How to turn a losing strategy into a winning

New comment
 

Hi all,

While working on an EA which was losing horribly, I thought well if it is losing big, why not turn it around and make it winning big!

I changed every buy condition into a sell condition and visa versa. So all my buy and sell orders should have been reversed and now every loss should be a win right?

Wrong! But why?

I expected the exact same trades, but instead of for instance for a buy with a loss I would instead sell with a profit because in the strategy tester the price during that period went down instead of up. But it seems to be making different traders all together ending up in an even bigger loss!

But why? Please help me figure this out...

Thanks for any thoughts on this.

 

> now every loss should be a win right? Wrong! But why?

Every system I've reversed has done the same, some worse than others -i'd just move on!

Start with a simple and sound system that you could trade manually, then automate it

-BB-

 
redindian:
I changed every buy condition into a sell condition and visa versa. So all my buy and sell orders should have been reversed and now every loss should be a win right? Wrong! But why?
You pay the spread no matter which direction the trades are. So every loss becomes less of a win and every win becomes more of a loss.
 
redindian:

But it seems to be making different traders all together ending up in an even bigger loss!

But why? Please help me figure this out...

Thanks for any thoughts on this.


It will be making different trades altogether if you are using the Bid and/or Ask as part of your trade criteria, ie if(Bid>MA)... or if (Ask<MA) ..something of that nature, because when in your original EA the Bid price was at the trigger point, when you switch the criteria, the same trade now reversed requires the Ask price to be there instead, and as it probably isnt, that trade signal will not be true anymore and a different set of circumstances will trigger the trade.

Overall what you are experiencing could probably be considered something similar to Newtons law, every action has an equal and opposite reaction, so every set of circumstances that triggers a profitable trade will at another time trigger a loss trade, over a long time period several thousand trades later this averages out to somewhere close to zero, minus the spread.

 
If your original system was losing more than the spread plus average slippage over a test period, the correct reversed system should make a profit. But you have to be careful. Entries and exits have to be shifted slightly to be on the other side of the spread. So if the original system would have bought at 1.0000 at a certain time with a presumed spread of 2, your reversed system needs to sell at 0.9998 at the same time.
 

Thanks for all the reactions. I actually left everything the same and just reversed the names of the buy/sell functions so it should really reverse everything. You still loose the spread of course. Will look at it again and see if I can find anything.

Still I'd never use it, but I was just wondering. :-)

New comment