 'loopy' so takes a long time to run. If there are any sas programmers on this forum make yourself known as I bet there are more efficient ways of writing the code (retain? lag? - I used macros).

Please also tell me what extra stats are useful. E.g. average length of time the ALR was active.

What do people think of the frequency I ask the question. Every hour might be a bit over the top and it would certainly speed up the code if I did every seventh hour for example (if I choose a number that 24 isn't divisible by then I will get a mixture of times of day).

Thoughts please. I have a thick skin so throw rocks at it. The more it is challenged the more I can improve it.

I need to start my 'day job' now so I will be quiet for the rest of the day.

Shona

p.s. Once I have my code perfected I will happily run it for other currency pairs and post the results here for everyone's benefit. It takes a long time to run but I would just set it going overnight.
You do not have the required permissions to view the files attached to this post.
Lagrange
Trader
Posts: 35
Joined: Fri Aug 01, 2014 1:21 pm

Automatic Loss Recovery System (ALR) - read this first

Post by Lagrange »

Hi TambaTrader,

this ALR system is in theory interesting.
I assume you did not read my posts how it behaved in the real trading arena.
Most of the time you can get out at scratch or even better or with a small loss, this is great of course.
But: once in a while you will face a MONSTER trade which did not follow all your backtests...
It will wipe all your profits and even worse.
Spread, news, brokers,LP's, when you push the ALR button.....be prepared.
edit:
also important: when you are in an ALR trade you cant afford it to ' miss the boat' so most EA's will fire a stoporder instead of a limitorder. Your broker / LP will like this!

all the best Lagrange.
blazecrown
Trader
Posts: 31
Joined: Thu Jul 31, 2014 10:20 am

Automatic Loss Recovery System (ALR) - read this first

Post by blazecrown »

The problem that needs solving
As I see it the excel calculators make assumptions about how often the SL or TP would have hit based on randomly moving prices. E.g. with a TP to SL ratio of 3:1, the probability of TP getting hit is 25%. This is the best you can do without backdata but it ignores important elements such as the volatility of the pair. You can't backtest an EA such as wealthmaster's in Empty4. You therefore need a better analysis tool. I can use SAS (a statistical analysis system) to analyse what really would have happened if the ALR had kicked in at a particular time.

The question I am asking multiple times in my analysis is "If the first ALR trade started a specific time how many turns would it have taken to hit the TP, given that the take profit level is X and the stop loss is Y".
hi Shona, this is a good question... but along with it, in addition, better questions for the SAS to solve would be:

1- "What drawdown percentage did it take to achieve TP, considering the number of turns taken to hit TP." --- this would help analyze minimal drawdown...
2. "Considering volatility, what is the best Profit in USD to use for the given profit level X and stop loss level Y." -- this would help analyze optimal profit...
TambaTrader
Trader
Posts: 51
Joined: Sat Jun 28, 2014 5:27 am

Automatic Loss Recovery System (ALR) - read this first

Post by TambaTrader »

Lagrange » Mon Oct 27, 2014 6:49 pm wrote:Hi TambaTrader,

this ALR system is in theory interesting.
I assume you did not read my posts how it behaved in the real trading arena.
Most of the time you can get out at scratch or even better or with a small loss, this is great of course.
But: once in a while you will face a MONSTER trade which did not follow all your backtests...
It will wipe all your profits and even worse.
Spread, news, brokers,LP's, when you push the ALR button.....be prepared.
edit:
also important: when you are in an ALR trade you cant afford it to ' miss the boat' so most EA's will fire a stoporder instead of a limitorder. Your broker / LP will like this!

all the best Lagrange.
Hi lagrange,

I am a very experienced analyst but a very inexperienced trader. I'm not sure what you mean in the above post? Do you mean that the loss you think you will get (by looking in the excel ALR calculator) is not a true reflection of real life because it doesn't account for everything. Why would news (for example) stop it working.

