Backtesting

Manual

Some chart related strategies or price-action based strategy could be done manually by reading historical charts.
Investment related strategies could easily be verified manually because of lesser amount of data.

Algorithmic Backtesting is most preferred by traders. It is less time consuming and results could easily be shown to others if you are working for an organization.

Next, we are going to show Algorithmic backtesting example.

Step-1 Choose a strategy[go to previous section, where we have discussed many strategies]

Step-2 Preparing for historical data

Step-3 Choose any coding language

Step-4 Optimizing the strategy

Step-5 Forward testing or paper trading

Frequently Asked Questions-Q&A

If I have backtested my strategy, does that mean it will give me similar result if I do actual trading?
- No, if you are a quant trader then you should understand the concept of overfitting or over-optimization. Backtesting just tells us that strategy is working on historical data not future data.
My strategies are giving very good accuracy during backtesting but perform poorly on actual trading. What am I missing?
- One of the solution for above problem is how do we prepare our data for backtesting. Because it is very important to test our strategy on out of sample data and forward testing on live data before doing actual trading.
- If you are a data scientist then you should have heard of training data and testing data. We reserve some data to test our backtested model. For more details read our backtesting page.

Page updated

Google Sites

Report abuse