Rating Systems in the Computer Era

I was reading the section on rating systems in Kenneth Harkness' Official Chess Handbook last night. What struck me is that all of the rating systems described, including his and the Elo system with various tweaks, all derive from the pre-computer era when the only usable (objective) date available was win, loss or draw and game history of same. Some suggestions:

A) Following the recent World Cup on the chess24 site, I noticed that at the end of each game there was a computer evaluation of the percentage of correct (I forget the exact term they use) moves made by each player. Some games displayed a wider gap than others between the winner and the loser in the quality of the moves played. 

Perhaps a system that adjusts ratings on the basis of win/loss/draw tempered by the percentage of good bad-bad moves would help. If you have a player rated 400 points higher than an opponent who wins their game, but barely, given the percentage of good moves made by his/her lower rated opponent, the rating adjustment might give the losing player a few rating points for playing an excellent game even though he/she lost. The winning, higher rated opponent would of course receive some points based on winning and the "quality" of the win. 

B) Another thing that I hope will be considered now independent of the above computer percentage of good moves calculation, is how to treat a loss, under the current system, by a player who is rated perhaps 300 points lower than their opponent. Given that there is little chance that a much lower player will defeat the much higher rated player, I suggest that in the event of a loss by the lower rated player, that player should not lose any rating points and the winning higher rated player should not receive any points. In the event of a draw, each player should receive some points based on a formula to be determined or not.