SuperStochs EA - Page 4

shahin · 05-06-2008, 02:09 AM

Very good -- I took the settings from kokas (kokas: what broker's data you use? I use Alpari Russia) -- my results are not as good
I like to see less SL and close some more positions that are in profit.

kokas · 05-06-2008, 02:58 PM

Hello shahin,

I'm using Alpari data too. Very strange because if you look at the curves they are equal...

Well, regarding the second question, if you notice your statement and mine, only short positions are closed with loss. I'm really curious about the performance on other pairs... As soon as the optimization ends for this I will make another for other pair, any suggestions??

cfabian · 05-06-2008, 04:23 PM

Hello All
I've been demoing this EA with encountered results. I'm using it on G/Y, E/U and U/C. In 2 days I had 51 trades (all buy), surprisingly all in profit. Using 0.1 lots, they accounted for 20% return

But since yesterday, I have 9 open trades, all with considerable drawdowns (200 pips each). Non has closed, but ALL had the oportunity to be closed in profit. I'll let them run to see if can recover. All profits from the 51 trades are potentially lost.

I wonder what is the criteria to close trades in this EA?
I also wanted to point out that trailing stop is not fuctioning. This would avoid having such drawdown and had all open trades in profit.

I'll try to get some ideas as to how to improve this EA. I can see lots of potential. Many thanks to all your efforts.

Cheers

kokas · 05-06-2008, 05:30 PM

Well, I have no idea for TP, I was thinking on using ATR. For now, the best solution that I have is optimization.

Any sugestions?

Regarding the Trail I think that it is working, at least I see it working on backtest... I think... will check it.

roger

cfabian · 05-07-2008, 04:51 PM

Quote:

Originally Posted by kokas

Well, I have no idea for TP, I was thinking on using ATR. For now, the best solution that I have is optimization.

Any sugestions?

Regarding the Trail I think that it is working, at least I see it working on backtest... I think... will check it.

roger

Hi Kokas,

ATR will be a good option to use, as if we set it as a % value, brokers will not have a straight forward stop value. I'm not really sure about that been an issue, but heard many people talk about it.

As per suggestions, probably a hedge option could be a good idea. Use it in case trend changes and avoid huge drawdowns.

I've continue demoing v3 EA, now I'm sure the trailing stop is not working although you say it works on backtest. When a trade is open, the SL value is placed immediately, but as soon as the trade goes into profit by more than the trail value, it does not move SL figure.

I also agree stevensign's comments that "The version 3 code only compares each of the two lines for a given stochastic, not any crossovers or actual levels. It may be useful for example if all the stochastics were above 50 or 80 before placing a sell order or below 50 or 20 before placing a buy order. Or enter when the 5 mniute Stochastic crosses the 20/80 line or at least when the two lines on the 5 minute chart cross."

The above statement for sure will help us get into better trades.

Thanks for your good work and shares.

Best

henrietta · 05-07-2008, 06:25 PM

Quote:

Originally Posted by cfabian

I also agree stevensign's comments that "The version 3 code only compares each of the two lines for a given stochastic, not any crossovers or actual levels. It may be useful for example if all the stochastics were above 50 or 80 before placing a sell order or below 50 or 20 before placing a buy order. Or enter when the 5 mniute Stochastic crosses the 20/80 line or at least when the two lines on the 5 minute chart cross."

The above statement for sure will help us get into better trades.
Best

The following code will reduce the number of trades, but will only activate when the S1M (the 5-minute stochastic(5,3,3) Main) is less (for longs) or greater than (for shorts) the desired threshold:

Code:

extern int stoch_threshold = 20 ;
...
Signal="NONE";
if( S1M>S1S && S2M>S2S && S3M>S3S && S4M>S4S && Buy && S1M<=stoch_threshold )
{
   Signal="LONG";
}
if( S1M<S1S && S2M<S2S && S3M<S3S && S4M<S4S && Sell && S1M>=(100-stoch_threshold) )
{
   Signal="SHORT";
}

Attached is a gif image of trading the EA on EURUSD with the above modification both long & short since 01 March 2008 using optimized values of stoch_threshold = 29 and TrailStop = 65 in a USD $1000 account.

21.4% profit in 2+ months is acceptable by my standards, especially given the minimal drawdowns involved.

My feelings are that the market is incredibly dynamic, and new optimizations need to be made quite frequently -- monthly if not weekly. For this reason I'll only optimize for the most recent two months. Optimizing over the past couple of years is counter-productive in my humble opinion -- it takes far too much time and doesn't adequately recognize current market conditions.

kokas · 05-08-2008, 02:32 PM

Quote:

Originally Posted by henrietta

My feelings are that the market is incredibly dynamic, and new optimizations need to be made quite frequently -- monthly if not weekly. For this reason I'll only optimize for the most recent two months. Optimizing over the past couple of years is counter-productive in my humble opinion -- it takes far too much time and doesn't adequately recognize current market conditions.

Could not agree more with you henrietta

I have implemented this piece of code on the EA, and I'm running some tests.

I have another idea that I need to code, as soon as it is completed I will post the EA.

Many thanks,
roger

cfabian · 05-08-2008, 06:10 PM

Hello there
I could finally made trailstop work. I was trying to use it with value of 9 pips (as commented on the first post). But changed to 19 and is now working. But still, having the option to use ATR will be great.

Let me tell you I'm really impresed by this EA, I'm starting to love it. But as mentioned before, hedging option could be great idea. So please consider as my example as follows:

I'm attaching a chart, in which you can see many short trades right when the price is moving up. The DD so far is 12% of the equity. If we have that trade in place, and the EA can confirm that M5 and M30 stochs are going UP (going against the trade), and crossed 20, placing hedged long trades will reduce DD and probably profit more. Same thing if long trade in place and M5 and M30 go down and crossed 80.

Just some thought as to how to reduce risk and increase profitability.

Best

kokas · 05-08-2008, 06:36 PM

Thinkiing on the logic of this:

Every time:

1 - Check if a trade as a negative balance
2 - Check stochs to open a possible hedge trade
3 - Open this hedge trade if possible and link it to the trade on point 1
4 - Continue to check the balance of the two until sum reaches 0
5 - Close both trades

I do not see have way to exit this hedge with both positive... I think that this is impossible... at least on the same currency... maybe later we could implement one hedge on another correlated currency. In your case CHFJPY would be the perfect choice.

What do you think?

cfabian · 05-08-2008, 08:17 PM

It sounds like a good hedge option. Probably we can determine to allow a custom number of negative pips (not balance) per pair before placing hedge, as trades always start negative. If that number is reached, start placing cumulative hedge trades and until pip sum reaches 0 or whatever possitive number (to cover trading fees), close just negatives, and continue trading the positive as have potential to keep adding pips.

What I just said isn't it too smoked?