Martingale with "low" margin hit

[DragosDanescu]

I thought I had it. Had Aud/Usd set up to buy with a levelsoffset of 12, stop level of 2 and tp of 6. Trade was opened with a stop loss of 32. Trade info shows levels offset of 16.1.

It's correct! Ea had found an ib/Engulfing bar that has 32 pips range, therefore LevelsOffset is half of that (16.1), and TP is 6 times the LevelsOffset, therefore 96. That's dynamically calculated according to the trigger candle size. In case of Time trigger, it's fixed(at the specified input) or also dynamically calculated according to the previous day range(if you select AutoLevelsOffset=true, that will override fixed value).
For Time option, I've mentioned earlier, use only hours, no minutes.It uses as reference H1 candle, that's why.So, instead of 2:30, use 02:00. Hours has to do with how many hours after opening the order you want to force closing it.

Also, I reccomend to use only one option, using both doesn't seems to work, it looks only for Engulfing.Have to figure out why....

[blahn]

How about grabbing 7-bit's button code from his snowball grid ea so we can manually start/stop the grid without waiting for triggers?

https://sites.google.com/site/prof7bit/snowball

[DragosDanescu]

Look at what I'm working now...
I'm still using a trigger(seems a better choice from my perspective), Ea is starting a straddle with 0.01 lots then, if the result of the straddle becomes negative ( a user input amount, I use 2*Expected Profit)
EA is "shrinking" the straddle to as less as 5 pips, as you can see in the picture, but starts to autocalculate neccesary lots to recover the loss, based on initial TP.
So, basically, if the straddle (range between BUY and SELL) was 40 pips, meaning a TP of x*40, EA automatically reverse this at 5 pips, but maintaining the same TP level.
This one was closed at 8th reversal, at max lotsize of 0.03 (base lot 0.01).

[spyderman]

Look at what I'm working now...
I'm still using a trigger(seems a better choice from my perspective), Ea is starting a straddle with 0.01 lots then, if the result of the straddle becomes negative ( a user input amount, I use 2*Expected Profit)
EA is "shrinking" the straddle to as less as 5 pips, as you can see in the picture, but starts to autocalculate neccesary lots to recover the loss, based on initial TP.
So, basically, if the straddle (range between BUY and SELL) was 40 pips, meaning a TP of x*40, EA automatically reverse this at 5 pips, but maintaining the same TP level.
This one was closed at 8th reversal, at max lotsize of 0.03 (base lot 0.01).

Hmmm...You'll have to keep us updated. I'm still testing your EA and trading manually at the same time.
I think this type of system will be hard to monitor in terms of profitability as you get quite a drawdown as the cycles progress and then a big positive if you hit your bail out. I've had one trade so far that went 10 cycles and I bailed at a loss. Other then that all trades have hit TP or BE.

We'll see.

[skoda2008]

Look at what I'm working now...
I'm still using a trigger(seems a better choice from my perspective), Ea is starting a straddle with 0.01 lots then, if the result of the straddle becomes negative ( a user input amount, I use 2*Expected Profit)
EA is "shrinking" the straddle to as less as 5 pips, as you can see in the picture, but starts to autocalculate neccesary lots to recover the loss, based on initial TP.
So, basically, if the straddle (range between BUY and SELL) was 40 pips, meaning a TP of x*40, EA automatically reverse this at 5 pips, but maintaining the same TP level.
This one was closed at 8th reversal, at max lotsize of 0.03 (base lot 0.01).

Interesting idea. I have spent more time than I care to mention trying to perfect cost averaging / hedging / semi-martingale systems over the last few months, and had all but given up. I found that the long raging periods happen a lot more than I expected, and that if you keep the lot sizes small enough to survive that the profitability is just not worth the risk. I'm keen to see if you guys manage to come up with something.

I can also help with coding, but tend not to use Steve's shell, as I understand my own programming logic better. Also, Empty4 is starting to drive me nuts with it's lack of real backtesting capability, not being able to test multi-currency strategies etc. I have decided to teach myself Dukascopy / JForex instead.

