Better System Trader

Better System Trader

Better System Trader is the podcast and blog dedicated to systematic traders, providing practical tips from trading experts around the world.

Blast Buy & Hold with this simple Bollinger Band strategy

In Episode 4 of the Better System Trader podcast, Nick Radge discusses some trading ideas he’s used to create profitable systems. He mentions a Bollinger Band idea which is also published in his book Unholy Grails. Nick says:

the strategy that we did test and showed very promising results was an entry using a Bollinger band and an exit using the opposite Bollinger band, but we use 3 standard deviations for the entry and 1 standard deviation for the exit, just to keep the trailing stop a little bit tighter.”

In Unholy Grails the strategy is used on the Australian stockmarket but in this article we’re going to test it on the Nasdaq 100 instead to determine if the strategy has potential in other markets.

The trading rules

Firstly, here are the test parameters:

  • Period: Daily charts
  • Universe: Nasdaq 100, using historical constituents to eliminate survivorship bias, data from Premium Data
  • Test period: From 1/1/2005 to 1/1/2015. This period was chosen because it has a mix of bull and bear markets, along with high and low volatility
  • Starting equity: $100,000
  • Maximum number of simultaneous trades: 6
  • Position size: Each position will be 1/6th of $100,000
  • Compounding profits: No
  • Commissions: $10 each way
  • Leverage: 0%

Now for the entry and exit rules. In Nicks book, he uses 100 period Bollinger Bands so we’ll do the same. The upper Bollinger Band will be 3 deviations from the central line, the lower Bollinger Band will be 1 deviation below the central line.
Entry: Buy on the Open the day after a stock closes above the top Bollinger Band
Exit: Exit on the Open the day after a stock closes below the lower Bollinger Band

Here is an example of an entry (10/05/2007) and exit for AAPL:

ChartExample-AAPL
Comparing the results of the basic Bollinger Band strategy to Buy & Hold:

Buy & Hold Basic strategy
Annual Return (CAR)% 9.88% 11.83%
Max Drawdown (MDD) -52.81% -21.80%
CAR/MDD 0.19 0.54
Transactions 1 81
Win % n/a 71.60%
Win/Loss Ratio n/a 2.40

The annual return of the basic strategy is almost 20% better than Buy & Hold with less than 1/2 the drawdown. The equity curve of the basic strategy shows a general rise in equity with a few periods of drawdown:

BasicStrategy-EquityCurve

Adding a market filter

A market filter is used to switch a strategy on or off based on broader market conditions. As this is a long only system we probably don’t want to enter trades in a bear market so we’ll only enter trades when the index is rising. With the S&P 500 the most frequently used index by financial professionals, we’re going to use that for the index filter.

In this test a bull market will be defined as the index closing above the 100 day simple moving average; when the index closes below the 100-day moving average it is a bear market and we won’t enter trades until prices closes back above the 100 day moving average. The 100 day moving average was chosen to match the Bollinger Band value, other moving average lengths may work better but will need to be tested. The results:

Buy & Hold Basic strategy Basic strategy with Index filter
Annual Return (CAR)% 9.88% 11.83% 13.01%
Max Drawdown (MDD) -52.81% -21.80% -17.95%
CAR/MDD 0.19 0.54 0.73
Transactions 1 81 73
Win % n/a 71.60% 64.38%
Win/Loss Ratio n/a 2.40 3.66

BasicStrategyWithIndexFilter-EquityCurve

The index filter has improved the quality of the strategy, with a higher return, lower drawdown and higher win/loss ratio with fewer trades.

There are periods during the test where more trade entry signals are presented than we can take using a maximum of 6 positions, so we need to decide which stocks to choose when this happens.

Let’s try a basic ranking strategy in order to systematise the selection process.

Ranking

When a number of stock entries occur on the same day we need to make a decision on which ones to take. We could choose them randomly but we would need to run monte carlo simulations to get a better indication of the possible variations using this method. I prefer to add a simple ranking system to the strategy so stock selection is completely systematic.

The ranking strategy I’m going to use here is based on what I think the strategies strength is. I expect the strategy will perform best either just after a bear market or a period of consolidation, entering at the start of a new bull market or breaking out of consolidation and riding it higher. In this case, we’re going to try ranking by Rate of Change over the last 90 days, so the stocks with the smallest Rate of Change will have a higher priority than those with a large Rate of Change. Logically it makes sense but what do the results tell us?

