Testing new Breakout EA | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Testing the G.TN. London Daybreak Breakout Strategy 0 replies

New Trading Guide for the New Year (Need Help Testing) 86 replies

Strategy Testing - Testing 2EAs simultaneously 5 replies

Testing Parabolic breakout 24 replies

Testing the breakout. 14 replies

Attachments: Testing new Breakout EA
Exit Attachments

Testing new Breakout EA

  • #21
  • May 17, 2015 9:49am May 17, 2015 9:49am
Quoting ibthescottyb
{quote} the EA should work on any demo account. Only if it is placed on a real account should "The EA is not on an authorized account" appear. Or, wait until the market opens and the EA will connect with the broker for a clean initialization.
It could also be possible that the MT4 platform you are running the strategy tester on is on a live account. The EA will see that and not run the test.
  • #22
  • May 17, 2015 10:08am May 17, 2015 10:08am
Hey ScottyB,

Sent you a PM forgot to mention that those code snippets allow to test with 99% modelling quality using those bar types.

Best Regards
  • #23
  • May 17, 2015 10:09am May 17, 2015 10:09am
Unfortunately, I am not a programmer. I designed this EA but paid someone to code it for me. This is the reason I am hopeful our fellow traders can find a set file that works with it.

I would recommend using the EA in manual mode on the chart configuration you recommend. Post your results in the forum for all to see and comment on. If it proves consistently profitable, we can begin to explore making that code change.

If we cannot find a profitable set file, then I will begin to explore giving the source code to someone here to make changes based on other trader's ideas. Then it becomes a team project that all can benefit from and we are maintaining a limited sense of version control.

I think a horizontal channel can work in many different instances, but I also want to keep it here in a public forum on FF and not have source code spreading all over the internet.

Thank you for joining the forum and if you find my EA works on your chart style, please share
  • #24
  • May 17, 2015 12:18pm May 17, 2015 12:18pm
Scottyb, I am getting a following message when trying to back test, "FX Rogue not on authorized account -please contact administrator"

Do you have any idea why?
  • #25
  • May 17, 2015 12:22pm May 17, 2015 12:22pm
Quoting OnTheEdge
Scottyb, I am getting a following message when trying to back test, "FX Rogue not on authorized account -please contact administrator" Do you have any idea why?
the EA should work on any demo account. Only if it is placed on a real account should "The EA is not on an authorized account" appear. Or, wait until the market opens and the EA will connect with the broker for a clean initialization. It could also be possible that the MT4 platform you are running the strategy tester on is on a live account. The EA will see that and not run the test
  • #26
  • May 18, 2015 7:44am May 18, 2015 7:44am
Iam in!
I did set it up on IC Markets demo acc. on EURUSD, GBPUSD and USDCHF 5 minute charts.
Here's the myfxbook link to the acc.:

https://www.myfxbook.com/portfolio/m...out-ea/1251075

I will check how it performs during the next 2 or 3 weeks and if it is quiet acceptable I let it run on my micro live account to see how it's doin in the live market environment.
  • #27
  • May 18, 2015 10:20am May 18, 2015 10:20am
Quoting Mr.Ma
Iam in! I did set it up on IC Markets demo acc. on EURUSD, GBPUSD and USDCHF 5 minute charts. Here's the myfxbook link to the acc.: https://www.myfxbook.com/portfolio/m...out-ea/1251075 I will check how it performs during the next 2 or 3 weeks and if it is quiet acceptable I let it run on my micro live account to see how it's doin in the live market environment.
Mr. Ma - Welcome. Are you testing the EA "as-is" or do you think there is a better time and place to use it? I am playing with some manual entries to see how it does since I am not 100% convinced that an automated version works consistently.
  • #28
  • May 18, 2015 10:58am May 18, 2015 10:58am
Quoting ibthescottyb
{quote} Mr. Ma - Welcome. Are you testing the EA "as-is" or do you think there is a better time and place to use it? I am playing with some manual entries to see how it does since I am not 100% convinced that an automated version works consistently.
Hi Scotty,

