Reversi Endgame Strategy — Parity, Disc Counting, and the Final Sequence
How to win the last 15–20 moves of a reversi game through parity control, region analysis, and the shift from positional play to disc maximization.
What Changes in the Endgame
The endgame in reversi — roughly the last 15–20 moves — plays fundamentally differently from the opening and midgame. The strategic shift is dramatic:
| Phase | Priority | Key Concept |
|---|---|---|
| Opening (moves 1–15) | Position | Openings, central control |
| Midgame (moves 16–44) | Mobility | Move count, frontiers, corner access |
| Endgame (moves 45–60) | Disc count | Parity, region analysis, greedy play |
In the midgame, you play quietly to minimize your frontier and maximize mobility. In the endgame, you do the opposite — flip as many discs as possible on every turn, because there aren’t enough remaining moves for the opponent to flip them back.
The player who played superior mobility strategy in the midgame will typically have a better position entering the endgame — more corners, more stable discs, and a favorable layout for the final sequence.
Parity: The Endgame’s Core Concept
Parity is the most important concept in the reversi endgame. It refers to whether the number of empty squares in a region is odd or even — and who benefits from each.
The Basic Principle
The player who makes the last move in a region has an advantage. Their final disc flip cannot be answered — the opponent gets no chance to respond in that area.
Since Black moves first in reversi:
- If the total number of empty squares is odd, Black makes the last move (first + last)
- If the total number of empty squares is even, White makes the last move
But the real power of parity comes from analyzing separate regions, not just the whole board.
Regions
As the endgame begins, the board often splits into distinct regions — groups of connected empty squares separated by filled areas. Example: there might be 4 empty squares in one corner area, 3 in another, and 8 scattered along an edge, totaling 15, but divided into three separate regions.
The player who makes the last move in each individual region has the advantage in that area.
Controlling Parity
The key insight: you want to play last in as many regions as possible, especially the larger ones. You control parity by:
- Playing in even-count regions first — this doesn’t affect who plays last in the odd-count regions
- Forcing your opponent to open new regions — when they create a new empty area by passing or making certain moves
- Splitting regions — a single move can sometimes divide one region into two smaller ones, changing who plays last in each
Parity Example
Situation: 11 empty squares remain, split into three regions:
- Region A: 5 empty squares (odd)
- Region B: 4 empty squares (even)
- Region C: 2 empty squares (even)
It’s your turn. If you play in Region A first, you force your opponent to respond in A. You alternate until the region fills. Since A has 5 squares (odd), you play last in A — advantage.
Your opponent then plays first in Region B (4 squares, even), meaning you play last in B too.
You play first in Region C (2 squares, even) — your opponent plays last in C.
Result: you get the last move in A and B (9 squares), your opponent gets the last move in C (2 squares). You have a massive parity advantage.
The Transition: When to Switch
One of the hardest decisions in reversi is knowing when to stop playing for mobility and start playing for discs. Switch too early and you blow your position. Switch too late and you leave points on the table.
Signs the Endgame Has Begun
- 15–20 empty squares remain (rough guideline)
- Most or all corners are taken
- The board is divided into distinct regions
- Your opponent has minimal mobility regardless of what you do
- Edge formations are locked in
How to Switch
Stop prioritizing quiet moves. Instead:
- Choose moves that flip the most discs in favorable directions
- Consider parity — play in regions where you’ll get the last move
- Protect your stable discs — don’t let the opponent flip through your stable walls
- Calculate sequences — in the late endgame (~8 empty squares), try to read the exact sequence of remaining moves
The Common Mistakes
Switching too early (move ~35): You start playing greedily while your opponent still has mobility. They use your frontier discs against you, take corners you should have won, and enter the endgame with a better position. Stay patient.
Switching too late (move ~55): You’re still making quiet moves when only 9 squares remain. Every quiet move now is a wasted opportunity to flip discs. At this stage, greedy play is almost always correct.
For more on timing mistakes, see the common mistakes guide.
Reading the Endgame
As the endgame progresses and fewer squares remain, you can increasingly calculate the exact sequence of remaining moves. This is called “reading the endgame” and it’s the most precise form of reversi analysis.
