This is a popular enough theory that it's been tested multiple times. In the successful cases, the rival league still ended up trying to poach the top players and ended up merging with the other top league. In the less successful cases, the league went out of business.
In software, that's not the way it works. Or at least, the competitive business aspect of it is largely out of engineer's hands. Few businesses ultimately succeed because they wrote better code, IMO. Though interestingly, a bunch of startups only get big in the first place by having great code that allows them great speed/flexibility/etc. Stripe itself got big originally because its API was really great, which I'd attribute more to product than to engineers, except in this case it's kind of a merger of the two.
"March Madness" is an anomaly, a single-elimination tournament with 48 games in one weekend with a history of upsets and buzzer-beaters. Additionally, college basketball and football in the US benefit greatly from the local and alumni network affinity for the programs, that doesn't exist with, say, minor league baseball or the NBA D-League.
The AFL was actually relatively successful since it forced a merger with the NFL. Meanwhile, at least according to a 30 for 30 documentary, the original USFL had a potentially viable niche as a second-tier spring league for awhile until they made the bizarre decision to switch to the fall football season and try to force an NFL merger.
I genuinely believe that a league made of the 500-1000 rank players would be just as (or at least almost as) entertaining as one with the top 500 players. Relative rank is what matters and it's why NCAA March Madness is so popular despite being made up of relatively low ranked players.