Relative delay of reinforcement and choice.

Quilitch et al. (1973) tested how pigeons choose between two keys. Each key paid off on its own schedule. The team held the overall payoff rate steady while they changed only the delay to food.

They wanted a clean number to predict the birds' split. They tried the inverse of delay squared.

The simple formula worked. Choice ratios almost perfectly matched the inverse of relative delay squared.

Even when payoff rate never changed, delay alone steered the birds.

PLISKOFF (1963) showed pigeons match response ratios to payoff ratios. R et al. keep the matching idea but swap in delay instead of rate.

Neuringer et al. (1968) added a change-over delay and still saw matching. R et al. prove the delay piece itself drives choice, not just the payoff rate.

Sarber et al. (1983) later tweaked the math with a square-root twist for chained schedules. The field keeps refining the same core idea: choice follows a tidy equation.

When you set up concurrent reinforcement, think delay, not just rate. A slightly longer wait to praise on one task can shift a client's whole effort to the other side. Try trimming the delay on the task you want to boost and see the split change.