MischiefSC, on 31 August 2014 - 12:01 AM, said:
Honestly?
I'd be happy with an official announcement that sync dropping is breaking the rules and can be reported. Even if it's only investigated in egregious examples that's fine.
Discouragement is fine. It really comes down to 'what is the point of the queue'. If it's solo players... then sync dropping is against the rules there. That's exactly the whole reason why the group/solo queues got created - to deal with sync dropping. Phases 3 and 4 are exactly what's being discussed as a way to stop sync dropping and groups rolling against pugs.
We're now in phase 4. I'm all for their being more important issues right now but a non-response is the worst possible answer. If the answer is 'Yeah, it happens and we're okay with it' then fine - You can get a decent estimate of relative Elo scores and with a bit of trial and error get the detail of drop order to best sync with in an evening and actually do so effectively. If it's not okay... then say it's not okay. Will people still do it? Sure but only people who are legit cheats. The current approach of yes/no/maybe is bad.
You got the first part of the math correct. For 1 player there needs to be 69 other unique players to play 3 matches with out seeing the same person twice. However you need to carry that out to all 24 players.
Match one 24 players
Match 2 for each player no repeats requires 24x24=576 players dropping at around the same time. With approximately the same ELO and proper mech classes.
Match 3 for each of those 576 you need to make sure that they have not played with anyone else in their previous 3 matches 576*24=13,824. For no repeats from the last match and again for the match before =331,776 players playing at around the same time. Now it is actually less, then this but that is above my math to get the exact number. However I'm just trying to get the idea across of how many players would be required and how much data and processing would be involved in a matchmaking process that already has problems creating matches often enough.