the only changes I made are:
I have set a Take Profit of 25 pips and a Stop Loss of 15 Pips.
And I have set the "AutoChannelMult" to 0.75 so the EA starts to trail the SL when price reached 75% profit of the size of the channel.
Plus I let it start trading one hour prior to Frankfurt open and let it stop trading when New York closes.
  • #29
  • May 18, 2015 12:39pm May 18, 2015 12:39pm
i am getting some message when i am trying to backtest it can you tell my why? i am following your all rules.
  • #30
  • May 18, 2015 2:12pm May 18, 2015 2:12pm
Quoting mdrasel
i am getting some message when i am trying to backtest it can you tell my why? i am following your all rules.
Hi mdrasel, could you tell me what the message is?
  • #31
  • May 20, 2015 9:42am May 20, 2015 9:42am
Hi scotty
I am testing EA in almost original setup on 5m EU.
Result looks promising.
If this week testing finishes with same accuracy it would be interesting to test it on live account to test broker slippage and other conditions.

Regards
Attached File(s)
File Type: zip FXRouge.zip   9 KB | 655 downloads
  • #32
  • May 20, 2015 12:55pm May 20, 2015 12:55pm
Quoting GoForPip
Hi scotty I am testing EA in almost original setup on 5m EU. Result looks promising. If this week testing finishes with same accuracy it would be interesting to test it on live account to test broker slippage and other conditions. Regards {file}
Would you mind sharing your settings?
Iam testing it on M5 also, but I definately don't have the trade history like you have
  • #33
  • Edited 3:28pm May 20, 2015 2:41pm | Edited 3:28pm
Quoting GoForPip
Hi scotty I am testing EA in almost original setup on 5m EU. Result looks promising. If this week testing finishes with same accuracy it would be interesting to test it on live account to test broker slippage and other conditions. Regards {file}
Looks good. My backtest was on EURUSD 1m. with a test spread of .2 pips. I tested up to .6 with similar results. After that is failed. I tested on FXCM that has similar spreads and the results were terrible. After seeing yours, I am trying on EURUSD5m. We'll see - thank you for posting!
  • #34
  • May 20, 2015 3:36pm May 20, 2015 3:36pm
Quoting Mr.Ma
{quote} Would you mind sharing your settings? Iam testing it on M5 also, but I definately don't have the trade history like you have
I use sleep time/ true nothing else is changed
  • #35
  • May 20, 2015 3:45pm May 20, 2015 3:45pm
Quoting ibthescottyb
{quote} Looks good. My backtest was on EURUSD 1m. with a test spread of .2 pips. I tested up to .6 with similar results. After that is failed. I tested on FXCM that has similar spreads and the results were terrible. After seeing yours, I am trying on EURUSD5m. We'll see - thank you for posting!
I also am testing on one minute at second platform with settings from your backtest. Results are not as good and it seems EA closing one trade on SL and opening another after SL is not exact, so I sometimes have buy and sell open simultaneously. As for 5 min setup it results are far above my expectations but further testing is needed especially on live account. RR might be an issue but for now it doesn't affect the result and EA has very good "sniping optics". My regular spread on EU is in average 0.2 and I am running EA at home and on VPS.
  • #36
  • Edited 5:34pm May 20, 2015 4:56pm | Edited 5:34pm
Quoting GoForPip
{quote} I also am testing on one minute at second platform with settings from your backtest. Results are not as good and it seems EA closing one trade on SL and opening another after SL is not exact, so I sometimes have buy and sell open simultaneously. As for 5 min setup it results are far above my expectations but further testing is needed especially on live account. RR might be an issue but for now it doesn't affect the result and EA has very good "sniping optics". My regular spread on EU is in average 0.2 and I am running EA at home and on VPS....

