"Velocity Change and Fatality Risk in a Crash–A Rule of Thumb" by Joksch. When collisions result in a Delta-V of 65mph, collisions are on the order of 70% fatal.

The fitted curve predicts the fatality rate to be ((delta-v / 71) ^ 4), suggesting a 100% death rate at 71mph.

Dropping down to 65mph drops your fatalities to 70%. Dropping down another 5mph to 60mph drops you to 50% fatality rate. The 5mph after that is a 36% chance of fatality.

So the 65mph limit is a rather magical number in terms of surviving a worst-case scenario. Going above that, fatalities go up dramatically. Going below that... fatalities are significantly lower.

it would also imply that a collision with someone where both of you are going 32.5 miles per hour and hit each other head on would end in near 100% fatality.

That said, I would believe that a head-on collision at around those speeds has a high chance of fatality being that it's a 130mph collision.

Mythbusters did an entire episode on this.

They crashed a 100mph car into a wall, a 50mph car into a wall, and then two 50mph cars into each other.

I suggest watching the videos to visually see the proof. The 50mph vs wall is similar to two 50mph cars crashing into each other.

32mph vs 32mph car would have a very low chance of fatality, although the two cars would probably be totaled. When I say ~65mph as a problem, that is the 65mph vs 65mph cars hitting each other head on... or 65mph vs a tree / wall / bridge (which has an equivalent chance of death)

Remember, the kinetic energy of you and your car goes up with the square of your speed.

You made me think for a few more moments on the issue.

The reason 2 cars hitting each other is similar is more due to their mutual crumple zones, so they're helping each other slow down at the same time.

I suspect this would also be dependent upon the car, since newer cars have better crumple zones to help survive accidents better. There's a lot of details that go into actual damage. Most people, when they're in an accident, don't go headlong into a wall. They hit a pothole and spin out, or nudge another car and go spinning, or hit a side wall at an angle, where their speed toward the wall isn't actually 65 mph, and so on.

At the end of the day, it's not that 65mph is the near-death number, as far as car speed is concerned, it's more, as everyone else has been saying, the delta-v. The easiest way to hit 65mph in that way is hitting a tree head-on, or wall head-on. On the freeways, this isn't likely the case, and even the edges of roads where two roads split and there's a barricade are protected by several crunch zones in the road itself.

Edit: ah, I didn't notice that you had been saying delta-v as well. Either way, I still think I follow what you're saying now

I think he's talking about 65mph delta-V, referring to the change in velocity at the (nearly instantaneous) time of collision. Two cars rubbing like that won't produce a 65mph delta-V in the way that a brick wall will. It might be 5-10 mph, which will be enough to shock people but not fatal. He's right that 65 mph delta-V is almost uniformly fatal.

Likewise, if you jump out of a moving vehicle at 65mph, you'll get banged up badly but might not die, because you don't lose 65mph of speed the first time you hit the ground, since you're hitting it obliquely. On the other hand, if you jump from enough height to reach 65mph and hit the ground directly, you're almost certainly dead.

Right, and most crashes on the highway aren't fatal, and aren't in to immovable objects. So, I kind of don't the point of the comment.

Surely sometimes they are in to immovable objects, of course, and those ones are pretty bad.

Even worse are side impacts on trees, electrical poles and light posts.