This is exactly how we do it. Simple to assess. Much less subjective than thinking about likelihood. Although proper risk thinking is important, if you ignore the things that are unlikely, you end up with Texas City or Deepwater Horizon (depending on the nature of your business of course).

Hi Sarah

Our industry tends to look at High Impact Low Probability (HILP) events, of which there is a wealth of literature. For high hazard industries HILP focuses on catastrophic events (multiple fatalities).

You are on the right track - I'm reassured by the fact that you have used the word realistic as in my experience HILP pushback conversations tend be dramatic and unrealistic (such as Godzilla attacks etc) or of a magnitude that would never be able to be handled by any in-house contingency planning (complete failure of regional electricity infrastructure).

How are you assessing these risks if you do not take likelihood into account?
if one of your assessed risks is the a fatal vehicle accident, how do you treat the potential consequence, given that any accident may result in a fatality?

By not including likelihood in the equation, you end up with things such as crossing the road or driving a car as a high potential event, which isn't very useful.

Depends on the lingo you use. I am busy with a project on Critical Risk Management and setting thresholds that are acceptable and not acceptable

Sarah started this conversation with a question about "events", not risk. If we must discuss risk, use the ISO definition of risk: "effect of uncertainty on objectives". Ask how uncertain someone is. Or how certain they are. Or use a simple word, how sure are you? This does wonders for nailing down how uncertain people are - how little they actually know.
Learn to use probability. Ask for a consensus view about ranges. "Are you 90-95% sure this tank will fail in the next 10 years?"
Gather data about similar events of concern. Why did they happen? What's different.
Don't use likelihood words. They mean whatever a reader wants them to mean. Or as Lewis Carroll wrote:
“When I use a word”, Humpty Dumpty said, in rather a scornful tone, “it means just what I choose it to mean – neither more nor less.” (p 196, Through the Looking-Glass).

Gather data about similar events of concern. Why did they happen? What's different.
Don't use likelihood words. They mean whatever a reader wants them to mean. — Chris Peace

I have always liked the use of likelihood descriptors if using a risk matrix approach - in the past for construction projects I have found the type of "a similar event has occurred within this; plant, site, company, industry, etc." to be much more consistent than the "highly likely - highly unlikely" scale... especially for the poor person that has to read the assessment/report in the future.

Regarding the OP, I always think "why are we trying to classify incidents like this" - not in a negative way, but thinking about what the purpose of the categories will be and how they will be used.

Often in the case of HiPo events (or critical risk) it is a shorthand for evaluating the amount of effort spent on investigating the event - and on the surface it makes sense, if you only have the time to investigate one event you'd choose the one that could have caused the most harm, right?

The issue is that the amount of learning an incident can provide isn't directly linked to the potential severity of that event - you will likely learn less from a missing scaffold handrail on a 10m scaffold than you will from a badly managed design change in a process plant that resulted in discharging hot (clean) cooling water straight into the stormwater drain.

The catch 22 is how do you know what is worthwhile to investigate before you investigate it??...

And to answer the OP - from memory of the last company I worked in that used a HiPo classification was essentially anything above an LTI or a Notifiable Event, but that company had very little risk of their staff being involved in a multiple fatality type event so could have skewed their classification lower so that there was stuff to investigate for continuous improvement.