For most of trading history, one skill mattered above the rest: reading the market. Everything else — the execution, the record-keeping, the analysis — was either manual or handled by someone else. That’s no longer true, and the traders coming up now are being shaped by a very different set of expectations than the ones who came before them.
Finance is still the foundation, and it always will be. Understanding risk, position sizing, what actually moves an instrument, how leverage cuts both ways — none of that goes out of style just because more of the process is automated. If anything, this knowledge matters more once execution is automated, because a bad idea coded well just fails faster and with more conviction behind it. We still see this every cohort: students who can write clean code but have no real grasp of risk are often more dangerous to their own capital than the ones who trade slowly and manually, because their mistakes scale.
Coding is the newer requirement, and it’s the one causing the most friction for traders who came up the traditional way. It’s not about becoming a software engineer. It’s about being able to translate a rule in your head — “only take this setup when volatility is below this level and I haven’t already lost twice today” — into something that executes exactly that way, every time, without needing you to be present or emotionally steady enough to follow your own rule. Traders who skip this step end up dependent on someone else to build what they can’t build themselves, which means every change to their idea goes through a bottleneck they don’t control.
Data is the skill people underestimate the most. Having a strategy idea is easy. Knowing whether it actually holds up requires being able to pull historical data, structure it correctly, account for things like slippage and survivorship bias, and interpret a backtest without fooling yourself. This is where we see the most well-intentioned mistakes — someone builds a strategy, runs one backtest that looks great, and goes live without understanding that the backtest was quietly overfit to exactly the period it was tested on. Reading data honestly is its own discipline, separate from both finance knowledge and coding ability.
The traders who are thriving right now tend to have some working level of all three, even if they’re not equally strong in each. A finance-first trader who picks up enough coding to automate their own rules. A coder who develops enough market instinct to know which ideas are even worth testing. Someone who’s strong in data and treats every strategy as a hypothesis to be disproven rather than a belief to be defended. The combination is what compounds — each skill checks the blind spots of the other two.
None of this means the old single-skill trader disappears overnight. But the ceiling for that path is getting lower every year, simply because the tools available to everyone else keep getting better. The next generation of traders isn’t going to be defined by which one of these three skills they picked. It’s going to be defined by how quickly they stopped treating this as a choice at all.
For more details:
