Christopher,

have you tested stop loss order placement based on recent volatility using ATR, or BBands, and then trailing the stop as the position moves in your favor? You would probably get a slightly higher win ratio, however your average loss would increase as well, but it may improve the overall results depending on the characteristics of your system.

My advice to you Anthony is the following: No one can extract the last possible cent of profit from a trading system. You may be able to improve your results but you will always leave some money on the table.

In your system, 33% of the time TP is hit, 4:1 RR ratio, but what is the % of time SL is hit?. I think you need to rework this part. Let me know ur comments. Also, where is your entry level, always the S/R.?

you're not alone.

I am coming to the conclusion that I should be happy with whatever I can get and not think too much about what I left on the table.

1 way I found around this is to enter 2 trades, 1st has a fixed profit target , the other moves to be when the first closes and then trails price action. That way I have banked some, and psychologically I can accept whatever happens from then on.

Tightening stops is always one of those "damned if you do, damned if you don't" deals. Did you quantify the probability of the trade going on to become profitable after being stopped at break even ? Compare the amount of profit you would have made in those circumstances with the amount lost if you maintained the original stop. Another thing to do is, depending on your testing platform, actually show a points "path" of each trade starting from zero up to its exit. It may help you fine-tune your tightening process. It may also be very market dependent so one hat may not fit all markets and especially if liquidity is a factor. Some markets get a nice little kick on once they gather steam while others may meet brick walls. Sorry to be vague as these issues usually come down to testing and experimenting. But the caveat of course is to be careful of overfitting. Maybe try a tightening process that is more adaptive to the volatility, time of day, market, higher time frames ( so even if you are trading intraday the price action may also look to daily or even weekly price action for guidance), movements in correlated markets etc.

Hi Christopher,

I am a forex trader - Ive only been trading for a few weeks - but as a mathematician at heart and very intense I've covered a lot of ground. Under my instruction I've had an EA built to work on MT4 - its now in it's 28th iteration. It grew in complexity - as I grew to understand the "true" meaning of the fx technical indicators and how they played together - only to then simplify and focus on a handful of core principals.

I'd like to share my thoughts regarding your question:

Ill try and structure my thoughts:

1. I place 2 orders per trade - both with the same SL but one with a TP set at 1:1 ratio. That way one order may hit the TP whilst the other is left to run.

You could try this with 50% of your lot size placed on each trade - take one off the table once your TP is hit and let the other breath and see where it goes.

2. I trade 1H 4H and 1D. Since the higher time frames need more room to breath (ie a candle can cover more ground in the higher frames) then they require larger SL - ie more risk. So I reduce my lot sizes as follows - 1H = 1 4H = 0.5 1D = 0.3 to manage my risk

Your SL of 15/18 then moving to 10 once BE is hit sounds too tight to me. My min SL is 50 - which is set as trailing so any gains reduce the hit. Setting too tight a SL will only stop you out when you could have picked up more pips.

So I have order 1 SL and TP = 50 with SL BE = 50 once TP is hit for order 2

I also use a SLSTEP which resets the SL to 50 trail every time 30 pips profit is made - again just to allow the order to breath and follow that long trend

Please note - the above figures I use on the 1H - I use greater SL and TPs on 4H and 1D with lower lots sizes to compensate.

3. Lastly - I use SMA/EMA cross as part of my order entry system - I also use the same SMA/EMA cross as part of my exit system. ie if the SL has not been hit then the cross does a good job at calculating when the trend has run its course.

I did try to exit using a different SMA/EMA cross to the one I entered with but I now use the same.

Hope that helps?

May I ask - why do you only trade in the morning - are you hitting the 8am London open for volatility. I used to do the same with 1H but refined my system so that it can run 24/7 - 4H and 1D can run 24/7 anyway.

Sean :)

Thank you all for your valuable comments.

Rob I have used ATR in the past, and thought about that last night. I will give that a try, and also BBands. Thanks for the advice.

