FrankenPip's C4 EA --- Would appreciate help | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Noob here - Appreciate calendar 3 replies

help, help, help (templates) 5 replies

Would someone please help me with writing a strategy? 4 replies

I would appreciate some help with EA 5 replies

Attachments: FrankenPip's C4 EA --- Would appreciate help
Exit Attachments

FrankenPip's C4 EA --- Would appreciate help

  • #21
  • Aug 25, 2009 9:01am Aug 25, 2009 9:01am
Quoting lotabulstool
Actually when I look at the chart I see winning short trades according to the rules. But a rule from this recent post by FrankenPip should be incorporated into the ea because it is a logical trade filter that will increase the success on any trade setup.
Hi lotabulstool:

That is not a rule. It is "uncommon trading sense". It could just as well filter out a big winner. Sometimes, you have to trust your gut just do not make a habit of it.
If you only trade long, then you can't always be wrong.
  • #22
  • Aug 25, 2009 9:13am Aug 25, 2009 9:13am
Quoting hayseed
hey kostas and pardy........

Hayseed,
If you have something for release, we can help testing it and find out how it behaves in different situations.

Kostas.
  • #23
  • Aug 25, 2009 12:35pm Aug 25, 2009 12:35pm
Quoting FrankenPip
HI Pardy:

Only trade in one direction. Do not have the EA trading both ways at the same time in the same account. The rat will beat you.

Also, make sure the EA does not over trade. Set a daily or session goal, hit it, quit.
Agree with that, Frank.

Regards
Pardy
  • #24
  • Aug 25, 2009 12:47pm Aug 25, 2009 12:47pm
Quoting hayseed
hey kostas and pardy..... after rereading franks deal i believe your correct, he suggests entering while the bar is unfinished.....

if your going to use unfinished bars in this situation, odds are it will manual trading only...... not sure of the outcome......h

Hayseed,

One interesting filter would be length of previous bar (it seems that the longer the better the trade on the current bar).

Kostas
  • #25
  • Aug 25, 2009 12:51pm Aug 25, 2009 12:51pm
Quoting hayseed
hey kostas and pardy..... after rereading franks deal i believe your correct, he suggests entering while the bar is unfinished.....

if that's the case, the coloring of the bars is doing nothing more than misspainting the true live picture..... of course it's painting the historical picure correctly, but then so does the zigzag.....

look at the latest picture he has posted..... it's easy to see, we all want to see, the trade frank is speaking...
Hi Hayseed,

If you allow the Ea to take only one trade at a time, then the whipsaw will not happen. this trade must close for a profit or loss before the next trade is taken according to the rules.

Sure a candle will change color many times during its term, important point is was it the right color when it reached the .00 or .01 level. If so the trade is taken for a profit or loss regardless. that happens in manual trading too anyway, and is what price does. Probability suggests that it will go 5 pips, then in most cases u will be right!

to identify the candles, if candles colors are confusing, maybe use statements incorporating open and close price for the ea to work out if its a long or short candle. Each time a candle closes therefore, the ea knows if its a long or short candle, therefore it must take a trde only in the opposite direction. It takes the trade based upon price reaching the upper/lower point of the grid the candle is currently in, this is the trigger to place the order.

Not a programmer, but we have basically 10 grids which we can specify in the ea. therefore, a command would recognise which grid the candle is in when it closes. So if close price is < open price and it closes in grid1 ( say 00-10) then the entry trigger specifies a trde long at .10. If it closes short in grid 3, then the entry trigget specifies a trade at .40. and so on.

That would be the logic, I suppose. Hope it helps.

regards
Pardy
  • #26
  • Aug 25, 2009 1:42pm Aug 25, 2009 1:42pm
hey kostas and pardy..... both yall have explained the system perfectly.... for some reason my mind remembered using closing bars, perhaps that was c4 thread.....

rewote everything....... now opens the trade as price hits the grid line on an unfinished bar after the prior bar met it's rule.....

here are some examples.... running it in 2 ibfx mini's, will post the account numbers and read only password later....


//----------

at first we need to take every trade reguardless of how many will be taken later, that's to help assure no potential trades were missed..... otherwise you might be along time between trades..... profittargets and stoplosses are high, again due to us only concerenerd with entry's....

the grid lines are placed by the ea for ease of confirming the trades..... they actually have no bearing otherwise.... the ea is calculating the math in a separate function to place the trades.....h
Attached Image (click to enlarge)
Click to Enlarge Name: kostas7.gif Size: 15 KB
Attached Images
to trade and code, keep both simple... no call to impress....h
  • #27
  • Edited Aug 26, 2009 12:05am Aug 25, 2009 2:19pm | Edited Aug 26, 2009 12:05am
Quoting hayseed
hey kostas and pardy..... both yall have explained the system perfectly.... for some reason my mind remembered using closing bars, perhaps that was c4 thread.....
.h

Kudos to Hayseed. We got an EA!
Dude, I hope you have patience with our suggestions now...
  • #28
  • Aug 25, 2009 2:25pm Aug 25, 2009 2:25pm
Quoting hayseed
hey kostas and pardy..... both yall have explained the system perfectly.... for some reason my mind remembered using closing bars, perhaps that was c4 thread.....

rewote everything....... now opens the trade as price hits the grid line on an unfinished bar after the prior bar met it's rule.....

here are some examples.... running it in 2 ibfx mini's, will post the account numbers and read only password later....


