FREE EA - Zig and Zag (FIFO Compliant Yay!) | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Zig Zag Expert Advisor Problem 5 replies

Programing the most powerful Zig Zag 3 replies

Zig Zag Help 1 reply

Zig and Zag - anyone tried this indicator? 5 replies

Zig Zag 3 replies

Attachments: FREE EA - Zig and Zag (FIFO Compliant Yay!)
Exit Attachments

FREE EA - Zig and Zag (FIFO Compliant Yay!)

  • #41
  • May 6, 2016 8:20pm May 6, 2016 8:20pm
Quoting mojepole
Hi Daniel, quick test on default settings at DE30 1M period 21/April-April DD 4,3 % account USD 10k Looks nice, but i think that if i close trades manually with higher profit results would be better. Thanks for Your EA. Will be testing on 5M as well. {image}
Yeah, it's a tricky one. I tried the 1min chart as well and it's not too bad in some instances. I have thought about manually closing trades too but then in some cases there is a lot missed for pips. It can trade nice long trades by itself. What I really need out of this EA is a filter that can get me out from trading these bad trades without missing the good ones. I don't expect miracles but if I could filter 50% of the bad trades even by losing 10% of the good ones, the EA would be phenomenal. It is a bit premature of me to have shared it with FF but I think it needed to be shared so we can all brainstorm about this. I humbly accept that I don't think I will come with an ideal solution on my own because indicators and/or indicator strategies are not my strong point.
  • #42
  • May 7, 2016 2:38am May 7, 2016 2:38am
Here are the results after changing the code to use 4H by default and using 1M control point data
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Screen Shot 2016-05-07 at 08.36.06.png Size: 79 KB
Click to Enlarge Name: Screen Shot 2016-05-07 at 08.35.57.png Size: 83 KB
Click to Enlarge Name: Screen Shot 2016-05-07 at 08.35.48.png Size: 64 KB
So many in the Commercial section want to make me rich! :o
  • #43
  • May 7, 2016 8:47am May 7, 2016 8:47am
Quoting FX-Flux
Here are the results after changing the code to use 4H by default and using 1M control point data {image} {image} {image}
Oh right, control point method.

I think you are missing a crucial point indicated in the MT4 backtester. It says "results are not to be considered."

Does anybody know the use of control point backtesting if MT4 themselves say this?
  • #44
  • May 7, 2016 8:55am May 7, 2016 8:55am
Excuse me if I'm talking about some repetitive thing, because the Google translation is not 100% efficient.

I believe that the EA has to have a TP, TS, and BE to better administer our account ... as I said I will be testing during the course of the week in TF H4 with the parameters of the EA.

Good weekend for all.
  • #45
  • May 7, 2016 9:22am May 7, 2016 9:22am
Quoting Carneiro
Excuse me if I'm talking about some repetitive thing, because the Google translation is not 100% efficient. I believe that the EA has to have a TP, TS, and BE to better administer our account ... as I said I will be testing during the course of the week in TF H4 with the parameters of the EA. Good weekend for all.
I'm on the case Carneiro. I have been playing around with the EA today and it seems to me that even though the entry signals are semi-efficient, there is better profit and less losses to be made using trailing stops and fixed TPs it would seem. I still need to figure out a way how to filter losing trades but I will be studying a very long backtest from 2008 to present to see if I can get some insight on this.

Revised EA to be with this thread soon.
  • #46
  • May 7, 2016 9:24am May 7, 2016 9:24am
Quoting keydcuk
{quote} Oh right, control point method. I think you are missing a crucial point indicated in the MT4 backtester. It says "results are not to be considered." Does anybody know the use of control point backtesting if MT4 themselves say this?
Control points are fine on 1m if you have large tp and sl. Since the tester is feeding in open high low close, it is unlikely your stops will be hit.
So many in the Commercial section want to make me rich! :o
  • #47
  • May 7, 2016 9:39am May 7, 2016 9:39am
Quoting keydcuk
{quote} I'm on the case Carneiro. I have been playing around with the EA today and it seems to me that even though the entry signals are semi-efficient, there is better profit and less losses to be made using trailing stops and fixed TPs it would seem. I still need to figure out a way how to filter losing trades but I will be studying a very long backtest from 2008 to present to see if I can get some insight on this. Revised EA to be with this thread soon.
I do not want to be here giving suggestions,, all the time,, but see what you can do .... or try to do

EA have only valid input,, after the H4 candle is closed and toward the candle D1 for example.

EA made an entry in the H4 pair GBPUSD BUY (negative -38 pips), but note D1 candles are red.

See Sailing Week (W1) denied the SMA 34 and closed red.

These entries have to be only in the direction of the largest TFs. we have fewer entries, but with a very good chance to win ... and win many pips.

