On the other han, euro crash tests are absolute (as far as I know). You can see the decrease in safety in smaller car models, down to which part of the body is getting crashed in which type of collision in a smaller car, but is intact in a larger car.

Euro crash tests include large weighted components of safety to others outside the car.

So it's totally possible for a car more safe then another for its passengers, in all conditions, to be nonetheless ranked lower because of lower pedestrian safety or something.

You can also check out how the car does in specific tests. So, if you want to ignore risk to pedestrians, you can.