You don't have to agree with me on everything to get value from this website or to become a client. Some agreement helps, but it isn't essential because I'm here to help you do what you want to do. With that caveat out of the way, what do I really think?

Favourite themes

To give you a first sense of what this website is based on, here are some of my favourite themes:

• Psychology

• Design

• Innovation – often based on insights gained from critical analysis

• Clarity – especially in writing and thinking

• Usefulness

These are themes that often struggle for attention in the world of internal control and risk management, which is one reason this website is unique.

Winners and losers

This table should give you a sense of what comes out well on this website and what comes out badly.

Some points that might surprise you

Finally, these are some of a number of interesting conclusions I have reached over the years that may surprise you.

• There is no such thing as 'the' probability of something happening. Probabilities are conditional so it depends what conditions you take into account. I also agree with the Bayesians that it also depends on what you thought before you processed the last bit of evidence.

• Risks are not real. Risks are events whose occurrence is uncertain. If we knew they would happen or not happen, with certainty, they would not be risks. So, logically, their existence is only in our minds. Danger is real, the limits of our knowledge are real, but these things we call 'risks' are part of our thinking. Risk analysis is really a process of structuring the fog of uncertainty that surrounds our thinking into something we can think about. (That's the aspiration anyway.) Accordingly, it is misleading to talk of 'identifying' risks; we define them.

• Thinking about risks is not a particularly effective way to think under limited knowledge. Very often it is quicker and just as effective to see that there is a lot that we don't know in a particular area of our work and respond to that fact. It is usually more helpful to define uncertainties (typically, as variables in a mental model) than to itemise risks, which are just sets of possible values for those uncertainties. It is almost always more helpful to think in terms of probabilities, distributions, and values than to get embroiled in thinking about risks.

• Apparent 'risk aversion' can be explained by common sense, and we don't need to hypothesize an aversion to risk or uncertainty, or utility curves. Obvious, everyday effects can explain most and perhaps all of our apparent aversion to risk.

• The main way that risk control can be improved in most organizations is by improving ideas for risk responses, not pouring more effort into analysis. Design skill is where the main opportunities lie.