Stop Loss experiment

[ChattiFX]

Hi,

Having worked back through May's trades I am running an experiment through June to hopefully tweak my initial Stop Loss value. I just thought I would share the reasons behind this, how I am running it and the results at the end of the month. The purpose is not to give anyone a rule for their stop losses, as it all depends on the individual system obviously, it is more to illustrate how I test and adapt my strategy over time. Hopefully it will be of interest to some and maybe help others.

I won't go in to the numbers too much as this is more about the principle of developing a robust system.

I trade a basket of 10 Forex pairs, using an automated system, 5 min chart for the signals. Risking 1% per trade, worked out from the size of the Stop Loss. i.e. when the SL is hit I lose 1% of my capital. By using 10 pairs I think it speeds up the evaluation process as during the month some will be trending others will be sideways, others will be against the larger trend etc etc...

It follows that the size of the SL dictates the size of the position I take. So bigger SL = smaller position and vice versa.

The Reason:
In May I had a return of ~27%. Which I am pretty happy with. Of that I had 40 winning trades, 36 Breakeven & 134 Losing. Average winning trades roughly four times more profitable than losing trades.

My SL is a multiple of the ATR, which means it takes in to account the volatility of the individual pair.

I am looking to reduce the size or amount of loses.

The Theory:
Amongst a million other ideas, I have two thoughts on how to increase profitability.

1) Increase my SL value so that it is hit less in a sideways market, trapping less losses, but on the downside reducing the value of my wins.
2) Decrease my SL so that although it will be hit more, my winning positions will be bigger.

The Experiment:
I run three Accounts. Two demo's and one real money.
Real - Continue current strategy (Benchmark) - SL=5xATR
Demo 1 - Increase SL value - SL = 7.5xATR
Demo 2 - Decrease SL value - SL = 2.5xATR

All other parameters, entry's and money management are the same.

At the end of the month I will post the results.

For the purposes of the post I'm not interested in sharing the ins and outs of the system, this is purely a post to hopefully give an idea of how I develop my strategy over time.

Any thoughts or comments welcome.

Cheers.

CFX

 

[NVP]

Hi,

Having worked back through May's trades I am running an experiment through June to hopefully tweak my initial Stop Loss value. I just thought I would share the reasons behind this, how I am running it and the results at the end of the month. The purpose is not to give anyone a rule for their stop losses, as it all depends on the individual system obviously, it is more to illustrate how I test and adapt my strategy over time. Hopefully it will be of interest to some and maybe help others.

I won't go in to the numbers too much as this is more about the principle of developing a robust system.

I trade a basket of 10 Forex pairs, using an automated system, 5 min chart for the signals. Risking 1% per trade, worked out from the size of the Stop Loss. i.e. when the SL is hit I lose 1% of my capital. By using 10 pairs I think it speeds up the evaluation process as during the month some will be trending others will be sideways, others will be against the larger trend etc etc...

It follows that the size of the SL dictates the size of the position I take. So bigger SL = smaller position and vice versa.

The Reason:
In May I had a return of ~27%. Which I am pretty happy with. Of that I had 40 winning trades, 36 Breakeven & 134 Losing. Average winning trades roughly four times more profitable than losing trades.

My SL is a multiple of the ATR, which means it takes in to account the volatility of the individual pair.

I am looking to reduce the size or amount of loses.

The Theory:
Amongst a million other ideas, I have two thoughts on how to increase profitability.

1) Increase my SL value so that it is hit less in a sideways market, trapping less losses, but on the downside reducing the value of my wins.
2) Decrease my SL so that although it will be hit more, my winning positions will be bigger.

The Experiment:
I run three Accounts. Two demo's and one real money.
Real - Continue current strategy (Benchmark) - SL=5xATR
Demo 1 - Increase SL value - SL = 7.5xATR
Demo 2 - Decrease SL value - SL = 2.5xATR

All other parameters, entry's and money management are the same.

At the end of the month I will post the results.

For the purposes of the post I'm not interested in sharing the ins and outs of the system, this is purely a post to hopefully give an idea of how I develop my strategy over time.

Any thoughts or comments welcome.

Cheers.

CFX

hey Dude ...............sure perhaps tweaking S/L may help ....but you'll probably make more money by looking at the winning vs losing performance of the pairs being traded at any one time and evaluating their correlation factors
just a thought :smart:
N

 

[c6ackp]

When you say you risk 1% - is that for the whole basket, or per individual trade?

