Hello,

While playing with Standard Error Indicator (difference between price and its linear regression), I created this controversial system that I’m going to post today.
It’s been running for several months with minimal adjustment of the variables. You can see current statistics on the attached picture.

Logic is very simple.
- Breakout of the Standard Error through the trigger line initiates trades.
- Direction of the trade is determined by the Linear Regression Moving Average, color of the trigger line is changing depending on the LRMA being up or down.
- System tries to take 14 pips (user defined) of profit in the direction of the breakout.
- User has an option of exiting either when target price hits Close or when it crosses High/Low.
- Defaults are for EURUSD M5.

And now the controversial part, I’ll try to keep it in the boundaries of this forum’s rules.
The system does NOT exit if the signal in opposite direction is generated. It exits only when either Stop or Target is hit. Nothing new about that, I just decided to validate some claims of people who advocate that almost random (but systematic) entries can generate a good profitable system. If you do some web search on “grid scalping”, you’ll get an idea of what it’s all about. The approach has no appeal to me but if someone can make money doing that, I wish them all the best.

The system is trying to answer the following question:
If entries are not random but with high probability of price moving certain amount of pips in the direction of the trade, and stops are for catastrophic protection only, can the scalping system be profitable?

I don’t have a certain answer. If the system performs with 90%+ win/loss ratio (as it did), there is no problem. But you need to remember that a couple of losses can wipe out 10 profitable trades. The screenshot shows that currently average Win is around 17 pips, average Loss is 81 pips.

I wanted to shelve the system giving it some time to rest (because in my opinion with all the shiny curve-fitted profits and extraordinary win/loss ratio it’s not ready for prime time yet), but then decided to expose it to other traders. Hopefully it’ll generate some bright new ideas.
Enjoy!

Good Luck,
S

This looks quite good, nice and simple.
could you please let me know what your original setting were, when the screen shot was taken.



Cheers


Beno

Hello Beno,

Default settings of the system are the same as on the screenshot. In the last 3 months I tried to keep them as stable as possible. 4 weeks ago Trigger value was 0.0007 (now 0.0006) and stop of 70 pips was sufficient (now 75).

Regards,
S.

Hi Samony,

Very impressive, thank you for sharing it with us.

Regards
merk

You are welcome, Merk.

Just remember that nerves of steel are required to observe how one loss wipes out gain of several profitable trades. Still thinking of adding something extra to make system more "normal". Nothing worked so far that didn't impede system's performance.

So let's forward test it, and see if it survives.

Regards,
S.

You could try incorporating Donchian midpoint indicator and see what happens. This it the system that VTS coded for me but take what you need.
Have you tried a trailing stop as well as the hard stop.

Does anyone else have some input.


Cheers


BenoClick to view attachment

Sorry Beno. Incorporated Middle Line of the Donchian Channel but it produced no improvements to the system.

Speaking of Donchian Channels, in the past I had better results with variations of traditional Upper/Lower channel breakouts. When you use Middle Line, you are trying to play the "trend", not a "breakout" game.

Anyway, thanks for your suggestion.

Regards,
S.

hmmm, 2 cents

on 5 min bars , and a 90%win ratio ,, what about a trailing stop as suggested but an intrabar trailing stop , yes VT only works on close of bar ,but there is TS trailing stop that works by the tick for vt ,,, manual trading only but incorperating a stop like that to get your profits could work well .. maybe!
hemp

Hello Hemp,

I understand your suggestion. It might have worked for a "normal" system Let me demonstrate the nature of abnormality by giving you these numbers:
- Target 14 pips, Stop - 75 pips , Win/Loss - around 90% (something like 55/5 on live account on 5000 bars);
- Target 14 pips, Stop - 50 pips , Win/Loss - around 60-70% (which is already unacceptable due do the negative risk/reward ratio).
Trailing stop of any kind will exacerbate this issue.

BTW, all info for this system is for Live account only. There's such a big discrepancy with Demo in the last 2 weeks, that I had to stop using demo account for backtesting, just for validation of automated functionality.

Regards,
S.

You got that right , I have stopped using the demo completely for exactly the same reason,s ...
I punched in the #,S into a Equity Curve Random Generator , does say it would be profitable but also have a high probability of ruin if you happen to have a few 75 pip hits early on ...
That being said , again just another thought , what about a resting hedge stop order at say 40 to 50 pips ... In effect bringing your stoploss to 45 pips ,S instead of 75 .. At only 1 out of 10 trades using it , it would make the ts ,S probability of profit statistically way better .. ?
You would have to baby sit the hedge order once filled though if the market turned around after being filled ..
Peace
Hemp

That might work, Hemp.
I just need to figure out how to implement that.
Thank you,
S.

