EMA/RSI Trend Spotter system | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

EMA & RSI Intraday M15 system 42 replies

Heiken Ashi Doji Spotter Alert 21 replies

10 Ema and 200 Ema cross system 6 replies

Di Napoli, RSI, EMA 5 min system 14 replies

Fx1618's Journal - EMA|RSI|Trend 34 replies

Attachments: EMA/RSI Trend Spotter system
Exit Attachments

EMA/RSI Trend Spotter system

  • #41
  • Jun 26, 2010 12:10am Jun 26, 2010 12:10am
Quoting eagle4x
I requested a programmer friend of mine to review the indicator from post #1 and he said it repaints:

"The original one scans signal from right to left, and plot arrow 3 bars to the left (giving an early sign impression) after signal confirmed. That's why its signal looks perfect."

He revised it so it should not repaint. All I have done with it so far is attach to a chart and noticed not nearly as many signals as with the orignal one - so, if you desire more signals, will have to adjust the settings. If anyone comes up with...

Thanks Eagle4x, lets all take a look at it an see what we can come up with. There has to be some optimal settings and rules to make the strategy work. As RR always says, its not about the entry, but the rather the exit!

I have been strategy testing, and Ill get 70 and 80 pips winners turn to BE or a small loss. Got to figure a way to catch the runs, but also not give away good profit
  • #42
  • Jun 26, 2010 2:15am Jun 26, 2010 2:15am
When I put the original indi on the same chart as the new one on a strategy tester with the same settings, the orignal triggered multiple trades and the new one only triggered 1 or 2. There is something not quite right
  • #43
  • Jun 26, 2010 4:04am Jun 26, 2010 4:04am
Quoting eagle4x
I requested a programmer friend of mine to review the indicator from post #1 and he said it repaints:

"The original one scans signal from right to left, and plot arrow 3 bars to the left (giving an early sign impression) after signal confirmed. That's why its signal looks perfect."

He revised it so it should not repaint. All I have done with it so far is attach to a chart and noticed not nearly as many signals as with the orignal one - so, if you desire more signals, will have to adjust the settings. If anyone comes up with...

Sorry to disappoint you guys, but this v2 indy is still not right...
The "programmer" had it right to spot that it was incorrectly plotting from right to left, and also correctly removed the use of the "after" values (which would litterally mean that the indicator could read one bar in the FUTURE... ).
But he also removed one major aspect of the indicator: it was "remembering" that an EMA cross or RSI cross had occured, and would plot a signal when the other (EMA or RSI) cross occurs, several bars later.
V2 considers the signal only when both crosses occur on the SAME BAR, which is not what we want... and will very hardly happen!
This is why you are getting virtually no signals with V2.

V3 attached has this fixed.

SQ
Attached File(s)
File Type: mq4 EMA_Cross_RSI_Trend_Spotter_v3.mq4   6 KB | 1,962 downloads
  • #44
  • Jun 26, 2010 4:16am Jun 26, 2010 4:16am
sidus_v3 gives the same results with a simpler coding approach...
thanks pudseybear.

SQ
  • #45
  • Jun 26, 2010 5:23am Jun 26, 2010 5:23am
HI squalou

EMA CROSS RSI V3 = EMA CROSSVER SIGNAL ... FOR YOUR INFORMATION.

A QUESTION PLEASE... SIDIUS V3 REPAINTING ?????

THANKS FRIEND.
  • #46
  • Jun 26, 2010 7:29am Jun 26, 2010 7:29am
Quoting porfirio
EMA CROSS RSI V3 = EMA CROSSVER SIGNAL ... FOR YOUR INFORMATION.
It is a EMA crossover (which actually is nothing more than a simple MACD)
but combined with a RSI +/-50 filter.
Just like Sidus V3.

