Skip to content

The Science of Poker Study

A practical, evidence-based guide to actually getting better at poker — built on what cognitive science says about how skill is really learned, then applied directly to how you study the game.

The core idea

Most poker "study" is low-value: re-watching solver outputs, reviewing hands you already understand, grinding volume and hoping. The methods that feel productive are usually the ones that don't work. This wiki shows you the methods that do — and exactly how to apply them to solvers, drilling, databases, and your own game.

How this wiki is organized

  • How Learning Works — the science: the handful of principles that actually build durable skill (retrieval, spacing, interleaving, deliberate practice…). Read this first; it's why everything else works.
  • Studying Poker — the application: how to use solvers, trainers, Anki, trackers, hand-reading and the mental game the way the science says you should.
  • Build Your Routine — put it together into a study system you actually follow.

The one-minute version

Principle What it means at the table/desk
Test, don't review Drill spots closed-book (predict before you reveal the solver) — don't just read solutions
Space it out Many short sessions beat marathons; use Anki for ranges & heuristics
Mix it up Interleave spot types (random-spot drilling), don't grind one node 200×
Trust the difficulty If study feels easy and smooth, you're probably not learning
Extract the why Pull 3–5 heuristics from a sim; don't memorize mixed frequencies
Test after a delay The honest check of what stuck — not how it felt while studying
Process over results Variance means you can play perfectly and lose; judge decisions, not outcomes

New here?

Read The Big Idea and Retrieval Practice first (5 min), then jump to Drilling Spots to see it applied. In a hurry? Go straight to Build Your Routine.


Every principle here is grounded in peer-reviewed research and applied to poker study. Pages link to their sources; where the science is debated, the page says so.