Reverse Engineering a Profitable System (REPS) | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Very Profitable System - Roi system 355 replies

Reverse engineering order flow from naked chart 18 replies

MA, Gann and Ichimoku system (profitable system!) 60 replies

Fractal Reverse Engineering Journal 21 replies

METATRADER REVERSE ENGINEERING 3 replies

Attachments: Reverse Engineering a Profitable System (REPS)
Exit Attachments

Reverse Engineering a Profitable System (REPS)

  • #61
  • Aug 10, 2015 10:05pm Aug 10, 2015 10:05pm
Quoting spyderman
{quote} MK, Appreciate your thoughts. Have to say I disagree however. You speak of randomness producing trends as if that would somehow discount the opportunity. There's much debate here on FF and elsewhere as to how random the market is or isn't. As I stated earlier, I just don't care. It is immaterial how/why long tail market movement happens to me, the point is that it can be traded. Those trends happened, random or not, and they will happen again. The "stats" that are important to me is what the price did in fact do and, at least in my mind,...
Sure, randomness will happen again without a doubt. My point was strictly about a starting point for system development. The stats you speak of that are important to you....well I haven't seen you post any yet. I have seen many many people developing a system like this: grab some random hindsight data that looks good and they wish they could have partaken in, develop a set of guidelines/rules that seems to work on their cherry picked examples, walk it forward in real-time or worse trade it with real money, after some time has passed they wonder why the system is not profitable and it doesn't come close to matching their exploration.

What you are doing in this thread, is the common way I have seen traders develop a method/system. If instead you started with a proven statistical market tendency, you can at least start the long and time consuming process of developing a system with a chance. I've been full-time trading for 12 years now and have seen this play out many a time and been subject to it myself many a time as well.

I'm not trying to insult you or poopoo on your efforts. I'm trying to save you some time so you can re-direct efforts towards something with a chance.

Nothing else to do but go back into shadow I guess. Good luck ahead.
  • #62
  • Aug 11, 2015 2:18am Aug 11, 2015 2:18am
Quoting MidKnight
{quote} Sure, randomness will happen again without a doubt. My point was strictly about a starting point for system development. The stats you speak of that are important to you....well I haven't seen you post any yet. I have seen many many people developing a system like this: grab some random hindsight data that looks good and they wish they could have partaken in, develop a set of guidelines/rules that seems to work on their cherry picked examples, walk it forward in real-time or worse trade it with real money, after some time has passed they...


Hello MidKnight,

i would like to step in in your favour.
The thing is that you're right, he's picking up a cherry picked scenario. But i can also say that this isn't that bad.
You want a system that works in such an environment because you want to extract as many pips as you can from THAT scenario.
No doubt that the system won't work with a ranging or non trending market. Actually, it won't be succesful also in a weak trending scenario. But those are the periods of time in which you just want to preserve your capital.

Does anyone remember Graeme's thread? It was illuminating. The only thing you need is to keep your risk controlled. Then the inevitable will happen.
Of course i've been Always skeptical about his methods, but they can theoretically work.

I also would like to point out that we could further improve our odds of being right. How? Using monthly pivot points. Heres a little part of my findings (only on GJ, so Im assuming something that might be wrong based only on this CCY pair)

The probability that the month will be green If Open> PP (and viceversa) is slightly above 50%.

Also, 75% of the times the PP is touched. This means that if we open higher than it we need to know that 75% of the times the price will go down and touch the PP.

Another thing to exploit is R1/S1.

These are, also visually, points in which usually the price stalls or retraces. So why should we open there a position knowing that probably the next day we will be able to open at a lower price?
Were away from S1,R1? Ok lets open at break of DC. We are on them? Lets wait. Simple huh?

I have an excel sheet with all these statistics if youre interested.

Also, please tell me if this contribution is not appreciated.
  • #63
  • Aug 11, 2015 2:29am Aug 11, 2015 2:29am
Quoting Copernicus
Hi spyderman. Nice thread. :-) We are on similar missions. The shorter term .....
Great post, thanks for sharing.

I didn't understand one thing. When you talk of 200 do you mean the DC 200 on daily charts? Essentially you're right, but if AUDUSD changes trend right now we won't be able to jump on it until 40 weeks.

About the rest, i think you're essentially in line with the following: NO RISK NO REWARD.

