they are nothing crazy just small ideas that would improve LD so much.
1. For random play divide the teams based on the top 2 highest dmg dealers or most defends. I say the most defends because that is a good indicator of a good player for captures you can simply just stand there.
2. The ability to do Guild Lockdown with 4 players(like random play), only being able to play with 6 players exclusively is outrageous.
I think OOO needs to implement some meaningful player rankings before #1 could be effective. Unfortunately implementing good player rankings for LD involves working with statistics that I'm pretty sure are beyond the capabilities of any of OOO's current staff (judging by the hilariously bad implementation they have currently).
Ideally the player ranking should represent how influential a given player is to their team's victory. That is, assuming the rest of the team are equally matched with each other, a player with a higher ranking should be observed to win more often than not.
A basic sketch for how the ranking should work:
Players all start off with a neutral ranking score at the beginning of the ranking period.
Upon forming a match, the system applies a prediction function on the game based on the ranking score composition of each team. This function gives a probability of team 1 achieving victory over team 2 (or vv).
Upon completion of the match, the actual result is compared to the predicted result, compensated for players who quit or were excluded for idling, and rank points are distributed based on how unexpected the actual result was. For example, if Team 1 was predicted to win over Team 2 with a high degree of certainty, but team 2 won, then team 1/2 would lose/gain a large number of rank points respectively; if the prediction had a low certainty, only a moderate number of points would be awarded/lost, and if team 1 won when predicted with a high certainty then only a few points would be awarded/lost.
Unfortunately it's not possible for me to give exact numbers on how many points to award, or what the prediction function should look like, given that I don't have access to LD win/loss records.
Anyway, once this sort of ranking system is implemented, it would be much more viable to use it to match teams more evenly, or even use it to hand out awards for top ranking players. (kind of like the old brawl crown)
That makes sense.
Let me shorten it, to see if I understand;
It's based off win/loss, BUT;
It can see other peoples ranking, and predict what team has the edge
If a team has the edge AND wins, they get a few points, becuase it was wrestler killing a midget (psh, easy, I could kill that guy id I wanted)
If a team is weaker AND wins, they get allot more points, becuase its like a midget killing a wrestler (WHOA, HE'S GOT SKILLS, HE KILLED A WRESTLER)
And vice versa for if you lose
ONE PROBLEM;
People who play more will have more points. Though that's pretty fixable.
@Orangeo People who play more will have more points only if they win more than they lose, which is sort of the point. But further play will have diminishing returns as the predicted outcomes converge to the actual outcomes.
I vote down number 1, but up number 2. Teambalance does need tewaking, but captures are certianaly more important than damage. I'd just base it off of equipment, to some extent. A little randomness is good too though. The 6 player only guild lockdown is kinda lame though, I'd like to do 4 sometimes.