Mining systems with a Take Profit: A source of additional diversification

Up until recently I sincerely hadn’t been a fan of using take profits when trading (you can read more about that here). In essence I considered that take profits limit the upside potential of trading systems and provide an additional degree of curve-fitting that is hardly justified for most strategies. However after generating strategies using automated low data-mining bias processes it became clear that these strategies offer a degree of diversification that simply isn’t possible when you use strategies that avoid this type of exits. This is most likely because take profits succeed under conditions where gain potential is limited but smaller movements can be predicted with more certainty — therefore TP containing strategies favor ranging markets, especially if reward to risk ratio values tend towards a value of one.


After we started mining for TP containing systems at Asirikuy it became evident that these systems were very different from systems available within our trading repository. It took us a coupled of months to gather more than a few hundred strategies and we now have enough to say that the TP containing strategies within our repository have an overall lower correlation with the rest of our trading strategies. Out of sample trading results have also confirmed this with the results of the TP and non-TP containing strategies rarely aligning through any week. It is expected that under some sets of market conditions both system types would be profitable – otherwise it wouldn’t be possible for both types to achieve R² values above 0.95 in 30 year tests – but when market conditions are unfavorable for one type or the other they seem to do a good job hedging each other’s bets.

It is also interesting to note that their distribution is very different when it comes to the symbols where they can be found. While strategies that lack take profits – which generally tend towards being trend following – appear most commonly in the EURUSD it was evident after a few weeks of mining that for take profit containing strategies it was very difficult to add strategies for this symbol. It is also worth noting that the problem is not that the strategies weren’t hard to find in the EURUSD at all but our data mining bias assessment process reveals that the probability that such systems come from random chance is very high. It seems to be very easy to get profitable results just out of chance when using a TP in the EURUSD in the same manner as obtaining non-TP containing strategies out of chance on the GBPUSD is very high. If inefficiencies are spurious then the chance that they alpha-decay extremely rapidly is much higher, so when using a TP it is especially important to be careful with DMB analysis, even more if the TP is lower than the SL — and therefore naturally favored by chance.


When it comes to TP containing strategies most of the systems we have found belong to GBP containing symbols, most commonly the GBPJPY. It has become clear that in this symbol these strategies are very prevalent and it is relative easy to find low DMB processes that can generate these systems. The first image above represents the overall total composition of our repository two weeks ago while the second image above represents the composition of TP-only systems within it. At this point 77% of our TP containing strategies belonged to the GBPJPY while the EURJPY and GBPUSD contributed in second and third place to this section of our trading repository. The contributions from these strategies have helped reduce the overall reliance of our repository in EURUSD containing strategies and is also helping our hedging since TP-containing strategies – as I mentioned above – behave very different in terms of correlation with the rest of the repository.

The simple TP is only the beginning. Right now I have finished the implementation of functionally adjusted TP mechanisms which will expand the array of potential exits that we can achieve within our mining using a TP. Strategies that use functions to adjust TP values as a function of time give us the possibility to explore strategies that expect different price action evolution and that may close upon gains much faster than traditional non-TP containing strategies that do so by means of functionally adjusted trailing stops. With the concept of the “trailing profit” we seek to find arrays of strategies that will be intermediate between strategies that use no-TP and a fixed TP that isn’t moved through the strategy’s life, potentially increasing the symbols where we can find strategies and the quality of strategies we’re able to find.


Even though I was fairly skeptical of implementing TP mechanisms – due to the reasons I previously mentioned – out-of-sample trading and mining are showing that the introduction of a TP does a lot to diversify our trading system repository. Diversifying exit mechanisms seems to do a lot more for strategy diversification than attempting to diversify trade entries by increasing/decreasing complexity or performing similar modifications. The implementation of these changes plus some other changes – that I’m leaving for a future post – will do a lot to increase the level of diversification within our trading. If you would like to learn more about our trading repository and how you too can trade using thousands of trading strategies coming from our GPU based mining process please consider joining, a website filled with educational videos, trading systems, development and a sound, honest and transparent approach towards automated trading.strategies.

Print Friendly, PDF & Email
You can leave a response, or trackback from your own site.

One Response to “Mining systems with a Take Profit: A source of additional diversification”

  1. josh Osborne says:

    This is awesome. Through the years I have been so interested in what has been learned through your thorough experimentation, and very hard work.
    I have still never seen any reliable science backed research in this area other than what Asirikuy is doing, even up to today, there is simply nothing worthwhile.
    Thank you Daniel for your hard work and seemingly undying dedication to what seemed at many times like the impossible task.

    Josh Osborne

Leave a Reply

WordPress › Error

The site is experiencing technical difficulties.