Daily breakout followed by money management

Hi, people.

I saw this strategy in a babypips thread. I tried this live for about 10 days, I grew my account 60 percent.

You place buy and sell orders around daily candle and wait for at least 2 more daily candles to form within that master candle. When the next candle after the 2 candles breaks through the upper or lower boundary a position is automatically opened. Up to here everything is taken from that thread. But I added the following:

After the position is opened(lets say a buy is opened) you place sl at the other side of the master candle and cancel the sell order.
Lets say sl is 100 pips you also set trailing stop to 100 pips( of course the trailing stop and sl can vary depending on the pair, these are for eur/usd) and lets say your position size is 10000 , you divide both your sl and position size into 4. So you have 25 pips and 2500.
Now when the buy position moves 25 pips in your favor you close 2500 of your position size. and when it moves another 25 pips you close another 2500 ps and when it moves 25 pips more you close another 2500. After it has moved 75 pips you move your hard stop to breakeve+25 pips and let the last position ride with 100 pips trailing stop.

The reason this worked ( I made money on gp/yen , eur/usd, …) is daily breakouts give you a minumum of 20 to 35 pips when they break out especially when the range is stuck in the boundaries of a candle for at least 2 days.
Now my reason for posting this is to see if one of you experienced ones could come up with a better money management for this strategy as for(INITIAL SL, INITIAL TRAILING STOP AND WHEN AND HOW MUCH TO MOVE THE HARD SL AND HOW TO DIVIDE THE POSITIONS BETTER[I DIVIDE THEM INTO FOUR]ALSO MAYBE SOMEONE COULD CREATE AND EA FOR THIS STRATEGY WITH ALL THE MM. ANOTHER VARIABLE COULD BE A DYNAMIC TRAILING STOP MEANING THE TRAILING STOP COULD DIMINISH BY A PERCENTAGE AS THE MOVE GOES CERTAIN AMOUNT OF PIPS ETC.)
There could be numerous amount of variances and all would be about ps , sl , tls, mm that could make this strategy better.
The EA would make it possible to test different pairs with different variables to come up with the best setup.

Thanks in advance.

seems like a nice strat, will give it a shot next week.
also hoping some of the vets will chime in.

Question: If the master candle is not breached up or down after 2 more candles form, Do you select the most recent candle as the master?

Good stuff Hibra thanks for the post

good question chartzard. I think that could be added to the system but I also think break outs from the original master candle area would yield bigger moves. But this is just an opinion, thats why this should be back and forward tested with different pairs via an EA.

Also I forgot to mention that this strategy is great for guys like me who has to be in all the time. Its such a solid system that we are using daily candels nothing smaller yet we have 28 pairs to look for setups. Thats alot of pairs, if one doesnt have a setup another will which is good for impatient players. Its also good that we get to take out profits early in the moves.
I think a strategy should fit the trader. When have 3 or 4 setups waiting during the week 1 of them definetely plays out and that keeps me from being bored.:slight_smile:

I also dont want people thinking that I’m claiming that I have great succes. Its just a new and simple strategy, I might have been just lucky which is what I’m trying to find out.

This is a [B]very [/B]basic Inside Bar trading strategy.
Your success @ trading this way will be [B][I]directly [/I][/B]related to whether or not you go the extra length to understand:

a) What this formation is representative of
b) What types of market [U]environments [/U]are best for this type of formation
c) What instruments are best for this type of formation
d) Why (in the long run) [B]simply [/B]trading the markets as you lay out in you original post will never be profitable

You got in before me, and said the exact things that occurred to me, on reading the OP, so I’ll say “good post” instead. :slight_smile:

I can add a couple of things … (well, three, actually …)

  1. The underlying basis of this is a very sound, reasonable and good one, and it certainly [B][U]is[/U][/B] possible to construct reliable methods along these lines, but it will take a [I]lot[/I] of research and testing. As the saying goes, “the devil is in the detail”.

  2. Don’t confuse an “entry-system” with a trading system. The money-management (including SL’s, but targets also) will ultimately determine whether or not it’s profitable, overall, and you’ll need methodically and meticulouosly to test [B][U]very[/U][/B] big sample-sizes for your findings from each to be evidential and/or predictive. There’s not much point “theorising” about the possibilities: if you want to make something tradable out of it, you need to do the investigative work, too (and that means knowing how to do it, including which parameters to monitor, and having the right software and data). The partial-closure method you’ve mentioned above sounds interesting and complicated, but I don’t instinctively see any reason at all for it being a good idea, so I’d certainly want to see [I]very[/I] convincing explanations for it being a sensible one, myself, before being willing to put in the work required to investigate that.