Buy & Hold Basic strategy Basic strategy with Index filter Basic strategy with Index filter and smallest ROC ranking
Annual Return (CAR)% 9.88% 11.83% 13.01% 13.82%
Max Drawdown (MDD) -52.81% -21.80% -17.95% -17.54%
CAR/MDD 0.19 0.54 0.73 0.79
Transactions 1 81 73 68
Win % n/a 71.60% 64.38% 72.06%
Win/Loss Ratio n/a 2.40 3.66 3.31

BasicStrategyWithIndexFilterRanking-EquityCurve

The ranking strategy produced a higher annual return, with lower drawdown, a lower number of trades and a higher win %. It may have not impacted too many trades so the addition of the ranking may not be statistically significant but it does provide a systematic method to choosing stocks when multiple opportunities present themselves.

The power of Compounding

So far we’ve seen the basic strategy slightly outperforms Buy & Hold but with considerably lower drawdowns. The inclusion of an Index Filter and ranking by smallest ROC has improved the strategy although the results aren’t outstanding.

Let’s see how compounding profits impacts strategy results:

Buy & Hold Basic strategy Basic strategy with Index filter Basic strategy with Index filter and smallest ROC ranking Basic strategy with Index filter and ROC ranking & profits compounded
Annual Return (CAR)% 9.88% 11.83% 13.01% 13.82% 20.20%
Max Drawdown (MDD) -52.81% -21.80% -17.95% -17.54% -21.93%
CAR/MDD 0.19 0.54 0.73 0.79 0.92
Transactions 1 81 73 68 60
Win % n/a 71.60% 64.38% 72.06% 73.33%
Win/Loss Ratio n/a 2.40 3.66 3.31 3.33

BasicStrategyWithIndexFilterRankingProfitsCompounded-EquityCurve

BasicStrategyWithIndexFilterRankingProfitsCompounded-Drawdown

BasicStrategyWithIndexFilterRankingProfitsCompounded-MonthlyReturnsTable

Update 27/4 – As requested by Rick, here is a histogram of distributions, with the majority of trades in the -25 to +70% range and a few trades with 100% and higher:

BasicStrategyWithIndexFilterRankingProfitsCompounded-Distribution

With compounded profits we now have a strategy that produces more than double the returns of Buy & Hold with only half the drawdown. The win rate of 73.33% and win/loss ratio of 3.33 are also good for a trend following system.

It appears the strategy has some potential and warrants further investigation. Some areas of consideration could be:

  • The length of the Bollinger Bands,
  • Different market filters,
  • More adaptive trailing stops,
  • Ranking based on other metrics,
  • Suitability to other markets.

Like a copy of the AmiBroker code?

Want to get the latest updates automatically?

The best way to get notified when new stuff is released is to sign up to the email list below and we’ll be sure to let you know:

[mc4wp_form]

Related Posts

Comments

  1. Dear,

    Thanks for the very interesting article “Blast Buy & Hold with this simple Bollinger Band strategy”. I’m using Amibroker as well. Is it possible to post (or send me) the code of this system?

    Thanks in advance.

    Kind regards,
    Hans van der Helm

  2. Andrew- Thanks for the write up. I am interested in the afl. Appreciate your work on this.

    • Hi Casey, I’ve only tested it on stocks however it may work on futures/forex etc.

      I can provide the AmiBroker code if you want to test it for yourself?

      Regards,
      Andrew.

  6. Hi Andrew

    Thanks for that interview with Nick and the analysis of his Bollinger Band system.
    Looking forward to the AFL code.
    Very interesting.

    Cheers
    John

  7. Hi Andrew,

    Really enjoying the podcasts and the great information you provide. Could you please send through the AFL code.

    Thanks

    • Thanks Paul, glad you’re enjoying the podcast.

      I’ve just emailed you the AFL.

      Cheers,
      Andrew.

      • Thanks a lot, got it!

        Two things: (a) Could you post results from index start? Although there is a downtrend, the chose period has two uptrends.

        (b) What was the contribution of AAPL and GOOG in the results? If you were to remove those companies from the index, what would be the result? I’m saying this because it is unlike;y to have similar companies in the near future. To what degree are your results influenced by a few outliers?

        Best.

        • Hey Rick,

          Great point about considering the outliers.