But at this point, having so much risk, are we actually sure that a simple entry method based on engulfing, inside bar, Flying buddhas isn't better?

We would have much lower risk per trade, and we could only select opportunities in a trending environment. Example:

Break of the DC but than retraces, IB/OB/FB bullish and we jump in.

If we want to not overtrade with just 4/6 pos per ccy pairs this is really suitable.

The good thing of entering at every break of the DC is the immense scaling in we have. Even if we loose on the last positions are we sure this isn't still the most profitable way?
  • #64
  • Edited 3:49am Aug 11, 2015 3:30am | Edited 3:49am
Quoting lemming78
{quote} Great post, thanks for sharing. I didn't understand one thing. When you talk of 200 do you mean the DC 200 on daily charts? Essentially you're right, but if AUDUSD changes trend right now we won't be able to jump on it until 40 weeks. About the rest, i think you're essentially in line with the following: NO RISK NO REWARD. But at this point, having so much risk, are we actually sure that a simple entry method based on engulfing, inside bar, Flying buddhas isn't better? We would have much lower risk per trade, and we could only select opportunities...
Hi lemming :-)

The 200 D1 periods relate to the average extent to which indicators used to detect persistent price anomalies (trends) need to encompass to ensure the sample size is sufficiently large to "on balance of probability" pick longer term sustainable trends. This figure comes from some of the recent trend studies using extensive price series and also seems to agree with some extensive backtesting of a large number of instruments that I have undertaken with a very talented EA developer from this forum....*just in case you are looking*...

Inside bar breakouts are very appealing to me (and to Peter Crown et al), but in reality, you can set your distance to stop at any arbitrary value to maximise leverage provided you are trading in the direction of price bias, have miserely stops and work on the principle that the Law of Large Numbers is on your side. I am not so dependent on price patterns on the short term timescale as I find it exceptionally difficult to determine what is random price movement and what is not....not discouraging others, but this is just my own personal findings. Edit.....If you look at your patterns and adopt the same principle in reducing R (distance to stop) without adhering to the patterns themselves but establishing strict entry rules on 'random entry' into a trend....then I totally agree with you...... In essence you are seeking to maximise your Reward to risk ratio trading in the direction of price bias and being prepared to significantly reduce your win:loss ratio in doing so. My current Win:Loss ratio is barely reaching 30% but the trade expectancy is excellent....so I think we are on the same page here.

One of the key reasons I like a Donchian break is that you 'ring fence' price and ensure you catch every trend of substance...

What I am really struggling with at the moment is not trend detection, but trade management during trends. In riding great trends there are numerous times you close out of the trend after watching a peak unrealised profit disappear while waiting for the trend to bend. Of course if you harvest these peaks you potentially cut the trend short etc......and I am just having to suck in the fact that this is a requirement of trend trading....*grrrrr*.

A further significant struggle is in seeking to flatten the equity curve with trend trading. As we all know....the times between trends can be significant.....and the drawdowns can be painful......however are there are other ways to manage drawdown without having to compromise your trend trading technique by diluting it's potential through traditional risk management measures such as reducing your position size etc. For example, with a portfolio, you have the ability to 'harvest some successful' trending instruments to flatten the equity curve....however once again you are interfering with the profit performance of those successful instruments. I am really wrestling with this at the moment and wonder if anyone has good ideas on this.

One of the things I am thinking is that you may be able to run tandem strategies such as a trend following strategy and a mean reversion (shorter term) strategy so hopefully while one doesn't perform, the other does....but there must be better ways to do this without interfering with the power of the traditional trend trade.....or am I just being too greedy :-)

Anyway, I don't want to derail the thread but some of these issues will need to be addressed by Spidey as he develops this technique.
  • #65
  • Aug 11, 2015 4:19am Aug 11, 2015 4:19am
Quoting Copernicus;8428864[B
but there must be better ways to do this without interfering with the power of the traditional trend trade.....or am I just being too greedy :-) [/b]Anyway, I don't want to derail the thread but some of these issues will need to be addressed by Spidey as he develops this technique.
It's all a matter of priorities.
Do you want to NOT interfere with the trend? Just don't do it Open a position, no stoploss, no takeprofit. Just have FAITH in the trend until you're proved you are wrong.

Do you want to preserve capital? Just put sl to be and let it go.