Hi GoForPip - When I started live demo testing on a ECN account with small spread (.1-.5 pips), I also saw that error with the SL being set outside the channel causing either two trades to be opened, or worse, not allowing the next trade in sequence to open due to hedging rules. I sent to my coder to fix and have posted an update on POST #1. This version also shows on the screen SPREAD in decimals instead of whole numbers to show a more accurate reading. All other coding is the same.
  • #37
  • May 21, 2015 2:55am May 21, 2015 2:55am
I have found one more issue with EA. If it fails to modify, order as it did for me ( on one platform it has successfully modified it and on other it didn't) EA should check condition after order-modify is sent and try again. So one EA has successfully closed the trade with "AutochannelTS" and other did not and has the order still open. Obviously this has happens during increased volatility.
  • #38
  • May 21, 2015 6:22am May 21, 2015 6:22am
Quoting GoForPip
I have found one more issue with EA. If it fails to modify, order as it did for me ( on one platform it has successfully modified it and on other it didn't) EA should check condition after order-modify is sent and try again. So one EA has successfully closed the trade with "AutochannelTS" and other did not and has the order still open. Obviously this has happens during increased volatility.
Thank you for the post. We can add that to the bug fixes. Is the order still open in profit? or Negative? What was not modified? the TP/SL?

I have had that occur if I have ecnBroker = TRUE on a non ECN broker. Or if I had it to FALSE on an ECN broker.
  • #39
  • May 21, 2015 7:15am May 21, 2015 7:15am
Hi.
It happened on ECN broker. Settings for ECN / false but with this setting it was working without any issues until this one, I also must mention that it happened on my home PC, while on VPS order was modified just fine and considering that connection between my home PC and server is about 120 ms and between VPS and server is 1 ms, so that difference in speed would justify discrepancy.
Order has later closed without any issues in profit when price hit the line.
Modification was for I think for AutochannelTS, if I presume correctly, that function moves TS/SL right behind the price when condition is met.
Attached Image (click to enlarge)
Click to Enlarge Name: 2015-05-21_1401.png Size: 9 KB
  • #40
  • May 21, 2015 7:42am May 21, 2015 7:42am
Quoting GoForPip
Hi. It happened on ECN broker. Settings for ECN / false but with this setting it was working without any issues until this one, I also must mention that it happened on my home PC, while on VPS order was modified just fine and considering that connection between my home PC and server is about 120 ms and between VPS and server is 1 ms, so that difference in speed would justify discrepancy. Order has later closed without any issues in profit when price hit the line. Modification was for I think for AutochannelTS, if I presume correctly, that function...
Thank you for the image. It is not connected to the AutoChannelTS as that is only for trailing stop and is not triggered until the price crosses the settings in AutoChannelMult variable.

I have seen this on a couple occasions. It always occurs during order open as it does on your image.

In this case, the modification was attempting to load the stop loss at the opposing side of the channel, depending on the latency, the market could have moved back near the opposing side of the channel making the market price too close to price the SL was to be placed. This can happen if your broker has limits of distance for modifying orders. But possibly the latency in sending the modification could also be the culprit. There were 36 seconds between order entry and modification.

I know in the past that a true ECN environment, usually it takes two mt4 sends to complete the order. The first send the order itself (Size and direction) then a second send immediately after with TP/SL information. If, ECN equals FALSE, it will send all the information in one send. But, on an ECN broker, speed is an issue which can cause this issue. Some ECN brokers accept 1 send, but the EA still has to modify TP, SL and channelStopLoss once complete. Latency can be a major factor in this.

The only real way around this would be to change the EA entirely to place pending orders instead opening order on the fly. But that would take a complete re-work of the code and if the problem is solved with a VPS than that is the way to go.

That is a major problem in back testing. There is no latency, requotes, or communication issues. It is a perfect world. It looks like the challange is to find the right setup between set file, VPS and broker to make this magic happen. Otherwise, it is just an exercise.
0 traders viewing now