how does hard stops reduce reward / risk significantly? id imagine adding a trailing stop would reduce risk significantly and reduce reward partially (20$ in your case). the end result increasing the reward / risk ratio? are you sure how you calculate risk is accurate?
frostengine, Great posts... Could you answer this: When your bot finds a signal, do you enter at market or via a limit order? I know from manual trading, the ER2 has some fill issues. You previously said you are at 1 contract, so I guess this isn't a problem.
walter, The way I am looking at my reward/risk is the profit factor. For instance I am dividing the average win divided by the average loss..... That number currently sits at around 1.98... so close to a 2 to 1 reward/risk ratio... When I add in hard stops, depending on how big of a hard stop I put in the ratio drops down to as low as 1.1.. which doesn't have to be a BAD thing if my win% compensates.. in this case the increase in win% isn't enough to make it appealing to me. Another thing to consider, is as it stands the average loss is in the neighborhood of 1.3 PTS without a hard stop. The trailing stop on the other hand.. not sure why it performs SOO badly.... the trailing stop I actually used in that test was using the parabolic sar.. I may have the speed on it too fast... Will have to try some other alternatives for trailing stops to get a better picture.
whackamole, I am using market orders. On 3 different "tests" I guess you can say.. #1 being the 2 days of live account orders, #2 being 4 months of running on IB's demo account which actually I believe handles estimating slipage VERY well and is probably on the side of OVERESTIMATING slipage, and #3 being an order matching test mode i built into my trading platform for slipage estimation. In all 3 of these environments the slipage incurred did not affect the system much at all. After reasonable slipage is taken into account the system still has room for about 2 and 1/2 ticks of slipage per direction before the profitabitity of the system will be in jeapordy. Using Limit orders may increase the profitability somewhat, but it runs a high risk of not getting filled on some of its best entries... This may not be a problem for some systems but in mine its a big problem.
<i>"However, every trailing stop, profit target, and hard stop I have ever tested with it has shown a GREAT decrease in the expectancy and profitabiltiy of the system."</i> I've run literally 1,000s of system tests and realized the same facts myself. That said, do you find any hard stop that keeps average $$ profit per trade similar, yet greatly reduces max drawdown? You might find that hard stops shave a tad of overall profit BUT likewise slash your max historical drawdown OR the real one somewhere in your future. * Stick with market orders in the ER. Partial fills on orders above 5-lot are the norm, not exception. Partial fills will kill you much faster than average slippage will. Best Wishes
Austin, Glad to know that its not just my strategy that does not do well with stops. I would like to thank everyone for their input thus far and I hope this thread continues to be constructive.
Have you guys found that not having stops etc in your systems requires a lot more capital to handle the swings? do your systems support / dont support large amounts of leveraging? 200x leverage can make a huge difference for a system that has a good win / loss ratio.
Walter, Well my strategy is trading ER2 contracts so inherently it is utilizing TREMENDOUS leverage. I force my system to have at least $7,000 in capital per contract which is about 4k over required margin here at IB. Even if I used stops and my system had lower historical draw downs I think I would still want at least 4k of wiggle room simply because you never know what may happen. My largest draw down is probably going to be at some point in the future. So I do not believe that not using stops forces my system to need more cash or to use less leverage. One thing I really should consider is at least a 5 pt stop. This stop would hardly ever get hit, but would possibly serve as a type of protection against a huge market shaking type of event. However I would imagine such an event it wont matter if you have stops or not as the market would fall through them. probably actually better NOT to have a stop in for such an event as you will more than likely get filled at more extreme prices. However I have no data to back up such a claim. Its only what I would assume would happen.
The profit factor is actually the sum profits of all the profitable trades divided by the sum losses of all losing trades. Which is much different from what you're using. Personally, I look at the reward/risk from the point of view of net profit divided by the largest drawdown. For me this gives me a more immediate sense of how much drawdown (i.e. pain) I need to suffer for a certain profit (i.e. gain) in the worst case. By the way, I'm also auto-trading the ER2. I try not to sweat the day to day variations which might tempt me to overide the system. Good luck with your auto-trading. Richard
Great thread Frosty... I'm about 2 months behind you as I prepare to trade an automated currency algo Good Luck!!!