          I checked the trade results and the trades with the highest returns weren’t actually AAPL or GOOG. In fact, the strategy didn’t take a trade in GOOG at all and AAPL was only the 3rd highest, here are the top 5:

          GILD: +213.98%
          BIDU: +137.33%
          AAPL: +107.10%
          EXPE: +103.48%
          QVCA: +98.10%

          If I remove all the AAPL trades, the Annual Return is 18.43% and DD is -23.60% so the returns are slightly lower but who’s to know what will happen in the future – AAPL may continue higher, another stock could take over, this strategy may fail miserably tomorrow, we just never know.

          Regards,
          Andrew.

          • Thanks! Still I am concerned about outliers. I would be nice if you could add to the blog a histogram of returns per stock traded, maybe at least the top 30 ones. Then it will be clear if the performance was due to a few random outliers or due to the method. Apologies for the request but I do not have the data to do it, otherwise I would.

          • Hi Rick, I’ve added a chart showing the returns. The bulk of the trades are in the range of -25 to +70%, with a few +100% or higher. I hope that answers your questions.

            Please keep in mind that this research is not a complete trading system, it is just a starting point. The purpose of the research was to determine if the strategy that Nick mentions in the podcast has potential in other markets. It appears it may but further investigation obviously has to be completed before taking it further.

            If you can I highly recommend getting some data and running some of these tests yourself, I’m sure the strategy could be improved so I’d be interested to hear your results.

            Regards,
            Andrew.

  9. Great write up. It’s amazing what a a simple system with just a few tweaks can do. I’d appreciate a copy of the AFL code so I can see if I can make a few other tweaks that might help. Thanks.

    • Hey Gav, glad you enjoyed it.

      Yes, I’ve found the simple systems are often the best, I look forward to hearing what you uncover in your testing.

      The AFL is on its way.

      Regards,
      Andrew.

  11. Nice article, can u please mail me the code so i can test it on instruments i trade.
    thank you
    mandeep

  12. Hello Andrew, Thank you for the great Information and Podcasts. I also enjoyed the interview with Cesar Alvarez. I´m looking forward for more Podcasts with real Traders. Could you please send through the AFL code.

    • Thanks Juergen, I’ve just sent you the AFL code.

      Keep an eye out for the next few episodes, we have some awesome traders coming – Kevin Davey, Gary Stone, Rob Hanna, Howard Bandy and more.

      Cheers,
      Andrew.

      • Hi Andrew,
        Simple works best and that’s what this system is. I like many of the concepts used in this system.

        Some of my concerns or ideas for improvement are:
        1) The high winning percentage. I would expect this to come down over time and be closer to 50%.
        2) A sample size of 60 isn’t really enough to draw too many wide ranging conclusions.
        3) Also having a max of 6 open trades is a bit on the risky side for me. It really doesn’t provide enough protection against specific risk.
        4) Dividing capital equally among the 6 open positions could be improved to something along the lines of fixed fractional or volatility adjusted position sizing to help control risk better.

        Love the website and podcasts. Keep up the great work!

        Could you send me the AFL code?

        Thanks,
        Gregg

        • Hi Gregg,

          All valid concerns which need to be investigated further, I’ve sent you the code.

          Keep in mind that Nick Radge uses this strategy on the ASX and it produces better results than the results in this test, which tells me the strategy is robust. If you want more details I highly recommend you check out his book Unholy Grails.

          Regards,
          Andrew.

  13. Hi Andrew. Loving the podcasts! I’d like to hear some more about your own background and trading style in a future podcast.
    Can you please send me the AFL. I’m a huge fan of Nick’s work
    Thanks,
    Ashley

    • Hey Ashley, glad you’re enjoying the podcast.

      You can read a little about my background and trading in the About page on the website or in podcast episode 000 available in iTunes but it’s a good idea to include it in a future podcast episode so I’ll do that too.

      I’ve just emailed you a copy of the AFL code.

      Regards,
      Andrew.

  15. Hi Andrew, your doing an excellent job with the podcasts and I’m really enjoying them, keep up the great work and when you have a chance could you please send me the AFL code.

    • Thanks Glenn, glad you’re enjoying the podcasts, I just emailed you the AFL code.

      Regards,
      Andrew.

  16. I’d like the AFL code too please. Just getting started with AFL and need all the good sample code I can read.

    I’m especially interested in your thoughts on the 100 day moving average as a long-term bull-bear indicator… do you have any other long-term bull-bear-sideways indicators that you use, or have references to others using?

    • Hi Paul, I’ve just sent you a copy of the AFL code.