//----------

at first we need to take every trade reguardless of how many will be taken later, that's to help assure no potential...
Great stuff Hayseed. Well impressed. Please check your pm.

Regards

Pardy
  • #29
  • Edited 4:48pm Aug 25, 2009 4:27pm | Edited 4:48pm
One filter I've been using with good results is the HMA color nrp. To only go long or short if the HMA color is also the same (green for long, red for short). Seems to eliminate some of those whipsaws. Miss some winners and eliminate some losers. Could be it doesn't even matter, but I like it.

Settings:

8
3
3
true
2
0
Attached File(s)
File Type: ex4 HMA Color nrp.ex4   6 KB | 225 downloads
  • #30
  • Aug 25, 2009 5:01pm Aug 25, 2009 5:01pm
Quoting todd960960
One filter I've been using with good results is the HMA color nrp. To only go long or short if the HMA color is also the same (green for long, red for short). Seems to eliminate some of those whipsaws. Miss some winners and eliminate some losers. Could be it doesn't even matter, but I like it.

Settings:

8
3
3
true
2
0
Price action of course comes first. By the time the hma will change colour, you would have banked those pips! If you stay with Franks suggestion on better quality trades, I find your hit rate is pretty decent. You might however want to stay in a trade longer perhaps.

I think once we test the ea, then we'll have a much better fix on improving hit rates, if that is possible, whether using hma kind of indicators or maybe obos ones.
  • #31
  • Aug 25, 2009 6:12pm Aug 25, 2009 6:12pm
I agree with pardy,

I'm just starting to test this strategy but I find that Frank's "filter" which is making sure a train goes at least 2 stations in a particular direction b4 being considered a signal candle is a good way to stay away from false signals. You could always up this minimum to 3 or more stations if you wanna play more conservative.

Could you guys possibly post the EA here so that others, including myself, can test it; or even PM me with it so I could use it in my testing.

Thank you
  • #32
  • Aug 25, 2009 6:37pm Aug 25, 2009 6:37pm
Great thread guys! kudos to Kostas and Hayseed for getting together to code this EA. I am a coder too and I have some ideas as to how we can improve it.
In my diabolic search for timing better entries I have come up with some concepts regarding tick timing which seem to work very well. I will quickly enumerate the main concept here.
I define the Tick Interval or TI as the time between one tick and the subsequent tick in milliseconds (ms). The TI concept basically consists of 3 features.
1. Reference TI - typically calculated as a rolling buffer for last 200 ticks. This typically is around 600-1000 ms.
2. Sample TI - same as #1 but only for last 10 ticks.
3. Pip count - this simply tells us if the market is going up or down during the sample 10 ticks by counting the up or down pips since tick #1 to tick #10. This could be +3, +5, -2, -4 as and example.

To put everything together we now combine the 3 above as follows. If the sample TI (#2) is smaller than (#1) this means we have a fast market and the momentum is increasing. We have a fertile condition for a possible entry.
If we have a positive Pip count higher than lets say +3 then we look to enter long. Vice versa for shorts.
This concept combined with FrankenPip Train concept will filter entries and qualify only those that meet the momentum/acceleration criteria therefore allowing us to reach the +5 TP faster. Newton second law basically states that a body in motion will continue as long as it is not opposed by a greater opposite force.
I would like to be a part of testing your EA and perhaps add my tick timing functions to it to make it really robust. Please post it as soon as feasible or contact me by PM so that I can participate in its development and testing as well.
Looking forward working with the group

Best luck to you all
If it backtests good run it live!
  • #33
  • Aug 25, 2009 6:39pm Aug 25, 2009 6:39pm
hey todd..... you are correct..... the hma, if properly applied, has the potiental to raise the chance of success..... one slight drawback is that it does repaint occassionally...... seldom, but it does.....

once we get it running, the need for filtering will be painfully obvious.... if it isn't already.....h
to trade and code, keep both simple... no call to impress....h
  • #34
  • Aug 25, 2009 7:00pm Aug 25, 2009 7:00pm
Quoting fxspeedster
Great thread guys! kudos to Kostas and Hayseed for getting together to code this EA. I am a coder too and I have some ideas as to how we can improve it.
In my diabolic search for timing better entries I have come up with some concepts regarding tick timing which seem to work very well. I will quickly enumerate the main concept here.
I define the Tick Interval or TI as the time between one tick and the subsequent tick in milliseconds (ms). The TI concept basically consists of 3 features.
1. Reference TI - typically calculated as a rolling buffer...
Fxspeedster,That is very interesting indeed. I believe Tex is working on something similar on the C4 thread, also that RR had gotten involved at some stage. The concept is sound of course. Have you created an indicator for this to start with, just so I can understand the nuances.

