The Science of Chess Improvement
Thousands of studies have examined chess improvement. The findings are remarkably consistent, and they challenge many popular beliefs about how to improve at chess. Here is what the research says, distilled into actionable advice.
Deliberate practice is the key
Charness et al. (2005) found that the single best predictor of chess rating is hours of deliberate practice, focused, effortful study with feedback. Not casual play, not watching videos, not reading articles. Deliberate practice means: solving challenging puzzles (not easy ones), analysing your own games with an engine, and studying positions that stretch your understanding.
What counts as deliberate practice
- High ROI: Puzzle solving (with deep thought, not speed-solving), game analysis (your own games with engine comparison), endgame study, and playing slow games.
- Medium ROI: Studying master games, opening preparation, reading chess books.
- Low ROI: Watching streams passively, playing only bullet/blitz, reading articles without applying the ideas.
The role of talent
Research shows that talent (measured by IQ, spatial ability, or working memory) explains only 5-15% of the variance in chess rating. The remaining 85-95% is explained by training volume, training quality, and starting age. Translation: anyone can reach 2000 with the right training. See Why 90% Never Reach 2000.
Your evidence-based improvement plan
Based on the research:
- 30 minutes of daily puzzle solving (deliberate, not casual).
- One slow game per day with full post-game analysis.
- One endgame or strategy study session per week.
- Track your progress with objective metrics (puzzle accuracy, game accuracy, rating). Follow the 90-day plan or the 0-to-2400 blueprint.