I doubt that 80% figure. Do you have a citation? Furthermore, riding in the roadway is safer than off to the side, which may go against your intuition, but there it is.

Percentage of collision direction is going to be EXTREMELY subject to local riding conditions.

On the route I ride, it would be almost impossible for someone to hit me from the front. I'm riding miles and miles down high speed roads with very rare driveways and intersections.

On routes in cities that involve a lot of pull-outs and left turns, I'd imagine that being hit from the front or being pulled out in front of would be much more common.

There's just no way that a single number can quantify risks for all cyclists. You could probably create a stat that indicated that 80% of all bicycle accidents were caused by protruding tree roots if you sampled the right cyclists.

The driver training program is something that is created outside the company. We used to use "Smith System" but a couple years ago they switched to another program that is accessible to the employees via the internet. Sorry but I forget who or what company produces it. Who knows where they get their figures. (?) I doubt the ( 80% ) number is representative of all jurisdictions. If you live in a rural setting the number is probably quite different. Oh, almost forgot to mention; the 80% number had to do with "motorized traffic". I only mentioned it because like cars, bikes use the roads too. How that 80% number relates to bicycle accidents is anyone's guess. The point I was making though was that motorist often times aren't paying attention to "what's in front of them".

Tom, about that "riding in the road is safer"...( lol ) I'm not sure I can agree with what you said. It might be true when approaching intersections or something like that but on roads with high speed traffic I'm going to listen to my gut that's telling me to ride on that very nicely paved shoulder.

Most traffic statistics are on a state or national level, and that would smooth out the problem of small sizes like the one you're speaking of. You make a valid point which is that a well sampled survey may not apply to an individual's situation, but statistics are useful anyway, especially if, over your lifetime, you are going to be riding in a variety of conditions. I bet most people won't do most of their cycling on divided highways.

My first year of commuting was on a bike that has cheap hub-powered lights with no flash mode.

My 2nd bike didn't come with lights so I got a set of Lezyne USB-charged lights, which do have flash modes on both head and tail. I've noticed drivers hesitating to pull in front of me with this setup, whereas when I've ridden the other bike, they just pull in front like I'm not even there. I generally prefer riding on the street over the sidewalk, but I do have to ride on sidewalks/MUPs that cross streets and parking lot entrances for small segments of my commute, in order to access bridges.

Not sure if there's a difference between slow and fast flash modes.

I have a colleague who doubles up on his lights - two in the front, two in the back.