Your PnL is lying to you.

Baseball had the same problem before Moneyball changed everything.

Scouts watched a player and saw what they wanted to see. A good swing. An athletic build. A highlight reel. The stats backed it up — batting average, home runs, RBIs. Everyone agreed. The player was elite.

Billy Beane asked a different question. Not what does this player look like. What does this player actually produce when you strip away the noise?

On-base percentage. Walks. Pitch selection. The ugly, invisible stuff that never made the back page but predicted winning more than anything the scouts were measuring.

Dealing rooms are still running the old scouting model.

PnL is batting average. It tells you something. It doesn't tell you enough. And it definitely doesn't tell you whether the process behind it is sustainable or deteriorating.

Two traders. Same month-end number. One executed with precision across every regime the market offered. The other survived a drawdown through position averaging and a late recovery that had nothing to do with skill.

Same batting average. Completely different player.

A real performance metric would sit across three layers simultaneously.

Execution quality relative to the market regime at the point of commitment. Not after. Not averaged. At the moment the decision was made. That's on-base percentage.

PnL trajectory — not where it is but how it got there. A rising book built on clean decisions is a different asset to the same number built on recovery and luck. That's the difference between a walk and a swinging bunt that found a gap.

Behavioural consistency — is the trader executing the same way at plus twenty as they are at minus five? If not, the edge isn't structural. It's conditional. That's a player who only performs with no one on base.

Most desks are still reading the scoreboard. Very few have a live feedback loop connecting what the market offered, how the trader responded, and whether the pattern of execution is holding or shifting.

By the time the PnL tells you something is wrong, the drift has been compounding for months.

Beane didn't find better players. He found better questions.

But here's what Moneyball didn't solve. Knowing what to measure is the diagnostic. It detects the drift. It doesn't correct it. Oakland could identify undervalued players. They still couldn't turn a declining one back into what he was.

The measurement layer is one part of a larger architecture. Beneath it sits the structural work — what protects the decision-maker's judgment under sustained load, what holds execution steady when the pressure compounds across months and years. And around that sits the environment — whether the desk itself is engineered to sustain its best people or just expects them to carry the weight alone.

Most firms have none of these layers. Some have one. Almost nobody connects all three.

The firms that build the full structure won't just see the drift earlier. They'll have something in place to answer it.