Here's a reality that is virtually the same concept.
I used to keep times for a 10K run here. There are thousands of racers. Let's look at predictions from two points of view. The first is what is the gender of the first place runner going to be? I can virtually guarantee it will be a male. The first three will be male. PROBABLY the first 10. I will be right 99.9999% of the time basing this prediction on gender.
But lets try another prediction. Let's say I pick two of the runners at random and try to predict which will come in before the other based ONLY on gender. If I pick the male, I'll be right about 50.1% of the time. If I pick the female, I'll be right about 49.9% of the time. Why? Because there are other factors. So how useful is the assumption that a given male will beat a given female? Not very.
So, if I want to be the winner of a benchrest contest, it would behoove me to shoot a bolt action because it is more likely to be in the top .0001% than a semi-auto. On the other hand if I'm picking a gun off the rack and assume that it will be accurate simply because it is a bolt action, I'll only be right about 52% of the time. If I assume it's accurate because it's a semi-auto, I'll be right 48% of the time.
I'm honestly not trying to argue with you and am probably saying what you intend to say, however the point is if you want top of the line, tack driving accuracy, you're much more likely to be able to get that out of a bolt action. If you want an accurate off the shelf gun, the odds are only a little better with the bolt action. -Michael Bush