I use an ECN broker so am thinking that I don't need to worry about stop loss hunting. Why do brokers like stop orders? Is that true of ECNs too?

Shona
TambaTrader
Trader
Posts: 51
Joined: Sat Jun 28, 2014 5:27 am

Automatic Loss Recovery System (ALR) - read this first

Post by TambaTrader »

blazecrown » Wed Oct 29, 2014 2:59 am wrote:
hi Shona, this is a good question... but along with it, in addition, better questions for the SAS to solve would be:

1- "What drawdown percentage did it take to achieve TP, considering the number of turns taken to hit TP." --- this would help analyze minimal drawdown...
2. "Considering volatility, what is the best Profit in USD to use for the given profit level X and stop loss level Y." -- this would help analyze optimal profit...
Hi blazecrown,

Agreed they would also be good things to look at. Thanks for the feedback.

I will add things like lot sizes, drawdown, profits etc. It's relatively easy to do. Unfortunately I only have evenings to work on trading until mid December (when I finish my current full time contract).

Will post here when I get some results. Might be a while though. Progress is slow :/

Shona
Chubbly
Posts: 3
Joined: Wed Jun 11, 2014 4:16 am

Automatic Loss Recovery System (ALR) - read this first

Post by Chubbly »

[/quote]


I use an ECN broker so am thinking that I don't need to worry about stop loss hunting. Why do brokers like stop orders? Is that true of ECNs too?

Shona[/quote]


If you use an ECN that charges only for commissions and not on spreads then you will be fine. I use Dukascopy and never had this issue people claim.

I am also playing with this concept but using ATR as the "recovery" area
blazecrown
Trader
Posts: 31
Joined: Thu Jul 31, 2014 10:20 am

Automatic Loss Recovery System (ALR) - read this first

Post by blazecrown »

Agreed they would also be good things to look at. Thanks for the feedback.

I will add things like lot sizes, drawdown, profits etc. It's relatively easy to do. Unfortunately I only have evenings to work on trading until mid December (when I finish my current full time contract).

Will post here when I get some results. Might be a while though. Progress is slow :/

Shona
Hi Shona,

i was thinking if i may add, before those two previous questions, a good question also for the SAS to answer would be:

"Considering volatility, what would be the best profit level X and stop loss level Y to use to maximize Profit in USD but minimize the number of turns taken to hit TP"

as you will notice the higher you go with the number of turns, the riskier it gets approaching its risk of ruin...ideally, i'd like to see number of turns to be less than 5... maximum of 3 turns would be superb...
TheInsider
Trader
Posts: 20
Joined: Wed Aug 06, 2014 10:14 pm

Automatic Loss Recovery System (ALR) - read this first

Post by TheInsider »

Hi guys!

I have now updated my version of the ALR-calculator - ALR-Calculations v. 1.4. It can be downloaded from page #28 in this thread (in order to avoid double posting).

The main difference from the other ALR-Calculators is that it now also offers lot size calculations based on a dynamic TakeProfit reduction, which will make sure that the basket of open hedged positions are closed much faster, and that this TP-reduction can be changed as well as a Minimum TP to be used, and it will present the dynamic TP both graphically and in Points. There's also two examples, one with standard fix TakeProfit and one with this dynamic TP-reduction.

I have also a working ALR-Manager that can manage manually an open position, as well as run backtests and forward tests (set a date and time and it will open an order and manage it), with the option to trail profit. I have also made several hundreds of tests to evaluate ALR as such, and my ALR-manager, and my conclusion is, that this system (ALR) is working extremely well during periods when prices moves enough. But if TP and SL (ALRZone) is to wide it can be very dangerous during periods of flat/horisontal markets. Also, it requires proper settings depending on how the market is moving.

I plan to release this manager, but I'm hesitant as I understand its potential, why making it free for anyone to download and use is not an option. So some kind of registration will be required. Stay tuned!
Post Reply

Return to “Automated Loss Recovery”