  3. From my own experience (and I traded a similarly based method on both Cable the Euro for over 3 years, with very few and small losing months during that period), I’ll be [U]extremely[/U] surprised if a [I]trailing[/I] stop turns out to be the best approach to this.

Good luck!

Hi, tnx for the reply.

The trailing stop is actually is equal to the hard stop in the beginning of the trade. The trailing stop really becomes help full after the 3rd tp position takes place and we are left with 1 fourth of the trade running with trailing stop.
As I said before I think this method works well because its a daily break out(higher timeframe) and we take partial and early profits out of the market.
But coming back to what I want to learn from experienced traders is that what works best? a hard sl at the other side of the master candle or at midway or whatever.
I think what will determine this is the percent of success. If after the break out the market goes straight back down right away more than it goes in the direction of the breakout than its certainly a losing system.
So I will keep posting about this system and use it as it is. But I wish someone could try different variations and report back. I mean u can change Sl, trailing stop, the way you divide take profits, the percentage of tp and so on. A nice strategy could be built on the main idea of a daily master candle breakout. And also keep in mind its not just a simple daily breakout. We wait for 2 daily candles to be inside the range of the first master daily candle. The longer the tight range the bigger and more sudden the breakout.
tnx for the interest.

Forex unlimited tnx for the reply as well.
Since you said "Why (in the long run) simply trading the markets as you lay out in you original post will never be profitable"
then you know the answer already. Please feel free to explain it to me why?
And u must also know the opposite;“why another strategy will always be profitable” please kindly give me that strategy as well.
thanks again

Sure, the answer is very simple, however, you need to quote the exact statement that I made:

Bottom-line; if it were as easy as you lay out, the washout rate of mostly every trader wouldn’t be where it is. In other words, everyone would be making money hand over fist with no worries in the world.

No problem- I’m of the opinion that algo-trading and any other type of automated strategy can never be profitable in the long run, if employed time-and-time-again the exact way it was originally designed. Why? Markets are constantly changing based on the global landscape. 1 out of every 100 strategies may be able to brave the storms, but over time, they all will fail.

Again, it’s also in my opinion that the only truly long-term profitable strategy a retailer can employ is based on understanding price action. Being able to read price action and what the underlying order flow is indicative of provides a real-time connection w/ exactly what is happening, regardless of any macro, micro factors. The current market price reflects every single piece of information possibly available to every single trader baked in.

Have you [I]demonstrated[/I] and [I]proved[/I] that, over a statistically significant data-sample, or is it just your [B]guess[/B] from what you’ve seen so far?

Do you [U]know[/U], for example, that doing that is more profitable over the long-term than dividing into (say) three lots and letting the third lot run until it crosses (say) a displaced (right-shifted) SMA of the median price? I ask because my money, from my own backtesting experience, would be firmly on the latter, and I suspect your assessment criteria are very different from and very much more subjective than mine.

I used to think that, as well, but discovered on testing that it actually worked much better on a shorter time-frame than that, taking the RTH of different regions into account, partly because the trading opportunities were very significantly higher, that way.

Sometimes it does and sometimes it doesn’t.

The point is that you’d be well advised to research and test that for your own exact system, rather than trying to make assumptions based on others’ experience of what [I]they’re[/I] doing.

I know it sounds uncooperative and unhelpful (and apologies for that), but [I]that’s the way it is[/I], I’m afraid, and - respectfully - it sounds like that’s something you need to learn about trading-system design, before you go too much further.

This I disagree with more or less completely. In fact I’ll be astounded if win-rates are the outcome-determining factor for a method of this kind (and I’ve backtested a [U]lot[/U] of them, very methodically and meticulously, over large data sample-sizes).

There’s no “certainly” about it, I’m afraid. You might easily find that it turns against you 60%/65% of the time, but the remaining 35%/40% of winning trades are each, on average, three or four times the size of the average loss.

If you start out with the [I]fixed belief[/I] that “if, after the breakout, the market goes straight back down right away more often than it goes in the direction of the breakout, then it’s [U]certainly[/U] a losing system”, you’re arbitrarily going to exclude a lot of profitable systems based on this overall method from your consideration.

Indeed.

That’s among the tasks that trading institutions employ highly qualified, experienced analysts to do, because it’s a hugely time-consuming and rigorous investigation-process. No impoliteness expected, but to make a profitable system out of this, you’re almost certainly going to have to be willing to learn how to do that for yourself. It’s a [B]hugely[/B] worthwhile skill to add to your armoury.