Hi
it is very interesting system. Great. Coul you tell me parameters like maximal drawdown and RRR ratio?
Thanks a lot
regards
Tom

Hi Tom,

I can't give you the "official" parameters, because I don't have ones due to the limited backtesting capabilities of VT. By observing system for several months I can say that maximum drawdown was around 150-170 pips (2 losses in a row).
Based on the initial Target of 14 pips and Stop of 75, RRR is around 5:1 (too much risk for a little reward).
Win/Loss ratio has been around 9:1, it balanced poor RRR.

You also need to realize that randomness of initial entry (whatever is the first signal on the left side of the chart), and controversial exits (described in post 1), make backtesting a little bit tricky. That's why forward testing is the only way to see if it lives up to expectations.

With all that said,
Enjoy and make it better
S.

Hi Samony,

I,ve made some modification to your system.
It' looks very encouraging, but we need more testing.
GBP5m Life Acc 5000 bars.

Regards
Merk

Hi Merk,

Great modifications!!!
I noticed that you added extra trend identification with HMAs and trailing stop when minimum profit has been achieved. That should help and make it more predictable (hopefully).
Did you try it on other pairs?

Let's keep testing and think about more improvements (like dynamic trigger line, etc.).

Thank you for great contribution.
Regards,
S.

Hi Samony,

Thanks.

This is the best I could get out of EUR.
I haven't tried any other pares. JPY is to volatile.



Regards
Merk

hi merk, i have question:
1. your system automated trading or not?
2. or your system just an automatic signal and not an automatic trading?
3. is the result (total profitinpips) is true?using real live account? have you counting the bid/ask spread condition so the result in profit display (total profitinpips/inspect button) have no discrepancy with the result in account statement (demo and real live account)?
sorry if my english is not good, im from indonesia.thanks.

Could you please share you parameters with us all as that 100w to 8L is amazing

Cheers

Beno

Hi Harto,

Firs,
We should all thank Samony for allowing as to use his system.
I takes lot of time and effort to make a good system.

The system is fully automated.
I've stoped using demo Acc for testing there is big discrepancy between demo and Life.
I use demo Acc only to test if the system takes proper orders.
Profit display is on Life ACC 5000bars and I Think is real.

Thanks

Hi Samony,

What do you mean by " dynamic trigger line" I've search the Google and couldn't find one.
If you have one can you post it.

Regards.
merk

Hi Beno

Today is 110w 6L.
The system I've posted is pre set. Life ACC 5000bars.
We know the past but we don't know the future so lets hope.


Thanks
merk

Note to admin: if duplicate, please disregard (couldn't submit, got HTTP error).



Hi Merk,

Probably I invented this phrase ("dynamic trigger line") if you couldn't find it on the web . I have several systems when parameters that trigger trades are not static but dynamically changing depending on some fluctuating variables (mainly volatility of the underlying security) or smoothed instead of being just a straight line.

In case of the Standard Error System the trigger line is static.
SEHline:= SE_Trigger; - straight line that triggers trades when Standard Error values break through it.
As a result, we need to adjust the value of SE_Trigger if market conditions change. Ideally dynamic trigger line would "self-adjust" itself depending on changes in market conditions.

There are many ways of doing it; with limited time that I spent, nothing improved system's performance. But you can consider substituting static trigger line with:

- Moving Average of the Standard Error adjusted by some value or by its Standard Deviation.
- Line that represents the difference of HHV and LLV of Standard Error during defined lookback periods adjusted by a certain value.
- Changing number of Lookback periods depending on Standard Error's volatility. Something like that.
NormalizedStDev:= Stdev(SE,stdevPeriods) / Mov(Stdev(SE,stdevPeriods),StdevMAperiods,StdevMAtype);
SEDynamicPeriods:= Int(StaticPeriods/NormalizedStDev);
and then creating a Moving Average using dynamic periods of SE.

Hope it'll stimulate your creativity.
Best Regards,
S.

Hi Samony,

Thank you very much.
That clarified my thoughts. I am not en expert or programmer but slowly learning.
I am trying lot of different things.Some are woring others not.

Best Regards,
merk

Hi Samony,

Can you call me on
Skype - MirekJu - Melbourne Australia
I have something interesting for you.

Best Regards
Merk

Hi Merk,

Done, please check your Skype account.
BTW, under the name of "samony" you can find me on other forums where PMs are allowed, including Polish Forex forum (although my Polish is limited to a couple of bad words )

Regards,
S.

