Try My EA! Need Group Feedback...

That leaves you in a reasonably good position (if you have followed the advice of 0.01 lot per 1000k)

Sometimes you need a lot of patience with this method - ThePipster mentioned he had to wait over 10 days for decent breakout before it closed in profit on a recent cycle...

All the trades have a 0.03 lot size on an account of 2700. One cycled closed earlier at a loss of 200.

I don't think the risk reward is good with this system. 200 loss to gain 24 seems too much doesn't it?

It shouldn't have closed at all, that's the thing. Hopefully this update will stop that kind of thing in the future.

Lot size is OK but you must read the thread, then you will realise that the EA is still in beta form

This means it has some bugs such as cycles closing in loss or problems when changing timeframe or restarting MT4 etc.

It's a work in progress which is why ThePipster shared it with us and dgodfrey is doing a stellar job ironing out the kinks

Well I've burned my fingers! Hopefully the new version can recover what I lost :-)

http://www.kitco.com/images/live/nyg...07787682455684

Don't worry about the DD...it is normal for this EA. You just have to be patient for price to move far enough to profit. You may go a week without seeing trades closed, if the market is flat.

Here's the new release. Because of the issue of the arbitrary shutdowns and not being able to do normal things without messing it up, I would advise you all to try this version. It's my hope that all those stability problems are fixed.

Please read the explanation carefully. There are some important changes.



The variables:
TradeAmount: The percentage of your account balance to trade in dollars. The program uses this to calculate the correct lot size for your boroker. This should work regardless of what your broker calls 1 lot.
Example: $1000 account, you want to make the trade size $1, then put 0.1 here, 1/1000 of your account. This will scale with your balance, so it is naturally going to compound. You can adjust it as needed.
NOTE: If you put in a value that results in a lot size less than what your broker allows, the program will correct your entry to the lowest amount allowable.
Another NOTE: The compounding occurs in $1000 steps, i.e. your balance rounded down to the nearest $1000.

Unique: This is what should fix the stability issues. Unique is used to identify this instance of this program running on this computer. You need to put in a number you can remember to assign a unique global variable to your computer's memory which stores the beginning balance at grid initiation. You can use this program on as many MT4 platforms on a computer as you want (I have 5 going right now) to check brokers, inputs, etc. Just make sure each one has a different unique number. And don't forget it in case you have to restart it for some reason. The program can then be turned off, timeframes changed, all that stuff without danger of resetting the beginning balance figure and taking a loss, or for corrupting another instance of the program running on the same computer, causing it to recycle unexpectedly.

Spread, Buffer and Gridstep: Enter in cents. Even if you have a 3 digit broker, the program will correct for that and give the proper steps.

UseSlippage: Setting this to false will make market orders of your trades, and you will get the fill on exit that you get, regardless of what the price is by the time the trade clears. Leaving false here avoids orphans, because the program will redundantly search for orphans until all the trades are closed. If you set this to true and the value in slippage will not allow the trade to be executed, you'll get an alert that the grid is corrupted and you'll see the orphaned trades. Just be aware in a fast market market orders might help or hurt a lot. Of course, being stuck in some trades on the wrong side of a runaway market isn't going to feel too great either if you use slippage, and the program won't run again until all the trades in the ending cycle are closed.

Target: Enter the percentage of your account you want for a target gain.
Example: Target=1 will set a target of $10 on a $1000 account.

UseTimeToStart: Set this flag to true to stop and start at specified times given below the flag, or to false to ignore the time settings.

And that's all for now. I'm going to try to make a dynamic grid in the next version that will reach out to the BE and Target lines if necessary. As you can see here:



only a few more buys and my BE and Target on the sell side will be beyond the grid, and also will be inaccurate if the grid isn't present. I'll also work in the actual P/L so the numbers will be more accurate.

Give feedback on problems you have so we can incorporate and fix them as they arise

Thanks for updating the EA!
Your account is doing nice!
Mine did alright today.
Have a good weekend, everybody.

Outstanding work dgodfrey, thanks!

dgodfrey,

are the attached image is your current settings?

Just looking at the settngs, has the option to trade fixed lot sizes really been removed?

Yes, other than the unique number, but those are just some settings. Spread should be reflective of what your broker's spread really is. Buffer is just where you want the grid to start, it could be small for close up to the current price, or large if you want more wiggle room. GridStep can be anything, just try stuff to see what has the best effect.

I should say that a couple of the big gains were because I was doing testing and the trades didn't close on target, but ran away in my favor and I closed them manually. In my first demo I had the target set at 10% of the account balance by mistake and made over 100% gain in a day. But then I got boxed in during volatility and gave most of it back. Getting in a hedge box is easy if your target is too large, but too large is relative to GridStep and the number of steps and a lot of things.

