Trading the anti-grid with the snowball EA

I'm posting this thread to illustrate how I use my snowball EA. The snowball EA itself has its thread here: http://www.forexfactory.com/showthread.php?t=226059 where you will find news about new versions and new features and technical issues can be discussed there while I will post screenshots of the actual usage of the EA in this Journal thread here.

I will start with a screenshot of what i have currently running on EURUSD, I started a new cycle in LONG direction on Tuesday, the exact time is marked with a vertical red line. The goal is having price cross the cloud of little green dashes that mark the break-even level. In the upper left corner are some statistics about the current cycle: losses accumulated since start on Tuesday and current floating and total profit. So far I have accumulated 397.30$ losses which might look a bit scary for the average retail trader but you will soon see what snowball can do and I will also post realtime equity curves over time that will put all these numbers into relation.

I decided to run it in long mode because I believe we will break (or have already broken) the trend line on the H4 and will see a few hundred pips long soon.

A you can see: This method needs some patience and I have much time to do other things while waiting. Yesterday it went down and hit a few more stops but also trailed down the first entry a few pips, the green break-even line was also trailed down. I now have accumulated 490$ losses¹, still nothing to worry about (there are 14000$ on this account):



¹) in case you wonder why 2 stops increase the losses by more than 60$: If you look closely you will see that there was a pause/resume (accidentially re-compiled the running EA with an error in the auto-tp function and then had to resume from there) on wednesday and paue/resume will naturally move all profit&loss from floating to realized. This will distort the realized&floating reading a bit but not the total profit or the break even calculation. Every stoploss will cost me 30$ in total and since yesterday I had two stops.

Hi,

I notice you are using a grid of 30 pips. Why 30? Is this an optimum spacing and if so... how can you tell.

I can understand that something like 5 pip grid is too small because of the many spike-outs and lost spreads.

As a test... I am trying the EA on a setting of 20 pip, account balance 5000 and lot size at 0.1.

Tim.

ps: great improvement on the the EA update.

[quote=7bit;3776665]A you can see: This method needs some patience and I have much time to do other things while waiting. Yesterday it went down and hit a few more stops but also trailed down the first entry a few pips, the green break-even line was also trailed down. I now have accumulated 490$ losses¹, still nothing to worry about (there are 14000$ on this account):

Another question:

I see the auto TP is set to two levels. If your auto TP is hit, will this make up for the 490$ loss you have already incurred?

Thanks,

Tim.

auto-tp 2 levels means 2 levels above break even. It will compensate all losses plus the additional profit from being 60 pips above break even. This additional profit (the total net profit of the cycle after all losses and gains) is estimated and displayed in the info block.

i have found that using bigger distances slightly improves the profit-factor, the break even will move away a bit slower, you will have less frequent trades and less spread costs. But I have not found a clear optimum for this value, 30 pips is just a good guess and everything less than 10 pip would probably bee too small.

I have now paused the bot for the weekend, i don't want to risk gaps bigger than the stop_distance. It was at level 0, no trades open, so nothing had to be closed. Losses so far $800. I will continue this cycle on Monday.

Thanks for the specific answers...

In the -$800 scenario you illustrate, would it not have been better to enter bi-directionally... since when we get near support/resistance/trendlines... we know that a reaction is possible... either break or bounce.

In other words, what is there an advantage of setting the grid for only longs or shorts?

On a 15 min chart... I have found I-Regr indicator a useful addition to existing support/resistance/trend lines. If the market is near the edge of the I-Regr bands... AND... it is near a confirmed support/resistance/trendline... then a reaction is near... and we would want to set our fish-traps bi-directionally.

Conversely... do not trade when the market is diddling around the centre line.

As always... constructive comments are welcome.

Tim

Yes, in hindsight you are right. I too think I should not have used the long mode, but for some reason I had the very strong belief that it would make a considerable up-move very soon and using the trailing entries in long mode would allow me to catch the absolute bottom.

The very first version of this EA had only the bidirectional mode and I implemented the unidirectional modes after I got some considerable drawdown with a nasty two weeks sideways range in GU where price kept oscillating around the entry level for nearly two weeks. I would have been profitable with either long or short in this specific case in just a few days and so I implemented the unidirectional modes and use them very often lately.

