Tree Risk Assessment Qualification (TRAQ) – ISA Best Management Practices

[Acer ventura]

I'm following so far - the diagrams are very helpful, thanks.

 

Hi Paul

 

Thanks for the feedback. When I was first looking at this I tried to describe 'Betweeness' in an email to a number of people, without diagrams. As I'm sure you can imagine, it was not an easy thing for them to grasp.

 

Did you manage to follow the logic to the end?

 

Cheers

 

Acer ventura

[Acer ventura]

Risk Matrix Betweeness

 

In the Risk Matrix there is one example of Betweeness.

 

 

• Somewhat likely + Minor = Low
  

jumps directly to

• Likely + Significant = High
  

A Low risk rating jumps to High, skipping over Moderate, with marginal increases in Likelihood that takes Somewhat likely to Likely, and Consequences that takes Minor to Significant.

 

In addition to the Betweeness from the Likelihood Matrix where the upper value of Unlikely (Number 1) and Somewhat likely (Number 2), is the same upper value as Likely (Number 3). The upper value of Low must be the same as the upper value of Moderate, otherwise it could not go directly to High without passing through Moderate.

 

The top value of Low = top value of Moderate

 

 

Whatever the risk rating is for the ‘Moderate’ range, is also the highest risk rating for the ‘Low’ range, and its respective cells.

 

In the Risk Matrix, the only cells with a ‘Low’ risk ranking in, after adjustment for Betweeness in the Likelihood Matrix, is the Negligible column on the Consequences axis.

 

Here’s the Risk Matrix, expressed in probability ranges, accounting for Betweeness in both the Likelihood Matrix and Risk Matrix.

 

 

Has everyone been able to follow this? Any points of clarification needed?

 

Please post here, or pm/email me if you'd rather not go public with an opinion.

 

Cheers

 

Aver ventura

[sloth]

Hi Acer, I'd just like to thank you for taking the time to illustrate this thread. I may not be contributing, but I am following it, and, I think, grasping it!

 

Sent from my GT-I9100 using Arbtalk mobile app

[daltontrees]

Could you define 'betweenness' please?

[Acer ventura]

Risk Contours

 

We saw these two diagrams on the Betweeness section to illustrate that the risk gradient increases from bottom left to top right.

 

 

• The lowest value for each cell is the bottom left corner.
  
• The highest value for each cell is the top right corner.
  

Here’s what the Risk Contours typically look like if the Likelihood and Consequences axis were of a similar scale.

 

 

The Risk Contours represent a line of equal risk. In the same way that contours on an Ordinance Survey are points of equal elevation.

 

The top right contour is the highest level of risk and left the lowest.

 

The contours are ‘illustrative’ only to demonstrate a problem that all risk matrices suffer from.

 

The geometry of the Risk Contours v cell structure has implications for the actual risk values of the risk rankings.

 

 

Risk A = High rating

 

Risk B = Low rating

 

 

Overlay with illustrative Risk Contours.

 

 

Remove all but one Risk Contour

 

Risk A, the High rating, lies to the left of the Risk Contour

 

Risk B, the Low rating, lies to the right of the Risk Contour

 

However the actual value of the risk is lower to the left of the Risk Contour, and higher to the right of the Risk Contour.

 

Therefore Risk B, the Low rating in the Risk Matrix, is a higher risk than Risk A, with the High rating.

 

The portion of the Low Risk B cell that lies above the contour, is a higher risk, than the portion of the High Risk A cell that lies below the contour.

 

Low risk ratings can be higher levels of risk than High risk ratings.

 

High risk ratings can be lower levels of risk than Low risk ratings.

 

And so forth with Moderate risk ratings.

 

The problem that all risk matrices face is Risk Contours are negatively sloped, running top left to bottom right, whereas the cells are infinitely sloped being constructed of vertical and horizontal lines. Consequently, cells of different qualitative rankings will be bisected by Risk Contours of the same value pitching anything to the left of the contour in a cell at a lower actual risk value than anything to the right of the contour, which will be higher.

 

Cheers

 

Acer ventura

[Acer ventura]

With the Risk BMP Risk Matrix, there is such intense ‘range compression’, overlap, and lack of identity of the ranges in the Likelihood of Failure and Impact axis that it’s not possible to plot the Risk Contours. My best Scientific Wild Ass Guess would be that they are extremely negatively sloped, running almost vertical like the stripes on a zebra because of the probabilities for the risk of a death.

 

 

This is far as I’ve got in terms of analysis. There's some off-forum discussions going on, and I will attempt a summary latter this week to include elements of those.

 

As always, any views would be very welcome on or off forum.

 

Cheers

 

Acer ventura

[Acer ventura]

Risk Matrix Literature

 

For those of you who are interested in reading a bit more about risk matrices and getting a better understanding of their intrinsic flaws from people who have forgotten more about the subject than I know, here’s some links.

 

The ground-breaking seminal paper that first looked at matrices in detail is by Tony Cox, ‘What’s Wrong with Risk Matrices’

 

What's Wrong with Risk Matrices? - Anthony (Tony)Cox - 2008 - Risk Analysis - Wiley Online Library

 

The paper is hard work and very mathematical in parts.

 

Kailash Awati does a very good job of distilling the essence of the Cox’s work on his ‘Eight to Late’ blog here.

 

Cox?s risk matrix theorem and its implications for project risk management | Eight to Late

 

Douglas Hubbard, ‘The Failure of Risk Management’

Chapter 7 - Worse than Useless: The Most Popular Risk Assessment Method and Why it Doesn't Work.

 

Here’s an Amazon.com link, which I’ve gone for because it has more reviews than the UK site.

 

The Failure of Risk Management: Why It's Broken and How to Fix It: Douglas W. Hubbard: 9780470387955: Amazon.com: Books

 

Alexander Pickering & Stephen P Cowley, ‘Risk Matrices: Implied Accuracy

and False Assumptions, is a good paper and pretty easy to grasp.

 

OHS SJ Internals 9-16

 

Cheers

 

Acer ventura

[daltontrees]

When I was first looking at this I tried to describe 'Betweeness' in an email to a number of people, without diagrams. As I'm sure you can imagine, it was not an easy thing for them to grasp.

Do try please? We might grasp it. Otherwise the significance of the posting is not clear.

[Acer ventura]

Could you define 'betweenness' please?

 

Hi Jules

 

I thought I had done, and at some length.

 

This might help

 

Betweeness

 

Cheers

 

Acer ventura

[daltontrees]

I genuinely can't find a definition of betweenness, your link is helpful but ony takes me to betweenness risk matrices. I was hoping that there would be a simple definition for the word 'betweenness' in isolation.