Duration discrimination: effects of probability of stimulus presentation.

Two monkeys sat in a cage with two levers. A light stayed on for either 2 or 8 seconds.

If the monkey pressed the left lever after the 2-second light, it got food every time. If it pressed after the 8-second light, it got food only half the time.

The researchers slowly changed the odds. Sometimes the 8-second side paid off more. They watched how the monkeys’ lever picks tracked the new payoff rates.

The monkeys moved their choices almost perfectly. When the 8-second side paid more, they pressed that lever more until their overall food per hour was as high as possible.

Their behavior looked like a calculator: choice share matched the ideal math for maximizing snacks.

Spanoudis et al. (2011) later showed the same fast math in pigeons looking at colors, proving the rule works across species and cues.

Yuwiler et al. (1992) seems to disagree. Their pigeons on concurrent VI schedules followed the matching law even when it paid less, not the maximizing rule. The gap is the schedule: F used clear win-stay cues, while A used time-based VI schedules where maximizing and matching predictions split.

Davison et al. (1989) added that overall rate feedback alone can’t explain matching, backing the idea that local cues, not just totals, steer choice.

Your client’s world is full of shifting payoff odds—staff attention, peer reactions, token boards. This study reminds you that learners quickly sense which response pays better and drift toward it. If you change reinforcement rates (thinner schedules, differential attention), expect the allocation to follow within minutes. Track the proportions; they are telling you what the learner thinks is “best.”