7 Things a Trader Should Understand to Do Proper Backtesting

Backtesting is a process of Testing the trading conditions with respect to the past historical data, evaluating not only the profitability of the system but the underlying risk factor associated with the Trading/Investing Model. Proper Backtesting gives belief and enough confidence to a trader to trade a set of rules. But are the newbie traders really doing proper backtesting?

The Answer is Nope! Some of the common misunderstanding among the newbie traders are

1)Backtest a trading system with limited dataset and if the system gives 60-80% winning ratio or 10% guaranteed returns every month then the newbie trader expect the same to extend in the future. In reality you cannot simply beat the market every month on month consistently and get guaranteed returns. Moreover the shorter time duration of backtesting (3-6 months) doesn’t cover most of the trading scenarios. Testing a strategy with small test data is dangerous and more possibe that it could provide misleading results.

2)Doing a backtest without position sizing : This is one among the mistake made by most of the fellow newbie traders. They dont give priority to position size and so by default most of the trading analysis softwares utilizes 100% of capital everytime and the 100% of returns are reinvested in each and every trade in default backtesting scenario without position sizing. Most of the times such kind of backtesting reports provides misleading, compounded and exotic results. In a real world scenario most of the traders dont want to reinvest their profits or they opt for fixed/variable quantity trading randomly (atleast a newbie traders) mostly non systematic and treating profits as their part of additional income. But the backtesting results are mostly systematic end results not suitable for discrete trading most of the time.

3)Focusing on Profits rather than Risk: Most of the fellow newbie traders focus on strategies that makes money for them however they failed to concentrate on the risk involved in trading system. Some of the questions a trader has to ask himself before trading a trading system. Here are few

i)Am I enough leveraged to trade such strategy?
ii)Does this trading strategy suits my trading style?
iii)Does this strategy can be traded by a Part time or seasonal/Lunch Time Trader?
iv)What is the worst case scenario happened in the last 5-10 years of historical data?
v)Which is the right timeframe to trade this strategy?
vi)What is the worst case scenario involved in trading the strategy?
vii)How many consecutive losses this strategy had made the past? What if the strategy is going to give 10-12 consecutive losses?
viii)How much Iam willing to loose in a worst case scenario?
ix)Does the strategy has enough risk-reward ratio?
x)Can i handle a drawdown of more than 3 months mentally?
xi)How much drawdown I can withstand trading a system before shutting them down?
xii)Should I have to book partial profits? Should I have to fix a trailing stoploss/fixed stoploss? Which is the most efficient one?

Backtesting gives the enough information about the risk involved in any sort of trading system. There is a saying that once the risk is controlled in your trading system the profitability is take care by itself by the trading system.

4)Testing the Trading System Visually : This is again a dangerous game played by the newbie traders which most of the times ends with misjudging a trading strategy. Too many human errors are involved in visual backtesting and moreover it is time consuming and so the manual paper trading requires hell a lot of patience to test each and every trading conditions visually and more accurately. To speed up the process and to avoid most of the human errors it is recommended to do statistical backtesting where in a touch of button you can get the complete backtest report for years within few seconds provided the trading conditions are set right. There are few other factors which we mentioned in our article 20 reasons why your buy sell signal indicator will fail

5)Not including Trading Cost : This is yet another mistake committed by fellow traders. If you are serious about trading your strategy then you should consider including your trading cost (Trading Software, datafeed, Internet Charges, Brokerage, Transaction Cost, Impact Cost, Slippages ,rollover cost..etc) to get more realistic results. These cost may seems negligible to you but in the long run its affects the overall performance of the trading system if not given enough consideration in your backtesting results.

6)Too much of Curve fitting : This is the mistake committed by newbie and fellow experienced traders. Once you know how to do optimization then next step people want to do is to find the best parameters which works in the past with great results. And keep on adding more optimization parameters to fine tune the system is yet again a dangerous game and often leads to curve fitting. It is often recommended to use not more than 2 optimization variables in any trading system.

7)Last but not least backtest a trading system and if the system gives descent results then the next step of a newbie trader is to trade live without validating the model. There are various ways to validate the model. One is doing a simple forward testing and checking how the trading system performs in the live market withtout trading the system and taking all the trading cost in consideration(the actual paper trading). The other methodologies to validate a model are performing walkforward testing & monte carlo simulation which we will discuss more in out upcoming articles.