Scoobs, in his BOJ bots, uses a similar combination, but with re-enforced RSI filter: it uses the same RSI +/-50 filter, but applied on multiple timeframes, which is supposed to give even stronger entries.
We could add these extra TF RSI filters to Sidus to make it stronger.
You will get less signals, but possibly better ones.
You could also use the CCI +/-50 filter instead of RSI, or addition to RSI, which also gives good results.

Quoting porfirio
A QUESTION PLEASE... SIDIUS V3 REPAINTING ?????
Not for past bars.
However while the current bar is forming, EMAs and RSI are calculated at each tick, and their values will change with price action, and the signal could toggle (appear+disappear) during that bar.
Once a new bar starts, the signals plotted on the previous bars will not be changed anymore.
You should wait for the current bar to close before taking the signal.

SQ
  • #47
  • Jun 26, 2010 8:47am Jun 26, 2010 8:47am
Why did you comment that v2 has less signals, then add v3 that has much higher settings that generates even less signals than v2?


Quoting squalou
Sorry to disappoint you guys, but this v2 indy is still not right...
The "programmer" had it right to spot that it was incorrectly plotting from right to left, and also correctly removed the use of the "after" values (which would litterally mean that the indicator could read one bar in the FUTURE... ).
But he also removed one major aspect of the indicator: it was "remembering" that an EMA cross or RSI cross had occured, and would plot a signal when the other (EMA or RSI) cross occurs, several bars later.
V2 considers the signal...
  • #48
  • Jun 26, 2010 10:15am Jun 26, 2010 10:15am
Quoting squalou
Sorry to disappoint you guys, but this v2 indy is still not right...
The "programmer" had it right to spot that it was incorrectly plotting from right to left, and also correctly removed the use of the "after" values (which would litterally mean that the indicator could read one bar in the FUTURE... ).
But he also removed one major aspect of the indicator: it was "remembering" that an EMA cross or RSI cross had occured, and would plot a signal when the other (EMA or RSI) cross occurs, several bars later.
V2 considers...
I believe you were extracting the rules from inside the original indicator. When two or more programmer, including that "programmer", trying to extract the rules from the existing indicator which can have logic flaw, they could have different interpretation. And you should not make a quick judgement that you are correct, and others are incorrect.

Anyway, the only post from OP that clearly state how he want the indicator works is in this post
http://www.forexfactory.com/showpost...5&postcount=29
Quote
Seems to me that RSI helps in filtering trades. I guess that can be confirmed visually. It is nice though when the signal is plotted when RSI crosses too, makes it a little easier than confirming visually. Just a thought.
One note, you should not count on global variables (not GlobalVariables) for what you said as "remembering" indicators state (RSICrossedUp) inside the for loop. It is the reason why you use "if(i<=2 && Alerts && !downalert)", isn't it? Look at sidus's code, he uses trend[i] and not a simple global variable LastTrend for "remembering" last trend.

Hope this help
  • #49
  • Jun 26, 2010 1:31pm Jun 26, 2010 1:31pm
Agreed. I didn't specify any rules when I sent the indicator to him. I just sent it and requested if he could revise so it wouldn't repaint. He had no way to "interpret what we want".


Quoting sangmane
I believe you were extracting the rules from inside the original indicator. When two or more programmer, including that "programmer", trying to extract the rules from the existing indicator which can have logic flaw, they could have different interpretation. And you should not make a quick judgement that you are correct, and others are incorrect.

Anyway, the only post from OP that clearly state how he want the indicator works is in this post
http://www.forexfactory.com/showpost...5&postcount=29
One note, you should not count...
  • #50
  • Jun 26, 2010 4:20pm Jun 26, 2010 4:20pm
Quoting squalou
sidus_v3 gives the same results with a simpler coding approach...
thanks pudseybear.

SQ
Thanks for spotting the issue with the indicator. I don't think the issues are a problem with the programmer as much as it probably wasn't specified to him what we needed.

