Matchmaking poll

[dweiler]

Since it is less than a month until next season, I think it is time to see how we should move forward.

This poll is "only" about random teams vs balanced teams. In the case that random team is chosen, there will be a new poll about the Elo/points subject.

[Don_Killuminati]

maybe put a 3rd option ;
random teams with 2highers elo on both teams.

[TinSoldier]

another option, added onto Don would be captains mode for all 4's with 2 high elos as captain

[Jamo]

No idea what captains mode is, tbh
I like random teams, and I just noticed, that yellow and red are not highest elos anymore? That true? I think that was nice and should be established again and also in the vote, like Don already suggested!

[FadingSuns]

Random teams rocks, maybe the elo issue should be fixed (i dont care about the 15/15 but i understand the points of the ones who care) but definitely the random teams are OP.

I really think that captain mode is ok when there is a challenge etc, but those games of 29/1 are always nice to play (and to win), I dont really see it as an option for every game...

[pewpew lasergun]

yes random teams is nice, just adjust/scale a bit the elo

[dweiler]

maybe put a 3rd option ;
random teams with 2highers elo on both teams.

yes random teams is nice, just adjust/scale a bit the elo

Those options will certainly be part in the second poll if random teams are chosen

[nabo.]

If you guys prefer +/- 1 or 10 points per w/l games instead of the current 1000 elo base +/- 15 elo....we can work it out.

[Dong]

If you guys prefer +/- 1 or 10 points per w/l games instead of the current 1000 elo base +/- 15 elo....we can work it out.

From what i can see, the thing the community wants most, is scaling elo win/loss, with limits, so the elo won or loss compares more to the chances you have you to win or lose. But not like it was before, where games could be 29.70/0.30 etc.. but with limits like, say, 20/10 something like that, the details will be worked on later i suppose

Dong-

[nabo.]

Thing is, elo has never been a good accurate measure of skill nor your overall performance.

What do you mean "shuffle certain times"? It just randomly shuffles the players. Every permutation is equally likely to be selected.

Current random setting makes it so the likelihood of you playing with certain person is same with another and the number of games played with another would be same IF enough number of games are played or the more games are played.

There are many factors that are not considered with the previous elo such as teamwork, communication, leadership, etc. If you won the game, you won it as a team. You were part of the team, thus you contributed on your win. Assuming this is the case for all games on average, scaling isnt easy. You all know who are the stronger/weaker players. If your rating is higher than another because you played more games and you won more game...thats great and I think the points you got is deserving. If you are complaining about one or few particular imbalanced team games (considering your own judgement of each player's skill level), that is an issue because not enough number of games were played.

Current system is not elo which attempted to represent your skill level (which it never did!)...it is just a point system relative to your win rate.

If you guys want to have a captain game or want to label certain slots as "game caller" or captain slots, we can do that and make it so these people get few more/less elo than others in the team depending on win/loss. This is only if you guys want to decide on one person calling the final shots in a game on each team. If the calls will be more based on team discussion and team decision, prob not needed.

[Dong]

personally, i just like the random teams.

[pewpew lasergun]

all systems are good, i think most of us enjoyed this season system. and either elo or no elo both are fine.

But except very few exceptions, i think the previous elo system is very reflective of skill.
if u check the !top and !bot of all seasons, those would be in fact the best and worst players of lihl.

if elo system is to be kept maybe implement a function like this with random teams:
win = 15 elo points
loss = 15 elo points

if Team West elo points >= 25 then Team West {win= 20 and loss= 10 elo points}
and vice versa

so basically its always random teams with 15/15, except when elo is 25+ on 1 team then it becomes 20/10

[ILOCOS_NORTE]

For those who dont know,
the latest Elo formula had a smooth logarithmic cap for ELO w/l (for LIHL only).
https://entgaming.net/forum/viewtopic.php?f=207&t=57531&hilit=ilocos&start=30#p233432

For example:
if Elo difference was 0, games were 15:15
if Elo difference was 100, games were around 17:13
if Elo difference was 250, games were around 20,5:9,5
if Elo difference was 500, games were around 22:8
if Elo difference was 1000, games were around 24:6
if Elo difference was 10000, games were around 26,5:3,5

so, techniqually there were no such games as 28:2
I just put the smooth cap to prevent retards from grind 2k+ Elo

Most ppl didnt realize this cuz the "!scores" function did not show the real Elo w/l and uak made it clear not changing the !scores function for LIHL only.

So if you ask for a cap, we had this already ...

[Diablo_]

Yes, at least I know that. That function was imo ok for auto balanced games but for random teams it is still too extreme and should be about 20% flatter.

[dweiler]

Random teams won, a new poll will be made how the random team should look like.