I don't know which way it would go, but I'm interested in your thoughts. ...
Ok...here is a thought.
Let's first make some assumptions about a hypothetical pool universe. (although somewhat unrealistic, it'll be useful)
A1. There exist n+1 pool players in our hypothetical universe.
A2. All games played between players 1 through n are independent 50/50 coin-flip propositions with the break conferring no advantage to anyone.
A3. However, for the n+1 player (we'll call her Jane), when she breaks, she wins 60% of the time. And, when she does not break, she wins 50% of the time.
With these assumptions, we let the pool season begin. We'll have everyone play everyone a bunch of times, record the statistics, and compile them at the end of our hypothetical pool year.
When we analyze those statistics, we will discover some interesting findings.
F1. In aggregate, it will seem as if the break is to everyone's advantage.
F2. With winner break rules, we will see evidence of momentum. (large for Jane, small for everyone else)
F3. When compared with Alternate Break rules, Winner break will seem to confer a big advantage to the better player (Jane).
The explanation for this is quite simple.
By construction the break confers an advantage to Jane. However, since everyone plays Jane during the season, it will seem as if the break benefits everyone. Why? because when you're breaking, Jane isn't. This by itself is sufficient to generate the findings above.
Furthermore, this model can be generalized by relaxing assumption 2, and we can even have several Janes thrown into our pool universe. However, the basic result will still hold.
If the break is consequential in pool, and a few competitors have a significant advantage in breaking, you will find evidence of a break advantage for everyone, and with winner break rules, both momentum, and what will look like an amplification of skill variation.