Drag. Drop. Overfit.

Code is no longer the real barrier to systematic trading. Block builders and coding agents make building easier, but edge design is still the work.

For years the reason most traders gave for not giving systematic trading a whirl was: "I can't code."

Then block builders arrived (drag-and-drop platforms that assemble a strategy without a line of MQL or Python), and that "I can't code" began to hold less meaning. More recently, the rise of agentic coding agents, such as Anthropic, Cursor, Open AI et al. has lowered the barrier to entry even more.

The Real Barrier

Removing the code barrier just drops you faster at the wall that was always the real one: designing an edge that survives contact with a live market.

I build in multiple ways. I data mine, use AI agents to scour source material, and generate my own hypotheses to test. I used to lean on block-style tooling for the parts that would otherwise be a coding slog, however over the past few months 90% of my workflow has been redirected to utilising teams of agentic coding agents. What that buys me is time.

My time then goes into the work that actually decides whether a strategy lives: feature selection, signal logic, how the sizing responds when volatility shifts, where the exits actually sit. The design work. Hard because it's genuinely hard, not because the tooling is fiddly.

A Curve Is Not an Edge

The trap I watch beginners walk into a lot is that these platforms make unscoped building feel like you're making progress. You generate fifty candidate strategies, half of which have lovely curves, and it feels like you're getting somewhere. But a curve is not an edge. The platform can produce the curve. It cannot tell you why the edge should exist, which regime it's built for, or how it should behave when that regime flips. Those are design questions, and they're still entirely yours.

So a block builder or similar platform is a force multiplier on whatever judgement you already have. Point it at a clean strategy idea and it'll get you to a testable candidate fast. Point it at a vague one and it'll get you nowhere. It's not going to do the work for you.

Learn the Design Questions First

What I'd tell anyone picking one of these tools up: learn the design questions before you learn the platform.

What is your edge, in one sentence, and why should it exist?

What is it built for, and how should it behave when conditions turn against it?

If you can't answer those, the platform won't help you. It'll just help you be wrong faster, with a nicer-looking backtest.

The coding barrier coming down is genuinely good. More people who think in markets but not in syntax get to build. Just don't mistake the open door for the finish line. There's still a hell of a lot of work for you to do as the portfolio or strategy orchestrator.