-We are hoping to optimize the settings, add, takeaway, and do what we can with this system and see if we can create an EA for back testing, and forward testing if it works well enough. Maybe you can help with that?
  • #51
  • Jun 26, 2010 7:23pm Jun 26, 2010 7:23pm
Quoting techfan1128
It will do this from time to time, and I am not sure why exactly. Can any ciders explain this?
LMAO, a few ciders can explain many things :-)
  • #52
  • Jun 27, 2010 9:49pm Jun 27, 2010 9:49pm
As we approach monday morning and the opening of the European and london open, I will look for some opportunities and post them here. Ill be using Version 3 on 5,12,21 settings. Lets see what happens!
  • #53
  • Jun 27, 2010 10:40pm Jun 27, 2010 10:40pm
Entered trade on USD/CAD M5 Chart (Entry 1.0359, SL 25 pips). The market is in the Asian(and sunday) Range, but I took the trade because it looks like this could be a perfect senario of a bounce off the 38.2 Fib level on the 4 hour TF. Lets see what happens!
Attached Image (click to enlarge)
Click to Enlarge Name: example 11.gif Size: 10 KB
  • #54
  • Jun 27, 2010 11:05pm Jun 27, 2010 11:05pm
Heres another entry. Looking to reach fib level 23.6 then on down to 38.2. Lets see what happens
Attached Image (click to enlarge)
Click to Enlarge Name: example 12.gif Size: 10 KB
  • #55
  • Jun 28, 2010 1:28am Jun 28, 2010 1:28am
Just moved my stop to BE at the reaching of the 23.6 level. Now the next target is 38.2, if it reaches there I move my stop about 15-20 pips behind to protect profits
Attached Image (click to enlarge)
Click to Enlarge Name: example 12.gif Size: 9 KB
  • #56
  • Jun 28, 2010 1:34am Jun 28, 2010 1:34am
Just entered EUR/USD as we got a signal for the crossover, and waited from price to breakout of the Asian Range. SL 30 pips
Attached Image (click to enlarge)
Click to Enlarge Name: example 13.gif Size: 11 KB
  • #57
  • Jun 28, 2010 2:05am Jun 28, 2010 2:05am
Just BE on GPB/USD. Not know for sure if price would move to the 38.2 level I hung in there, GBP/USD BE. Im off for the night, I will update my other trades tomorrow if I get a chance
Attached Image (click to enlarge)
Click to Enlarge Name: example 14.gif Size: 9 KB
  • #58
  • Jun 28, 2010 3:39am Jun 28, 2010 3:39am
Quoting sangmane
I believe you were extracting the rules from inside the original indicator. When two or more programmer, including that "programmer", trying to extract the rules from the existing indicator which can have logic flaw, they could have different interpretation. And you should not make a quick judgement that you are correct, and others are incorrect.

Anyway, the only post from OP that clearly state how he want the indicator works is in this post
http://www.forexfactory.com/showpost...5&postcount=29

Look guys, i'm trying to be constructive here.
I did not to put any value judgement on the "programmer" himself.
At least this was not my intention.
You misinterpreted my words.
He's done good work on part of the code (and i mensionned it), but also introduced what is commonly called a "bug", which is that the indicator does not follow the strategy rules anymore.
By spotting this, i'm just trying to make things go in the right direction.
I also produce code that has bugs, but i don't take it as an offence when other people are spotting them. It it a normal development process, and this is how we progress.
If any of my codes have bugs (and i'm sure there are), then write a post, and i will be HAPPY to fix it.
And i will just be as happy if ANYONE ELSE is fixing it.
Because it goes in the right direction.

When a new bug is discovered, it has to be clearly spotted, so non-programmers won't just blindly go ahead and use a wrong indicator...
and you know what happens next on your account...

The rules as exposed by Techfan in post 1 are very clear: use this indicator that he found "somewhere", and see what we can make out of it.
This is exactly what we i am doing here.
The only way you can improve (and even USE) an indicator is by UNDERSTANDING WHAT IT DOES AND HOW IT WORKS.
If you don't, then you're heading right into the wall.
So, yes, i "discovered" the rules by the only way i had: read the code.


