What is the total population, what is the violent crime rate, what is the prevailing attitude of civilians towards police, how many police encounters are there, how well does the government record these statistics, etc...

You can't be shot by a police officer if you don't encounter one. In the US, most police officers know if you have a criminal record, a warrant, or LTC before they even approach your car. If your culture thinks arguing with cops is a good idea, your culture also might be more prone to violent clashes with police.

All those things factor into the individual situation, thus I would expect these would be represented in aggregate figures as well.

The problem is that none of these risk factors are considered and thus produce meaningless statistics.

As a surfer I see this same BS with shark attack stats. Most of them go like this: There were 150 shark attacks the the US last year and there's 300M people in the US, so your chances of getting attacked are 150/300,000,000.

I hope you see the obvious flaw in that. Some one in Arkansas isn't at risk. So a better number would be everyone who visits the ocean. But now you're counting people who probably only dipped a toe in the water or stood in waist deep water for 15 minutes.

The real people at risk are swimmers and surfers and to get real numbers you don't just need to know how many people are in the water, but also how much time they spend there. And more than that, you need to know what time of the day it is.

When you're done, the final numbers will be a table, not one neat number. Each number in the table will represent the time of day and be expressed as a probability per unit of time spent in the water.