Backtesting is extremely valuable because it allows you to see if a trading strategy has an edge, before you risk real money.
But there are many ways that you can make mistakes when backtesting. If you make any of the mistakes listed below, you won't have accurate data.
When you have faulty data, that can cause you to start trading a strategy that doesn't actually have an edge. This usually leads to frustration, losses and believing that backtesting doesn't work.
Avoid that headache by learning how to backtest properly. This video will show you how to avoid the 19 most common mistakes.
The Most Common Backtesting Mistakes
1. Not Having a Written Plan
A written plan is essential to backtesting success.
There will be times when you get caught up in backtesting and forget the rules of the strategy. This is especially true if the strategy you're testing has a lot of moving parts.
So write EVERYTHING down. Here are a few things that you should define:
- Entry signal rules
- Percent risk per trade
- Where to place the stop loss
- Profit target
- How to manage the trade
- Indicators used and settings
But there can be several other things to define, depending on the strategy you're testing. Take some time to figure out every single parameter that you would have to define to do an accurate backtest.
You can get my free trading plan worksheet here.
…or make your own. It doesn't matter which worksheet you use.
Just write down a detailed plan and follow it.
2. Altering Backtesting Trades to be Correct 90%+ of the Time
This is an interesting one.
Some traders want to have a really high win rate and will everything they can to get it.
So they fudge trades in backtesting to make themselves feel better. They roll back the chart and take trades on future knowledge of what will happen on the chart.
In reality, most trading strategies aren't going to have a really high win rate and you won't be able to catch an ideal entry all of the time.
The best strategies are robust. They make money consistently over long periods of time.
So trying to be a high win rate trader in backtesting doesn't help you develop a robust trading strategy.
You're just fooling yourself into thinking that the system is better than it really is.
Stick to the testing plan and get an honest result.
The more your backtesting simulates live trading, the better your chances of success.
3. Not Taking Enough Trades in Testing
I've seen several traders backtest 6 months worth of daily chart data and declare that a system has an edge.
Yeah, it has an edge in those 6 months, but what about the rest of the time?
Test your strategy with as much historical data as possible, and take as many trades as possible. Your goal is to have a robust trading strategy that will make money in different market conditions.
Optimizing a strategy to a specific time period is called curve fitting and will set you up for disaster.
4. Not Accounting for Your Emotional State
Everyone has a bad day here and there. At the very least, we have days when we are really tired.
Just like in live trading, your mood and your emotional state will affect your backtesting.
You'll miss trades and you just won't be very sharp.
Account for your emotional state when backtesting and don't test if you aren't feeling up to it.
Also test a system a few times on different days, to account for potential distractions or off days.
5. Quitting When the Results Aren't Immediately Spectacular
Some traders quit a backtest if it isn't profitable within the first few trades.
This doesn't make sense because the beginning of a test could just be period where the system has a normal drawdown.
So stick with a test until you hit a catastrophic drawdown that would not be acceptable in live trading.
A series of 7 losses in a row and a 7% loss is not catastrophic. It doesn't feel good, but any good trading system can overcome that level of drawdown.
6. Changing the System in the Middle of a Test
This one is very common…
“I'll just make one little tweak here.”
“Oh this might work better.”
When you change a system in the middle of testing, you have contaminated data.
It's like if someone mixed apple juice and orange juice together, then gave it to you to drink, and asked you which one tastes better.
Obviously, it would be impossible to tell because you're tasting both of them at the same time.
Likewise, when you change the rules of a trading system, you don't know how your new rules would have performed in the past and you don't know how your original rules would have performed after the change.
Test each system separately and evaluate each one with complete data.
7. Not Following the System
Some traders write down a trading system, then they don't follow it…at all.
I'm not sure why that happens. They might have a problem with authority or something.
But this almost goes without saying…almost.
Write down the rules, then test the rules as they are written.
Most traders can do this. But if you have a tendency to stray from the rules, find a way to remind yourself to stay on track.
8. Not Doing Enough Analysis After Testing and Not Having a Good Reporting System
Knowing the win rate and return of a trading system is not enough.
You have to know things like:
- Maximum drawdown
- Highest number of consecutive losing trades
- MAE and MFE
- Yearly return
- Monte Carlo simulation results
- And more!
You can get this information by using backtesting software that provides this information, or you can import your backtesting results into Excel or Google Sheets.
These additional stats will show you how to improve your system and what to expect in live trading.
9. Thinking That Live Results Will be Exactly the Same as Backtesting
Successful trading requires that we think in probabilities, not certainties.
A trading system that looks good in backtesting may not perform exactly as we expect in live trading. There can be a few reasons for this.
For starters, you may be getting nervous when there's real money on the line and you could be doing things that you didn't do in backtesting. You might be moving your stop loss, or you may cut your winners short.
Some traders hit a drawdown when they first start trading a system live and they think that the system is broken. But in reality, if they reviewed their backtesting, they might realize that it was a normal drawdown.
Third, there might be a variable that you aren't accounting for in your backtesting. That could be the spread, trading at the wrong time of day, or something else.
Finally, there might be something about the markets that's different from when you backtested. There might be unusual news events or the markets may have fundamentally changed.
Your live trading results should be similar, but not exactly the same as your backtesting. Use your backtesting as a guide, but understand that you also have to account for a certain degree of random market fluctuations.
10. Not Considering Portfolio and Correlation Risk
Strategies cannot be traded in a vacuum.
Even if a strategy works well on 2 Forex pairs separately, trading the strategy on both pairs at the same time might not be a good idea.
You may have to reduce your risk on each trade, or decide to only take the signals that look the best.