Indeed. Exactly so. There are many variables, and whoever does the testing will have to start by selecting a reasonably representative set of each, to look at, and refine the selection-process as they go along. And that’s a skill-set of its own. Otherwise, realistically, they’re just [I]guessing[/I] (which is honestly what you’re doing at the moment - sorry!).

Yes, my guess is that that’s probably true … and that you’ll be able to find a much “nicer” (more stable, more profitable and more frequently-trading) strategy using a smaller time-frame. I found this very consistently.

In its simplest terms, it’s much better, much safer and much more reliable ([I]as well as[/I] obviously much more profitable) to trade four times a day and make an average of 10 pips’ profit per trade than it is to trade once a day and make an average of 25 pips’ profit. It also, effectively, hugely reduces the sample-sizes that you need to analyse, to achieve statistical significance, and it makes your method more robust, too.

Good luck!

Thanks for your reply again.

[I][B][I]"No problem- I’m of the opinion that algo-trading and any other type of automated strategy can never be profitable in the long run, if employed time-and-time-again the exact way it was originally designed. Why? Markets are constantly changing based on the global landscape. 1 out of every 100 strategies may be able to brave the storms, but over time, they all will fail.

Again, it’s also in my opinion that the only truly long-term profitable strategy a retailer can employ is based on understanding price action. Being able to read price action and what the underlying order flow is indicative of provides a real-time connection w/ exactly what is happening, regardless of any macro, micro factors. The current market price reflects every single piece of information possibly available to every single trader baked in."[/I][/B][/I]

In your above statement I do not see any specific strategy that will be profitable in the long run let alone a strategy that will be profitable all the time.
These are run of the mill vague explanations of how a strategy should be or what a strategy should be based on.
tnx anway.

Quote Originally Posted by hibra68 View Post
The trailing stop really becomes help full after the 3rd tp position takes place and we are left with 1 fourth of the trade running with trailing stop.

[B]"Have you demonstrated and proved that, over a statistically significant data-sample, or is it just your guess from what you’ve seen so far?

Do you know, for example, that doing that is more profitable over the long-term than dividing into (say) three lots and letting the third lot run until it crosses (say) a displaced (right-shifted) SMA of the median price? I ask because my money, from my own backtesting experience, would be firmly on the latter, and I suspect your assessment criteria are very different from and very much more subjective than mine."[/B]

I do not know lexy. This is why in my post I stated that this worked for me and it was 10 days since I started using it and that I need people to come up with even better mm,sl,tp etc. And Since you backtested and came up with better results in dividing the position into 3 and letting the third run with sl that is a great input. Thats the kinda input I’m looking for.

Quote Originally Posted by hibra68 View Post
I think this method works well because its a daily break out(higher timeframe)

[B]“I used to think that, as well, but discovered on testing that it actually worked much better on a shorter time-frame than that, taking the RTH of different regions into account, partly because the trading opportunities were very significantly higher, that way.”[/B]

That is a great input as well. So I will test this on breakouts of different opens ie.(london, ny,…)
But when I look at my daily candles (I live in NY and I think my broker uses gmt-2)alot of times magically the price stop at or very near the previos days candle. That is why this strategy was interesting to me. But using London or NY open and close with 1hr candles or 4hr candles might work better.

Quote Originally Posted by hibra68 View Post
what I want to learn from experienced traders is that what works best? a hard sl at the other side of the master candle or at midway or whatever.

[B]"Sometimes it does and sometimes it doesn’t.

The point is that you’d be well advised to research and test that for your own exact system, rather than trying to make assumptions based on others’ experience of what they’re doing.

I know it sounds uncooperative and unhelpful (and apologies for that), but that’s the way it is, I’m afraid, and - respectfully - it sounds like that’s something you need to learn about trading-system design, before you go too much further."
[/B]
Yes well that is alot of programming that I cant do by myself. I’m certainly not going to go crazy and try to manual test with years of data. This is why In the original post I wanted someone to write an EA to test this strategy. But it could be segmented. Meaning someone could just write an EA that tests the above situation(How often does the price go all the way back down to the other side of the master candle when it breaks out as opposed to not) That will be tested on different pairs.etc. I can list the different parameters to consider for this if someone is interested. EX(the EA can exclude certain days and times of the week,month…)

Quote Originally Posted by hibra68 View Post
I think what will determine this is the percent of success.

[B]“This I disagree with more or less completely. In fact I’ll be astounded if win-rates are the outcome-determining factor for a method of this kind (and I’ve backtested a lot of them, very methodically and meticulously, over large data sample-sizes).”
[/B]
I do not know about that because I didn’t back tested (I still don’t have an EA) But I mentioned above in replying to your statement that if someone can write an EA we would know. I alos would like to know what is the determining factor. Is it %of risk or something else?

