Link: http://www.etc-architect.com/?p=276
From ETC-Architect
When designing any architecture as architects we often tend to underestimate some areas, while overestimating others. Typical examples in underestimations are around the data availability and documentation when designing interfaces or to overlock manual business continuity. Examples of overestimation are areas like terrorism or potential non-availability of less critical systems.
This under- or overestimation is called probability neglect and is usually not limited to our professional life.
Most experts usually will recommend that you will need to change your decision making by adhering to the normative rules of decision making based on recognised statistical reference situation. The problem however is often the complete lack of statistical reference data in our profession. So instead of trying to relate back on normative rules I try to avoid making judgement on a emotional basis. For that I follow the time tested approach on risk management by listing a scoring all risks.
However while doing the scoring I try to observe some additional rules. The first one is always to run against the main stream of opinion, as a lot of overestimation is the result of positive group feedback. The opposite applies for items with no group feedback as they are often neglected as they will not have a sponsor in the risk evaluation group. At the same time I also try to collect data to hopefully better the chances for a normative decision process in the future.
