Kristian Glass - Do I Smell Burning?

Mostly technical things

Words of Estimative Probability

Uncertainties are common, from “we’re almost certain this is the cause of the outage”, to “project X probably won’t finish before we need to start project Y”, and “the train will almost certainly be on time tomorrow”.

Broadly speaking, decision making consists of “get information, do a thing”, whether a structured framework like OODA (Observe/Orient vs Decide/Act) or “let’s get lunch – where’s good?”

Often, this involves multiple people – the person making a decision will rely on information and analysis from others. During an outage, an Incident Commander will be making decisions based on observations by Subject Matter Experts; a Project Owner is best placed to speak about the timeline of their project but resource allocation is done elsewhere; I am familiar with local public transport and my houseguests would like to know so that they can make plans.

From Wikipedia’s “Words of Estimative Probability”:

Words of estimative probability (WEP or WEPs) are terms used by intelligence analysts in the production of analytic reports to convey the likelihood of a future event occurring. They express the extent of their confidence in the finding.


A well-chosen WEP gives a decision maker a clear and unambiguous estimate upon which to base a decision. Ineffective WEPs are vague or misleading about the likelihood of an event. An ineffective WEP places the decision maker in the role of the analyst. The decision maker has to infer the prediction alone, thus increasing the likelihood of poor decision making or snap decision making.

There are “good WEPs” (e.g. Sherman Kent’s list – certain/almost certain/probable/etc.) that can clearly communicate probability, there are “bad WEPs” (weasel words like “might” and “could”) that are vague and likely to cause confusion.

Analysis, clearly communicated, helps the decision-maker. Telling me you’re “almost certain” about the cause of an outage helps me more than “it might be X”.

Next time you’re called upon to communicate some sort of uncertainty, consider the language you use and the analysis you offer.

Communication is hard. Uncertainty and probabilities are hard. Shared vocabulary helps.

Thanks to Eugene Kogan for telling me about these