Oscar I am coming to that conclusion as well. With a manual system, a bit of discretion seems to always make more than a fully automated system.

Salim the stop is hit 33% of the time, target 33% and breakeven 33%. So basically its a 1:1 W/L system. And with a reward which is greater than 2.5 times the risk, the strategy is solidly profitable. Its the 33% of breakeven trades that I am trying to make something on. Today is a perfect example...five breakeven (meaning there was at least 10 ticks profit) and one target. Target achieved 40 ticks, but all six trades combined, even with a target of 10 ticks, would have earned 60 ticks. The math over a larger sampling still favors the large target, however it would be nice to figure out a way to turn some of these breakeven trades into winners, without giving up too many of the full target winners.

Neville I understand that method. And the strategy as a scalp, short target system is also profitable. But its amazing how the larger targets improve the performance over time. Over 1000 trades, the single large target will outperform the short target by a minimum of 3 : 1, which is why we favor it. Its the Larry Williams "bet small, win big" philosophy. But like you, psychologically, we all would feel better winning 66% of the time, rather than 33% of the time!

Martin we have quantified the probabilities, and have determined moving to breakeven after 10 ticks (for most instruments) gives us the highest profit potential. Our platform does not allow me to plot a points path...I am not certain I understand the concept, but we can program pretty much anything. I would very much appreciate more info on how to plot such a path. And I agree, we are very careful not to curve fit, but do recognize if we are trading CL versus ES, we do not expect the same type of movement or targets. Its probably a good idea to check time of day, ATR and a higher time frame as well. We will give those a try...thank you!

Sean thank you for your detailed advice. We are actually trading Futures instruments, and find the morning session to be most profitable with the best price action. That said, the strategy is applicable to most actively traded markets...anything that moves in a decent range. We use a custom renko bar, set with a reversal bar anywhere from 8 to 16 ticks in size, depending on the market. This is the only "modification" to the system we make. In other words, we do not curve fit the strategy for a specific market. Instead we adjust the renko bar size to reflect the movement of a particular instrument. We have found overall, once a reversal bar fires, the trade will retrace, and our system is likely to fire an entry. Our current stop is hit 33% of the time, and we have tested stops up to 30 ticks, finding an 15 - 18 tick stop provides the right balance of risk/profit. Average time in a trade is just shy of 9 minutes, so its a fairly quick, day trading strategy. I actually considered using a SMA/EMA cross as an exit, but it seems to be somewhat hit or miss determining the period for these moving averages. Any suggestions are welcome.

Thank you all for your contribution to this thread. Its been quite helpful.

Whatever be the reason, human brain can perform better, if he innovatively combine technology/software and intelligence together. The most original super-computer is still human brain (though computer processor outsmarts the speed), probably the declining importance of human brain to computer-based software packages is also another threat to human brain`s dominance. Automated trading systems execute decisions without emotions, but any unexpected/untoward incidents/developments are outside the purview of that computerised system. As market dynamics keep changing, focus on skill development, that is the crux for financial market aspirants.

Rgds

Andrews W Salim

Hi Chris - path of the trade is simply a little line graph that shows the return or equity path of an individual trade. You can visually see its adverse or favourable excursions - a sort of visual representation of the life of a trade. So your graph axes would be time or bars versus return or profit. Now you can overlay all or most of your trades for a strategy on top of each other. It could help provide some guidance as to how aggressive a stop tightening program would need to be etc. Sometimes people test an entry idea by plotting these graphs with a certain hard time stop to examine the rough average life of a typical trade with that sort of entry. People also plot trade PL or returns against time as well - a scatter-like plot - that helps you check that their stops are effective while riding profitable trades to profit targets. I am sure you may have seen these in books. Stridman's book "Trading Systems that Work" shows some examples when demonstrating MAE and MFE.

Sounds like you are doing some good research - good luck.

