How to Backtest Multiple Parameters on TradingView

TradingView strategy tester runs one parameter set at a time. You input a value, run the backtest, write down the results, change one parameter, run again, compare. For 3 parameters with 10 values each, that is 1,000 separate backtests. Nobody does that manually. They test a handful of combinations and hope for the best.

The real issue is worse than time. Single-parameter testing assumes parameters operate independently. But in practice, the optimal RSI period depends on your overbought level. The best moving average crossover period depends on your stop loss distance. These interactions are invisible when you change one parameter at a time.

I fell into this trap for months. I would optimize RSI period while keeping overbought at 70 and oversold at 30, land on period 14 as best, then move on. Later, I ran a proper multi-parameter grid search and found that RSI period 9 with overbought 78 was substantially better. The interaction between period and overbought level was the real driver of performance, and single-parameter testing completely missed it.

Here is a concrete example of why multi-parameter testing matters. I was optimizing a Bollinger Bands strategy and found that MA period 20 with standard deviation 2.0 gave a profit factor of 1.4. But MA period 20 with standard deviation 2.5 gave a profit factor of 0.9. In isolation, you would think MA period 20 is neutral or slightly positive. But the interaction with standard deviation tells a different story.

The grid search output revealed that the real driver was the combination of MA period and standard deviation, not either parameter alone. The best cluster was MA period 18-22 with standard deviation 1.8-2.2. MA period 20 with standard deviation 2.5 was in the bottom quartile. A single-parameter test would have averaged these out and told me nothing useful.

This is why grid search is not just a time saver. It reveals structure in your strategy performance that single-parameter testing cannot detect.