Posting this idea as I think it would be a simple but effective solution to Elo manipulation in team games: calculate each individual’s Elo change based on average of opposing teams. This way if you partner someone of significantly worse Elo, then you won’t have any increase in yours when you win. Then normalise results to ensure that Elo gained / lost in a match is correct (average 15 per player, for example).
Worked example:
Team 1:
Player A, Elo 1500
Player B, Elo 1500
Team 2:
Player C, Elo 1000
Player D, Elo 2000
If Team 1 wins, news Elos will be:
Player A (1515), Player B (1515), Player C (1000), Player D (1970).
If Team 2 wins, news Elos will be:
Player A (1485), Player B (1485), Player C (1030), Player D (2000).
Any downsides to this? This won't "solve" smurfing, but I think it will completely stop the ridiculous TG Elo situation we have going on at the moment.
Worked example:
Team 1:
Player A, Elo 1500
Player B, Elo 1500
Team 2:
Player C, Elo 1000
Player D, Elo 2000
If Team 1 wins, news Elos will be:
Player A (1515), Player B (1515), Player C (1000), Player D (1970).
If Team 2 wins, news Elos will be:
Player A (1485), Player B (1485), Player C (1030), Player D (2000).
Any downsides to this? This won't "solve" smurfing, but I think it will completely stop the ridiculous TG Elo situation we have going on at the moment.