Your argument would apply if we were trying to ascertain a fairly accurate estimate of probability.
The fact is that being hit from behind during the day is so rare, the odds of it happening to anyone, regardless of their riding style/behavior, or any realistic/reasonable factors, are negligible.
In comparison, the incidence of being hit from elsewhere is so much higher that it is obvious that anyone's odds of behing hit from elsewhere are much higher than being hit from behind.
The irony is that the better you get at defending against hits-from-elsewhere, the lower becomes the difference between the two probabilities. But the hit-from-behind possibility remains fairly constant no matter what you do, short of not getting on the bike in the first place.
Why worry about something over which you have no practical or reasonable control, and the probability of which is neglible anyway?