I don't put them in mine either. We work on a binary system. Is the risk adequately controlled?; YES or NO

Personally I do use a risk matrix, but Im very much open to the idea of the binary system of YES / NO. Looking at the actual value of risk matrix one must ask themselves what is the benefits in using one, other than to rank and rate your hazards.

One advantage of the matrix is to show how close to "yes" or "no" the risk is - I use this to decide how frequently the risk assessment needs to be reviewed. This is in line with the Approved Code of Practise for my industry which states "An assessment must be reviewed and updated as necessary" and "Risk Assessment and control is a continual process. Hence, written risk assessments should be subject to periodic formal reviews to confirm the validity of the assessment and whether the risk controls are still effective and adequate."

As above, the Approved CoP for my industry recommends using a matrix, albeit an odd 4x3 matrix. If I decided not to use one I would have to be very certain that any alternate was "at least as effective as the guiodance in this CoP".

I would be intersted in looking at the CoP you refer to Colin. Decimo

'Tis large, about 9Mb http://www.dft.gov.uk/mca/coswp2010.pdf

I think that the matrix's are very subjective, Low Med High or 3 x 3 or even a 5 x 5 matrix is just as subjective. I think that its a requirement of the five steps to risk assessment document which the HSE produces in which a person must assess the risk and the 3 x3 or a 5 x 5 was a way to do that. Thats what I took from the document. Incidentally I was reviewing a FMEA document the other day for high velocity material being ejected and potentially striking someone. Still a very subjective document as its all possibilities based on limited data. Reminded me of the swiss cheese accident model and the old saying of three types of lie. Lies, damn lies, and statistics.