Quote
One note, you should not count on global variables (not GlobalVariables) for what you said as "remembering" indicators state (RSICrossedUp) inside the for loop. It is the reason why you use "if(i<=2 && Alerts && !downalert)", isn't it? Look at sidus's code, he uses trend[i] and not a simple global variable LastTrend for "remembering" last trend.

This part of the code is from the original indicator, not by me.
I just let it in because it is doing its duty as expected.

There's nothing wrong in using global variables here, and you, as a programmer yourself, should know that.
Even if you remove the indicator from the chart (so it looses the global variables values), and then reload it (with fresh zeroed variables), it will come back to the same state because it is calculating all the bars again when it loads. No flaws here. No "global variables fear" to have.

The Sidus indicator, as i mensionned also, gives the SAME results, with a simpler coding approach, which is not an optimization of any sort, by a correct understanding of the behaviour of the underlying indicators being used (EMA crosses, and RSI).
Sidus will give you the same result as V3, as long as you configure it with the equivalent inputs (EMA instead of LWMA, periods, and RSI period).

- I did not look at the alerting code, for which i have no interest at this moment.
If you found a bug in this area, then feel free to post a fix, everyone will be happy.


SQ.
  • #59
  • Jun 28, 2010 3:41am Jun 28, 2010 3:41am
Quoting eagle4x
Why did you comment that v2 has less signals, then add v3 that has much higher settings that generates even less signals than v2?

That's because you are not using the inputs suggested by Techfan in post 1, which are:
EMA 100 and EMA 110, and 50 for RSI.
You cannot compare indicators that are configured differently.

SQ
  • #60
  • Jun 28, 2010 7:03am Jun 28, 2010 7:03am
Quoting squalou
...When a new bug is discovered, it has to be clearly spotted, so non-programmers won't just blindly go ahead and use a wrong indicator...
and you know what happens next on your account...

The rules as exposed by Techfan in post 1 are very clear: use this indicator that he found "somewhere", and see what we can make out of it.
This is exactly what we i am doing here.
The only way you can improve (and even USE) an indicator is by UNDERSTANDING WHAT IT DOES AND HOW IT WORKS.
If you don't, then you're heading right into the wall.
So, yes, i "discovered"...
Squalou, sorry to dissapoint you. Your indicator that soon to be Version 4.0 repaints the past.
Try it on strategy tester to speed up debugging, use small TF like M5, use smaller EMA and RSI period to make signal shows faster.
Once you get a few arrow, especially when RSI cross 50 up and down in less than 3 bars between each cross, go to indicator's properties, hit ok to make it refresh (in real time, you can do this by switching TF). Some of the arrows will dissapear.

Do you know why it repaint? Let me tell you, it is because you use the same technique as original indicator, using global variables the wrong way. It's not about scaring to use global variables, but but it's about how to use the correct type of variables for the right objective. ExtMapBuffer[] is also a global variables.

It is normal to have bug in indicator, No one is perfect. What is not normal is make a quick (false) judgement to other's work while at the same time your work is worse. It is like point your index finger to someone, wile the remain four fingers pointed to you.

Quote
The Sidus indicator, as i mensionned also, gives the SAME results, with a simpler coding approach, which is not an optimization of any sort, by a correct understanding of the behaviour of the underlying indicators being used (EMA crosses, and RSI).
Sidus will give you the same result as V3, as long as you configure it with the equivalent inputs (EMA instead of LWMA, periods, and RSI period).

- I did not look at the alerting code, for which i have no interest at this moment.
If you found a bug in this area, then feel free to post a fix, everyone...
Nope, sidus' indicator doesn't repaints. Yours does. Read again my explanation about how to correctly "remember" previous state of indicator's buffer.

Beside understanding behaviour of indicators, rules, bla bla bla, etc, you should UNDERSTAND MQL FIRST..

Hope this help
0 traders viewing now