How to Improve Your Trading Results Through Testing

As most seasoned traders know, even with simple strategies, there are thousands of variables that can affect the profitability of their trading. And since no real world strategy is going to yield 100% positive results, no single trade can demonstrate the effectiveness of a strategy, therefore the only real solution is to run tests — many, many tests.

Improving your forex trading results using testing

For those of you who don’t know me, my name is Tim and I’m the guy who gets to code and test Casey and Nathan’s (and sometimes my) trading strategies here at Winner’s Edge Trading. So I can tell you in all too much painful detail how much testing the Winner’s Edge Trading strategies go through.

 

What are the Variables?

blog2

As an example of the many variables, consider a simple strategy like using a moving average as a trigger. Let’s just say that we will buy as the price moves above the given moving average and sell as price moves below the same moving average. These are some of the many considerations of this strategy:

 

  • Do we wait for a close before a buy or sell? (yes or no)
  • Do we use a Simple Moving Average (SMA) or an Exponential Moving Average (EMA) or an even more complex MA? (SMA, EMA or other)
  • How many periods should the moving average be? (7, 10, 20, 21, 30, 34, 40 or more)
  • What time frame should we be trading on? (1m, 5m, 15m, 60m, 4h, 12h, Daily, etc.)
  • How should we determine our trade size? (Fixed lot size, risk a fixed percentage of equity, some sliding value based on prior profitable trades or something else)
  • How should we determine our target? There are a host of considerations here:
    • Simple pip value (5 pips, 10 pips, 25 pips?)
    • Some dollar profit (that, of course, is just an extension of using a pip target – unless you are varying your lot size for risk purposes)
    • Some indicator (say ATR – that introduces its own group of parameters)
    • Some Support or Resistance level (whew – figuring out what is support and resistance)
  • How should we determine our stop loss? Again – a host of considerations – similar to the target determination:
    • Simple pip value
    • Dollar loss
    • Indicator
    • Support/Resistance
  • Should we only trade our strategy at certain times of the day (Say, only the first 2 hours of the New York trading session, or all of the London session, or 24 hours/day – just to name a few)
  • Should we only trade our strategy at certain times of the week (Say, never on Monday or Friday for example)
  • Should we take our trade off during big news events – or just move our stops to protect the position?
  • What currency pairs should we trade this on?
  • How many simultaneous currency pairs should we trade this strategy on?

 

Every single one of the items mentioned can cause your trading strategy to be very profitable or a total bomb. There’s no way to know just by looking at them. For example, you might choose a Take Profit/Stop Loss (Risk:Reward or R:R) ratio of 4:1. That way you can lose four trades for each win and still break even. But sometimes, the opposite is indicated; use a 1:4 ratio and win five trades for each loss (or 9 trades for each 2 losses) to be profitable. These things may be obvious, but a minute change in the market can overturn your successful trade in a heartbeat.

 

Trading is a statistical endeavor. In other words, you’re taking a strategy that seems like it will be profitable and testing it on actual market data to determine if, over the long haul, the strategy will prove to be profitable. One loss is meaningless. One win is equally as meaningless. The only solution is testing, testing and testing. Did I mention you need to test your strategy?

 

How Does Winner’s Edge Trading Test Strategies?

When we at Winner’s Edge Trading want to prove out a new strategy, we extensively backtest our strategy. Backtesting is testing on historical data. Once we nail down the broad rules of our strategy (what indications will be used to trigger our entry, what indications will be used to confirm our trigger, what conditions will we take trades under and so on), we code the strategy for Tradestation trading software. Why Tradestation you might ask? Because they have the most reliable historical currency pricing data that we could find. There are probably others, but we found and like Tradestation and that’s what we use. As you know, our indicators and Expert Advisors are typically only available for MT4. We would love to be able to use MT4’s Strategy Tester for all of our testing. It would save us from having to start with TradeStation code and rewrite the programs for MT4 (which is no small task.) While there are some solutions for reliable data in MT4, we just haven’t found a great combination of tools that does the job for us.

 

The Testing Process

blog3

As you know, a simple indicator won’t tell you much during testing unless you’re watching it, so at this point, the indicators for our systems are combined into a “robot” that actually triggers trades based on those indicators’ signals. Once we have a working robot in Tradestation, we set out to run our tests. This is a very meticulous and tedious process. We always keep copious detailed notes as to what parameters we are using, what time periods in which we are testing (such as from/to dates and times, etc) and so on. We vary everything within the parameters of the strategy’s rules, including varying the pairs and modifying the spreads to make the tests more “real world”. We even try some “out of the box” crazy numbers just to prove that they are truly crazy (sometimes they turn out to be valid tests.) I won’t go into great detail about Tradestation right now, since it’s beyond the scope of this article. Perhaps in the future, I can get into using Tradestation for backtesting and also, help you avoid some of the pitfalls of using MT4 for your backtesting.

