Experimental evaluation of matching via a commercially available basketball video game

Nine college students played a basketball video game for 30 minutes. The game paid 2 or 3 points per shot on random-ratio schedules the researchers secretly set.

Each player finished five conditions. The ratio of 2-point to 3-point payoffs changed every time. The team logged every shot the kids took.

Six of the nine players picked shots that matched the payoff odds almost perfectly. When 3-pointers paid off twice as often, those players hoisted more threes.

The other three kids showed no clear pattern. Their choices did not track the hidden odds.

GRADARDANO et al. (1964) saw the same thing in pigeons 56 years earlier. Birds pecked keys on matching random-ratio schedules and their response ratios lined up with payoff ratios. The new study shows the law holds for humans in a video game.

Laposa et al. (2017) found matching in grocery shopping. Higher-income shoppers picked brands that gave more reinforcement per dollar. Schenk et al. (2020) swap shopping aisles for a digital court and still see the same math.

Frank-Crawford et al. (2018) warn that preference can break down when work gets hard. They saw kids drop preferred toys once the task cost rose. Schenk’s gamers only had to press a button, so effort stayed low—maybe why matching stayed tight.

If you run a token economy or point system, check whether the pay rates kids face match the behavior you want. A 3:1 point ratio for on-task vs off-task behavior could pull choices toward the higher payoff just like those 3-pointers did. Try logging the exact ratio you deliver and graph it against the client’s response split—if they are not matching, tweak the odds first before you add more prompts or rules.