Bellman Update Missing Reward

// One Bellman backup for state s in value iteration
function bellmanUpdate(s, actions, transitions, reward, gamma, V) {
  let best = -Infinity;
  for (const a of actions) {
    let q = 0;
    for (const [sp, prob] of transitions(s, a)) {
      q += prob * (gamma * V[sp]);
    }
    best = Math.max(best, q);
  }
  return best;
}