Plus, we are not considering something which i think is very important for this type of trading, which is FUNDAMENTALS.
Let's all admit that all the best recent trends started with fundamentals. EU NU AU GJ UJ. Each one of them has started with a fundamental fact, because that's how you piggyback institutional traders on that timeframe scale.

You are looking to the shortcut for the undeniable assumption which has Always proven correct: No risk No reward. I honestly think this is not the way we should think about that.

In the optic of reverse engeneering a trade system: who would have made the most money in the last year and a half? Who strongly believed in the trend and on the fundamental shift between economies. The same one who started selling eu at the break of 1.4. Tell me if i'm wrong.

Let's think about the most vivid example of this kind of thinking. Jesse Levermore gained fortunes in a period in which i don't think stoplosses were used. And he lost them all. And he regained them.

As i said, it's all a matter of priorities. What do you want more? Risk less or get a higher reward? I don't think you can have them both ways.
  • #66
  • Edited 4:40am Aug 11, 2015 4:26am | Edited 4:40am
I hear you lemming

I have no doubt that fundamentals are driving the long term secular trend...but I am hopeless at making sense of it all and don't have the necessary inside info to capitalise on it so .......I cheat using the long term monthly trend as determining my bias.

I also do agree with your no risk...no reward statement and inevitably come to a trade-off point in that sense.......but I am a greedy bugger :-) For example speaking of AUDUSD, look how I have ring-fenced the bast***d on multiple timeframes....that is greed talking. I just hate letting them get away. Those horizontal lines represent my alternate approach to Donchians.....but whether you use Donchian breaks, Standard deviation channels, moving averages or whatever....the result is fairly similar for trend followers.


Attached Image (click to enlarge)
Click to Enlarge Name: AUDUSD.JPG Size: 367 KB
  • #67
  • Aug 11, 2015 6:35am Aug 11, 2015 6:35am
Quoting Copernicus
I hear you lemming I have no doubt that fundamentals are driving the long term secular trend...but I am hopeless at making sense of it all and don't have the necessary inside info to capitalise on it so .......I cheat using the long term monthly trend as determining my bias. I also do agree with your no risk...no reward statement and inevitably come to a trade-off point in that sense.......but I am a greedy bugger :-) For example speaking of AUDUSD, look how I have ring-fenced the bast***d on multiple timeframes....that is greed talking. I just...
And it's not going to be any different. You can jump in whenever you like. Doesn't change anything. The only thing that changes is the risk. How to minimize the risk? You want to be god damn sure of what market is doing. How do you do that? Trading the momentum.

But i think here we go offtopic. Sorry Spiderman.. i stop here
  • #68
  • Aug 11, 2015 8:10am Aug 11, 2015 8:10am
Copernicus...Lemming

Don't worry about derailment. The points you bring out are valid and interesting and I'm sure will help some in the process.

Copernicus,
Been thinking about some of your earlier comments about moving to BE and limiting trade entry. You are of course correct that by moving to BE you cut down on the breathing room of the trade and may well lose entries that if the trend continues on, would payoff handsomely. My concern, however, is that with the aggressive MM that I'm employing here of compounding our floating profit to finance ever bigger lot sizes, not taking a "loss" becomes more critical as it effectively takes us back to our starting point...after banking our 50% profit.

For example, remember this retracement in Red that took out a couple of our BE entries in the upper half of our trend.

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDiH4.png Size: 102 KB


It would have been nice to have those entries still there as the trend ended up continuing. However, that retracement would have taken out at least 1 trade at s/l (see post #43) which would have derailed our compounding to that point. So, yeah, it's a tough choice.


On the point of limiting our entries to the start of a trend I'm running into the same situation. Due to our aggressive compounding, it's those later entries at higher lot sizes, that can really pay off. Granted at the "end" of the trend your recent entry may be taken out for a loss (only 1 if we're moving to BE) but at least we're locking in a 50% payout. That last entry made late in the game for 6.23 lots had a + $14k payout and would never exist if we cut our trades short.
Again, it's a tough choice.
  • #69
  • Aug 11, 2015 8:18am Aug 11, 2015 8:18am
Lemming...

I think you mentioned earlier about an EA.
I don't think that would work on this. It's true that most aspects our mechanical, but there is still much discretion on trade entry point.
We're taking the break of the daily DC, but not every break. We have to have sufficient room from previous entries to move to BE, move s/l's down, etc.
Price is moving so fast in some areas that our s/l on the H4 would be further away than our BE on a previous trade.

