To illustrate, let's pretend all of the non-fatal data is obtainable to see how that might help determine if a cyclist is more likely to be killed from behind or elsewhere.
Assume:
- 10,000 car-bike collisions (all types) per year.
- 700 result in death.
- 10 are hit-from-behind during the day, and all 10 result in death.
- Therefore 9,990 (10,000 - 10) are hit-from-elsewhere, and 690 (700 - 10) of the 9,990 hits-from-elsewhere result in death.
Now, let's determine whether likelihood of death from hit-from-behind or hit-from-elsewhere is higher. What data do we need to look at?
According to my argument, all we need to know is:
- Number of deaths from hit-from-behind (10).
- Number of deaths from hit-from-elsewhere (690).
From this data alone we can conclude that a given cyclist is 69 times more likely to die from hit-from-elsewhere then from hit-from-behind. None of the data about non-fatal deaths is required to determine that a given cyclist is much more likely to die from hit-from-elsewhere than from hit-from-behind.
According to Gene, we also need to know "how many people have been hit from 'elsewhere' and simply walked away". Well, we do know this (in this hypothetical example). Of the 9,990 where the cyclist was hit-from-elsewhere, 690 resulted in death, so 9300 walked. How does that help us?
Now, in order to determine the likelihood of death
if you get hit from elsewhere, you do need those non-fatal numbers. That would be 690 / 9990, or .069 (6.9%). And according to our statistics, the likelihood of death
if you are hit from behind is 10 / 10, or 1.0 (100%). OK, so what? The likelihood of death,
if you are hit from behind, is much higher than the likelihood of death
if you are hit from elsewhere. OK, but that does not tell us which type of hit is more likely to kill you.
How does this additional non-fatal data help us determine whether a given cyclist is more likely to be killed from behind or elsewhere?
It doesn't at all. It tells us squat about that, because we already have all the data we need by just looking at total deaths from each type.
Does this make sense?
Do you recognize that you are more likely to die from a car collision than from a gunshot to the head? Now, if you are shot in the head the likelihood of death is very high, and the likelihood of death if you are in a car collision is significantly lower. And we do need the non-fatal statistics to determine what those likelihoods are. But in order to determine whether you are more likely to die from a car collision or from a gunshot to the head, all we have to do is look at the
fatality statistics for car collisions and gunshots to the head. The number of people who die from car collisions is much higher than the number of people who die from gunshots to the head. We therefore know that you are much more likely to die from a car collision than from a car collision, and the number of people who away from gunshot woulds and car collisions is totally irrelevant to this.
Now do you understand why we only need death statistics to determine whether a cyclist is more likely to die from a hit-from-behind in daylight or a hit-from-elsewhere?
Now do you understand why we know that a cyclist is much more likely to die from a hit-from-elsewhere than from a hit-from-behind, even though we don't have the non-fatal data?
If not, please reread this post a couple of times, and really give it some thought. Until your subsequent posts indicate that you have given this serious consideration, and you actually address the points made (or finally concede, of course), I'm not going to continue.