"Treat all police who shoot as suspects, says police union"

That is quite a headline. Police union capo Gerrit van der Kamp explained to a local news outlet:

"It is not the ideal solution but it is necessary to guarantee their safety and legal rights."

The police union apparatchik quoted above is based in the Netherlands. To this US blogger, who spends so much time trying to uncover when the police open fire at those they are bound to serve and protect, such a quote might as well have come from Mars. Things are very different in the Netherlands then they are in America. For one thing, police unions here routinely defend even the most outrageous police shootings of unarmed citizens. For another, Dutch police are required by law to inform the public when they shoot someone. In 2011, cops in the Netherlands killed a mere 5 people.

But its apples to oranges, you say! There's not even 17 million people in the Netherlands. There are more Americans at a Taylor Swift concert than there are in the Netherlands. Fair enough, let's look at Germany - they most populous country in the EU with a population of over 80 million. That is well over 1/4 of the population of the US, and, well, Germany is not known for flaunting law and order. So how many people did Germany's constabulary kill in 2011? Six people.

So the annual number of police killings here in the US should be somewhere in the neighborhood of 25. Right? Right?

During the same time period, the mean number of citizens killed by police officers was 1,150. Which means your chances of being shot by a cop are about 50X higher in the US than they are in Germany and 12X higher than in the Netherlands.

What about puppies? Well, in Germany they not only have to account for every killing but for every bullet fired. And 8812 of those bullets were for animals (compared to 85 for people). We don't yet track foreign puppycides. But 8812 bullets sounds like a lot! After all the entire public puppycide database only includes ~2500 killings, and that is for more than one year. The news is worse than it sounds.

Our database is just a sampling of a much larger number of killings; a sampling that is slow to grow because of our current lack of resources (we need your help! volunteer already!). We will be publishing a formal statistical analysis of our sample. For the time being, though, chew on this last tidbit of information. According to the Bureau of Justice Statistics, in 2013 there were 12,000 police departments operating in the United States. If we look at a mid-sized department that we have a complete subset of records for, the Palm Beach Sheriffs Department, in 2012 PBSO deputies shot 37 dogs and 2 cats. Some years some departments kill more than Palm Beach - Houston PD was averaging about 42 dogs during the same period. Other departments, like the NYPD and LAPD, both famed for the games they play with the violent crime statistics that make and break the careers of top cops, report numbers of dog killings that are so low they are almost certainly faked. Then there are a few small departments, like Davie Florida and El Cajon California, that simply appear to rarely kill pets: El Cajon has recorded one dog shooting in the last four years, while Davie one went six years without shooting a single pooch.

Calculating an average from these disparate sources is not a simple task. Where is the pattern among the 12,000 departments, with over 1.2 million officers, when behavior ranges from the openly violent to the dishonest to the prudent and responsible?

Consider how drastic the changes are in small departmental fluctuations. Palm Beach killed 39 pets; we do utilitarian calculus in our head and calm ourselves with the fetid justification that such a number is manageable. But if every department were to commit to the same treatment of animals as Palm Beach - if in one year every police department were to kill 39 animals - police would have a death toll over close to half a million. Half a million kids losing a best friend, half a million parents hosing blood off the drive way, half a million families explaining bullet holes in the side of their house. Geometries of scale can transform acceptable casualties into genocide. The problem appears bigger with each new record we add to our database.