Sure, no problem.
Say a firearms owner goes out 100 times (shopping whatever). Now statistics say that they will become a victim 10 times out that of 100, for example. However, what if on 20 of those outings, they were targeted by someone who was out to do no good but because that person knew they were carrying a firearm they chose not to engage on 10 of those. Well, they avoided being a victim 10 times. That number isn't recorded anywhere because no event took place. What we do know is that they did become a victim the 10 other times. In this example, a non firearms owner may have been victimized 15 of those 20 times. So they avoided it 5 times. Again, that 5 isn't recorded.
However, the firearm clearly aided the owner in avoiding more confrontations. We just can't nail down this exact number simply because there's no way to prove what didn't happen (proving a negative). But, to ignore that fact biases any conclusion that owning a firearm makes the owner less safe and makes that decision a bit one sided if you go by the victimization rates alone.
When statistics are being used to discuss this situation you are looking at what happens to a large group of people, dividing how often something happens by the number of people you looked at. The experience of different individuals will not all be the same. In the case of of being attacked, if you like in your example you should expect to be attacked 10 out of 100 times while shopping then some people will go shopping 100 times and never be attacked, some will be attacked far more often then 10 times. Say you were flipping a coin, if you flipped a coin 1000 times you should get roughly 500 heads and 500 tails. You have a 50% chance for either result, but it does not mean that it has to go heads,tails,heads,tails all the way to 1000.
For looking at statistics of gun carrying people vs non gun carrying people we consider that if we took 1000 people who carry a gun vs 1000 people who don't. If that is a big enough number then there should be roughly even amounts of people doing things like going shopping (unless you are trying to say that people with guns go shopping more often than people without guns). If the group of 1000 people each went shopping roughly the same amount of times at similar kinds of places then each group should have ran into about the same amount of people looking to attack someone while shopping. If having the gun made more of those potential attackers reconsider attacking that person then overall we should be able to notice a difference between how often a gun owner is attacked vs a non gun owner, so the average rate of being attacked between the two should be different.
Going back to your example, if the average is being attacked 10 out of 100 times, that also relates to a average number of people considering to attack you. In your example you say that person A, the gun owner had 20 people look at attacking him and only 10 attack, you have added another variable, the amount of times a bad guy is around to consider attacking you. If 1000 gun owners go shopping vs 1000 non gun owners why would there be more people looking to attack someone when the gun owners go shopping? So the second person in your example (the one without the gun) must also had 20 people consider attacking him. If the studies say that there is no difference in the average number of you times you are attacked regardless of if you carry a gun or not, then the gun makes no difference in how often those 20 people who considered attacking you did actually attack you.
You can dispute that the studies eskimospy is referring to did not account for some difference between the average gun owner vs the average non gun owner (if for instance gun owners went shopping twice as often than those without guns it would be unfair to compare the two based on how often a day they are attacked instead of based on how often per going shopping they are attacked). You would have to point out something specific though if you wanted to do that. The principles behind averaging are very sound and is a very valuable tool, trying to discredit those would require far more work than you could fit into a post here.
Facebook
Twitter