A saddle point in operations research (specifically in game theory) is a payoff in the game’s matrix that is the minimum in its row and at the same time the maximum in its column. When such a point exists, it gives the value of the game and indicates a stable optimal strategy for both players.

Simple definition (exam-friendly)

In a two‑person zero‑sum game with a payoff matrix:

  • Row player (A) wants to maximize the payoff.
  • Column player (B) wants to minimize the payoff.

A saddle point is an element of the payoff matrix that satisfies:

  • It is the smallest element in its row (row minimum), and
  • It is the largest element in its column (column maximum).

If such an element exists, then:

  • Its value = value of the game.
  • The corresponding row and column give the optimal (pure) strategies for both players.

Intuition (why “saddle”?)

  • The row player says: “I will choose the strategy that maximizes my minimum possible gain.” (Maximin strategy.)
  • The column player says: “I will choose the strategy that minimizes my maximum possible loss.” (Minimax strategy.)
  • If maximin value = minimax value , that common value is the saddle point value and the corresponding cell is the saddle point.

This “meets in the middle” like a saddle on a horse—high in one direction, low in another—hence the term saddle point.

How to check for a saddle point (stepwise)

  1. Find row minima
    • For every row, pick the smallest element.
    • Among these row minima, find the maximum.
    • This is the maximin value.
  1. Find column maxima
    • For every column, pick the largest element.
    • Among these column maxima, find the minimum.
    • This is the minimax value.
  1. Compare them
    • If maximin value = minimax value , then:
      • That value is the saddle point value.
      • The cell where it appears is the saddle point position in the matrix.
 * If they are **not equal** , there is **no saddle point** , and you need mixed strategies.

Quick mini‑example (conceptual)

Suppose a payoff matrix for player A:

  • Row minima → max of these = 3 (maximin = 3).
  • Column maxima → min of these = 3 (minimax = 3).

Because maximin = minimax = 3, the entry “3” at their intersection is the saddle point and the value of the game.

Why saddle point matters in operations research

  • It gives a clear, stable solution (pure strategies) for a competitive decision problem.
  • No player can improve their outcome by unilaterally changing strategy when playing at the saddle point.
  • It simplifies analysis compared to mixed‑strategy games where no saddle point exists and probabilities must be calculated.

SEO-style recap for your topic

  • Main keyword: what is saddle point in operation research
  • Core idea: A saddle point is the payoff that is row-minimum and column-maximum, where maximin = minimax, giving the game’s value and optimal pure strategies.

Information gathered from public forums or data available on the internet and portrayed here.