I apply a martingale strategy and an anti-martingale strategy based on pattern recognition. Hence run an engine to position longs and run engine to position shorts but determine entry level/stop loss/profit margin based on strength of trend

I'd suggest NOT using TP/SL as a part of yor enter/exit logic. The simple explanation (without prob theory) is that the market doesn't care about your wallet. So this exact price is either favorable or not in the moment. No matter you open or close. No matter you are profit or not. I'd suggest 'out-of-model' SL set per each instrument at the level or 2-3 sigma w.r.t. the distribution of losses for this instr. Such an SL is usually NOT hit at all, it only protects from instant market shifts, liquidity crisis and so on.

@Andrey. I like your way of thinking, protection is a most important factor in any trading. However the aim of this discussion is to extract every possible bit of profit out of a existing and workable system. To do this it is required that the protection margins be shaved down to a very slim level indeed. In other words, if you wish to obtain additional profits out of an existing system you will have to increase the levels of risk you take and decrease the levels of day to day protection you have built into the system. Eventually this exercise turns into increasing loss and decreasing profits. People get lost in the math of these discussions, the important thing is the sustainable level of profitability obtained out of the system.

That said, and providing a health insurance analogy, some level of catastrophic protection built into a system, something that you do not expect to ever have to trigger, will be a very good thing to have and will not affect the day to day operation of the system. If a black swan happens you may be very happy to have built this catastrophic protection layer into the trading system. Nice post Andrey, thank you.

Chistopher,

From the description of your system, I see that basically you have SL as the jump function of achieved PnL: ~(-15)bps when maximum achieved profit (MAP) is less than 10bps and 0 when profit achieves watermark of 10bps.

My simple proposal is to try more flexible setup. There are several variants of this:

1. Simplest and good in many cases - trailing stop-loss, e.g. SL always 15bps lower than achieved profit. Thus, when your trade achieves MAP of 35 bps (not achieving TP) and will breakeven with your system, it will give you 20 bps of profit with 15bps trailing stop-loss. On the other hand, when trade loses from the very beginning (MAP = 0), you have your standard (-15)bps SL. I think, this really can help to your system.

2. Try more flexible functions of MAP - e.g. SL is (-15)bps when MAP = 0, SL = 35 bps when trade achieves MAP 40 bps, and linear interpolation in-between.

And final proposition - why do you use fixed numbers for your SL/TP? Imho, better approach is to make them depend on volatility, thus you really condition your TP/SL on returns distribution quantiles and not some "all-vol fit" magic numbers.

When a trade stops out only to a few pips later turn and make a healthy profit, this is a very disheartening situation indeed.

But other than hedge there is really no way around this - thats why I like to run 2 orders at least money is taken off relatively early with a low TP on order 1 - which gives a nice mental boost.

I played poker for years which is all about mind games - forex is no different thats why, for me at least, an EA is the way to go.

My MT4 EA is starting to perform - Ive just picked up over 100 pips on short EURAUD and managed my way out of 2 trades. I did consider having no SL but you need to draw that line.

I don't use a stop loss that the market can see, instead I use a dual stop loss mechanism. The first stop loss warning is set at a particular percentage loss – let's say 14%. Next, what I'm looking for is a higher loss percentage of say 4% more. Once that is achieved, a market order is issued to sell immediately.

I'm using five-minute quotes and the percentage loss is calculated based on the stock option entry price. That's why my percentages might seem a little high to you – I trade options (puts and calls), not stocks.

I use the same mechanism to exit my trades regardless if the trade is profitable or not (of course there are a few exceptions). So if the trade rises 10%, and then falls, you can expect to lose at least 10%. If the trade rises 120%, you can expect your maximum profit to be a little more than 100%. I of course do decrease these percentages as the prophet levels increase, but the decreases are not very significant. I do this in order to "let my profits run" and to "cut my losses". There is no emotion involved in these trades – if a trade was looking good, it doesn't mean you're necessarily going to make any money on it. You must exit the trade when your system tells you to.

