You’ve got an idea for a trading strategy you want to backtest.
You have some ballpark idea about what the thresholds should be, but you’re not sure what the optimal settings should be.
That’s ok, you’ve got powerful backtesting software that can give you the answer for what the settings should be. (That’s the whole point of backtesting, right?)
It might take some time, but it’s doable, you tell yourself. Just use your intuition for the settings that should work for your strategy, run a backtest, see the results, make adjustments, and repeat.
This iterative process should hone in on “the answer.” (Or that’s what you believe.)
Your first backtest runs, and you end up with a lot more trades than you thought you would.
And the equity curve sucks.
That’s ok, it’s part of the process, you tell yourself.
Maybe I should add a relative volume filter to reduce the number of trades, you think.
You add the relative volume filter and run the backtest.
Wow, now you ended up with hardly ANY trades at all!
So you pick another threshold and run yet another backtest.
Pretty soon, you’ve run dozens of backtests, and you’re getting frustrated.
You’re nowhere close to finding the right thresholds for your strategy.
In fact, you’re probably further away than when you started because you’ve tried a bunch of settings and haven’t kept track of even what you’ve attempted.
This “guess and check” approach to backtesting is what most traders do, but it’s never going to work.
It’s like trying to hit a bullseye on a target that’s 400m away – with a pistol. You’ll fire a lot of shots and make adjustments each time, but a pistol is never going to be accurate at that distance.
If you hit the target, it’s going to be purely by accident.
Taking a different approach to your first backtest will save you a lot of frustration.
The first backtest for a strategy should be more like calibrating your scope on your sniper rifle.
More tomorrow.
-Dave
