Four Reliable EA's | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

How good/reliable are Forex eBooks? 23 replies

Best Reliable Sources for Daily Support and Resistance points. 2 replies

Most reliable mini platform 10 replies

MetaTrader BackTesting, Reliable or Not?? 17 replies

Reliable Data 14 replies

Attachments: Four Reliable EA's
Exit Attachments

Four Reliable EA's

  • #21
  • Feb 13, 2017 1:07pm Feb 13, 2017 1:07pm
Quoting fatolepig
{quote} Hey Dave, if I understand the video correctly, the optimizations should be rerun monthly using the previous 90 days of data. Is that correct?
Yes
  • #22
  • Feb 13, 2017 1:12pm Feb 13, 2017 1:12pm
Enclosed is an EA to automatically close all trades within a Magic Number range and Profit Target range and to disable Auto Trading. Attach it to any separate chart.
Attached File(s)
File Type: ex4 CloseAllTrades.ex4   8 KB | 239 downloads
File Type: mq4 CloseAllTrades.mq4   4 KB | 319 downloads
  • #23
  • Feb 13, 2017 2:20pm Feb 13, 2017 2:20pm
Thanks, Dave, for everything!
  • #24
  • Feb 14, 2017 10:40pm Feb 14, 2017 10:40pm
I have been running these EAs successfully on demo and live accounts after optimisation for what I thought was the full 2016 data set.

However, on closer inspection, it appears that the optimisation and back test only went from about July 2016.

Even though when back testing with other EAs, the full 2016 data set is utilized. So the tick data set kindly provided by Dave is fine.

I can get the EAs to use the full 2016 M5 data set, but just not the H1.

I have repeated the whole exercise with a newly opened MT4, but still have the same issue. I am running on Build 1045.

Has anyone else had these problems?

And one other question.

When we download the 3 months data set monthly, is the data kept cumulatively? In other words after my second download would I have 4 months of data or just the latest 3 months of data?
  • #25
  • Feb 15, 2017 3:36pm Feb 15, 2017 3:36pm
Quoting sydap2017
I have been running these EAs successfully on demo and live accounts after optimisation for what I thought was the full 2016 data set. However, on closer inspection, it appears that the optimisation and back test only went from about July 2016. Even though when back testing with other EAs, the full 2016 data set is utilized. So the tick data set kindly provided by Dave is fine. I can get the EAs to use the full 2016 M5 data set, but just not the H1. I have repeated the whole exercise with a newly opened MT4, but still have the same issue. I am running...
Free Tick Data will always have a minimum of 3 months data. For 2017, the month of January will have Nov Dec 2016, the month of Feb will have Dec 2016 and March onwards will have all of 2017's data.

Worth noting is that if you use 2016 data, the file name is exactly the same and that any 2017 data will be overwritten.

As I have stated in my last email to you, there is not any problem with 2016 60min data, the issue is your end as I have demonstrated.
  • #26
  • Edited 6:54pm Feb 15, 2017 3:38pm | Edited 6:54pm
Following on from the CloseAllTrades EA, this EA will close all trades during news periods, and restart them afterwards
Attached File(s)
File Type: ex4 AvoidNews.ex4   12 KB | 371 downloads
File Type: mq4 AvoidNews.mq4   4 KB | 387 downloads
  • #27
  • Feb 15, 2017 4:51pm Feb 15, 2017 4:51pm
Quoting Dave_C
{quote} Free Tick Data will always have a minimum of 3 months data. For 2017, the month of January will have Nov Dec 2016, the month of Feb will have Dec 2016 and March onwards will have all of 2017's data. Worth noting is that if you use 2016 data, the file name is exactly the same and that any 2017 data will be overwritten. As I have stated in my last email to you, there is not any problem with 2016 60min data, the issue is your end as I have demonstrated.
Thanks Dave. Yes I realise the issue is mine, so all I was asking is if there is anyone else who has had this issue, and if so have they found out how to fix it on their MT4.

As to my second question.
So if I have downloaded data for say Nov, Dec, Jan, then when I download new 3 months data at end of Feb (Dec, Jan, Feb), then that will overwrite my previous data set so I will have lost data for Nov. Is that correct?
If so, can that be avoided by naming the new download say AUDUSD60_0(2).fxt Then I would effectively have tick data for 4 months starting with Nov.
  • #28
  • Feb 15, 2017 5:09pm Feb 15, 2017 5:09pm
Dave, I just wanted to share results since Sunday - this is on Oanda demo account through today. l moved to live account last night and will post in a few days. I tried opening the demo account with $5000, but Oanda put it up to $1M. Oh well, the number of pips is what's important.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 10 KB
  • #29
  • Feb 15, 2017 5:17pm Feb 15, 2017 5:17pm
Quoting Dave_C
Following on from the CloseAllTrades EA, this EA will close all trades diring news periods, and restart them afterwards {file} {file}
Dave, is it okay to run all these EAs together in the same instance of MT4, correct?
  • #30
  • Feb 15, 2017 5:21pm Feb 15, 2017 5:21pm
Quoting fatolepig
{quote} Dave, is it okay to run all these EAs together in the same instance of MT4, correct?
Yes.

