Im not around enjoy to see behavior.
Then you are limited as to how much you can use behaviors. Thanks for including that. One behavior I will mention, if one is human aggressive he is on my "to go" list. I do not tolerate human aggressiveness.
What are your goals for the male? Decide on the traits you want, then start eliminating any that do not meet your criteria. It sounds like size is one of yours but if you are selecting for meat in the future I'd suggest looking at the size at the time you want to butcher them. To me, that is the size that matters. One of my goals is also meat, I like early maturity for that.
Once you set the traits you are looking for start removing any that do not live up to your desired traits. Other than size that may be feather color or pattern, comb type, or anything else that makes a difference to you. By the time you eliminate the easy ones you may be left with some pretty good choices so the last one or two may not be easy. That's good, they are probably both a good choice.
As Molpet said, you can't go that much by behaviors anyway until they are older. While they are in puberty they can easily be jerks toward the pullets. The pullets are going through puberty also which can make it even more confusing. You may be seeing some of that behavior at 14 weeks, it will probably get worse. By the time they reach more maturity things generally settle down. With my pullets that's generally about the time they start laying. It can vary quite a bit with the boys. I typically have older mature hens in my flock which changes the dynamics with the boys. Typically mine settle down around 7 months of age but I've had some earlier or later. With just pullets and no mature hens I'd think it could be fairly soon after they start laying.
There is a huge problem using behaviors as a criteria anyway. With multiple cockerels their behavior will feed off of each other. One may dominate the others and suppress behaviors. If you remove the dominant one then totally new behaviors may emerge. Or they may compete with each other and make behaviors worse.
My short answer is to determine your criteria and select based on that to the best of your ability. Even if you were there to observe behaviors you can still be left with a problem. Usually not but possibly. I don't know how you can do any better.
Good luck!