Hi Samony,
Witch Polish forum?

Regards Mirek

Hello,



Keep in mind URL's are not allowed in these forums.

Regards,
chris

The easy idea behind the random entry position is not really easy nor random.
Most of the technical analysis is based on the idea that the markets move on some periodicity, so that they are ciclic.
The random entry would be more like making a entry on fase with the market, so you could make almost any position and be profitable. Thats the solution and the problem at the same time. Finding the "market fase" is not something that's easy.
I read some books about Fourrier Transforms and they explain some good points on how to disamble a wave (on this case the market price/time wave) on its components, and that way you would be able to hopely know where youre sitting.
Other ideas that act as REALLY random systems dont work.
Just because there are some terrible price pikes and huge movements, so you would be scalping some pips maybe on a mayor 100 pips move, or you would be loosing 100 pips or more on a pike made by the news etc.
To end this, ALL technical analysis is based on study of history, and/or the periodicity of markets, so its never random.

Cya

Hello Cya,

Very good points.
Agree that absolutelly random entries won't work. Although the system was trying to eliminate randomness by selecting trades in the direction of the breakout, it defintelly requires refinement.

By the way, the anwser to the ironic question in the subtitle ("Can you lose money with 90% profitable system?") is "YES"


Regards,
S.

ARE THE RESULTS FROM EACH OF THESE TRADING SYSTEMS AT DEFAULT SETTINGS...IF NOT MAY I HAVE THE PARAMETERS THAT PRODUCED THESE RESULTS?

Hello,



Please try not to type in all capital letters. It is very difficult to read. Thank you.

Regards,
Chris

Im sorry about that Chris...it wont happen again.

Hello,
All results are with default settings at the time of publishing. That includes the original version and Merk's modifications.
If you still see good win/loss ratio, but the profits have been eaten by a few losses, feel free to go ahead and make the system better.
Regards,
S.

Samony,

Some pros take the trade when their favorite trend indicator indicates a trend in same direction, ths the chances for a winning trade %-wise actually increases. Your losers and big DDs, check to see if they have mostly occured when the "winds of" trends are against you (the direction the trade wants to go), if so, there is probably the basic problem, now the next "big problem" is what kind of trend indicator is viable enough for your application/coding.

jerry

Hello Jerry,

Actually the trades are taken in the direction of the immediate trend defined by Linear Regression Moving Average, which is a decent tool for that particular purpose.
The real lesson is that it's hard to trade a system with a low Risk/Reward ratio.

Regards,
S.

I found this system is very profitable some SL might help as the winning :loss ration is very high, but sometimes 1 lost can eat all the profit

Hi Samony,

I just read somewhere in the forum where you were trying to implement Trailing Stop after Initial Target reached. I will not mind if you can implement this on your system Samony Standard Error Breakout. I think the system will perform better as it will take the maximum profit out of the market not just the initial profit target.

Please attached the modified version.

Thank you

Hello M.

Implementation of Trailing Stop after Initial Target is reached was done for other systems that are currently in development.
If I come back to Standard Error Breakout system with this modification, I'll definitely publish it.

Best Regards,
S.

Hi All,

This system seems pretty tidy and neat, and more important it gives promissing results.

I have modified it from the original code to accomodate the symbolpoint, so it will cover JPY currency pairs.

One improvement need to be done is that the trigger value should be a dynamic one and not the static which will not cover different timeframe and different currency pair.

Regards
Ozkevin

Hi,

As mentioned ealier, I have change the trigger into dynamic value based on the largest value in the chart times a factor. In this way, there will much more trades opportunity.

Next step, the trailing stop of the standard templete will be added into the system.

Regards
Ozkevin

could someone help me with this TS?, i will like to add a session filter.

Thanx

Hello everybody,

Is anybody using this breakout system any more?

Has the behavior of the pairs so changed in the last one year?.. because I cant get any positive results, not to mention the extreme results shown at the begining of this thread.

I am prett new to this, so I am going through this place and try to learn, so any explanation or suggestion is appreciated.

elmer

Hi samony, merk and others,

I've been fiddling with vt for a few months now, but I'm not a programmer. I came across this post of samony's standard error breakout and the modifications and found it interesting because of the scalping idea and in testing it in demo mode for a couple of days, I've gotten great results.

I'm confused on one issue and had an idea for improvement, if you guys are (hopefully) still looking at new posts in here.

