Would you swap 3/3/3/3 class limits in the group queue for filling up all the class buckets before you can chose another of the same class.
- 1
- 1/1
- 1/1/1
- 1/1/1/1
- 2/1/1/1
- 2/2/1/1
- 2/2/1/1
- 2/2/2/1
- 2/2/2/2
- 3/2/2/2
- etc
FULL POST
For those not on twitter there has been some discussions there (as much as you can have in 140 words a post ..) about how to help the matchmaker and group queue dynamics.
Right now it is very easy to break the group queue especially in off peak times when there are less choices to put together teams. While groups need to conform to the 3/3/3/3 rule, this means larger groups tend to come in relatively balanced (apart from always bringing Daishi's and Madcats) but smaller groups cause the main issue.
When you have a group of 4 or so people tend to ton up to assaults and heavies often because they feel they need to have that tonnage advantage to help 'carry' their team as they simply do not know the quality of the other groups they will be matched with. People might also ton up simply because they want to output more damage too or are scared of the frailty of meds and lights just like the solo queue.
This means there is a matchmaker CLOGGED with smaller groups all running bigger mechs leading to matches that either are even but incredibly top heavy - or completely unbalanced tonnage wise.
If i can find my image i posted for proof i will but this may have been started by posting an image where my team had 1 assault i think and the enemy had 5 or 6 nearly all Daishi's. It does not matter if the other team had a lower Elo.
This happened 4-5 games in a row where the MM would break and just chuck groups together and for some reason my team always ended up undertonned. We only won one match where we were undertonned due to enemy bunching and being artied enough to scrap a victory through.
So that is the problem - and an example. Maybe you have seen the same?
The proposed solution:
Group would have to take one of every tonnage class before they can choose another of the same class.
So:
- A group of 2 would need to take one of any class, and one of another class
- A group of 3 would need to take three different classes of any type
- A group of 4 would be 1/1/1/1
- A group of 5 would be 2/1/1/1 where the 2 can be any class
- A group of 6 would be 2/2/1/1
- and so on.
As Russ stated it would be like putting a jigsaw together.
PROs:
- Shorter wait times
- More even class/tonnage
- Less strain on the MM overall
- More diversity
- More restrictions on what people can play
- More fiddling around when making your group drop deck
- No solution to still bringing the heaviest in each weight class
Alternative solution - Tonnage limits
This will probably not go over well but its worth considering to address the point of bringing the heaviest in each weight class.
What if we had a tonnage limit for group queues? I personally think it would not work well but its worth discussing at least.
If you put in a tonnage limit of 60T per person then you might solve the not bringing the heaviest in each weight class.
This would match tonnage exactly, but it comes with natural downsides
- Less choice in your team, someone takes a Daishi, you have to take a locust no 2 ways about it.
- LOTS more fiddling around to fit in the tonnage limit before a drop
- Lots of arguments about who takes what etc
The challenge of putting this into the current group queue would be immense because of the nature of F2P and so forth, however i would be interested to see if anyone had a solution to using tonnage limits for the group queue as to not be too fiddly or exclusive.
---
So what do we thing group players?
Personally i would love to see the modified class limits done ASAP. Why? Because i want matches to be close to tonnage balances with a range of different mechs on the field even if the meta mechs are most common. I want to find matches faster, and i want the matchmaker to be able to make better choices overall.
Mostly though, i think we are all wearing big boy pants in the group queue and will not have a toddler meltdown if we cannot always play out prefered mech in the interests of a better game.
Thoughts?
Edited by Asmudius Heng, 13 October 2014 - 02:52 PM.