Conventional wisdom: study endgames. Every coach says it. Every improvement guide repeats it. I committed 90 days to systematic endgame study to see what actually worked. Here's the honest report.
The plan
Three books across 90 days:
- "Silman's Complete Endgame Course" (covered to my rating level + one above)
- "100 Endgames You Must Know" (Jesús de la Villa)
- "Endgame Strategy" (Mikhail Shereshevsky) — lighter pass
Plus: solving all endgame puzzles in my training queue.
What actually worked
1. Pawn endings — measurable rating impact
The most undervalued part of endgame study. King-and-pawn endings happen in maybe 5-10% of my games but win or lose almost every time they occur. After 30 days of pawn-ending study, I won three games I would have drawn or lost previously.
The patterns that matter most:
- Opposition (direct, distant, diagonal)
- The square rule and corresponding squares
- Triangulation in basic positions
- Outside passed pawns and king position
Time invested: ~25 hours. Estimated rating impact: +20-30 points. Best return on time of any endgame study.
2. Rook endings — the bulk of endgame play
Rook endings occur in 30-40% of my games. Most are technical positions with known correct treatment. Studying them paid off, but slower than pawn endings.
The high-value patterns:
- Lucena and Philidor positions
- Rook activity over passive defence
- King-side cut-off techniques
- Pawn races with rooks on both sides
Time invested: ~40 hours. Estimated rating impact: +15-20 points.
3. Bishop vs. Knight understanding
Not specific endgame positions but understanding when to trade and when to keep. Affected my middlegame decisions as much as my endgame ones.
Time invested: ~10 hours. Estimated rating impact: ~10 points (mostly through better middlegame trades).
What didn't produce return
1. Memorizing complex theoretical positions
The "100 Endgames You Must Know" approach is correct in principle but the most-complex positions in those collections almost never appeared in my games. Rook + pawn vs. rook in specific theoretical positions: I memorized them; I never used them.
Time invested: ~20 hours. Estimated rating impact: ~0 points.
2. Strategic endgame study without concrete examples
"Endgame Strategy" was beautifully written but the concepts were too abstract for my level. Reading "principle of two weaknesses" didn't translate into actually exploiting two weaknesses in my games until much later. The book may pay off at master level; at 1900, the concrete pattern study was higher leverage.
Time invested: ~15 hours. Estimated rating impact: ~5 points (delayed and indirect).
3. Mass tactical solving in endgame positions
Solving lots of endgame puzzles helped less than expected. The puzzles were typically about finding tactics in endgames; my real-game weaknesses were technique-based. Different skill, different practice.
Time invested: ~25 hours. Estimated rating impact: minimal.
The pattern under what worked
The endgame study that produced returns shared features:
- It addressed positions that actually appear in my games
- It taught technique — what to do, in order — rather than abstract principles
- It was paired with immediate application in my own games
The endgame study that didn't produce returns also shared features:
- It addressed positions theoretically important but rarely seen at my level
- It was abstract or principle-focused without concrete pattern repetition
- It was disconnected from my actual game profile
What I'd do differently
- Study pawn endings first. Highest leverage by far.
- Study rook endings second. Most common endgame.
- Skip the rare theoretical positions until rating 2200+. Time better spent on more common scenarios.
- Pair every studied position with one of my own games. Apply or analogize within 24 hours.
- Drop the abstract strategy books until I've mastered the concrete patterns. Patience.
The number that mattered most
In the 60 days after intensive pawn-ending study, I won three games where my pre-study self would have drawn or lost. Three games. That's real. The remaining 86 games of the 90-day period had endgames where my study showed up at the margins.
Endgame study doesn't produce dramatic improvements. It produces small consistent improvements that compound. The key is targeting the small improvements that show up in your games, not the ones that appear in your books.
