After-action review (AAR)

A structured post-period review of one’s own trading: what worked, what didn’t, and what to change next. The term is borrowed from the military; Erik Smolinski reframes it as the trader’s version of a company’s quarterly earnings report — treat trading as a business, and the business reviews itself on a schedule smolinski-h1-2026-aar.

The idea

Some profit mechanisms are persistent and some aren’t — “some will make good money when they work, but then they won’t always work” — so the discipline is knowing when to stop doing a thing and pivot. An AAR is the recurring checkpoint where that call gets made against evidence rather than gut feel: performance attribution, regime shifts, and a candid methodology self-critique (where the data came from, how claims were tested, where they’re weakest).

Where it sits in the stack

The AAR is the evaluation layer performed by a human. Automated systems in this spoke evaluate a strategy with a backtest or an eval toolkit and tune it with hyperparameter search; a discretionary trader runs an AAR instead. Same job — validate and adjust the strategy — different instrument: a quarterly narrative review of a live book vs. a programmatic loop over history. The two rhyme: both are only as honest as the data and the guard against fooling yourself (the discretionary cousin of lookahead bias is over-reading a single episode, which Smolinski flags explicitly on his dip-buying slide).

For the non-professional

Smolinski’s scaled-down prescription: schedule a checkup quarterly or at least annually, and start with a benchmark panel — hold your returns against SPY / QQQ / IWM / TLT / GLD and ask whether you’re comfortable. The review’s value is as much in confirming “nothing to change” as in exposing excess fees or persistent underperformance smolinski-h1-2026-aar.

Related

smolinski-h1-2026-aar · market-regime-analysis · strategy-optimization · backtesting · erik-smolinski