One month's results could just be noise - can you backtest several years to check the viability of your automated system? the 4:1 avg r:r figure doesn't sound right - are you sure? What's your avg initial stop on eurodollar?

Out of interest, which API are you using?

Good luck with your project - the results should be interesting...

 

[ChattiFX]

hey Dude ...............sure perhaps tweaking S/L may help ....but you'll probably make more money by looking at the winning vs losing performance of the pairs being traded at any one time and evaluating their correlation factors
just a thought :smart:
N

Thank you for your reply and suggestion...

At the moment I'm just trying to focus on one "tweak", otherwise one thing affects another and so on... it becomes a nightmare to keep track of what is and isn't improving things.

I have a long list of potential tweaks and obviously this is going to take a lot of time. By using a basket of pairs I find it speeds the process up. Once I have looked at this and a few other ideas I will be looking at how to rule out pairs that are less likely to perform.

Thanks again for your reply... always interested in others ideas.

Cheers.

CFX

 

[ChattiFX]

When you say you risk 1% - is that for the whole basket, or per individual trade?

One month's results could just be noise - can you backtest several years to check the viability of your automated system? the 4:1 avg r:r figure doesn't sound right - are you sure? What's your avg initial stop on eurodollar?

Out of interest, which API are you using?

Good luck with your project - the results should be interesting...

Thanks for the reply, I'll try to answer as best I can...

- I risk 1% per trade... however... my basket is all combinations of 5 currencies, giving 10 pairs in total, so realistically when one currency is very strong/weak I am really risking 4%... I'm not sure if this is common practise and I wouldn't necessarily suggest it is a good long term strategy (It is on my list to prove/disprove this out), however for evaluating ideas I find it shows what's working and isn't working quite quickly as there are a lot of different patterns, trends, movements going on at once across the basket. Its not really a problem for me if a few pairs are losing as is it showing me market conditions where my strategy doesn't work, and giving me ideas for a strategy that would... Hope that makes sense :-/

- Agree one month isn't ideal, hence using 10 pairs. Also I should point out, only entry is automated my exit is manual so I cannot back-test accurately.

- R:R is correct.

- EUR/USD Initial Stop - For May average SL was 27 pts.

- API - I use an EA for MT4. I had it written for me based on my strategy. Works well, although I need to get my **** in to gear and learn how to code it myself. It took about a dozen iterations before I was happy with it.

Thanks again for the comments, I hope my answers make some sort of sense...

Cheers,

CFX

 

[BobbyBB]

1% risk per trade looks far too much to me, but it does depend on your good your mental game is.

134 losses in a month, which means it would be possible to have 10-20 in a row, which equates to a 10%-20% drawdown. Many say they can handle a 20% drawdown no problem, but if/when it comes they get scared and the whole system falls down on itself as trades aren't taken.

If you're expected to make around 20% a month why not consider dropping the risk to 0.50% and let compounding make up the difference over time?

Something to think about anyway.

 

[ChattiFX]

1% risk per trade looks far too much to me, but it does depend on your good your mental game is.

134 losses in a month, which means it would be possible to have 10-20 in a row, which equates to a 10%-20% drawdown. Many say they can handle a 20% drawdown no problem, but if/when it comes they get scared and the whole system falls down on itself as trades aren't taken.

If you're expected to make around 20% a month why not consider dropping the risk to 0.50% and let compounding make up the difference over time?

Something to think about anyway.

Hi Bobby, thanks for the thoughts...

You're estimate is right and I do get drawdowns of 10-20% at times. One of the things I like about using an automated approach is that it takes away the human element of opening a trade so this doesn't affect me too much, occasional on profit taking though I have to admit I can close to early... I am still struggling to automate the close... that's for another experiment though

Although I am trading quite a few pairs so the drawdown caused by one currency is normally being compensated by another. For example over the last few weeks the JPY hasn't been great for my system (breakeven at best) but the AUD has more than compensated.

The difficult bit is that losing trades happen quickly (5-60 mins) and winning trades happen over a longer period of time (24-72 hours) so whilst on paper my account is looking healthy in actual fact the unrealised profit is exactly that, unrealised, so until it is banked I am showing a fairly heavy loss.

Thanks again, all opinions welcome and considered

I'll be updating the post this afternoon to give results so far.

Cheers.

CFX

 

[NVP]

Thank you for your reply and suggestion...