I do have another mechanism that I use to override my stop loss mechanism, but it is a little difficult to explain. In simple terms if I receive a quote that is "out of line" with the previous quote or the quote before that, then, regardless of how much of a loss or profit the trade is sitting on, it is not sold yet. What normally happens in this case is that the price of the stock returns back into normal territory – this is because we are in an out-of-line situation or condition. The trade is tracked to make sure it continues back to where it should be – between the confidence interval lines. Many stop loss executions are avoided in this way.

Basically, the trade is given it's "head" like you would give a horse a little leeway when you're out in the pasture riding him. If the horse returns back to the trail on its own, then we don't need to do anything like steer him. Along the way back to the trail, the horse may continue to wander a little bit, but that is allowed to a certain extent. If the horse looks like it's going to start down a different horse trail (new trend starting), the trade is ended right there. Sometimes the losses using this mechanism are quite large (25 – 45%), but I can assure you the gains or loss recoup from wandering trades are significant.

I know, one of your main questions is: "what is out – of – line". All I can tell you is that it is a particular percentage change. This percentage is a factored percentage in that it is based on the stock price – the same percentage is used for all stocks regardless of its price. And this mechanism is also based on the change in the stock price alone, nothing to do with the option price. However, the option price is used to determine when the trade has returned to the trail or has decided to take another trail.

Well I want shooting!

Ive refined my EA to get into a trade as early as possible based upon only a few basic technicals - I was short on both EURJPY and EURAUD on 8th and 9th on the 1H chart - the AUD was quickly up to 60 pips the JPY didnt kick off until later - and what do I do?!!!

Both hit TPs on order 1 but I took TP order 2 at 100 pips on AUD and something low like 3/4 pips on JPY - only for them to slide further and make more pips!

So - from now on Im not going to play with my orders - either they will stop out or hit the sma /ema exit cross.

I've also set myself a rule of thumb to help me "not touch" my trades - here are the averaging running time of each trade:

1H - runs for 5 days

4H - 10 days

1D - 30 days

ie theres no need to keep checking my orders every 10 mins!!!

Live to Trade

Your rule of thump for exiting the trade ..1Hr-run for 5days.....is exciting to see. But have you really experimented the result on a monthly collective basis as what percentage of trade were successful as per ur defined parameters and trading plan?

Hi Christopher,

Your System is running 24Hr on VPS or is meant for specific session like London opening or US oprn or close specifically.? Is it for any specified currency sets ar is it not trading on Fridays or during News releases. What is the set time frame for your system? These infos can help us in fine tuning ur system, as one is handicapped to comment precisely.

Hey Christopher! Honestly it sounds like you have the beginnings to a great strategy. What's important is that your method is working for you on your time! It's great to get advice from experienced traders, but ultimately if you change your strategy drastically, your overall profits could plummet.

That being said, you could reanalyze your equity management. If you see you could risk a little more, try trading with bigger lot sizes (practice in a demo account first if you feel unsure). That could help you take back some of the money left on the table, but keep in mind, there will always be some money left on the table. That is the nature of trading. As long as your profits exceed your losses you are golden my friend.

Hi Chris, in my opinion your exit really depends on the type of strategy you are using. If it is a mean reverting system, you obviously want to exit when the price makes its reversion and that is not necessarily dependent on hard dollar stops and targets, but rather dependent on the reversion itself, which is, among other things dependent on current volatility. So with mean reversion you want to exit the trade before it reverses on you which is sometimes counter intuitive. With trend strategies, you can use your hard stops if you like, but again, you may want to explore a more dynamic stop that will adjust with volatility (like a multiple of ATR) The other consideration would be to buy two lots, taking one lot off early and letting one ride. If the market does not go in your favor, but you are still profitable on the lot you exited early, then hopefully you break even instead of booking a loss. Just some thoughts.