EA31337 released for free | Page 9 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

No Free Lunch but all the Free Coffee you can drink 736 replies

SWAP free, rollover free Broker? 15 replies

Hi...best time to place order after news released 0 replies

Flyer! Version 2.00 is released. 0 replies

Cad News Released 1 Min Before... ! 43 replies

Attachments: EA31337 released for free
Exit Attachments

EA31337 released for free

  • #161
  • Edited 5:01pm May 7, 2018 4:49pm | Edited 5:01pm
Quoting spojka
{quote} Thank you for quick reply. I tried to test on account with 50000 and EA works but results ere not good. There are very tight SL and TP with default settings. Can you hepl me with this problem? {image}
Based on your screenshot, you lost less than $5 out of $50000 balance in your demo account, so not big deal. Given the number of trades, that's actually a good result. You should wait a few days to see what happens.

In your screenshot, you cut the second half of the screen (below "----" line), you should see which strategy generates the losses. However, it's best, if you wait few days. If one strategy get out of the control with loses, it should be suspended after a while.

The behaviour sometimes is very inconsistent as some people report that v1.075 (see post #122) generate trades almost none, sometimes few trades per day (post #128) and you're reporting that it's generating too many trades. Unfortunately, I don't know the reason for that.

Secondly, @chrisdk2017 and @shiva, are testing v1.072, you're testing v1.075, so the logic, optimized settings, SL/TP, and amount of trades, can be different.
  • #162
  • May 7, 2018 11:26pm May 7, 2018 11:26pm
Quoting chrisdk2017
{quote} Looks odd, I have 200 pips SL on my trades.
Hi, can you send me settings which you are using? Please send it like *.set* file.
I thing that there will problem with another parameter.
What are your experiences with this EA?
  • #163
  • May 7, 2018 11:28pm May 7, 2018 11:28pm
Quoting kenorb
{quote} Based on your screenshot, you lost less than $5 out of $50000 balance in your demo account, so not big deal. Given the number of trades, that's actually a good result. You should wait a few days to see what happens. In your screenshot, you cut the second half of the screen (below "----" line), you should see which strategy generates the losses. However, it's best, if you wait few days. If one strategy get out of the control with loses, it should be suspended after a while. The behaviour sometimes is very inconsistent as some people report...
OK I will wait.
  • #164
  • May 8, 2018 3:12am May 8, 2018 3:12am
I just tried setting up 1.072 with correct set file at FXOpen and there I also get very small SL and bigger lots.

Not sure why the SL is set so low when using FXOpen?
Attached Image (click to enlarge)
Click to Enlarge Name: ea 1337 fxopen | Myfxbook - Google Chrome_205.png Size: 175 KB
  • #165
  • May 8, 2018 3:14am May 8, 2018 3:14am
The ea running at Ironfx abs zero account.
Attached Image (click to enlarge)
Click to Enlarge Name: ea 1337 | Myfxbook - Google Chrome_206.png Size: 162 KB
  • #166
  • May 8, 2018 6:28am May 8, 2018 6:28am
Quoting chrisdk2017
I just tried setting up 1.072 with the correct set file at FXOpen and there I also get very small SL and bigger lots. Not sure why the SL is set so low when using FXOpen? {image}
Small SL could be related to a low balance available (which is $1k in your case) given the calculated lot size (e.g. 0.03/0.05) and it is based on the calculated risk. For example, if your RiskMargin is set to 1% (which is by default in v1.072), for the given lot size such as 0.05, EA overrides existing strategy stop-loss value with the "safe" limit (which can be too low for low balance accounts), in order to not risk more than 1% of your balance. You can increase RiskMargin to 5% (or 0 for auto) and see if that helps. Normally EA should first lower the lot size to the minimum, then calculate the risk, so that logic for RiskMargin has been already fixed in the latest version which is calculated differently, so the old RiskMargin was replaced with RiskMarginPerOrder and RiskMarginTotal (which can be set to 0 for auto). So as already mentioned, the workaround for older version is to either increase the account balance or risk margin param, or disable boosting section for low balance accounts, or use the latest version which includes the logic fixes for better risk management.
  • #167
  • May 8, 2018 6:35am May 8, 2018 6:35am
Quoting kenorb
{quote} Small SL could be related to a low balance available (which is $1k in your case) given the lot size (e.g. 0.03/0.05) and it is based on the calculated risk. For example, if your RiskMargin is set to 1% (which is by default in v1.072), for the given lot size such as 0.05, EA overrides existing strategy stop-loss value with the "safe" limit (which can be too low for low balance accounts), in order to not risk more...
The FXOpen account balance is $1k.

Ironfx $1.5k and using smaller lots.

Minimum stop is different (FXOpen = 0 pips, Ironfx 3 pips I think).

Maybe the small lot at ironfx is caused by the minimum stop?

