For me, it is simple. I've had a risk management component to my job for over a decade, and spent some of that time as a certified risk assessor, so I analyze personal risk profiles the same way and using the same tools. I coach my defensive pistol and concealed carry students the same toolkit.
There are a million cliche's for this kind of discussion - "if you don't monitor it, you can't measure it, if you don't measure it, you can't manage it," or "if you don't monitor it, you can't mitigate it" - it really does start with an assessment of your personal exposure profile. As an example - I travel all over the country, go out for drinks with customers/clients/vendors where I can't be armed, well dressed, often paying in cash, spend a lot of time in major markets, stay in hotels between a hundred and hundred and fifty nights per year. My greatest exposure is assault out of the home. My wife, on the other hand, is a stay-at-home mother, only leaves our hometown a few times a month, typically into a SMALLER market or rural area visiting family, her greatest risks are home invasion (especially considering I'm gone so much) and getting assaulted at Walmart/Target/Dillons during a grocery run.
Once your exposure profile is nailed down, then tactical/operational level planning can happen (not tacticool, think Strategic, Logistic, and Tactics/operations in terms of levels of management). For inside the home, it's pretty straight forward to work out your floorplan, points of ingress, pinch points and traps, especially areas where invasion could happen without immediate alarm (ie. someone enters through a garage window at the back of the house, too far away from the living room to be readily heard during entry). It's also pretty simple to identify your exposure outside the home during routine or common activities. Once this skill is developed, it's not a challenge to adapt to new activities on the fly - for example - on a once in a lifetime vacation opportunity.
When considering which attacks are relevant for training and practice, a person needs to consider likelihood, balanced against severity of consequence. For example, keeping a loaded rifle at the ready in your car might prepare you for a mass gunman attacking you during gridlock traffic, but it puts you at a higher risk of a child finding your firearm and having an ND injury or death - and the latter is unfortunately much more common than the former. Grant Cunningham explains this risk management triage with the 3 P's: Possibility, Plausibility, and Probability. A great many things are possible, fewer still are plausible, and a very, very small number of things are probable.
Don't confuse, however, the Possible/Plausible/Probable funnel into the absurdist strawman some folks have tried to make it out to be. When something is "probable" in the context of an attack, it is NOT probable on any given day, for any given person. But rather if you consider the context to be: "If you are attacked... then it will probably in this manner..." This produces a more narrow spectrum of statistics - it doesn't play into "your odds of being attacked are very low, so you don't need any protection or preventative planning." It plays into - something bad IS happening, and it's most likely in this form. Your odds of having a car wreck are very low, but in the event of a life threatening car wreck, it is most probable to be at speeds over the average speed of other motorists on the road, so ways to reduce your risk are to wear your seat belt, have air bags, and to keep your speed in line with that of the traffic flow. Similarly, your odds of being in a car wreck within a mile of your home or work are higher than any other part of the average driver's mileage, especially during high traffic volumes, so putting on your seatbelt as soon as you sit down, avoiding momentary distractions like getting off of the phone, punching your route into your phone or GPS before you start moving, finishing that breakfast bagel before embarking, running for groceries outside of rush hour, etc, will help minimize your risk in those critical miles. It is POSSIBLE your particular fatal car accident will involve an escaped elephant running across the highway, but it's hardly plausible, and infinitesimally not probable - so I wouldn't worry too much about putting an elephant proof grill guard on your car.