Trading multiple markets with the same trading strategy can dramatically amplify your drawdowns, so be sure to factor that into your backtesting.
An easy way to do this is to merge the results of both tests into one spreadsheet and analyze the results together.
11. Only Testing on 1 FX Pair/Stock/Commodity and Assuming it Will Work in All Markets
I blame this one mostly on trading educators.
There are quite a few educators that will tell you that their trading system will work in any market.
Maybe. But I haven't found a case where that is true yet.
Each market and each individual stock, currency pair or futures contract has its own personality. They are influenced by different fundamental factors and market dynamics.
Therefore, thinking that you can test a strategy in one market and apply it to all the markets in the world is simply being lazy.
Test each market before you risk real money.
12. Being Distracted and Missing Trades
Yeah, I get it.
Backtesting can be tedious at times.
I used to play movies in the background, or listen to music while backtesting. Now I'm more conscious of what I'm doing when I backtest.
Watching movies is just a bad idea and you should be aware of the type of music you're listening to. Turn off your phone and eliminate as many potential distractions as possible.
If you want to listen to music, understand how it affects your concentration and mood. You might like to listen to Metallica in your car, but that type of music probably isn't the best when you want to focus.
…or maybe it is.
Awareness is always the key to improvement. Be aware of how different stimuli impact your mental state and eliminate the ones that don't help you.
At the same time, think about how to add things that do help. Maybe open a window to get some fresh air, add some decorations in your room, or listen to uplifting music.
13. Entering/Exiting Trades Optimally and Not Honestly
This one can actually go both ways.
There is something to be said for testing a system with an optimal entry.
What I mean by that is entering a trade at the point where it would make the most money and follows your plan, versus where you probably would have entered it in live market conditions.
Entering a trade according to your plan may seem black and white. But when you're testing a discretionary system, there can be big gray areas.
Many traders set the backtest speed to full-speed playback to get through as much data as possible.
Then they overrun the entry point and make trading decisions based on the data that they already saw.
You won't have the benefit of future information in live trading, so don't use it in your backtesting.
The only way that backing up and entering at the optimal point helps, is if you have very precise, mechanical entry rules.
If there is any discretion involved in your entry, like looking for a support/resistance level, then there's the temptation to scroll the chart back to the ideal entry.
To prevent this, look for backtesting software that only allows you to move forward on the charts. This will keep you honest.
14. Only Following Other People's Rules and Not Experimenting for Yourself
Predefined trading systems usually only work well for one trader.
…and that's not you.
Every single successful trader I've talked to or heard about has had to customize their strategies to fit their personality. Myself included.
So learn trading systems from successful traders. But consider those systems as a jumping off point.
When you find a system you like, be willing to customize it fits your personality and lifestyle. Play with different ideas and have fun with the process.
Your chances of success will increase dramatically when you start thinking for yourself.
15. Throwing Out a System Because it Doesn't Make X% Per Month
A monthly return goal is an expectation that may or may not be achievable by your trading system.
It can be possible to have monthly goals for day trading strategies. But for timeframes higher than that, it's almost impossible.
I think that it's much better to have a yearly goal because that gives your trading strategy room to go through a few winning and losing streaks.
So don't try to cram every trading strategy into your expectations. Your expectations are an arbitrary number anyway.
First find out if a strategy is profitable, per your plan.
If it passes that first test, then you can experiment with versions of the entry and exit to try to increase the return.
Also remember that you can potentially use that strategy on other timeframes and in other markets to increase the return even more.
Just be sure to backtest it first.
16. Trying to Over Optimize by Adding More Indicators/Conditions
This is a very common issue.
Humans are naturally built to want to make better versions of whatever they create.
So in search of perfection, we try to improve the win rate of trading strategies by adding more rules to a trading system, in the hope of filtering out losing trades.
Adding more indicators or conditions usually makes a trading system worse.
Your goal when developing a trading system is to find something that is as simple as possible, and is profitable over a long period of time.
Because you'll never develop something that's perfect.
17. Trying to be an Automated Trader When You Should be a Manual Trader and Vice Versa
Some people are better manual traders. Others are better at coding and are more likely to be successful trading automated programs.
Consider your skills and strengths.
I would say that most people are better off as discretionary traders. When they find a system that is profitable, then they can do Incremental Automation and hire a programmer to automate all or parts of their strategy.
However, if you're a programmer or engineer personality type, then fully automated trading could be for you. These types of people thrive on structure and can translate trading ideas into code.
18. Not Studying the Rules of a System That You're Testing (if learned from others)
This is one that's often overlooked.
I've caught myself a few times, where I was too eager to backtest a system and jumped into testing, without studying the rules of the system closely.
Upon further review, I discovered that I had missed a few key points that could have made the strategy profitable in my backtests.
So take your time when you're learning a trading strategy. Go over the material several times to be sure you haven't missed any details.
Then start testing.
19. Preset Bias to Prove or Disprove the System
Finally, some traders have a preconcieved notion about a trading system before they even start testing it.
For example, a trader might believe that most breakouts are destined to fail. So if they are testing a breakout system, they might cut the winners short because they want to capture profits as soon as possible.
The bottom line is to leave your biases at the door.
Follow the plan exactly and don't think that you know better.
Final Thoughts on Backtesting Mistakes
Even experienced traders can sometimes make these mistakes, so always be aware while you're testing and try to be as scientific as possible during the process.
If you follow these guidelines, you'll be rewarded with quality data that will help you improve your trading performance.
Break the rules and you could be spinning your wheels on the Trading Silodrome for years.
It's your choice.