The Ironfx performance is excellent, by the way. Maybe just pure coincidence.
  • #168
  • Edited 9:27am May 8, 2018 7:00am | Edited 9:27am
Quoting chrisdk2017
{quote} The FXOpen account balance is $1k. Ironfx $1.5k and using smaller lots. Minimum stop is different (FXOpen = 0 pips, Ironfx 3 pips I think). Maybe the small lot at ironfx is caused by the minimum stop? The Ironfx performance is excellent, by the way. Maybe just pure coincidence.
It is also worth to note, that each magic number corresponds to the different strategy. So if you compare the same numbers, they usually share similar SL/TP values, so 31352 corresponds to (31352-31337=15) 15th strategy from the ENUM_STRATEGY_TYPE enum (counting from 0) which is most likely ALLIGATOR30, so comparing SL between two screenshots (200 vs 5 pips for the same strategy), these small stop-losses for M30 doesn't really make sense, so I believe the value was cut by risk params (most likely, not related to the different minimum stop). So, that difference of $500 in balance (and calculated lot size) in overall could impact the risk management logic how stop-losses (SL) were calculated, unless it's some other bug or factor which I'm not aware of.
  • #169
  • May 8, 2018 12:09pm May 8, 2018 12:09pm
Hi all,

Can anyone share a compiled file for v1.075 ?
I'm having the initialise failed error for the one i downloaded from the github site.
Thanks.
  • #170
  • May 8, 2018 12:41pm May 8, 2018 12:41pm
Quoting Iry
Hi all, Can anyone share a compiled file for v1.075? I'm having the initialise failed error for the one i downloaded from the github site. Thanks.
The compiled version would work the same way. You need to read the error and post it, also make sure you're using Full one, since Backtest is only for backtesting purposes.
  • #171
  • May 8, 2018 4:57pm May 8, 2018 4:57pm
Thanks so much for sharing this EA, I'm having fun letting it run on demo for a few weeks here and trying to figure out how it works.
I did have a quick question for you. How would I go about figuring out which trade method is associated with a certain magic #?
I can view stats by magic # on fxblue, and would like to remove the only poor performing trade method in the near future.
Thanks for any help!
  • #172
  • Edited May 9, 2018 9:19am May 8, 2018 7:53pm | Edited May 9, 2018 9:19am
Quoting tazmaan
How would I go about figuring out which trade method is associated with a certain magic #?
Normally EA should report on the screen which strategy trades with (and daily/weekly/monthly stats which are the best and worse), sometimes it shows associated strategy to the order when hovering on the buy/sell arrow (MT4 is not always showing it for some reason), or comments (which is also not reliable).

Otherwise, if you'd like to convert the order magic number to strategy, it's a simple math: MagicNumber + sid, where sid is the strategy ID (sequential number starting from 0) defined in ENUM_STRATEGY_TYPE enum which can be found in ea-enums.mqh file (change tag according to your version in case the list is different).

So AC1 is 31337+0, AC5 is 31337+1, AC15 is 31337+2, AC30 is 31337+3 and so on (despite AC is not implemented yet, but it's on the list).

Basically, the magic number is assigned when the order is open for the given strategy (the signal is based on the bitwise filtering conditions which were matched), so EA knows how to manage it according to the strategy-specific params (even after restart it can continue where it finished). For example, when RSI30 is opened (31337MagicNumber+95sid=31432OrderMagic), based on its order's magic number EA can identify different strategies, so trailing TP/SL can be managed differently for each order accordingly to the associated strategy (e.g., what you've set for RSI and RSI30 specifically). Furthermore, each strategy knows on which timeframe it operates on based on its magic number (so it can view and analyse the market from the different angles), therefore each order can work independently and have its own logic when to open and close the position.