At 15–20 Empty Squares
- Identify regions and count their sizes
- Determine parity for each region
- Decide which region to play in first
- Prioritize playing last in the largest regions
At 8–12 Empty Squares
- Begin calculating exact move sequences
- For each of your possible moves, consider each opponent response, and your follow-up
- Look for moves that force passes (your opponent has no legal play), giving you consecutive turns
- Identify “forced wins” — sequences where optimal play by both sides leads to your victory
At 4–6 Empty Squares
- You should be able to calculate the entire remaining game
- Count exactly how many discs you’ll have at the end for each possible move
- Choose the sequence that maximizes your final disc count
- This is pure calculation — strategy and positional thinking are irrelevant at this point
Computer-Level Endgame Solving
Strong AI programs can solve the endgame perfectly from about 20–24 empty squares — calculating every possible move for both players using techniques like alpha-beta pruning and choosing the absolute best. Humans can’t match this precision, but practicing endgame reading from 8–10 empty squares dramatically improves your results.
Forced Passes in the Endgame
A forced pass occurs when one player has no legal move, giving the other player consecutive turns. Forced passes are extremely powerful in the endgame because:
- You get to play multiple moves in a row without opposition
- Each consecutive move flips more discs to your color
- The opponent can’t respond to any of your flips
Engineering Forced Passes
In the late endgame, you can sometimes engineer a forced pass by:
- Playing a move that leaves your opponent with no legal outflank
- Filling in the only region where your opponent had legal moves
- Creating a position where all potential opponent moves would require outflanking from a direction that doesn’t exist
When you force a pass and then play 2–3 consecutive moves, the disc swing can be enormous — 10 or more discs changing sides in a sequence your opponent can’t interrupt.
Endgame Disc Counting
In the final few moves, count discs precisely to make optimal decisions.
The Quick Count Method
- Count one color. Count all discs of one color (say Black). Call this number B.
- Current total is 64 minus empties. If there are 8 empty squares, B + W = 56.
- So W = 56 − B.
- For each candidate move, calculate the net change. If your move flips 4 discs, your count goes up by 5 (4 flipped + 1 placed) and your opponent’s goes down by 4.
Net Disc Swing Formula
For any move that flips F discs:
- Your gain: F + 1 (F flipped discs plus 1 placed disc)
- Opponent’s loss: F
- Net swing: 2F + 1 (the total disc differential changes by this amount)
A move that flips 4 discs creates a net swing of 9 in your favor. This is why endgame moves that flip many discs are so valuable — even a small difference in flip count creates a large disc swing.
Endgame Patterns
The Last-Move Advantage
In many endgame positions, the player who makes the very last move of the game has an advantage because their final flip goes unanswered. This is global parity — and it’s determined by whether the total remaining empty squares is odd (Black plays last) or even (White plays last).
The Swindle
A swindle is a move in a losing position that sets a trap — hoping the opponent will miscalculate and play a less-than-optimal response. Swindles are most effective when:
- Multiple regions remain
- Time pressure (in timed games)
- The losing player creates a complex position that’s hard to calculate
Even strong players fall for swindles in the endgame, especially online where move timers add pressure.
The Perfect Endgame
A “perfect endgame” is one where you calculate the absolute optimal move at every step, achieving the maximum possible disc count. Strong players strive for perfect play from at least 10 empty squares — this is where most games are decided.
Summary Table
| Endgame Principle | Application |
|---|---|
| Parity | Play last in the largest regions (odd-count regions) |
| Greedy play | Switch from quiet moves to maximum-flip moves |
| Region analysis | Identify separate empty-square groups and assess parity |
| Reading ahead | Calculate exact move sequences from 8–12 empties |
| Forced passes | Engineer positions where the opponent can’t move |
| Disc counting | Use net swing formula (2F + 1) to evaluate moves |
| Transition timing | Switch at ~15–20 empty squares, not before |
Connecting the Phases
The endgame doesn’t exist in isolation — it’s the payoff for everything that came before:
- Openings set the board structure that determines endgame regions
- Mobility strategy in the midgame secures the positional advantage you exploit in the endgame
- Corners provide stable foundations that anchor your endgame disc count
- Edge control determines which edges become stable walls in the final sequence
The best reversi players win the endgame before it starts — by building such a strong midgame position that the endgame merely executes an already-decided advantage.
Practice Your Endgame
Count the empty squares and play last in each region. Watch your win rate climb. Play reversi free against a real opponent.
Play Reversi Free