You're right, lets beat Goldman sachs at this game!!
Pls feel free to pm me or email me.
Thanks

Pardy
  • #35
  • Aug 25, 2009 7:06pm Aug 25, 2009 7:06pm
Quoting hayseed
hey todd..... you are correct..... the hma, if properly applied, has the potiental to raise the chance of success..... one slight drawback is that it does repaint occassionally...... seldom, but it does.....

once we get it running, the need for filtering will be painfully obvious.... if it isn't already.....h
Hi Hayseed,

I rather suspect that Frank will be right in his assertion - the rat will beat you with lagging filters.

However, there are certain nuances that can improve the performance.
I'd like to test the ea on my charts to test against those nuances and see how it performs.
What fxspeedster is proposing is actually quite good as well - just before entry!!
i have sent you my email by way of pm.

Thanks
Pardy
  • #36
  • Aug 25, 2009 7:11pm Aug 25, 2009 7:11pm
Quoting Pardy
Fxspeedster,That is very interesting indeed. I believe Tex is working on something similar on the C4 thread, also that RR had gotten involved at some stage. The concept is sound of course. Have you created an indicator for this to start with, just so I can understand the nuances.

You're right, lets beat Goldman sachs at this game!!
Pls feel free to pm me or email me.
Thanks

Pardy
Hi Pardy,
Thanks for you comments. I basically have these functions as part of my other EAs based on other concepts coded for a while but I don't have an indicator yet per se. During debugging I am displaying these values when I place the EA in the debug mode so that I can follow them.
I am sure that it could be adapted as an indicator in 10 min of coding at most. I am willing to share these and make them as part of the filter in the Train EA when the EA will be ready for further development. I originally shared my TI interval momentum concept with TudorGirl aka. CodeBreaker aka ?? and funny that Tex popped up with a similar concept. I feel we can definitely take the train EA concept to the next level for sure once this is implemented. There is no lag in the TI momentum concept and it will time our trades in a much improved fashion. You're right lets give GS a run for its money!!

Cheers
If it backtests good run it live!
  • #37
  • Aug 25, 2009 7:19pm Aug 25, 2009 7:19pm
Hayseed,

This suggestion of having at least the train traveling 2 stations seems to be a fairly good "filter" to consider.

This EA is a really great idea. Keep up the great work.

Looking forward to testing this out.

Bruce

Quoting Rbi
I agree with pardy,

I'm just starting to test this strategy but I find that Frank's "filter" which is making sure a train goes at least 2 stations in a particular direction b4 being considered a signal candle is a good way to stay away from false signals. You could always up this minimum to 3 or more stations if you wanna play more conservative.

Could you guys possibly post the EA here so that others, including myself, can test it; or even PM me with it so I could use it in my testing.

Thank you
  • #38
  • Aug 25, 2009 7:25pm Aug 25, 2009 7:25pm
I've also been working on programming an EA based on momentum-based scalping utilising both the first and second derivative of price action to try and discern entry points. I'd be very interested in part-taking in the testing and tweaking of this EA, hayseed.

Quoting fxspeedster
Great thread guys! kudos to Kostas and Hayseed for getting together to code this EA. I am a coder too and I have some ideas as to how we can improve it.
In my diabolic search for timing better entries I have come up with some concepts regarding tick timing which seem to work very well. I will quickly enumerate the main concept here.
Very interesting you should say all that fxspeedster - I've been formulating strategies based on a similar concept. I had not considered counting the time between ticks however - I was examining the velocity and acceleration of price changes purely as a function of time. Have you had any success writing something that reads and parses raw tick data along those lines?
  • #39
  • Aug 25, 2009 7:30pm Aug 25, 2009 7:30pm
Quoting fxspeedster
Hi Pardy,
Thanks for you comments. I basically have these functions as part of my other EAs based on other concepts coded for a while but I don't have an indicator yet per se. During debugging I am displaying these values when I place the EA in the debug mode so that I can follow them.
I am sure that it could be adapted as an indicator in 10 min of coding at most. I am willing to share these and make them as part of the filter in the Train EA when the EA will be ready for further development. I originally shared my TI interval momentum concept...
You're right there!!

Hope u'll write it, so I can have it in the next 10 mins!!

seriously tho' i'll pm you my email. I'd like to see how well this will fit in so we can work the nuances out, and we'll know enough then to add it in the ea, and at what point.

Regards
Pardy
  • #40
  • Aug 25, 2009 8:17pm Aug 25, 2009 8:17pm
Attached is a shot of what one of my EAs using the TI concept shows just FYI. Slow Asia market this is why the intervals are so long. Usually I stay out during this type of long intervals since the momentum has a low probability of getting you to +5 TP.
Attached Image (click to enlarge)
Click to Enlarge Name: TickIntervalConcept.jpg Size: 269 KB
If it backtests good run it live!
0 traders viewing now