Quoting tazmaan
I can view stats by magic # on fxblue, and would like to remove the only poor performing trade method in the near future.
After each run (on exit or restart), EA generates TXT report file (this is also printed in Journal log, but MT4 drops the lines) in your tester/files/ folder (open via menu: File, Open Data Folder). Once you open the file, it'll give you the list strategies and their success rate (profit factor). Based on this information, you can then decide whether you want to disable some strategies or just tweak/optimize them.
  • #173
  • Edited 10:35pm May 8, 2018 10:23pm | Edited 10:35pm
thanks for the quick reply!
I have them all written down however I see some trades for strategies that I wasn't aware I had enabled?
for example, I have magic numbers of 31381,31382,31383,31384 which appear to be 'Breakage m1, m5 and m15 and m30'
also have 31407,31410,31412 which is gator and ichimoko
How did those get enabled? my text file shows 17 strategies enabled, which I count correctly in my set file, however these aren't ones that should be trading.
A little confused now
Thanks again
  • #174
  • May 9, 2018 5:10am May 9, 2018 5:10am
Quoting tazmaan
thanks for the quick reply! I have them all written down however I see some trades for strategies that I wasn't aware I had enabled? for example, I have magic numbers of 31381,31382,31383,31384 which appear to be 'Breakage m1, m5 and m15 and m30' also have 31407,31410,31412 which is gator and ichimoko
I've checked and v1.072-v1.075 have the same list and numbers. So given this link to ea-enums.mqh , note the line numbers and that list starts at line 41, so:
31381(OrderMagic)-31337(MagicNo)+41(start-line)=85(lineNo)=(DEMARKER1); DeMarker-based strategy on M1
31382(OrderMagic)-31337(MagicNo)+41(start-line)=86(lineNo)=(DEMARKER5); DeMarker-based strategy on M5
31383(OrderMagic)-31337(MagicNo)+41(start-line)=87(lineNo)=(DEMARKER15); DeMarker-based strategy on M15
31384(OrderMagic)-31337(MagicNo)+41(start-line)=88(lineNo)=(DEMARKER30); DeMarker-based strategy on M30
31407(OrderMagic)-31337(MagicNo)+41(start-line)=111(lineNo)=(MA15); MA-based strategy on M15
31410(OrderMagic)-31337(MagicNo)+41(start-line)=114(lineNo)=(MACD5); MACD-based strategy on M5
31412(OrderMagic)-31337(MagicNo)+41(start-line)=116(lineNo)=(MACD30); MACD-based strategy on M30

I'm also attaching the list as the image (along with the line numbers in the file).
Attached Image (click to enlarge)
Click to Enlarge Name: Screen Shot 2018-05-09 at 11.02.57.png Size: 143 KB
  • #175
  • May 9, 2018 9:11am May 9, 2018 9:11am
thanks again. I see it was too late at night for me, just had simple math wrong (can't count)
  • #176
  • May 11, 2018 7:13am May 11, 2018 7:13am
One of my accounts is currently sitting in 40% drawdown so I wonder if that is considered normal?

It has reached the limit of open orders, so cannot trade more.

This is the 6k account I mentioned before which is approx 10k now (in balance).

Weird thing is another account with another data feed (different account type, but same broker, max orders 80 in chart) did not suffer from this drawdown, only 5% dd so far.

The preset and ea is the same (1.072).

Account with 5% dd has balance 1.6k now.

Could be a broker thing (max orders=15 in chart), or because of the data feed, or a vps issue?
  • #177
  • May 11, 2018 7:34am May 11, 2018 7:34am
Quoting chrisdk2017
One of my accounts is currently sitting in 40% drawdown so I wonder if that is considered normal? It has reached the limit of open orders, so cannot trade more. This is the 6k account I mentioned before which is approx 10k now (in balance). Weird thing is another account with another data feed (different account type, but same broker, max orders 80 in chart) did not suffer from this drawdown, only 5% dd so far. The preset and ea is the same (1.072). Account with 5% dd has balance 1.6k now. Could be a broker thing (max orders=15 in chart), or because...
I am losing more than 1000 usd in 10k demo too.
Im using ICmarket true ecn
  • #178
  • Edited 1:24pm May 11, 2018 10:06am | Edited 1:24pm
Quoting chrisdk2017
One of my accounts is currently sitting in 40% drawdown so I wonder if that is considered normal? It has reached the limit of open orders, so cannot trade more. This is the 6k account I mentioned before which is approx 10k now (in balance). Weird thing is another account with another data feed (different account type, but same broker, max orders 80 in chart) did not suffer from this drawdown, only 5% dd so far. The preset and ea is the same (1.072). Account with 5% dd has balance 1.6k now. Could be a broker thing (max orders=15 in chart), or because...
I know, it would be too good to be true to have 20-30% per every week.

I think today something is going on. All cryptos went down for 10-20% within last 24h, see on Binance (I'm mentioning cryptos, because this is what I trade currently). Probably everything else is also affected. For EURUSD I can see that the main trend changed (see H4).

In terms of EA. It basically doesn't monitor for breakages from support and resistance (yet), but it tries to calculate the current trend in order to avoid for some strategies to not trade against the main trend.
Trend detection has been introduced in v1.072 and since it's the new option for this version, could not work well. This feature can give the overview what's the current trend and how to trade with it. Although it improves performance comparing to the older versions, I've found later on that it really not work as expected and it doesn't adjust quick enough on trend changes, so this feature has been refactored in the latest version. In v1.072 version, the TrendMethod param is 113 set by default, but can be optimized (basically the higher, the lower timeframes are taken into the account, so the quicker EA can adjust to the trend changes, but you'll get more false-positive). The logic of GetTrend() can be found in Trade class. It basically counting pull and bear candles and decides whether the trend has changed or not (it should be reported on the screen what the current trend is according to the calculation). This can affect all strategies since v1.072 as by default all signals method only trade with the current calculated trend unless signal method param for the strategy is negative. So most likely now EA trades against the trend because didn't notice the change yet as per current value for TrendMethod (it could think it's still bearish trend by comparing the whole branch). You can try to optimize TrendMethod and find the optimal value, so EA will react quicker to the main trend changes. Code logic in v1.072:
curr_trend = Market::GetTrend(fabs(TrendMethod), TrendMethod < 0 ? PERIOD_M1 : (ENUM_TIMEFRAMES) NULL, _Symbol);