 

After we’ve narrowed the field down to a few (half dozen or so) very profitable parameters and on which pairs they show the most promise, we recode the programs for MT4 and confirm that they work identically to the Tradestation version of the software. Even though that only took a single sentence to describe, the actual process is probably the most extensive of the whole project. Now the REALLY boring stuff happens. We have several VPSs (Virtual Private Servers) that we use for forward testing our strategies on live data. Forward testing is exactly that; we put the programs on live demo platforms and let them run for months. Typically, we are running six to eight tests on each group of currency pairs for each strategy and reporting the data to MyFXBook for tracking purposes.

 

To forward test our strategies, we set up multiple MT4 platforms from the same broker on the same VPS (computer). We use the same VPS so all of the tests are equally enhanced or hampered by the computer’s performance and the network’s performance. We use the same broker for our testing series so the price feeds and spreads are the same for all tests. As you probably already know, every Forex broker has a different “opinion” of the current price of an instrument based upon the trading that takes place within their brokerage. Additionally, every broker sets the bid/ask spread on each instrument differently (this is one way the brokerage makes its money.) Due to these factors, we have to be sure that parallel testing takes place on the same broker so we are comparing “apples to apples” as the saying goes. As another consideration, we need to choose brokers that have unlimited demo accounts. Many brokers limit their demo accounts to 30 or 90 days or some other value. Since our forward testing is ongoing, we need to have accounts that will continue indefinitely as if they were live, real money accounts.

 

We will then set up each set of parameters in each test according to the results of the backtesting and run them on as many as 30 pairs on each platform. As mentioned, we attach each platform to our MyFXBook account which has the capability of analyzing the results in many different ways. We are able to view the results as a collection (to see the overall results of a certain group of parameters and compare it to the other parameter groups) or by individual pair (to give us an idea as to which strategies are best suited to which pairs.) We can also determine what time frames are best for each pair and what times of the day/week/year are ideal for the strategy. Many times, one of the tweaks we are testing will show tremendous potential. For example, one of the forward tests we implemented (based on our DX9 Strategy) had +10.9% difference in net gain compared to our baseline strategy from Oct 1 – Oct 22. That massive deviation was the result of one minor entry tweak to the base strategy.

blog4

One thing that Nathan stresses is not “curve-fitting” our strategies. Curve-fitting is the act of “dialing” a strategy in with tight, specific parameters so a certain time period on back-testing or forward-testing shows the highest profit. The market changes daily (even hourly) and will quickly kill a strategy that’s based upon curve-fitting. That’s why some people sell the greatest strategy ever that shows 10% daily profit for the last five years. It’s easy to dial the parameters of a strategy to show those results in a back-test. Don’t be fooled by that sales hype. Our best resulting strategies are the ones that have the widest “market-appeal”. In other words, the one that seems to work on more pairs and more different time periods and time frames.

 

Testing your strategy is an important way to improve your trading results, even if you use a strategy that was developed and proven by someone you trust. Testing will allow you to find out if the strategy is right for you and help you find the pitfalls between your new strategy and your personal trading style.

 

Want to See What We’ve Been Working On?

 

If you would like to take advantage of some of the testing we have been doing, sign up for our DX9 Signals Trial (it’s only one dollar) and test out the strategy for yourself. Year-to-date, we have won over 70% of our entries. Click Here to Sign up for the Trial.

blog5

**Note: We are also having a special training webinar for signal members that discuss some of the changes we are implementing to the strategy based on the testing to which I referred in this article–the webinar will teach you the money management strategy that is doing extraordinarily well for us right now.

The following two tabs change content below.
Winners Edge Trading was founded in 2009 and is working to create the most current and useful Forex information and training available on the internet.

Winner’s Edge Trading, as seen on:

Winner's Edge Trading in the news

  • raj I would like to do that it is a great question.

  • malenchelon

    There was a website i’m trading with. “Totradefx” best scalper and good leverage

  • Raj

    Hi Casey,
    Would you be kind enough to write an article about how we should choose the fx pairs that we would like to trade. I’m sorry to waste your time if this is a silly question.

  • Your welcome I am glad the article can be of help to you.

  • malenchelon

    Great work. I some times make testing on the strategies. Where they makes me to know some of the things which i ever knew

  • Donna

    Great work on what all goes into testing. I would like to see more testing done on particular times of day with different time periods on various currencies combined with a momentum indicator, such as rsi. thanks for addressing such an important and overlooked topic.