Reinforcement Randomiser
Spin a wheel or draw from a bag to deliver reinforcement on a variable schedule — consistently, without slipping into a pattern.
Each spin has a 33% chance. The wheel shows one green slice out of 3.
Shown in the result as a reminder of what to deliver.
This tool decides when to reinforce. Make sure you've already identified something that actually works as a reinforcer for this learner.
Delivering reinforcement on a variable schedule
Thinning reinforcement from "every time" to an intermittent schedule is one of the most common things a behaviour programme asks staff to do — and one of the hardest to do well by feel. Asked to reinforce "about every third response", most people drift: they fall into reinforcing exactly every third time, front-load early in the session, or quietly forget and reinforce far too little. This tool makes the call for you, so delivery stays close to the target and stays consistent from one team member to the next.
Spinner or bag — two ways to randomise
They look similar but behave differently, and the difference matters.
- Spinner (random ratio). Every opportunity is independent, with the same fixed chance of paying out — a 1-in-3 setting is a 33% chance each spin. This is the closest match to a true variable-ratio schedule: the average comes out right over time, but you can get short runs of nothing or two payouts close together. That unpredictability is exactly what makes variable schedules so durable.
- Draw from a bag (fixed ratio per block). The bag holds a set number of tokens — say one "reinforce" token in five — drawn without replacement until it's empty, then refilled and reshuffled. The order is random, but the rate is guaranteed: the learner always gets exactly one reinforcer in every five opportunities. Use this when a steady, predictable density matters more than true randomness, or when you want to be sure a learner who is close to giving up still gets paid out within each block.
How to use it
- Set your target — the ratio you want to thin to, or the number of reinforcers per block.
- Each time the learner gives you an opportunity (a response, a correct trial, a stretch of appropriate behaviour), spin or draw.
- If it lands on Reinforce, deliver straight away. If not, carry on and go again at the next opportunity.
- Glance at the actual rate now and then to check delivery is tracking your target.
Why randomise at all?
Two reasons. First, people are poor random-number generators: left to judgement, reinforcement delivery becomes predictable, and a learner who can predict exactly when the next reinforcer is coming has every reason to coast in between. Second, it keeps delivery consistent across a team — three staff members "winging" a thin schedule will run three different schedules, which muddies your data and the intervention. A shared randomiser removes that drift. Kept visible, the spinner can also work as a "mystery motivator" the learner watches, adding a bit of anticipation to each opportunity.
First, make sure it's actually a reinforcer
A randomiser only governs timing. It can't make an item or activity reinforcing — that's defined by whether it actually increases the behaviour it follows. If responding isn't improving, the schedule isn't the first thing to check; the reinforcer is. (See why a preferred item isn't automatically a reinforcer.) Pair this with a token board when you're running a token economy.
Frequently asked questions
What is a variable-ratio schedule?
A variable-ratio (VR) schedule delivers reinforcement after an unpredictable number of responses that averages out to a set figure — VR3 reinforces on average every third response, but any given reinforcer might come after the first or the sixth. The unpredictability is what makes VR schedules produce steady, persistent responding that's resistant to extinction. The spinner mode approximates VR by giving every opportunity the same fixed probability.
Which mode should I use?
Use the spinner when you want a genuine variable schedule and can tolerate the odd dry run. Use the bag when you need a guaranteed minimum density — for example, when a learner is fragile and you don't want a string of non-payouts to push them into giving up.
Does the learner need to see it?
No. It works purely as a staff decision aid. But showing it can help: an on-screen spin makes the moment salient and adds anticipation, which some learners find motivating in itself.
How do I thin the schedule over time?
Start dense — often continuous (reinforce every time) — and lean the schedule out gradually as the behaviour becomes established, nudging the ratio up only while responding holds steady. If responding drops off, you've thinned too fast; step back to a richer setting and move more slowly.