Hacker News new | ask | show | jobs
by cconover 3221 days ago
Spot on. In the Coast Guard we have a similar operational risk assessment model called General Assessment of Risk (GAR)[0], which uses six categories and a score of 1-10 in each to get a cumulative score, which corresponds to either Green (low risk), Amber (moderate risk), or Red (high risk). In a best-case scenario, all involved in the operation will conduct a GAR brief as a group, and the person doing the brief will solicit the group for what number they would give for each category, and the highest number anyone calls out is the one the whole group goes with. Typically, if somebody calls out a 5 or higher, they are asked to explain why they feel the risk in that category is so high. Once the reasons are identified, the group then "mitigates" the risk by discussing the identified reasons so that everyone understands the risks. The score is tallied, and the color (risk level) is identified.

That is the best-case, textbook way to run the model. Still very subjective pseudo-science, masquerading as objective risk management, but at least has something of a method to it.

In practice, the repetitiveness of the GAR model results in many crews blowing it off, and giving a vague, arbitrary cumulative score without any discussion around how they got there. This sounds bad, and by policy it is bad, but in practice I have observed no discernible difference in how crews approach risky missions and operations when they conduct a full GAR brief or just give a somewhat random score and move on. In other words, the GAR model does not seem to provide any tangible risk management benefit, and largely seems to serve as a bureaucratic CYA solution.

GAR was introduced in an attempt to reduce the number of mishaps occurring due to what was deemed to be excessive risk-taking. The statistics may demonstrate that it had that effect, though I would argue from my perspective that other training programs introduced to address problems related to risk assessment have far more deckplate-level impact and effectiveness. The problem with GAR is that it tries to objectively standardize something which is, by its very nature, dynamic and subjective. No two people, in the exact same situation and having the exact same capabilities and experience, assess the risk the same way. Assigning numbers to a series of broad categories and giving a color-coded risk level to the situation does not inform anyone of anything very useful. Discussion of risk factors is more helpful, but due to the way the system is structured, is a step frequently skipped.

What matters far more is focusing on continuous, dynamic training and education of those in billets for whom risk management is a critical part of the safe completion of their mission, and emphasizing clear communication unfettered by rank or positional authority to ensure that everyone has full situational awareness. Be respectful, but make sure that information can move freely between all involved.

[0] https://www.uscg.mil/hq/nsfweb/foscr/ASTFOSCRSeminar/Present...

1 comments

Shout out to a fellow Coastie.

I think what I learned from my superiors is that it doesnt matter so much what system is used for risk assessment, but whether a conversation about risk was had in a meaningful way. Sometimes quantification helps this, where complex systems can be analyzed and the consequences assessed. Usually if you are going to do that it has to happen well in advance of an operation. Othertimes the desire to get a number leads to a real ham fisted attempt to "quantify" things like how fatigued the crew is on a scale of 1-10, or to rate the environmental conditions. When the GAR model, which stands for Green - Amber - Red, was used to facilitate an honest conversation about how people felt about an operation as opposed to just checking a box so that it could be put in the logs. When I saw it being used as the former it absolutely made things safer, but that was very dependent on the attitude of those participating.

Agreed, and a more succinct way of saying what I was getting at. Attempts to quantify inherently subjective attempts aren't very productive in their own right. What matters far more is having a substantive conversation about risk, and work to promote a culture of open and honest communication.

Semper P

If you want to know how tired your crew is, wouldn't it suffice to pluck a random few of deck and ask them if they're tired as fuck?
How would you make them answer truthfully? Admitting to being tired either means you're showing personal weakness, or you're calling your superiors incompetent at scheduling. Neither possibility sounds good for your career. I think it would be better to measure reaction times, eg. by making them catch a falling ruler immediately after it's dropped and seeing how far it falls.
This is why good leaders create an environment where you can be honest about such things. You are basically talking about an inhumane system led by fear. Then again we are talking about the navy here...
This is part of the reason why they institute fatigue standards, to try to keep people from being over tired and still working, which led to people being hurt.
The hard part is that no one wants to admit that they are tired. Everyone wants to be "that guy" or "that girl" who is always ready to go. After all, most people Ive met in the military have a hard time putting themselves to bed because they have the next watch when its 0200 and we're pulling migrants off rafts or running helicopter ops to interdict $100 million worth of cocaine. During lots of operations you will have a lot of the crew stood up to run things, so balancing crew fatigue with surge operations is no trivial task. The SWO community could take a lot of pointers from the aviation folks - they are way better about making sure people get mandatory rest.