Quote Originally Posted by hibra68 View Post
If after the break out the market goes straight back down right away more than it goes in the direction of the breakout than its certainly a losing system.

[B]"There’s no “certainly” about it, I’m afraid. You might easily find that it turns against you 60%/65% of the time, but the remaining 35%/40% of winning trades are each, on average, three or four times the size of the average loss.

If you start out with the fixed belief that “if, after the breakout, the market goes straight back down right away more often than it goes in the direction of the breakout, then it’s certainly a losing system”, you’re arbitrarily going to exclude a lot of profitable systems based on this overall method from your consideration."[/B]

Again as I said above someone has to write an EA to test this out.( But what I meant was if it goes all the way back down and hit sl without breaking out far enough to reach our take profits and trailing stop to cover the breakeven. Again we could try many different scenarios for this but it has to be done with EA’s.

Quote Originally Posted by hibra68 View Post
I wish someone could try different variations and report back.

[B]"Indeed.

That’s among the tasks that trading institutions employ highly qualified, experienced analysts to do, because it’s a hugely time-consuming and rigorous investigation-process. No impoliteness expected, but to make a profitable system out of this, you’re almost certainly going to have to be willing to learn how to do that for yourself. It’s a hugely worthwhile skill to add to your armoury."[/B]

Yes kinda the whole point of my posting here. See I know all the above stuff u mentioned but there are actually people that exist in these forums that will spend time and write EA’s for these tests. I’m just waiting for one to see the post.:51:

Quote Originally Posted by hibra68 View Post
I mean u can change Sl, trailing stop, the way you divide take profits, the percentage of tp and so on.

I[B]ndeed. Exactly so. There are many variables, and whoever does the testing will have to start by selecting a reasonably representative set of each, to look at, and refine the selection-process as they go along. And that’s a skill-set of its own. Otherwise, realistically, they’re just guessing (which is honestly what you’re doing at the moment - sorry!).
[/B]

See above reply.

Quote Originally Posted by hibra68 View Post
A nice strategy could be built on the main idea of a daily master candle breakout.

[B]"Yes, my guess is that that’s probably true … and that you’ll be able to find a much “nicer” (more stable, more profitable and more frequently-trading) strategy using a smaller time-frame. I found this very consistently.

In its simplest terms, it’s much better, much safer and much more reliable (as well as obviously much more profitable) to trade four times a day and make an average of 10 pips’ profit per trade than it is to trade once a day and make an average of 25 pips’ profit. It also, effectively, hugely reduces the sample-sizes that you need to analyse, to achieve statistical significance, and it makes your method more robust, too.

Good luck!"[/B]

Yes but who has time to trade four times a day. As far as sample sizes i don’t know what u mean. You can test 365 candles/days vs (lets say 4hr strategy=6x365=1460) and there is significance in daily candle what is the significance in say every 4 hr candle? what does it represent? Market openings I understand very significant and I will try that.

thanks for the lengthy reply.

I hear you, but be aware that the only people who can do that are probably people who have backtested large numbers of variations of this kind of trading. And even then, it’s only going to be second-hand information, when you get it (not that that necessarily invalidates it, of course). Which is why the ultimate answer to these questions is always to learn to do it for yourself. You don’t have to be a “techie”, to do that, I promise you. You just need some forex backtesting software and adequate data (neither is expensive).

“Writing an EA”, in my opinion, is a positively dreadful way to test methods like this with multiple variables. It might tell you whether or not what you’ve got “works” (i.e. has an edge, though even that’s often far from trivial to tell) but it can’t tell you what’s better/best.

[B]Expectancy[/B] (not determinable by win-rate alone).

A method that works well this month may have lost money heavily last month. A method that works well this year may have lost heavily in other years (i.e. “and can again, tomorrow”). You need some sort of robustness-test, and one approach is to backtest data from different, randomly selected time-periods. Regarding working out expectancy, the higher the number of trades whose outcomes you include in your testing-sample-size, the more reliable the results will be.

An engineer, a physicist and a mathematician are travelling together by train from London to Edinburgh. As the train crosses the England/Scotland border, it goes past a brown cow in a field. “Oh look”, says the engineer, “All the cows in Scotland are brown”. “Over-generalised nonsense”, says the physicist - “You should simply observe that [I]some[/I] of the cows in Scotland are brown”. “That’s over-generalised too,” says the mathematician - “We should conclude only that there exists in Scotland at least one cow, [I]at least one side of which is brown[/I]”. :slight_smile: