Published on · Katrin Sommer · 7 min read
The brain looks for patterns; the machine produces none. Systems are born in that gap — and there they look right for a long time.
How unpredictable reward works
One of the most robust findings in behavioural research: the most persistent behaviour is not the one rewarded every time, but the one where it is unclear when the reward arrives. When a regular reward stops, the behaviour fades quickly; when an irregular one stops, it takes far longer.
A slot machine is the exact match for that description. This does not mean a machine "is addictive" — it means the moment to stop does not arrive by itself. That is precisely why the decision has to be made from outside, and beforehand.
The gambler's fallacy
"Nothing for a long time, so it is due." That sentence is the source of the single decision that systematically makes a session worse: carrying on to win the loss back. No game keeps a ledger of debt towards a player; every round is drawn independently of the last.
The same fallacy has an inverse: the "hot hand" feeling after a winning streak. Both come from the same error — assigning meaning to a sequence of independent events.
Why systems look like they work
Stake progressions are the most convincing, because they look good for a long time: many small wins, then one loss that erases them all.
| Round | Stake | Total at risk | Result |
|---|---|---|---|
| 1 | $1 | $1 | Loss |
| 2 | $2 | $3 | Loss |
| 3 | $4 | $7 | Loss |
| 4 | $8 | $15 | Loss |
| 5 | $16 | $31 | Table limit |
By round five, $31 is at risk to chase a $1 win — and the table ceiling arrives much sooner than expected. Filling the table in with your own numbers shows why the system fails better than any explanation.
The part fatigue plays
What is genuinely different late in the evening is not the game but the quality of decisions. Fatigue lowers the ability to stick to rules set earlier; without an interruption from outside, sessions run longer. Both raise turnover and therefore the expected loss.
Decisions made from outside
The practical conclusion fits in one sentence: decisions made during play are worse than decisions made before it. That is why the tools that work — deposit limit, time limit, a break — all take effect before the start.
How design reinforces this mechanism is in the article on sound and colour, the full list of tools in the article on live tables and in the fact check. The technical side of verifiability is in the article on provably fair.