At the moment I'm just trying to focus on one "tweak", otherwise one thing affects another and so on... it becomes a nightmare to keep track of what is and isn't improving things.

I have a long list of potential tweaks and obviously this is going to take a lot of time. By using a basket of pairs I find it speeds the process up. Once I have looked at this and a few other ideas I will be looking at how to rule out pairs that are less likely to perform.

Thanks again for your reply... always interested in others ideas.

Cheers.

CFX

and I assume yuo've looked at the basket trading work by Steve hopgood , Dreamliner , Hanover and some of the other guys at FF ?

N

 

[NVP]

The other problem usually with a basket approach is in truth a holistic stop loss is needed across the whole portfolio verses individual pairs S/L ............

individual stops will hit and you probably dont need it to happen as other trades will be sometimes compensating the losses

tough one to crack unless you have serious automation or just let her run without safety nets ....

N

 

[ChattiFX]

The other problem usually with a basket approach is in truth a holistic stop loss is needed across the whole portfolio verses individual pairs S/L ............

individual stops will hit and you probably dont need it to happen as other trades will be sometimes compensating the losses

tough one to crack unless you have serious automation or just let her run without safety nets ....

N

Thanks, I haven't looked in to the sources you mention... but will do now! thanks for the recommendation.

With regards to trading a basket of pairs. I'm not necessarily going to do this long term, I need to do a lot more research and analysis before then... the main purpose for doing this at the moment is that I find it highlights good and bad points of a strategy quite quickly.

So for example if I am testing ideas on one pair, a good example being EUR/AUD which I would say has followed two trends nicely since early Feb, then I might now think I have cracked it, however it isn't going to behave like that forever and as soon as the general "shape" (trend, sideways, extreme volatility) changes the chances are the strategy will fall apart, so it may take 6, 12, 18 months before I could test all shapes fully, then I'd have to make changes and then continue to test for months. By using 10 pairs I am testing against a broader range of movement.

So whilst the EUR/AUD is moving nicely at the moment (relative to my strategy at least), the EUR/GBP over the same period is less so and is a nice test against a sideways movement. GBP/JPY is behaving slightly different again, giving me some high wins but then taking it all back... I hope that makes sense...

Ideal world I can then adapt so as a pairs movement changes I am at least breakeven on sideways, maximising profit on strong trends and minimising loses and taking the odd profit on volatile movement. That's the goal anyway!!

Again thanks for the suggested research I will definitely be looking in to it more.

Cheers,

CFX

 

[ChattiFX]

Ok, quick update for the week... ideally I'd do it later but it's too nice out and the pub garden is calling, where I think I will probably end up staying for the rest of the day

Real - SL=5xATR - 45%
Demo 1 - SL = 7.5xATR - +67%
Demo 2 - SL = 2.5xATR - +109%

Ok, so a few comments.

Demo 2 performed great all week, in fact around midday today it was over 230% up!! HOWEVER it also looks like it will prove my gut thought that it will have horrendous drawdowns under the wrong conditions... it lost half of it's weekly wins in a little over 3 hours this afternoon!!

Demo 2 and the Real money account both performed nice and consistently all week with Demo 2 peeking at about + 95% before the JPY bounced at lunchtime.

I'd say its been an unusually good week for my strategy with the AUD, USD & JPY making some nice moves. I wouldn't expect these sort of returns to be consistent in the long run and strange as it sounds I could do with a poor week over the next few weeks, lots of sideways movement and volatility, to really expose the weaknesses. I might be wrong but I would have though a lot of people had good weeks this week?

At the start of the week I had in my head that the slightly larger stop on Demo 2 would be the best call long term, but we will see.

Anyway, good weekend to all... enjoy the sun

Cheers

CFX

 

[BobbyBB]

One of the things I like about using an automated approach is that it takes away the human element of opening a trade so this doesn't affect me too much

But who controls the computer, you or the computer?

You do, which is why you can override the system even if it wants to trade. Fear and greed then is no different than if you used no system and picked your trades straight from the chart.

I still maintain that a 20% drawdown is far too much risk to take. 10% should be the maximum for most people as successful trading is probably more a mental game than anything else assuming you have the skill. But as I said in another poster it does depend on your mind, some can handle 10% or even 30% no problem, others can't sleep etc.

