"Should I move from daily to intraday bars for backtesting?"

SMB Capital recently invited me to give a talk to their quant intern group. The SMB Capital internship is a very competitive program for young college grads to get hands-on experience on a trading desk.

I've received several follow-up questions from the presentation. Here's one from Varun (name used with permission, lightly edited for clarity, and I've removed the strategy details)

Varun:

Several things you said directly shaped what happened in the rest of our program.

Since the call: I shipped one strategy that held its pre-registered out-of-sample test.

I created an exit rule that came from doing exactly the chart reviews you pushed: losers reclaim the open early, winners never do. ~220 trades/yr, profit factor 1.39. A colleague independently reproduced it — his no-trigger version makes ~25% more gross but eats a -$10k COVID dip and a losing 2024 that mine skips; your "smooth is scalable" point in one A/B.

Which mostly confirmed something you said: quants undervalue base signal strength; every one of my failures was decoration on a signal without oomph.

One question:

When your daily-bar idea well runs dry in a universe (mine feels harvested — two real edges found, eight straight kills since), what's your signal for whether to go to finer data (minute bars), a new universe, or new idea sources entirely?

Dave:

Going from backtesting on daily bars to intraday bars is very likely to give you a lot more opportunities to trade the same universe.

Let's say you want to read To Kill a Mockingbird by Harper Lee.

You can read the Cliff Notes, but do you think you'd get the full experience?

Now imagine reading just a summary of the Cliff Notes.

How would that compare to reading the actual book?

That's what you're doing by using only daily bars.

So many of the details of how those bars were actually formed are just glossed over when you look at only the daily bars.

That's probably fine if you're holding positions for days or months, but if you're day trading and/or holding for short time periods, you're missing a lot of the details.

The biggest advantage is the variety of entry signals you can apply to the same universe.

It's not just more details that will lead you to overtrade.

It will lead to a deeper understanding of the strategy you're trading and different ways to enter and exit trades.

Thanks for the question, Varun, and nice work applying what you learned in the session!

-Dave

P.S. Do you wish you had a column library that would tell YOU how to make your strategy profitable? My column library is now included with MabeKit Get Instant Access