Based on above logic (counting the bars), I've found trend detection logic in v1.072 not working well, especially on trend changes (TrendMethod param), therefore it has been rewritten completely to use just RSI method which is more reliable. Current code logic:
double curr_rsi = iRSI(chart.GetSymbol(), TrendPeriod, RSI_Period, RSI_Applied_Price, 0);
curr_trend = fabs(curr_rsi - 50) > 10 ? (double) (1.0 / 50) * (curr_rsi - 50) : 0;

In addition, TrendMethod has been replaced with TrendPeriod.

I'm not saying these DD/losses can be caused by only trend detection, as there are too many factors which can play role, but this is my first guess. Also, this can give you further details how EA manages the trades. I hope that makes sense.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Screen Shot 2018-05-11 at 15.20.05.png Size: 65 KB
Click to Enlarge Name: Screen Shot 2018-05-11 at 15.28.17.png Size: 411 KB
  • #179
  • May 11, 2018 11:21am May 11, 2018 11:21am
Quoting kenorb
{quote} I know, it would be too good to be true to have 20-30% per every week. I think today something is going on. All cryptos went down for 10-20% within last 24h, see on Binance (I'm mentioning cryptos, because this is what I trade currently). Probably everything else is also affected. For EURUSD I can see that the main trend changed (see H4). In terms of EA. It basically doesn't monitor for breakages from support and resistance (yet), but it tries to calculate the current trend in order to avoid for some...
Ok, makes sense - maybe the war in the middle east is causing this.

Sitting on a lot of trades and then the trend changes, the account will go in big dd, a common ea problem.

I tried to compile 1.072 with the current classes but some files are missing, Arrays.mqh, Backtest.mqh etc. Do you have a copy of the old files? I couldn't find them in github either.

This was just to backport newest code to 1.072 to see if it behaves better with newest trend detection.
  • #180
  • May 11, 2018 12:50pm May 11, 2018 12:50pm
Quoting chrisdk2017
{quote} Ok, makes sense - maybe the war in the middle east is causing this. Sitting on a lot of trades and then the trend changes, the account will go in big dd, a common ea problem. I tried to compile 1.072 with the current classes but some files are missing, Arrays.mqh, Backtest.mqh etc. Do you have a copy of the old files? I couldn't find them in github either. This was just to backport newest code to 1.072 to see if it behaves better with newest trend detection.
Each strategy trade function (Trade_X) got this line:
result &= signal_method <= 0 || Convert::ValueToOp(curr_trend) == cmd;
so when X_SignalMethod is 0 or positive, it only trades with the current trend. Otherwise, when negative, it trades all the time. The easiest would be to run optimization and find the better TrendMethod which works for the current months, as EA was optimized using older years (and the market is always changing). If you'd like to compile v1.072 (tag), you need to clone the main repo recursively and find the right HEAD for class repo. For example, if you go to GitHub and you click on public-classes @ 93f39c2 folder, this is the version (93f39c2) of EA31337-classes files to use. Using git, you just checkout to that version (git checkout -f v1.072) and you should have the right files. But I don't think backporting will help, it could be not a simple change, but you can try. I think the easiest is to optimize the params one by one (starting from TrendMethod), or per group of params, so they can work better. Or find signal method values which can work despite the trend changes.

Quoting chrisdk2017
One of my accounts is currently sitting in 40% drawdown so I wonder if that is considered normal? It has reached the limit of open orders, so cannot trade more. This is the 6k account I mentioned before which is approx 10k now (in balance).
Unfortunately, losses play some role as well. So from 6k, you get 10k within 2-3 weeks with a bit of luck due to joining at the right time and trading along with the trend, only to prepare for the switch (or bad market conditions), so you can afford the losses which comes to the trend change. The risk is that the EA doesn't know whether the trend changed (nobody knows) as it also could be a false-positive. So by trial and error, the losses needs to be taken into the account, in order to adjust to the new market conditions. In the ideal scenario, you want to earn 2-3x more, then afford to lose less when the market goes bad. Successful trading comes with the right trend detection logic working well, otherwise trading lower timeframes against the main trend could be a suicide. Normally EA should detect the losses and take into the account, and lower the lot size to the minimum when the market conditions are bad (based on the profit factor per strategy), but it can take some time to adjust.
0 traders viewing now