      If you’re just getting started in AmiBroker, I recommend you take a look at the AmiBroker book by Howard Bandy here, it’s a free download for personal use.

      If you’re more interested in a course, check out the AmiBroker course by Cesar Alvarez here. Cesar has offered to open his AmiBroker Backtesting 101 course to Better System Trader podcast listeners. The course will teach you how to program strategies from beginning to end and even provides you with a strategy that averages 20%+ returns per year. The offer is only valid until May 20th and numbers are limited so get in quick.

      As for ways to determine bullish/bearish conditions, I’ve found simple methods are the most effective, like moving averages, volatility indicators etc. Keep an eye out for the podcast interviews we release on May 18th and May 25th where we discuss more about this.

      Regards,
      Andrew.

  18. Very impressive results for such a basic system. Certainly not over-optimized.
    I’m interested. Could you please send me the AFL code.

    Cheers,
    Peter

  19. I know I’m a pain but I insist that you remove from the list the top 4-5 stocks

    GILD: +213.98%
    BIDU: +137.33%
    AAPL: +107.10%
    EXPE: +103.48%

    How does CAR change?

    • Hi Rick, if I remove all GILD, BIDU, AAPL and EXPE trades the CAR has reduced as expected, down to +15.57% with -23.60% drawdown. It still beats Buy & Hold of +9.88% CAR with -52.81% drawdown but there is obviously more scope for improvement.

      I encourage you to use this strategy as a starting point only, it is by no means a complete strategy and I’m sure there are many ways it can be refined.

      Regards,
      Andrew.

  21. Hi Andrew, thanks for the podcast. Really appreciate your effort. Could you help with the code. Also wondering have you tried a long/short variation.

    rgds, Vivek

    • Hey Vivek, glad you like the podcast.

      I haven’t tried a long/short variation, please let us know how you go if you test it yourself.

      Have just sent you the AFL.

      Cheers,
      Andrew.

  22. Hi Andrew,
    I’m learning Amibroker and find it helpful working through other peoples code and understanding how the logic has been coded. Would love a copy of your AFL code for this system.
    Thanks

  24. Many thanks for this interesting article. Would it be possible to have a copy of the Amibroker code?
    Have a good one!
    Stephane

    • Hi Stephane, glad you liked it, have just sent you the AFL code.

      Regards,
      Andrew

      • Could you please send me the afl code. I am interested in further studies.

        Best regards

        Stefan

    • Hi Nir, click the blue ‘Download the FREE Bollinger Band code for AmiBroker’ button at the bottom of the article to download the code.

      Regards,
      Andrew.

  26. Thanks for the strategy. I have seen so many promising strategies that trade only LONG. I am yet to see a strategy trading profitable both LONG and SHORT. This would be a one up for your site if you could get an expert to write up an article on such a strategy. Regards.

    • Hi Perry, thanks for the comment.

      Check the blog next week, Rob Hanna provides a simple trading idea that does exactly that!

      Regards,
      Andrew.

  27. Hi Andrew,
    I’m really enjoying your podcasts and blog. May I please have a copy of,the AFL for this post.

    Best regards

    • Hey Growbuck, click the Blue download button just above the comments section to download the AFL code.

      Cheers,
      Andrew.

  28. Great post, i’m learning the amibroker to develop the trading strategy. Lucky to see your website. May i get the afl code for deep studying? Thanks.

    • Hey Phil, click the blue Download button just above the comments to get a copy of the code.

      Cheers,
      Andrew.

  30. Andrew,

    Just found your podcast and this was the first one I listened to – GREAT stuff. Thank you!

    Simon

  31. Hi Andrew,

    thanks for this great post. I would like to use the code to learn how to program an efficient strategy. Please send me the AFL code.

    Thanks:
    Philipp

  32. Hi Philipp, click the blue Download button just above the comments to get a copy of the code.

    Regards,
    Andrew.

  33. Dear Mr Andrew Swanscott

    Can you please send me the AFL code for this as I am using Amibroker and would like to back test the same

    Regards

    Ashok

Trackbacks

  1. […] Blast Buy & Hold with this simple Bollinger Band strategy [Better System Trader] In Episode 4 of the Better System Trader podcast, Nick Radge discusses some trading ideas hes used to create profitable systems. He mentions a Bollinger Band idea which is also published in his book Unholy Grails. Nick says: the strategy that we did test and showed very promising results was an entry using a Bollinger band and an exit using the opposite Bollinger band, but we use 3 standard […]