• 概述
  • 评论
  • 评论

Eurusd MX Trader

This is an experimental EA, still in the making. It works using an improved algorithm of our main EURUSD indicator, adjusted for the 1-hour chart specifically. The parameters used are too complicated to fine tune, and it will still take us months to do so, however early tests show that this EA is profitable over many hours to few days. It does take losses but seems to make them back.


Latest tests-settings:  This EA has been perfoming better on USDCAD (hourly chart, sensitivity at 1.000008), and average on EURUSD (30min chart -

sensitivity at 1.0000116). Testing is in progress, and we have 5 more variants based on this EA, to test.



All in all,  we are developing EAs for EURUSD,  USDJPY and GOLD.  It's the same algorithm, but optimized for each market individually, where the objective is to minimize losing trades and let the market give whatever profits are possible.  The good thing of this algorithm is that it doesn't degrade with time, as market conditions change. The refinement of parameters is most of thr work, and it is done through 3rd party machine learning tools. So that the user of an EA can change the perfomance of an EA simply by changing these settings.


This EA requires data from a wide range of EUROPEAN pairs, EURCHF,  as well as USDJPY,  these must be enable for the EA to work.


The EA may also work few other pairs, such as EURAUD or GBPUSD,  but it's not optimised for those pairs. We havent't tested for all pairs, there

may be a pair where it works best.  We will work this year to perfect this EA, for at least one pair, the workload is overwhelming, and the desired result is profitability over any randomly chosen 6 hour test!


90 % or even 99% Backtesting is not reliable enough for this EA, since it uses too many data, from all these pairs, instead we do tick by tick testing on random 6 hour intervals, and so far it is doing good.


We will be working on the settings furthermore and few other versions for this EA,  for now you can more or less use the default settings, especially

the sensitivity setting is so critical, the higher the number the fewer trades the EA makes.


Please make sure sensitivity is set to 1.00001 this will result in around 3-4 trades per 24 hours. It seems to yield good results. You can set it as low

as 1 and as high as 1.000018 or even 1.00002,  these values are under testing. Sensitivity factor works in an inverse way, the higher the number the less sensitive the EA becomes, the less it trades.


To achieve very good results in real trading, we may have to set sensitivity factor closer to 1.00002 or use variable sensitivity, which will be controlled by the EA itself and various parameters. The question is now whether the EA really works as it seems, and if it's worth dedicating one more year of development.


This EA trades one trade at a time,  for now,  it's too boring to watch live (only 4 trades per 24 hours), and attempts to make an indicator are not impressive, because the indicator is unusable at such slow trading, so it has to be an EA.  However, we are able to manually trade when the EA places a trade, by waiting for the market to move against our trade, sooner or later it will converge again, and we do one such additional manual trade to figure out how to include multiple trade capability in future versions.








无评论