Does Merk's modification of samony's original ts include a trailing stop that will exit the position only when it reaches a minimal profit margin, or was that still in the works? I was trying to program one in to another TS based on different indicators where we were opening trades manually and setting 5 pip limits that over the course of a month, in a demo account, gave us 100% wins. However, with such small limits, we'd have to increase our lot size much greater than reasonable in light of margin calls.

Anyway, my idea for improvement to protect against the catastrophic potential losses is whether someone can program the system to have a sort of level stop. That is, where you now have one stop at 77 pips, instead, you would have four different stops: a stop at 20 pips down where if you had say 4 lots on the open position, you closed only 1 lot. If the position fell another 20 to be down 40 pips, it would close another lot and leave you with only 2 lots from your originally four, and so on. That way, if the position fell down to 77 pips or whatever the bottom stop level is set at, it would only be for a fraction of your original lots. If this could be done, ideally, we could also add language that if the price went back in the profitable direction, and the TS gave a signal in that direction, the TS would reopen the closed lots on the way back.

I don't know how to program it, but I fiddled with the figures using simple arithmetic. You'd still have losses, but they would be much less, and if you could recover closed lots on the return trip, the loss would be even less.

Sorry for the length of this. Your input with a modification of the ts or at least an idea on how to program this concept would be appreciated.

This thread dead? I can't believe it, this TS looks too good! Sadly I came a little late but I would have positively contributed if this was still hot. In an hour or so I will test this,

anyway I think Merk mentioned that the performance of VT Trader Demo and Live accounts differ... What are the differences?? thanks!

Proupin

I've been using Merk's modified versions of the system to some success. However, the concerns of Samony and Merk are still present. You do well with small gains, and along comes a large loss and wipes it away. If you have any ideas you want to try, I'll test it on my end as well. To get it profitable, you're going to have to tweak the parameters and stay on top of them every week.

Sounds, really interessting. At first it reminded me alot of the BGX system with trigger lines etc.
But this is way more complex so I wonder. Why isn't there a BGX VTTrader TS? I use a MT4 Demo with a BGX tpl and I've been pulling 200-900 Pips out
of the market last week everyday. Sometimes caught 300+ Pips in one move.
No losses, tight stop. It gives you some bad signals but you can easily dismiss them with your entry rules.

Hello,

Does BGX happen to stand for "bunnygirl cross"? If so, there is already a system created based around her *original* set of rules from a few years back. I haven't been following her in recent times due to other commitments though.

Regards,
Chris

Yeah, thanks. I found that but it's really bad.
Bunny Girl is simple but the TS is lacking important things.

Anyway BGX is not the topic.
But I have another question.
I downloaded the original Standard Err. Breakout System and the modifed versions. But I don't get any Buy,Sell Exit Arrows on my charts. They just don't look like the screenshots posted here. I'm using the latest VTTrade.
Am I missing some additional Indicators? When I let it run on multiple pairs I sometimes get this "order open" or "limit hit" sound but it doesn't open any positions.

Hello,

Try opening your symbol palette in the Navigator Window>Drawing Toolbar. If it's blank (contains no symbols), exit VT Trader, reboot your PC, and then login again. Then reattach the system. This should resolve the issue. If it doesn't (and you have visible symbols in your palette), then something may be wrong with the system itself.

Regards,
Chris

Hi Chris/Guys,,,

This is a great ts,,, would like to know if it’s possible to modify the ts so to …

1. allow the stop loss to be adjusted at breakeven point once the market price allow.
( reason being, two continues loss wipe out all the profit gain earlier, it’s ok we don’t made any profit for this trade, more importantly... no losses for this trade,,, and next opportunity trade is just few bars away as time is at our side.)

2. allow trailing stop instead of fixed profit target.
( I notice there’s trailing stop in the ts,,, but some how doesn’t work e.g.: exit at 10 pips profit )

thx for the advise…

-willie-

Hey tjlpd,
Could you elaborate some on the BGX system? It seems to me that if you are getting 200-900 pips per week, you should stick with it. My own pip goal is only 80 per week, and sometimes it's a struggle because of the large stop loss. I'd love to make even 100-200 per week. If possible, and not restricted by the rules of this forum, can you post a txt, Word or Wordperfect file of the BGX code so we can compare it to Samony and Merk's efforts?

Like willcsk, I too do not find the trailer stop working right. In particular, I don't know why Samony or Merk added the "N times" parameter.

What I would like to know is if the software has some way of referring to the pip count on your closed positions. On several occasions, I've left the TS running auto overnight, it reached my weekly goal, then opened a position that wiped out most or all of my gains. I'd like to endsession automatically when I reach my pip weekly goal.

Thanks.