This is comparing murder to every instance of police killing someone, justified or not. Police murdering someone is much more rare. The ratio you are talking about probably is in fact well over 100:1.
You add up all those people and the ratio is about 100:1.
Now consider we have multiple data points of police committing murder and getting away with it. How can we argue that it's rare, given that we have several instances of police getting away with it and not being considered part of said recorded statistics?
Where we draw the line is debatable, but the person you were responding to wasn't really attempting to draw a line.