Understanding Walk Forward Optimization: A Key Technique for Reducing Overfitting in Backtests


As a trader, one of the most important tools at your disposal is a backtest trading engine. A backtest engine allows you to simulate the performance of your trading strategy on historical data, which can help you to evaluate the strategy’s potential profitability and risk. However, it is important to be aware of the limitations of backtesting and the potential for overfitting. Walk forward optimization is a technique that can help to mitigate the risk of overfitting and improve the reliability of backtest results.

What is Overfitting?

Overfitting occurs when a model or strategy is too closely tailored to the characteristics of a specific data set, and performs poorly when applied to real-world data. This can be a major problem in backtesting, as it can lead to unrealistic and unreliable results. For example, a strategy that is overfitted to a particular data set may appear to be highly profitable when tested on that data, but may struggle to perform when applied to new data.

How Does Walk Forward Optimization Work?

Walk forward optimization is a technique that helps to reduce the risk of overfitting by dividing the data into overlapping periods and optimizing the model on each period in turn. This process is repeated until the entire data set has been used. By using walk forward optimization, traders can get a more realistic and reliable sense of how their strategy is likely to perform in the real world, as it has been optimized on a wide range of market conditions rather than just a single data set.

One of the key benefits of walk forward optimization is that it allows traders to see how their strategy would have performed if it had been constantly reoptimized over time. This can help to identify any changes in the strategy’s performance that may have occurred due to changes in market conditions or other factors.

How to Implement Walk Forward Optimization

There are several steps involved in implementing walk forward optimization:

  1. Divide the data set into a series of overlapping periods.
  2. Train the model on the first period and test it on the second period.
  3. Optimize the model using the second period, and then test it on the third period.
  4. Repeat this process until the entire data set has been used.
  5. Analyze the results to see how the model performed on each period.

By following these steps, traders can use walk forward optimization to get a more realistic and reliable sense of how their strategy is likely to perform in the real world.


In summary, walk forward optimization is a powerful technique for reducing the risk of overfitting in backtests. By using WFO, traders can make more informed decisions about their trading strategies and increase their chances of success.

For now, you can do WFO on manually by selecting each backtest period. However, in the coming months, a new feature will allow you to do this automatically.

Joins us:
Runbot Twitter
Runbot Discord

Share This :