Jump to content

Lance Composition Question


3 replies to this topic

#1 Miodog

    Member

  • PipPipPip
  • Philanthropist
  • 91 posts

Posted 11 April 2016 - 11:48 AM

As currently constructed the MM tries to fill out a Light Lance, Medium Lance, and an Assault Lance. Once MM has to extend its search this becomes less rigid and you get a Light lance with a Marauder in it.

My question is why do we have this rigid weight class needs to be with weight class? Since for the most part the lances start some distance away from each other on maps and I'm running under the assumption I have no intel to deploy my forces, I would deploy 3 lances composed of all weight classes.

Hypothetical Deployment.

Assault
Heavy
Medium
Light

Assault
Heavy
Heavy
Light

Assault
Heavy
Medium
Light

This allows greater flexibility in what each lance can accomplish, splits up deathball tactics, reduces NASCAR matches. Each Lance should feel like it can meet any other lance and have a fighting chance of taking on the OPFOR lance.

Also and this depends on the granularity of the MM but Lance composition could also be augmented by what weapons each Mech is carrying. For instance a Raven with Narc would be paired up in a lance with a medium, heavy, or assault mech carrying LRMs as an example.

Any other thoughts or ideas that could make this work?
Has this already been suggested and been shot down?

#2 Afuldan McKronik

    Member

  • PipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 1,331 posts
  • LocationCanada

Posted 11 April 2016 - 11:56 AM

It would slow down the MM. As it is, it just tries TIERS>WEIGHT. Then it goes to TIERS<WEIGHT. Then if that fails it goes READY MECH>ANY RESEMBELANCE TO MM

#3 Mawai

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,495 posts

Posted 11 April 2016 - 12:29 PM

The matchmaker does the following ...

It starts off with a goal of making two teams with 3/3/3/3 weight class distribution.

3 lights/medium/heavy/assault.

This is divided into 3 lances of ...

light/med med/heavy heavy/assault

However, the queue is NEVER perfectly 25% in each weight class. So the matchmaker MUST change the proportions to match what is in the queue or the wait times become huge very very quickly.

With the current event, which rewards play of 4 heavy mech chassis, the queue is strongly skewed to heavy mechs. I have seen it over 70% in this event and it probably has been higher.

This means that the matchmaker ends up forming matches with as many as 70% heavies (on both teams) since the two teams must have identical weight class distributions. This could be 8 or 9 heavies and with that many heavy mechs ... you can get some in every lance.

Now ... why do they use the current system of mech distribution?

1) It is easier to quickly sort mechs by weight class and then run down the list assigning lances. It works for any combination of mechs that end up in a match no matter what the input distribution.

2) Mechs in similar weight classes tend to have more similar movement capabilities. Assaults and heavies tend to be slower while lights and mediums faster. Grouping by weight class means that the mechs are more likely to stick together in a lance rather than having the faster ones run off on their own leaving the lone assault to fend for itself. It also makes it easier for the team to group up since the slowest mechs tend to be in one lance ... so everyone can (if they want) move toward that lance rather than having slow mechs in every lance slow down the grouping up process.(Though from your post you seem think this would be good :) ).




Edit: more on matchmaker

1) Starts with Seed player who has been in the wait queue the longest - this player defines the intial PSR score range to be used for matchmaking.
2) Start with goal of 3/3/3/3 weight class distribution. If the queue wait times are out of sync then adjust the 3/3/3/3 to reduce the longer queue times.
3) Start collecting players for each side ... aim for the desired weight class distribution within the PSR band.
4) After a certain length of time then release the initial PSR constraint a bit
5) After more time ... release PSR constraint further.
6) After more time ... release PSR constraint again.

The acceptable PSR band does reach a finite value and that limit that does NOT include the entire player base ... for example tier 1 should not be in a PUG match with tier 4+ ... but it requires quite a long time for a match being formed before this happens.

Edited by Mawai, 11 April 2016 - 12:41 PM.


#4 Roadkill

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,610 posts

Posted 11 April 2016 - 12:44 PM

Under ideal circumstances, which means 3/3/3/3, the matchmaker tries to make the following:

The "light lance" (Alpha) is 3 lights and a medium. As near as I can tell, the matchmaker actually tries to select the fastest medium on the team to toss into the light lance.

On the opposite end, the "assault lance" (Charlie) is ideally 3 assaults and a heavy. In this case, as near as I can tell, the matchmaker actually tries to select the slowest heavy on the team.

Beta lance is ideally 2 mediums and 2 heavies.

Due to the release valves in the matchmaker, though, it simply does its best to create one "fast" lance, one "slow" lance, and one with whatever else is available. With this past weekend's focus on heavies, though, the "fast" lance often contained a heavy Mech. I was in the "fast" lance both in my Warhammer and in my Marauder. In the Marauder match we had a Pirate's Bane as our only light, and a Stormcrow as our only medium. (I don't remember what our other heavy was in my lance.)





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users