Lets just see what happens next week, -800$ is still not very much compared to the nearly 5000$ net profit that it made during the last few weeks and I have seen bigger drawdowns in earlier experiments that finally ended up in impressive profits almost as big as these drawdowns just a few hours later.

These trades are mirrored with a proprietary mt4-oanda bridge on my oanda live account and there I still use relatively small lot sizes. I trade real money but not nearly as much as is displayed in metatrader which only serves me as a platform to run the code and plot the equity charts. I have a factor that I multiply with profit and loss when I look at these numbers so i know how much it is really worth on my live account. I hope that by the time I have brought my oanda account up into the regions where I can trade such big lot sizes live I will have found ways to make better entries.

Interestingly enough it seems this EA allows me to survive and even make some profit even with the worst possible entries one could think of.

Interestingly enough it seems this EA allows me to survive and even make some profit even with the worst possible entries one could think of.[/quote]

For now I will keep going with 20 pip grid and bi-direction... 0.1 lots on a $5000 account. I will keep you posted on this week's results.

I have noticed how the EA nibbles away in a flat market... but sooner or later the moves and the profits come.

I am interested in what bridge you use between MT4 and Oanda. Oanda happens to be my favourite broker for many reasons... and I would like to know how you bridge this EA. (FX Spyder?)

Best

Tim.

from what i have read, we need to take directional trades and avoid bi-directional trades to minimize drawdowns.

we need a proven trend-following system that will give us reliable directional entries.

i have found such a system. it is called 2 MA MACD Combo and has great accuracy in choosing entries which are trend directional.

Combining your anti-martingale ea with the directional entry alerts from the 2 MA MACD Combo ea may be a powerful combination.

Trying to choose tops and bottoms is called kamikaze and sometimes you kamikaze 4 or 5 times before choosing the actual top or bottom. it is better to trade in the direction of the ongoing trend until it reverses.

Hi Wheelbarrow,

On what time frame would you use 2 MA MACD Combo?

And in regard to the Martingale word doc posted earlier... I found it hard to make sense of it... could you explain it in simpler terms?

Best,

Tim.

I just resumed it a few minutes ago, ~30 pips below Friday. BE is now another 30 pips lower.

Here are two new screen shots, break even is now around 1.2120 The second chart is the equity curve, the red line is where I started the current cycle:



equity:

This is a good test case... requiring patience to see it out of draw-down. I applaud the visual green break-even point and the offline equity chart. Both help to keep the eye on the ball.

By the way... any information you can post on bridging MT4 with Oanda would be appreciated.

Tim

It consists of mql code that constantly monitors the open positions list and detects any changes in the net exposore (order was filled or a trade was opened or closed by an EA or the user) and if it has detected a change then it will create a ticket (a file) containing the information and dump it into a certain folder. An AutoIT script will constantly monitor this folder and as soon as it finds a ticket it will read it, take appropriate action by sending mouse click and keyboard events to the running oanda applet, and if successful delete the ticket afterwards or if an error occurs that it cannot handle it will make a big noise and wake me up. The delay is usually around 2 seconds but it seems profit and loss due to delay and slippage cancel out each other in the long run.

This is one of the few pieces of code that I won't share publicly, I don't want to make oanda angry.

Somewhere on my todo list is trying to use their web interface for mobile phones instead of sending mouse clicks to the oanda applet, this should be more robust.

still waiting...


...the equity curve does not yet give any reason for being nervous:

Update for today (still waiting):

equity (H1):

The recommended account size when doing this with 0.1 lot is a few dozen kilo-dollars and when using 0.01 lot you should have at least a few thousand assigned to this strategy.

Since the beginning of this cycle on 1st of June I have burned 1141$ in stoplosses:


and the partial equity assigned to this strategy is down from initial 9807$ to currently 8697$

I also have another one running on USDJPY. I traded it short until yesterday and then reversed it to long while at level 3. Now it is at level 4 long. The magenta colored trend line has "stop" assigned to it (takeprofit):

equity:

I have put a fixed target at 1.2200