I think experimenting with a wide range of settings would be instructive, on demo accounts of course.


I did I can see how I should have left that, I'll put it back in the next release. I was just trying to simplify the differences in brokers and what lots are for each. Guess that's my own biases creeping in. I like compounding. I'll put it back in and post the new version this weekend. Sorry!

Here's a version with fixed lot size if you want to go that route.

This release is identical to 2.1 except for this minor change. Be aware that if you use a fixed lot size less than the broker allows this program will set it to the minimum.

You are a gentleman and scholar, thanks!

No problem Shouldn't have taken it out in the first place lol.

One thing I noticed, there's a code remnant that will cause an alert box and reset if you make the TradeAmount = 0. Even if you use fixed lots, leave a value in TradeAmount. It will be ignored.

Thank you for the updates, dgodfrey! You've managed to contribute so much in such a short time.

Can't wait until tomorrow to see the latest version working. Not to mention the dynamic grid - when it arrives, it will be pure fun to watch!

A similar idea has been hovering in my mind for some time now. Not to say we should try and push things too far, too fast, but it's surely something to bear in mind until we're sure the EA is perfectly stable and bug-free.

The next step could be to try and apply some optional (indicator based) add-on triggers, which would activate the EA (or notify the trader) only when the market is at its extremes. Of course, anything indicator based will only be approximate, but with the use of manual confirmation, there's always room for discretion to pinpoint the entry.

In effect, the scale-in grid will become a very precise sniper, with minimal DD. By adjusting grid spacing to the timeframe, it'd be up to the trader, whether to use it for day- or swing trading.

Mulling over the numbers today I came up with some analysis of what we're doing here. Just for your information.

An interesting thing is that one effect of a constantly growing grid is that the returns are exponential in nature.



Assume that the grid spacing is set to $0.50 and the amount traded is $10 on a $10k account (0.1 lots for 1lot=$100 brokers).
A move either way of only 5 grid steps will hit a 1% target, ie $3.00 away from the origin, assuming there is a $1.00 buffer from the bid when it was started and no hedging took place. Sounds good. But you will reach 2% gain before you get to the 8th step, and a 10% gain is reached in just over 18 steps. The growth of your return is exponential.

Make the same assumption, and do a grid with the setting at $0.30 steps. Now, to get a 1% return you will have to move about 6 steps, but that's only about $2.50 from the origin because the steps are smaller. W00T! No brainer, you think. Here's the catch. Your box is much smaller, around $20 top to bottom, with 30 cent steps. With the step size at 50 cents its about $30 top to bottom. The smaller step size is much more likely to build a boxed hedge. The closer steps will also build an inside hedge loss faster in a sideways market as the price wiggles around. With the steps closer together you're going to get that hedge loss going faster.

And one other caveat that came up from this: The bigger the step, the bigger the DD if you get boxed. You need to consider that risk. Early on Pipster told us the risk of loss building a perfect box hedge was 30% DD, BUT that's with 0.30 as the step size. (Actually with the above scenario and a $1 dollar buffer it's around 32% for 30 cent steps) But if you increase the step size to $0.50, that same max DD figure goes to almost 50%!

Rule of thumb: step size, in cents, is roughly equivalent to the percent DD if the hedge you build with it is boxed perfectly @ 30 steps. And yes, it does get almost to complete DD when the step size is $1.00, a 100% DD. Of course by that time your hedge is $60+ from top to bottom, so what's the likelihood of that?

Upshot: Smaller stepsize: easier to get to the target, easier to build hedge loss in sideways markets, easier to build a max DD boxed hedge, less risk if you do. Larger stepsize: harder to get to the target, harder to build hedge loss in sideways markets, harder to build a max DD hedged box, more risk if you do. And we may be shorting ourselves by making the targets too small. There is probably some optimal range of numbers - lets find them

I was observing this same thing, Kosmo. How nicely you take profits originating at the pivots in price, how miserable it is in the middle of that range. There must be something we could do about that. I got creamed during that Yellen announcement, made a perfect boxed hedge. At either end of it I would have been golden.

Another thought, weekend gaps where you might have the following scenario:

Friday night close
You entered at $1995 with $0.50 gridsteps earlier in the day so you have a grid going from $1980 to $1210
Price had moved from $1995 to $1998 and you picked up 6 buys
Price had also moved from $1995 to $1993 so you have 4 sells
Market closes at $1997
Market opens with a gap down to $1985 and 16 sell stops all get triggered at that price
Market then reverses back up
Ouch...