 

[the hare]

I'm surprised no ones asked which timeframe your using for ATR

Just out of interest, are the optimum ATR levels reasonably consistent across all pairs or are they differing significantly, or is it too early to tell ?

 

[ChattiFX]

But who controls the computer, you or the computer?

You do, which is why you can override the system even if it wants to trade. Fear and greed then is no different than if you used no system and picked your trades straight from the chart.

I still maintain that a 20% drawdown is far too much risk to take. 10% should be the maximum for most people as successful trading is probably more a mental game than anything else assuming you have the skill. But as I said in another poster it does depend on your mind, some can handle 10% or even 30% no problem, others can't sleep etc.

I do understand what you are saying, although I do slightly disagree that trading is more a mental game though.

Actually I take that back to an extent, for some it maybe for others not so much. My approach is to try and be as logical as possible, set the rules, let the computer follow them and if they don't work adapt them one at a time until the system works, ideally I want to get to a fully automated system. I am still using manual exits although I have a few ideas how to automate, I'll open a couple more demo accounts and run alongside.

CFX

 

[ChattiFX]

I'm surprised no ones asked which timeframe your using for ATR

Just out of interest, are the optimum ATR levels reasonably consistent across all pairs or are they differing significantly, or is it too early to tell ?

Ha yer I'm surprised too now you come to mention it... 5 min timeframe for the ATR.

I'm also surprised I haven't been asked what I'm selling (nothing btw) or to prove my results as that seems to be the normal reaction to most posts

Re optimum ATR levels, it is a good question but too early to tell... however from a purely visual point of view I'd say it is fairly constant, although I've done no proper analysis of this (yet) but it is doesn't look to be wildly different across the various pairs.

Cheers,

CFX

 

[ChattiFX]

Ha... you do get what you wish for sometimes!! So today has been a pretty flat day so far for my pairs.

Real and Demo 2 are fairly stagnant, a little bit down... Demo 2 however is on its ****! to put a few numbers on it.

Demo 2:
Starting capital - $6,500
Peak after 5 days - $23,000
Current balance - $8,500

And the majority of that drawdown has come in the past day and a bit!! Wins are nice but there are just so many (small) losses that any type of sideways movement cripples it.

As I said I asked for it and at least I am getting some useful information. Demo 1 looks to be performing as I want.

Cheers,

CFX

 

[robocop]

Can you post some equity graphs for the demo and other statistics?
Maybe you can use mt4i.com to give some more in depth statistics of your system.

Cheers

 

[BobbyBB]

I do understand what you are saying, although I do slightly disagree that trading is more a mental game though.

Actually I take that back to an extent, for some it maybe for others not so much.

Well, if you don't agree that it's mental (for you) and it's all about method then you are one of the few and if you can deliver a half decent method (it doesn't have to be any better than that if the metnal side is no so relevant) you really have a great future ahead of you, acres of Diamonds are coming your way so to speak.

But my bet is the mental side will get you at some stage and it will sneak up on you, then hit like a locomotive and it will stop you achieving the results your strategy should deliver (taking the mental side out). This is not personal directed at you, I'd say it to anyone.

It's all a good debate anyway. Good luck :cheesy:

 

[ChattiFX]

Can you post some equity graphs for the demo and other statistics?
Maybe you can use mt4i.com to give some more in depth statistics of your system.

Cheers

I have been using myFXbook, which I quite liked but then recently they have started having problems with updating so mt4i.com looks interesting, thanks for suggesting!

I'll have a look, although the reason for my post is more about sharing how I test and adapt my system rather than the actual results, if you get what I mean... however fair point and I will have a look.

Cheers.

CFX

 

[ChattiFX]

Well, if you don't agree that it's mental (for you) and it's all about method then you are one of the few and if you can deliver a half decent method (it doesn't have to be any better than that if the metnal side is no so relevant) you really have a great future ahead of you, acres of Diamonds are coming your way so to speak.

But my bet is the mental side will get you at some stage and it will sneak up on you, then hit like a locomotive and it will stop you achieving the results your strategy should deliver (taking the mental side out). This is not personal directed at you, I'd say it to anyone.

It's all a good debate anyway. Good luck :cheesy:

No problem, I'm not taking it personally... and in fact the very reason for trying to develop a fully automated system is to negate the mental side, as I identified something along those lines as one of my personal weakness a while ago... so I do pretty much agree with what you're saying...

Thank you for your thoughts, always interested... and of course good luck to you too!!

CFX