Notice the circled area here.

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDiDaily.png Size: 67 KB
  • #70
  • Edited 8:40am Aug 11, 2015 8:21am | Edited 8:40am
I see exactly where you are coming from Spidey. That area you highlighted is the bane of my existence at the moment as it represents a great harvest opportunity but that can only be said in hindsight. I am tending to think that your move to BE is perhaps a good technique when dealing with multiple entries given that some of the early entries may be able to get through those trend retracements. Have you thought of perhaps dividing your trades into trend trades and swing trades. For example let's assume you have released 4 trades on a particular trend. Could you reserve half of them for swing trades that have a profit target installed while reserving the balance of trades for the long term trend ride with no set profit target to let the profits run.

In fact when thinking about this say the initial trades entered are reserved for the entire trend but the more recent trade entries are employed to ride the swing segments with profit targets. That way you might be able to at least get the benefit of those inevitable retracements you get during the long trend trades......just thoughts :-)

For example those strong exponential trend movements (eg. XAUUSD chart below) are rarely captured by the trend follower as the exits wait until the trend bends. With a combination of both trend trades and swing trades on an underlying trend, you perhaps could take advantage of these crescendo moves at the end of trends without seeing this unrealised profit vapourise in front of your eyes.

Attached Image (click to enlarge)
Click to Enlarge Name: XAUUSD.JPG Size: 122 KB
  • #71
  • Aug 11, 2015 8:25am Aug 11, 2015 8:25am
Well...

Right, wrong or indifferent, I've taken my 1st entry on the Eur/Cad.

Here's the weekly. It actually broke the DC a couple of weeks ago, but obviously, we we're around yet.

Attached Image (click to enlarge)
Click to Enlarge Name: EURCADiWeekly.png Size: 62 KB


Here's the entry point on the Daily

Attached Image (click to enlarge)
Click to Enlarge Name: EURCADiDaily.png Size: 70 KB


And the s/l on H4

Attached Image (click to enlarge)
Click to Enlarge Name: EURCADiH4.png Size: 62 KB


And yes, it's a live account I'll be attaching an explorer.
  • #72
  • Aug 11, 2015 8:30am Aug 11, 2015 8:30am
Quoting spyderman
Lemming... I think you mentioned earlier about an EA. I don't think that would work on this. It's true that most aspects our mechanical, but there is still much discretion on trade entry point. We're taking the break of the daily DC, but not every break. We have to have sufficient room from previous entries to move to BE, move s/l's down, etc. Price is moving so fast in some areas that our s/l on the H4 would be further away than our BE on a previous trade. Notice the circled area here. {image}

Exactly, an ea wouldn't work i realized that after.
  • #73
  • Aug 11, 2015 8:31am Aug 11, 2015 8:31am
Quoting Copernicus
I see exactly where you are coming from Spidey. That area you highlighted is the bane of my existence at the moment as it represents a great harvest opportunity but that can only be said in hindsight. I am tending to think that your move to BE is perhaps a good technique when dealing with multiple entries given that some of the early entries may be able to get through those trend retracements. Have you thought of perhaps dividing your trades into trend trades and swing trades. For example let's assume you have released 4 trades on a particular trend....
Well, it depends. Are you sure your second strategy would have a positive expectation?
I wouldn't be that sure without data backing you up.
  • #74
  • Aug 11, 2015 8:39am Aug 11, 2015 8:39am
Quoting lemming78
{quote} Well, it depends. Are you sure your second strategy would have a positive expectation? I wouldn't be that sure without data backing you up.
Ah now that's the question..............but this is a brain storming session isn't it?
  • #75
  • Aug 11, 2015 9:05am Aug 11, 2015 9:05am
Quoting spyderman
Well... Right, wrong or indifferent, I've taken my 1st entry on the Eur/Cad. Here's the weekly. It actually broke the DC a couple of weeks ago, but obviously, we we're around yet. {image} Here's the entry point on the Daily {image} And the s/l on H4 {image} And yes, it's a live account I'll be attaching an explorer.
Into the breach you go Spidey. Best of luck. :-)
  • #76
  • Aug 11, 2015 9:06am Aug 11, 2015 9:06am
Quoting Copernicus
{quote} Ah now that's the question..............but this is a brain storming session isn't it?
Yes, of course. In that spirit, here's what i think:
Using the rules we've given so far, that strategy won't work. If we use stop orders we need momentum to back us up. If we don't get it we can't go anywhere. We would be in markets hands.
The hold indefenitely strategy would work because of the astonishing risk:reward implicit ratio. But this second one would need very far target in order to be positive in expectancy.