We have to go with the tide ... trend is our friend.
  • #48
  • May 7, 2016 11:01am May 7, 2016 11:01am
Quoting Carneiro
{quote} I do not want to be here giving suggestions,, all the time,, but see what you can do .... or try to do EA have only valid input,, after the H4 candle is closed and toward the candle D1 for example. EA made an entry in the H4 pair GBPUSD BUY (negative -38 pips), but note D1 candles are red. See Sailing Week (W1) denied the SMA 34 and closed red. These entries have to be only in the direction of the largest TFs. we have fewer entries, but with a very good chance to win ... and win many pips. We have to go with the tide ... trend is our friend....
Nah, it's not of anything you said. I literally was testing today and found that the close conditions are no good. As for the idea you mention above, I will look into it. I am off to work soon so can't do anything about this now.
  • #49
  • May 7, 2016 11:04am May 7, 2016 11:04am
POST #1 HAS BEEN UPDATED

I have scrapped the close conditions as they are not as good as what I have been playing with.

I have now Zig and Zag v2 because of this major change. Also the inputs are a bit more user friendly and I have updated post 1 with a screenshot of a trade condition that will hopefully help.

Any confusion, let me know.

Happy Pipping!

Daniel
  • #50
  • May 7, 2016 8:07pm May 7, 2016 8:07pm
For those of you who have downloaded in the interim of this post, please note I have found better settings than the defaults in which I noticed 37 of you have downloaded the v2 EA. They are simply SL 60, TP 0, Trailing Stop 60 & Breakeven 60. I have also posted a new backtest to show it seems much better.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot_3.png Size: 17 KB
  • #51
  • May 8, 2016 2:13am May 8, 2016 2:13am
What time period did you test over? And did you use tick or control point?

I only ask because i want to test a period you've not tested yet so we can build a bigger picture since it takes so long to test.
So many in the Commercial section want to make me rich! :o
  • #52
  • May 8, 2016 7:06am May 8, 2016 7:06am
Quoting FX-Flux
What time period did you test over? And did you use tick or control point? I only ask because i want to test a period you've not tested yet so we can build a bigger picture since it takes so long to test.
Preferably I would use tickmode but the indicators and EA are too time consuming. I use openbar and feel this is acceptable as the signals being used do not work on current bar but starts looking from previous bar. The test that you saw I think I started in January 2013. There is a lot of work to do with this I think as it can go 4 months with great results and by the end of the year going through great losses. It is still of my opinion that through filtering trades and possibly taking better entries that this EA could be something really big.

It really boggles my mind that considering it is 5min bars that 4 months I can triple my account only to find it dwindling down in the rest of the year.

More chart studying on my part and hopefully by others to come.
  • #53
  • May 8, 2016 9:36am May 8, 2016 9:36am
Well i have it on ticks at the moment and I'm testing GBPJPY 5M 2015-2016. The test has completed 3 months so far:
Attached Image (click to enlarge)
Click to Enlarge Name: Screen Shot 2016-05-08 at 15.35.07.png Size: 79 KB
So many in the Commercial section want to make me rich! :o
  • #54
  • May 8, 2016 11:50am May 8, 2016 11:50am
Hi keydcuk,
First let me say that I think this EA of yours has a lot of potential. I tried to do a backtest last night starting Jan 1st, 2015. I think it took 4-5 hours to do the first month alone so I stopped it as I needed to head to bed. Today I have been looking at the results of that first month and at first glance it seems that Carneiro's suggestion of looking at higher TFs candle to decide if the trade should be taken could have some merits. Two problems with that:

1. Correct me if I am wrong but I think that if it is implemented it would be impossible to backtest as MT4 can't handle multiple TF EAs; and
2. The time it takes to run a backtest. The backtest is so slow that testing different settings would take forever. Carneiro is right when he says traders are lazy and impatient but this is taken it to another level altogether. Someday suggested that different people could test different months to speed up testing but the problem is we all use different brokers, leverages, spreads, etc. which makes the combined results useless.

So I decided to look at the code instead in order to understand why it is so slow. Initially I thought it was the EA causing the problem and it might still be one of the cause however now I believe the indicators are the main culprits. I am not a great coder so maybe somebody else can confirm my findings. It seems like both indicators uses the function Bars to look back which deals with the number of bars in history. Most, if not all, of us have a very large number in our MT4 settings for both Max # of Bars in history and for the current chart. Both indicators loop through that number of bars and this happens on every tick. Furthermore your EA calls the Stream ZZ 10 times as it loops through your Look Back setting (default is 10). So if I understand this correctly both indicators are slowing down the backtest considerably as they both do a lot of calculations on every tick.

Maybe I am totally wrong on these findings. I am planning to make changes to the indicators to see if it can be sped up without affecting the results too much. Hopefully this way we'll be able to speed up the backtests and therefore test more settings. I think this is essential before we can start looking at your problem of filtering bad trades. I'm hoping better coders can first confirm my findings and second make changes to the indicators a lot faster than I can.

Hope this helps!

Edit: Sorry it looks like the Stream ZZ EA uses 600 as value to loop which is still quite big I think. The other EA uses Bars function.
  • #55
  • May 8, 2016 12:43pm May 8, 2016 12:43pm
Quoting jpyp99
1. Correct me if I am wrong but I think that if it is implemented it would be impossible to backtest as MT4 can't handle multiple TF EAs
You can use multiple time frames in your test.

