The pin hole is either counter sunk or reamed with a tapered reamer (the latter being the better of the two options imo). Because you have a taper that travels most or all of the distance of the hole in your scale, you will end up with an hour glass shaped pin that will not come out. If you just peen without tapering then the whole pin is swelled until it will not move in the hole, this pin can be removed a bit easier.
If you twist a full tang knife from the tip of the blade and the end of the handle, you will get a certain amount of flex in the tang. There's just no way around that, unless you're making all of your knives with very short, thick tangs. A real world example of how this force could be applied to a regular use knife is when someone is using the blade to pierce and crack a rib cage, not at all unheard of in the hunting world. Every deer, elk or whatever that gets skinned with the knife will work on that epoxy bond until the only thing left holding the knife together is your pins.
Meanwhile, you've got three different materials (handle, pin and tang) that all expand and contract at different rates in the cold winter deer season, as well as react differently to humidity changes and exposure to fluids. As these are expanding and contracting they are breaking down and weakening your pins' glue bond and creating a slip-fit pin.
A really easy way to see how much different metals will expand and contract is to take an old Canadian "toonie" coin and put it in the freezer over night. You can pop the center right out of one while it's cold. We can then easily apply this to brass/steel or even different types of steel in the case of 300 and 400 series pins and air hardening or carbon knife steels. In the case of the glue bond, once it's broken, it's gone, whereas the pins shrinking or swelling is just a temporary change.
Here's something to consider - I met a collector last year that had a modest Buster Warenski collection. One of the items he had was a set of Yin-Yang nunchaku that he had commissioned. He told the story of how he'd ordered the piece which took several years to complete as it was more of a friendly commission. One day while talking with him, Mr. Warenski explained the problem he was having interlocking all of the parts that made the Yin-Yang end caps and mechanically attaching them to the ivory shafts. The customer said "Just glue them, it's never coming out of my display case". His response was that 30 years after they were both dead and gone, someone might come by and pick up those chucks and start whirling them around. The last thing he wanted was an old glue bond to fail and have his reputation ruined years after he was in his grave, because who would be left to explain that they were never meant to be used?
The anecdotal experience of taking an epoxied knife apart doesn't account for test of time and temperature change, nor does one or two knives constitute a valid data set. I've not seen any good data out there that shows modern adhesives providing bonds significantly superior to mechanical ones, with or without including the time, temperature and flex variables we're talking about.
Reaming and swelling your pins takes just a few extra minutes to provide a lifetime of assurance. Of all the time consuming steps we take to make a knife, this is one of the easiest and well worth the time to me.
Anyway, I do want my knives to be the best that they can be. I've destructfully tested the ones that I've made that were just epoxied up with non-peened pins and they held up very well. The blade should break before your handle does. 