Assume just this:
200 pips stoploss (might be average using 4h channel as stoploss).
Breakout entry with stop orders (imho this will turn in a 30% success, which i mean price to keep going with momentum without stalling).

In this case you would loose 200 pips 70% of the time and how much you need to gain to be BE? 466 pips per trade. Just to make breakeven.
Of course would not be 30% success, because might be slighly better given the times price comes back without touching stoploss and break above in gain again.

Does this make sense to you guys?
  • #77
  • Aug 11, 2015 9:07am Aug 11, 2015 9:07am
Quoting spyderman
Well... Right, wrong or indifferent, I've taken my 1st entry on the Eur/Cad. Here's the weekly. It actually broke the DC a couple of weeks ago, but obviously, we we're around yet. {image} Here's the entry point on the Daily {image} And the s/l on H4 {image} And yes, it's a live account I'll be attaching an explorer.
Good luck man, really. I hope you'll succeed
  • #78
  • Aug 11, 2015 9:25am Aug 11, 2015 9:25am
Quoting lemming78
{quote} Yes, of course. In that spirit, here's what i think: Using the rules we've given so far, that strategy won't work. If we use stop orders we need momentum to back us up. If we don't get it we can't go anywhere. We would be in markets hands. The hold indefenitely strategy would work because of the astonishing risk:reward implicit ratio. But this second one would need very far target in order to be positive in expectancy. Assume just this: 200 pips stoploss (might be average using 4h channel as stoploss). Breakout entry with stop orders (imho...
It probably overcomplicates things. Simple usually is best. I must say however that I am not a fan of pyramiding entries later in a trend and need to get em' done earlier in trend development to focus punching power up front and then let it ride....unless you use them to augment profit performance with less ambitious targets:-)
  • #79
  • Aug 11, 2015 4:16pm Aug 11, 2015 4:16pm
Quoting lemming78
{quote} Does anyone remember Graeme's thread? It was illuminating. The only thing you need is to keep your risk controlled. Then the inevitable will happen. Of course i've been Always skeptical about his methods, but they can theoretically work.

Loved Graeme's thread. That's what really got me started thinking about longer term trading and stacking trades.
  • #80
  • Edited 6:44pm Aug 11, 2015 6:29pm | Edited 6:44pm
Quoting spyderman
{quote} Loved Graeme's thread. That's what really got me started thinking about longer term trading and stacking trades.
Spidey

You might be interested in applying an anti-martingale progression to your trends such as the Snowball which was developed by a very talented chap here at FF (7bit). Unlike a Martingale, it manages risk well and is excellent as a discretionary tool for trend trading. Like a move to breakeven, losing trades are closed out for a single level minor loss while trade entries that survive trend retracements prosper. It is effectively like Graeme's equity milipede approach applied to progressions in releasing lots of small trades with strict trade risk parameters hopefully allowing a few to get through for the long ride and can really boost profitability without significantly compromising drawdowns.

Attached is an example on EURUSD that visually demonstrates how it works. The EA is used as a discretionary tool for the manual trader and can be found here https://sites.google.com/site/prof7bit/snowball with a forum discussion here.

The chart below uses your Donchian but translates these to D1 timeframe so we have a 20period donchian on W1 represented as a 140 periood donchian for entry on D1 and a standard 20 period D1 donchian for exit. You simply turn on the progression on a break of the 140 period donchian and turn off the progression on exit.

The approach is almost a mirror to yours but more aggressively pyramids profits. Better still, you could turn off the progression (rather than keeping it going the whole way along the trend) as you get deeper into the trend to further manage potential drawdown impacts using this addition.

PS I should have exited earlier (as I just missed the Donchian 20 exit) but for the purposes of showing you how it works, I think you will get the picture.

Attached Image (click to enlarge)
Click to Enlarge Name: SnowballVisual.JPG Size: 145 KB


Attached Image (click to enlarge)
Click to Enlarge Name: SnowballEquityCurve.JPG Size: 126 KB
0 traders viewing now