[garyfritz]

I agree the long ranging periods happen more often than you expect. I wonder, what might happen if you started widening the channels after each iteration? Your losses would become bigger but maybe you'd have less of them.

I also agree about Empty4's backtesting. I'm used to Tradestation, which has some limitations (e.g. no portfolio testing, though supposedly they're fixing that), but overall it's a pretty powerful platform. Don't know anything about Jforex but maybe I should learn. It looks like it's limited to Dukascopy and forex? Tradestation is limited to Tradestation Securities (though you can call something like NinjaTrader to issue orders on almost any other platform), but it can be used for forex, stocks, futures, options, etc. And TS seems to be a pretty decent brokerage.

[jb68]

I agree the long ranging periods happen more often than you expect. I wonder, what might happen if you started widening the channels after each iteration? Your losses would become bigger but maybe you'd have less of them.

I also agree about Empty4's backtesting. I'm used to Tradestation, which has some limitations (e.g. no portfolio testing, though supposedly they're fixing that), but overall it's a pretty powerful platform. Don't know anything about Jforex but maybe I should learn. It looks like it's limited to Dukascopy and forex? Tradestation is limited to Tradestation Securities (though you can call something like NinjaTrader to issue orders on almost any other platform), but it can be used for forex, stocks, futures, options, etc. And TS seems to be a pretty decent brokerage.

Hi,

I noticed you have some math background so I'd like to point out a few ideas..
Why should we limit to only one way of martingale by doubling our bet? This has serious limitations like it requires equal W/L ratios.

Math speaking, martingale is a way of betting the outcome distributions. Any strategy based on betting the outcome distribution is a martingale. Es an extreme if we know after each 3 losses we will have one small winner and then a big winner we can make money with a special crafted martingale. But we don't...

Most of strategies presented here are not suitable for the doubling martingale. However there are different ways of adding martingale like using different coefficients like Fibonacci numbers and not have the martingale reset after the first win but let's say after the account is back to 80% or 95% or 100%.

[garyfritz]

The idea behind most Martingale recovery systems you see in the Empty4 world is simple: "I lost that time, so I'll bet twice as much next time. Whoops, lost again, I'll double my bet again. Whoops, ..."

There are plenty of variations but they all have that same basic idea. Most of them revert to normal position size after one or a few wins. I like your idea of getting up to a max fixed size, then stay there until you build the account back to a new equity high -- except that does the opposite of what you generally really want to do. When you're losing, you want to *reduce* your bet size. If you increase your bet size and keep it there, and you enter into a long losing streak, you could be in a lot of trouble. But you'll probably be better off than you would have been with an unbounded Martingale.

[jb68]

except that does the opposite of what you generally really want to do. When you're losing, you want to *reduce* your bet size. If you increase your bet size and keep it there, and you enter into a long losing streak, you could be in a lot of trouble. But you'll probably be better off than you would have been with an unbounded Martingale.

In my testing any martingale added to an EA will make things worse. In my tests is better to just play a bigger lot size instead of martingale.
What I mean is instead of running only one EA with bigger lots would be better to run the same EA on 2-3 pairs and having a mild martingale coefficient. Unfortunately Murfy's law said it will always happen an all 3 in the same time.

Lowering betting size after you lose money have a few grounds like:
- losing money will affect your making decision
- market is in bad mood for your system
- you need to preserve money in order to have available money to bet
On first point an EA doesn't have feelings but the rest remain valid points. The bad thing about the market is there is no way to know when our system will start making money again.

Martingale bet on knowing the outcome distribution so is not important if you lose. I see martingale as an EA by itself. In order to have a martingale we should have a simple EA with a hi win ratio (no necessary in money) and low consecutive loss.
A simple example would be a system with TP=100 and SL=200 to produce at least 60% win ratio . This doesn't mean it would be profitable, however with a martingale it could work.

[jb68]

Reducing bet-size is a martingale too... Some may call it anti - martingale.

Suppose next distribution: W L L L B L L L W
(W- win, B- Break Even, L - lose)

After each win we can reduce next bet to half. This will reduce DD and increase profits. (if any in first place)