Reward and return

Reward is feedback for one transition; return aggregates rewards across future steps. They are different learning targets.

Ask the Reinforcement learning assistant 1 min read · Updated September 9, 2026

Definition

Reward is feedback for one transition; return aggregates rewards across future steps. They are different learning targets.

Intuition

A small immediate reward can lead to a poor overall outcome. Discounting expresses how future rewards contribute, but also changes the task objective.

Worked example

For rewards 1, 2, 3 and discount 0.5, the return is 1 + 0.5 × 2 + 0.25 × 3 = 2.75.

The math

The discounted return is $G_t=\sum_{k=0}^{\infty}\gamma^k R_{t+k+1}$ . Finite episodic sums stop at termination.

In practice

Check the timing and scale of rewards before tuning a model; a sign or off-by-one error can reverse the learned behavior.

Go deeper

Sources

Sutton & Barto, Reinforcement Learning

More in Reinforcement learning

Assembled from the ReLU.chat curated knowledge base. These explanations are concise on purpose; check the sources for anything important.