Points differential has its limitation as a valid parameter.
Just today in my league, two teams were playing for first place in the other conference. Their points differential is not exactly similar, but close enough. The problem was that current leader played against a bot (winning 25-0) and the other played against a very weak team.
It was very prausiible that the other team will win with over 25 points difference and take the lead, and the leader could do nothing about it.
Just a simple example.
Still, as I said in a former message, since this parameter requires simpler algorithm, it makes sense to use it even if it is not the best one, since it is not that important.
Just don't try to convince me that the best parameters are used, especially since in a case of a tie in points differences - you prefer the team with more points.
If one team plays against a bot more times during the season and score 25 points per game, according to this parameter, this team is weaker. Is it really?
And if one plays "run and gun" and the other "push forwards", does it make the first one better just because the general score is higher?
This second parameter is completely worthless.
So if the problem is simplicity, stick with the current system, but if you find a way to treat the 'head to head' parameter and still, keep it simple, please go for it, cause by all means - it is a better parameter.