Both stops and take profits should represent any idea. In other words, you should understand why you need to get out of your trade. And it should be not because you just want to, or because you read somewhere that stops should be tight, or because you project the price target here or there. Normally market doesn't care about your projections, therefore you should have an exit strategy — about which you can at any moment say why you exited.

No wonder that trailing stops are better for what you're trying to develop. Have you tried to create a strategy without stops at all? I'm sure it will be very edifying.

Are the 3 strategies same base strategy using a different profit taking ? Or different strategies using different profit taking ?

Same system with same stop loss

Success belong to those who make their research to improve.

dear Steve, it looks like you think that 800k of datalines is a lot. I have no clue (as you do not disclose it) what these represent, but to give you an idea : we run models on Fx currency pairs in a 5 minute timeframe and with a database going back to late 2004. Well, on one curpair, this represents at present 865k datalines. Now, would it be statistically correct to base an overall statement on the analysis of just 1 currency pair (while we track a dozen others) ? I don't think so. So, please elaborate on your work providing us some more "beef" to chew on

Something like momentum indicator could work.

RSI divergence for example.

Here's an idea, take t(0)=time now

t(-10)=10 bars back

Take DeltaRSI(RSI(0) -RSI(-10))

Take deltaPrice(close(0) -close(-10))

if(DeltaRSI*DeltaPrice>0, stay in trade, else, exit)

Test it, might work...

Also, You probably already did, but optimizing stops are good, I use both a trailing stop + a fixed profit target

Steve, you will find that in trend following no pain=no gain, less pain=less gain, more pain=more gain. So, having a target means a higher profit ratio and less psychological pain but less gain. Having trailing stops means you'll see lots of profits vanish which is painful but more gain in the long-term. More loose (tight) stops means more (less) pain, but also more (less) gain long-term.

With good research, this equilibrium can be improved slightly but don't expect a massive difference. Tens of thousands of systematic researchers are doing just that every day and they have yet to disturb this equilibrium in a significant way.

Hi Steve if you have some time search Charles LeBeau exit strategies for stocks and futures. The author has given really good thought on how to keep trailing and stop loss. I have been using his techniques and I got good result after using it.

Hi Steve , since u hv tested yr strategy for entire data for several parameter and u mte hve taken the best with good graph and better Sharpe ratio ie. minimising yr drawdawn and maximising yr return....but do u think Is this the right method to do or will the back test data hold ....??

I guess u should divide yr entire data set in some parts and ex : if u taking 5 year data , divide the same in 5 parts say 1 yr

...so for 1st year test data n optimise yr parameter ..and run the same parameter for nxt year so u ll get some desired result and simultaneously test data n optimise parameter for 2nd year and run d same fr next year ...keep doing ti u get result fr 5th year ...and now chk the cumulative run time for all the for years ...if u ll able to get good graph and good Sharpe ratio den it is assumed yr strategy is consistent ....

Hi All, thanks for the feedback. Very helpful. Wish you all a great year and many to follow.

Hi there, On what basis did you set targets? Cheers pete

I spent so much time trying to find the best way to do it and ended up with a simular result, and still use every day. Good job man!

Backtested trading o

Backtested trading on interest rate futures 10 years - cross market - using conintegration tests-on the assumptions that spreads between two markets mean revert, used Zscore indicators for signal.

There is an oxford MSc dissertation paper that goes into more technical details. I used Matlab and SAS to optimism entry and exit triggers. I avoided Newton R. optimisation because of possible saddle points.

Nice idea Steve. I'd like to share my trading strategy. I also use trend following style. I define target profit and stop loss based on trend line and/or SR (support and resistance). To secure the profit, I use 2 kind of trailing at the same time : trailing profit in money and trailing by points/pips. Instead of using tight stop loss, I use averaging system. I'm very strict in defining my entry point and I use follow the trend, so I'm 100% believe I'm on the right track. If I'm wrong, the stop loss is there to secure me.

Good stuff. I am working on a time based stop as well now. idea is that the trail should different as the day progresses to protect profits as we approach the close. Not much difference on the gains but a small improvement on the losses especially when the mkt turns towards the close.

If the trailing stops are working for you, take a look a chandelier stops (a mod that may even work better). Beware overfitting as always, but volatility adjusted stops are the best in my view. They tighten up when when you need them too, and implicitly identify s/r levels in real time.

Thanks Drew. Been a long time since I've looked at chandelier stops. Will definitely revisit.