Quoting jpyp99
It seems like both indicators uses the function Bars to look back which deals with the number of bars in history.
I think they way the ea checks back through the history may be the problem - looking too far back when it doesn't need to. As this EA was built using a generator it will be very difficult for any of us to make simple changes, as the code it creates isn't very streamline.
So many in the Commercial section want to make me rich! :o
  • #56
  • May 8, 2016 1:05pm May 8, 2016 1:05pm
Quoting FX-Flux
Well i have it on ticks at the moment and I'm testing GBPJPY 5M 2015-2016. The test has completed 3 months so far: {image}
Hi,i tested it since 2015 and like other counter EA's..when the market is trending the ea bust the account..
Dream, believe,fight for it and you will achieve and NEVER EVER GIVE UP!
  • #57
  • May 8, 2016 1:15pm May 8, 2016 1:15pm
Quoting jpyp99
Hi keydcuk, First let me say that I think this EA of yours has a lot of potential. I tried to do a backtest last night starting Jan 1st, 2015. I think it took 4-5 hours to do the first month alone so I stopped it as I needed to head to bed. Today I have been looking at the results of that first month and at first glance it seems that Carneiro's suggestion of looking at higher TFs candle to decide if the trade should be taken could have some merits. Two problems with that: 1. Correct me if I am wrong but I think that if it is implemented it would...
I think you are 100% right in what you are saying. It is the indicators that slow down the backtest as I have no problem with using tickmode in my Piggy n The Middles EA.

I've been looking at this MTF issue and it is not that I wish not to do it but I fear the worse case as it cannot perform backtest.

I am thinking, ok forward test are the best. This is all fine but how long do you wait for a live account? I could have 4 months of brilliant forward test and then one or two bad months could blow the account. It's a risky situation. I test it forward for a year? I don't know if I have the patience for this. I am looking to make something that I can trade live at the moment. I would easily trade my 1st EA live and if anything else I just think it lacks profitability even though the system is good enough for live. Yet again though, I tested 3 years and someone else tested 8 years. I found max DD to be about £600-£700 and this other FF member has found it to be $3000 because he tested it longer. I would have blown my account if I though that I could trade 0.01 lot per £1000.

So what do we do? Your answer is as good as mine.

I am looking at 2 things at the moment with this EA.

1. There is an original entry method I was using in the development of this that did not require the streamline method. It seems that streamline is good(ish) on a 5 min chart but is not great on other TFs. I don't understand the settings enough to know what I would be doing adjusting the numbers.

2. I am looking at the 4H TF. Attached is a backtest of the original method on a 4H chart using an optimal trailing stop that I find to be 120 pips. To me it is reasonably impressive I think.

The only reason I am using the 3 level Semafor EA is simply because it is the only indicator that I can use were I could program wave conditions as if I was trading a manual method if trading. i.e Looking for HH and LL, looking for SWING HIGHS AND SWING LOWS. Even though it repaints, the principle is clear. Look for HH wait for swing high retrace that misses HH and enter on break of first swing low of HH after the swing high retrace. This is a method of manual trading I had learned from one of my mentors that works pretty well. It is a shame I have to use a 3 Level ZZ Semafor to program my EA with but there are no indicators out there that I found that I can use to automate the trade method.

I DO BELIEVE I AM GETTING SOMEWHERE WITH THIS! At the moment, it's a shame this screenshot attached still has a max DD of about £1600.

Daniel
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot_1.png Size: 19 KB
  • #58
  • May 8, 2016 1:16pm May 8, 2016 1:16pm
Quoting FX-Flux
{quote} You can use multiple time frames in your test. {quote} I think they way the ea checks back through the history may be the problem - looking too far back when it doesn't need to. As this EA was built using a generator it will be very difficult for any of us to make simple changes, as the code it creates isn't very streamline.
MTF EAs cannot be backtested as far as I am aware.
  • #59
  • May 8, 2016 1:20pm May 8, 2016 1:20pm
Well good news. I made a quick change to the 3 Level ZZ indi as it was the easier to change and the time require to backtest the same month (Jan 2015) went down from 4-5 hours to around 30 mins. What I did was change Bars function to 500. So then I change it again to 200 and it went down to around 17 minutes. As there are approximately 21 trading days in a month it means it was doing each day in a little less than a minute. I can't remember exactly the return I got with the original settings but it seems going down to 500 didn't affect anything. Going down from 500 to 200 was exactly the same return so good to know.

I'll try more changes with that indicator and then I'll tackle the Stream ZZ indi but that is a bit more work as it involves changing the EA as well. By the end of today I'm hoping to be able to post the new versions so people can use them to test a lot faster and then maybe tackle the filtering of bad trades.

Please bear with me for a day or so.
  • #60
  • May 8, 2016 1:23pm May 8, 2016 1:23pm
Here is something slightly off the topic that I think a lot of you will find useful.

"You can save your MT4 reports as excel spreadsheets!!!!!!"

Simply change the extension of the file name to .xls when you are saving the file. Use the dropdown menu to "All Files (*.*)"

THAT'S IT!!

I am so excited about this...
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot_2.png Size: 88 KB
0 traders viewing now