GridTechinque why not? - page 3
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Strategy Tester has never been that reliable. Even when modelling quality is up around 90%, it does not always translate into an accurate picture of performance. I also think that Strategy Tester has a hard time dealing with pending orders like buystops and sellstops. And grid trading uses a lot of pending orders.
I have also been experimenting with grids for a while. I like them, but I am still trying to find a good way to deal with the drawdown.
@BC Brett; what EA were you using for your backtests?
My First EA
@BC Brett; what EA were you using for your backtests?
What? You mean I could have downloaded a pre-built Grid System EA off the internet somewhere? If I only knew!
No - this baby's 100% my very own brain child, start-to-finish (system design and MQL4 coding).
I was hoping to do multiple testing of variations on the original design to try and find the best set of techniques and parameters to use
for maximum profit - but when the Strategy Tester tells me my EA will yield 1,864% annual retern, why bother wasting my time tweeking the code? I mean, I think the ST results are waaaaaaaay off!
Still, I'm not too discouraged by this. I'm going to give live trading a go after I make a few more additions to the code like:
- Automatic adjustment of lot size at specific intervals.
- Keeping an eye on margin available.
- Closing out "deep out-of-the-money" trades.
Eric, if you meant "what MT build was I using for the backtests" - it was MT4 build 191. Maybe the new build 192, being released this week will give more accurate ST results. From the looks of it (http://www.metaquotes.net/forum/1884/) the new build should be a major improvement. We'll see.
I have also been experimenting with grids for a while. I like them, but I am still trying to find a good way to deal with the drawdown.
Eric, this is going to sound critical, but it's not meant that way...
Grids don't cause big drawdowns, people cause big drawdowns. If you don't like the size of your drawdowns, "Get Small". With some MT brokers you can trade as little as $0.01/pip.
Of course, that also reduces your rate of return, but most people (not specifically referring to you) expect ridiculous returns on their money.
So, "Get Small" and the drawdown problems go away. I've demoed grids that had negative floats of 100,000 pips (not a misprint). It didn't matter as I was trading that at $0.01/pip. I still had potential returns in the area of 2-5% / month, and $1000 drawdown is easy to tolerate.
Just to disclose, I haven't gone live with these grids yet, as I'm still testing. I need to have complete faith that a script will function properly before I'm going to put real money on the line. By that, I'm not referring to profitability. I'm talking about the coding/processing aspect.
Anyway, my main point of this post was to just say, "Get Small"!
Keris
Keris
Keris,
I totally understand what you're saying. You are right about using proper lot sizing, even if it means trading penny pips. People will do themselves a favor by trading smaller. My live account is with interbankfx and I often trade in micro-lots.
Maybe I should have been clearer. I am testing various ways to limit the massive "negative floats", not necessarily drawdown. Basically, I would like to reset at certain points and start the grid back up again. That is what I am experimenting with now. I have some ideas, but I am still studying codersguru's courses, so at present, I am not able to try to program my ideas yet. Hopefully soon I will force myself to devote the necessary time to learning enough to make it happen.
Eric
First of all, I know modelling quality is low, but what is that anyway?
I don't understand what modelling quality really means.
All I know is I followed CG's instructions for setting up ST and ran the test.
If it says modilling quality is low, whose fault is that?
Besides, I'm using a 25 pip grid. If modelling quality is refering to the simulated ticks, I doubt that would make much difference to the overall results since they are derived from 1 minute data. I don't think there would be more than a very low percentage of 1 minute bars with a range >= 25 pips, so how many false crossovers could ST be generating?I think you need to get 1 min data (from Alpari for example) import it and convert it to all timeframes, like in the instructions that are available somewhere.
Then you need to test on only that period of time where you got the imported data. And do the test with the "every tick method". Like this you should get the 90%, it always works for me this way.
But of course, it's probably still not entirely accurate this way, even with 90%. But it can't hurt to get the maximum possible accuracy.
Been There - Done That
I think you need to get 1 min data (from Alpari for example) import it and convert it to all timeframes, like in the instructions that are available somewhere.
I followed CG's instructions for setting up ST and ran the test.
We'll see if MT4 build 192 generates a more believable ST report.
Yes, i did not read your posts carefully enough obviously
but what's also important is that you only test for the period where you got the imported data for. Testing for a period that exceeds the "good" data will worsen the quality. It's strange because i have never seen an EA where it was not possible to get 90 or 89%. Do you get it with another EA?
Good luck anyway.
build grid EA
Lets build a grid EA. Sounds simple enough.
Lets build a grid EA. Sounds simple enough.
Cardio,
Yes I think we should try to put something together.
Here is a basic grid EA that I have been doing some experimenting with. I have been thinking about adding three additional features to this EA.
1)You can see that you have to set the direction that the grid sets up pending orders in the inputs. What I would like to do is have the option of using a simple indicator to determine whether it sets up long grids or short grids. The exact point at which you start a long or short grid is not terribly important, but you want to generally be in the direction of the trend. What I have been using is the HMA indicator that igorad programmed. So if the slope of HMA is up, all pending shorts are deleted and the EA sets up a long grid, if slope is down, all pending longs are deleted and it sets up a short grid. (see picture)
So for example, what I am thinking is in addition to being able to input the grid direction by using the Grid Direction input, have the option of using HMA (true or false) to determine the direction automatically.
2) Then, also have the option of having a switch in HMA slope cancel all opposite pending orders (true or false) and also have the option of having the change in slope cancel all open positions (true or false) in addition to being able to do it manually by using the KillOrders and KillPositions input.
3) The third thing is what I have been thinking the most about. Having a point where you close everything and basically reset the grid. I think this should be done from time to time because you will accumulate so many orders. It will be nice to start fresh from time to time. My thought is this: have the EA monitor the Equity of the account. If you had an option to set a level of equity at which the EA would close everything and reset itself. So if you were trading a $1000 mini account with dime pips, you could have the EA close everything when you accumulated a net profit of 1000 pips by having the EA close all when equity reaches $1100.
These are my thoughts on a grid EA. I have been experimenting with grids for a while (i started a thread several months ago on the MakeGrid EA that never went very far) I have been slowly learning a bit about programming in mql4, so eventually I will modify this EA to incorporate these features I'm talking about, but if someone else wanted to work on it too, that would be great!
Sounds good
Hi Eric
This sounds do-able. Please post the HMA indicator.
I am trying to finish another EA now - so I/m maxed out time wise for the next few days - but I might get a few hours to work on it.
And half the work is already done.
Thanks