Back-testing trading strategies

 

The mbacktestajority of retail traders do not seem to place high emphasis on back-testing, and I certainly also was like that for a long period of time.  It seems we, the retail traders, are simply too eager to just get on with live trading, and to collect our strategy performance data from our live trading as we go along.  This does not sound smart.

I wanted to write down my thoughts as to why traders, and most likely retail traders in particular, avoid properly testing their strategies, as well as point the big benefits that can be gained from doing it in a professional manner.

What is back-testing?  In very simple terms,  it encompasses testing a defined and documented trading strategy over historical data.  This can be done manually using an application such as ForexTester, in a common application such as Microsoft Excel, or using more advanced programming applications.  Accurate historical data is also a prerequisite.  Search this blog using ‘testing’ category for numerous blog posts relating to testing.

What are some common reasons for not engaging in backtesting?  Well, I can give you some excuses that I have used in the past:

  • The strategy does not lend itself well to testing.  There are too many variables that i consider in trading it live, that i just could not replicate in a testing environment.
  • The strategy is based on sound principles – for example, buying strength and selling weakness, on a pullback – thus the strategy should work.  There’s no real need to test it.
  • Somebody else is already trading the strategy profitably.  Thus, testing is not required, because I will just be doing the same thing that the other trader does.
  • Trading live gives me real trading experience under my belt and helps me to develop battle wits.  Demo trading or back-testing is not going to give me such real battle experiences.
  • I don’t have time to backtest.  I need to make money now.

If I think carefully about each of these rationalizations, then I would find that out each reason highlights at least one big issue, that if unaddressed, is likely to pose significant challenges in the road to profitability and good trading.

There are salient benefits to backtesting such as:

  • Estimation of strategy’s trading frequency
  • Estimation of strategy’s edge (if any), and likely profit & loss performance
  • Estimating typical MAE and MFE information
  • Estimating strategy statistics such as win rate and typical reward to risk ratios
  • Helping to further develop/optimize strategy setup definition and/or trade management rules

If one cannot produce a profit in backtesting, then one surely cannot expect to make a profit in real/live trading.  The converse is probably not true – it’s quite realistic that one makes profits in back-testing that cannot be crystallized in live trading.  Further benefits of back-testing:

  • Back-testing, depending on how it is done, can be a form of practice – in the same way that sports player (whether professional or recreational) practice and train ahead of competitions.  The more practice one performs, the better one should become.  Refer to blog post on deliberate practice.
  • As pointed out by one of my colleagues, this entire process helps to build self-confidence in the trading processes.

There are things to be said about finding a balance between back-testing and live trading, but it seems to be that back-testing is an essential component of a successful trader’s arsenal.

Several traders that I respect and try to follow, place a high emphasis on testing such as Adey (note how much he harps on about testing in his “I wish I had known” [paraphrased] post), or Lance Beggs (see his simple diagram for developing and adjusting a strategy over time).  At the hedge fund that I am working next to, there are several systematic traders – meaning that their entire trading activity is automated – from turning on their strategies in the morning, position sizing, trade entry, managing and exiting.  They have a myriad of trading and programming skills.  They work full-time in the office.  If everything is automated, then what do they spend a big portion of their time on? You guessed it – backtesting!  They test their strategies over as much historical price data that they can get their hands on.

That said, on the whole, there probably is a lack of education/guidance available on this topic – or at least people who are able to take aspiring traders through this process.  I lucked out to a large degree by getting help from someone who contacted after reading my blog – though even then the process is still a difficult one.

That’s probably enough harping on about testing for now.  There will likely be a few more blog posts on the same topic in the coming days and weeks.

This entry was posted in Testing. Bookmark the permalink.

5 Responses to Back-testing trading strategies

  1. For me, the greatest value in back testing is breaking a system- most people try to confirm their ideas and tweak their system to fit the data. The reality, so often, is that many good trades will be missed but 100% of the losers will be taken unless you are 100% mechanical. Nobody has found a system that works even 50%,and it’s hard to deal with the inevitable 10 consecutive losers. Nobody should underestimate how hard directional trading is,and ‘random’ trading may be just as effective. I used to joke that Tuesdays were always down days, and I recall some years ago there were 17 consecutive ‘UP’ Tuesdays in US and UK markets-this probably redressed the balance. It’s too easy however to shy away from real trading in the guise of backtesting,leading to ‘analysis paralysis’

    Liked by 1 person

  2. Pingback: Back-testing and trading confidence | Trick or Trade

  3. Pingback: Mid-May Trading Update | Trick or Trade

  4. Pingback: End-of-May Trading Update | Trick or Trade

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s