Use a different Magic Number for each instance.
  • #31
  • Feb 15, 2017 5:21pm Feb 15, 2017 5:21pm
Quoting Dave_C
I have had a few queries relating to the BaseCurrency input. The base currency is only used to determine the correct lot size for percentage of equity in consideration of the currency being used with your broker. You can change that to whatever currency is your own base currency. For example if you were buying EURUSD and your base currency is AUD, the lot size will be determined by the exchange rate between AUD and EUR. That co-relation will be determined automatically even if the reverse order (EURAUD) exists. If there is no relevant rate in your...
Dave, should we not change the base currency in your .set files? I did have to change some of the magic numbers as 2-3 were the same, and if I understand correctly, they all need to be unique.
  • #32
  • Feb 15, 2017 5:29pm Feb 15, 2017 5:29pm
Quoting sydap2017
{quote} Thanks Dave. Yes I realise the issue is mine, so all I was asking is if there is anyone else who has had this issue, and if so have they found out how to fix it on their MT4. As to my second question. So if I have downloaded data for say Nov, Dec, Jan, then when I download new 3 months data at end of Feb (Dec, Jan, Feb), then that will overwrite my previous data set so I will have lost data for Nov. Is that correct? If so, can that be avoided by naming the new download say AUDUSD60_0(2).fxt Then I would effectively have tick data for 4 months...
Until the end of April, you will only have the most recent 3 months of data.

The other alternative is to use 2016 data. I definitely do not recommend any more than the most recent 3 months of data. Why would you choose to backtest on old news? You need to comprehend that the market is driven on fundamentals not technicals.
  • #33
  • Feb 15, 2017 5:31pm Feb 15, 2017 5:31pm
Quoting Dave_C
{quote} Yes. Use a different Magic Number for each instance.
By "instance", you mean instance of MT4, right? That would mean all EAs would have the same magic number.
  • #34
  • Feb 15, 2017 5:32pm Feb 15, 2017 5:32pm
Quoting fatolepig
{quote} Dave, should we not change the base currency in your .set files? I did have to change some of the magic numbers as 2-3 were the same, and if I understand correctly, they all need to be unique.
The Base Currency nomination is used to get the lot size technically correct. For instance if your are trading EURUSD and you deal with your broker in AUD then the exchange rate of EURAUD is sourced to get the correct lot size. In this case you need to nominate AUD as your base currency.
  • #35
  • Feb 15, 2017 5:35pm Feb 15, 2017 5:35pm
Quoting Dave_C
{quote} The Base Currency nomination is used to get the lot size technically correct. For instance if your are trading EURUSD and you deal with your broker in AUD then the exchange rate of EURAUD is sourced to get the correct lot size. In this case you need to nominate AUD as your base currency.
Okay, because I'm in the US and use USD with Oanda, I need to change my base currency to USD then in each EA. Thanks!
  • #36
  • Feb 15, 2017 5:35pm Feb 15, 2017 5:35pm
Quoting fatolepig
{quote} By "instance", you mean instance of MT4, right? That would mean all EAs would have the same magic number.
Instance means a particular chart. In other words a particular instance of the Expert Advisor(s)
  • #37
  • Feb 15, 2017 5:36pm Feb 15, 2017 5:36pm
Quoting fatolepig
{quote} Okay, because I'm in the US and use USD with Oanda, I need to change my base currency to USD then in each EA. Thanks!
Yes
  • #38
  • Feb 15, 2017 5:38pm Feb 15, 2017 5:38pm
Quoting Dave_C
{quote} Yes
Sorry to be a PITA, Dave. Thanks for all your guidance!
  • #39
  • Feb 15, 2017 7:09pm Feb 15, 2017 7:09pm
Quoting Dave_C
{quote} Until the end of April, you will only have the most recent 3 months of data. The other alternative is to use 2016 data. I definitely do not recommend any more than the most recent 3 months of data. Why would you choose to backtest on old news? You need to comprehend that the market is driven on fundamentals not technicals.
As you know I am conducting a 12 months experiment where I will compare performance of a once only 12 months optimised SET files vs monthly re-optimised SET files. Should be interesting so I will share results since I am agnostic either way.

Yes, markets are driven by the fundamentals in terms of overall direction, but not so much in terms of volatility once the initial impact of the announcement has been made. How that will play out for these EAs is not clear to me. Whatever past period we choose will never guarantee future success because we don't know what future fundamentals will occur. Also, in 3 months that are not all that many trades in the optimisation run so the settings are not statistically all that significant?

So the reason for choosing a longer period for optimisation is to argue that if an EA can survive and prosper over all the fundamental changes that took place in the previous year(s), and we have more data/trades on which to base the settings, then it is more likely to be able to cope with whatever changes the future may bring.

I will have a progress report in six months. (Assuming I can solve my issue for the H1 timeframe in back testing! :-) )

Thanks for the Close All and News EAs. And the free tick data. Very helpful. You are most generous.
  • #40
  • Feb 15, 2017 7:27pm Feb 15, 2017 7:27pm
Quoting sydap2017
{quote} As you know I am conducting a 12 months experiment where I will compare performance of a once only 12 months optimised SET files vs monthly re-optimised SET files. Should be interesting so I will share results since I am agnostic either way. Yes, markets are driven by the fundamentals in terms of overall direction, but not so much in terms of volatility once the initial impact of the announcement has been made. How that will play out for these EAs is not clear to me. Whatever past period we choose will never guarantee future success because...
I appreciate your commitment but market behaviour also changes over time.

A good example was the bahaviour of the AUD today when the unemployment figures were released. Twelve months ago there would have been a rise in the AUD but lately the market seems to factor in an expectation on news releases and unless the data is vastly different there is little movement.
0 traders viewing now