Runs of results: what chance does not compensate for
A supplier makes three consecutive non-compliant deliveries: should the next delivery be considered more likely to meet requirements? The managerial question is not whether the run must end, but whether the available observations justify changing the forecast.
What the literature establishes
Amos Tversky and Daniel Kahneman (1971, Psychological Bulletin) describe the belief that a small sample should faithfully reproduce the properties of the population from which it is drawn. Their observations among researchers show that scientific training does not necessarily protect against this excessive expectation of representativeness. Applied to a sequence of outcomes, this belief can make an unbalanced run seem more surprising than it actually is.
Why a return to balance seems plausible
Daniel Kahneman and Amos Tversky (1972, Cognitive Psychology) examine probability judgements based on representativeness: an outcome seems likely when it resembles what people expect from the process producing it. A short sequence alternating between good and bad outcomes can therefore seem more consistent with randomness than a sequence containing several identical outcomes. An intuitively plausible pattern of alternation may then be mistaken for a well-founded forecast.
The premature conclusion
It would nevertheless be wrong to conclude that previous outcomes should never influence a decision. If events are independent and their probability is constant, a run of failures does not make the next success more likely; with a supplier, these conditions cannot simply be assumed. Successive incidents may indicate a persistent malfunction, while a documented correction may justify revising the forecast. (our executive and employee training programmes)
What these studies cannot decide
These studies concern probability judgements, not the effectiveness of a quality-control policy in a company on the Swiss Riviera. They provide neither a universal alert threshold nor a number of deliveries after which a supplier should be replaced. The weakness of common practice emerges when an expected return to normal is invoked without specifying what has changed in the process.
A practical check in Vevey
As part of SHR’s decision-making programme, a purchasing team at the headquarters of a global food group or at an SME on the Swiss Riviera could pilot a post-incident forecasting log in Vevey. Before the next delivery, it would record the estimated risk of non-compliance, recent outcomes and any documented process change. Over a period defined in advance, it would calculate the proportion of downward risk revisions justified solely by the idea that “the run must end”, then compare forecasts with observed outcomes. This measure makes a line of reasoning verifiable; it does not, on its own, demonstrate improved quality. To go further: explore the Decision-making training in Vevey in the canton of Vaud, or browse our executive and employee training programmes in Switzerland.
In pictures: Decision-making in Vevey
- decision-making
- Vevey
- research
- training vevey