29 replies to this topic

[Noesis]

Hate it, sorry.

Lack of consistency to game play and as mentioned could be easily manipulated.

And I don't mean this in a derogatory sense but learning from defeats is an important part of Mechwarrior imho.

Roll on tonnage limits, lack of damage potential with lower numbers of heavier Mechs will help with survivability and might make the snowball effect less of a comparative issue? This is all relative however but some mistakes having more forgiveness to some roles and platforms as a result.

And ultimately like it or loathe it, 12 people will lose and 12 people will win if it isn't a rare draw. Which interestingly might be one useful metric to measure the absence of stomps. i.e. the frequency of draws.

[Prezimonto]

IceSerpent, on 03 February 2014 - 12:02 PM, said:

I would say not instead of Elo, but in addition to properly implemented Elo. If you use Elo difference ranges Paul mentioned in his last Command Chair post and convert them to the actual expected chances to win, you will instantly see why Elo "doesn't work" - it really does what it's supposed to be doing, but it's set up in a way that produces no results whatsoever. It's similar to setting up RL speed limits to 300mph - technically concept still works, but serves no useful purpose any longer.



Tonnage as a separate metric is not necessary here IMHO - it's included in BV by definition, as less tonnage implies less stuff you can pack into your mech. Ideally, heavier mechs would get BV increase due to more weapons / equipment / armor, and BV decrease due to lower speed. Lighter mechs are the other way around. Depending on the exact loadout, a "good" lighter mech should be equal in terms of BV to a "bad" heavier mech.

Oh Boy, someone else here understands statistics and/or math. I'm glad I'm not the only person that sees just how awful their implementation has become.

I still don't understand why they don't match to a narrow ELO for each player for each match, with an exception that increments the range wider for anyone in the outer 10% (either side) of the ELO distribution. This would give those edge cases longer match times, but not silly huge.

My fear is that this was done because the player base is too small for the total range of possible ELO. There's different fixes for that, but would still need some special conditions for the wings.

Edited by Prezimonto, 03 February 2014 - 07:33 PM.

[Sandpit]

I think if they'd actually have an elo for each mech instead of weight class there would be a shift. It would be interesting because a player can easily take out their locust and run around losing right now since there's no reason beyond cbills to win (this could change with CW) and then go out and own in their jenner resulting in an "average" elo so when they're dropping "seriously" in their jenner they're owning it up.

I think the main reason things are the way they are is because of player population. When you think about all of the options that go into a match, it gets to be a lot that MM has to factor in. Tonnage restrictions will alleviate that a bit but right now MM has to figure
elo
weight class
game mode
# of slots to fill
premades
and all while trying to find something that matches both teams equally. Game modes are one of the hardest I would imagine. Any obviously slides into the first open slots but when you're not on any then it has to find all of that to fit into that specific game mode.
Even with a decent population you either have to do one of two things. You have to either hope enough are on in any given elo bracket at any given time for a specific game mode or you have to allow that elo to broaden and encompass a wider range. If you don't do that you have the dreaded "Failed to find a match" which is the worst scenario because then you have players who want to play the game unable to simply because their elo and game mode aren't full enough. So taht doesn't mean the population here is "small" but it does mean that the population might be too small for certain ideas to work regardless of how good they are

[Duncan Fisher]

This isn't just a bad idea, it's a horrible idea. Competitive play would be all about stripping enemy mechs of weapons and damaging them as much as possible without actually finishing them off, then, assuming the other team was employing the same tactics, combat ineffective mechs would be forced to suicide or be team-killed to put points on the other teams board.

I'm sorry, I can't even tell you that you put some good effort into thinking this up, nothing about the idea is viable.

[NextGame]

Every aspect of this idea sounds horrendous. Losing sides should not be rewarded for their incompetence. Winning sides should not be penalised for successfully working towards achieving the objective of the game.

I don't even see any reason to try to be constructive about it.

Man up. Play better.

[IceSerpent]

Prezimonto, on 03 February 2014 - 07:32 PM, said:

I still don't understand why they don't match to a narrow ELO for each player for each match, with an exception that increments the range wider for anyone in the outer 10% (either side) of the ELO distribution. This would give those edge cases longer match times, but not silly huge.

Wider range should be a setting in the game options, something like a "I want short queue times at the expense of loose Elo matching" checkbox. Players shouldn't be forced into that.

Quote

My fear is that this was done because the player base is too small for the total range of possible ELO.

That's most likely how PGI views it. Problem is that it's a "catch 22" kind of situation - they lose players because of Elo not being done right (a.k.a "stomps") and they don't feel like tightening Elo matching because playerbase is small.

On a side note, same exact thing happened to 3PV implementation. They didn't feel that playerbase is large enough to give us an option to split the queues (1PV only / 3PV only / both) and lost quite a few players as a result.

Sandpit, on 03 February 2014 - 07:40 PM, said:

I think if they'd actually have an elo for each mech instead of weight class there would be a shift. It would be interesting because a player can easily take out their locust and run around losing right now since there's no reason beyond cbills to win (this could change with CW) and then go out and own in their jenner resulting in an "average" elo so when they're dropping "seriously" in their jenner they're owning it up.

I don't see how this would change anything - this same player can easily just lose a bunch of matches in a Jenner if they desire to do so.

Quote

I think the main reason things are the way they are is because of player population. When you think about all of the options that go into a match, it gets to be a lot that MM has to factor in. Tonnage restrictions will alleviate that a bit

I highly doubt that tonnage restrictions would do anything at all, as mechs don't generally follow "heavier = better" scheme. I.e. in current meta Highlander is better than Atlas despite being lighter. Not to mention that not all configs of the same mech are equally good.

Quote

Even with a decent population you either have to do one of two things. You have to either hope enough are on in any given elo bracket at any given time for a specific game mode or you have to allow that elo to broaden and encompass a wider range.

There is a much better way to deal with it - give players an option to specify if short queue times are more important to them than even matches.
In a perfect world we would have matching by Elo and some sort of BV with uneven teams possible (just to give MM more matching options - a smaller high Elo/BV team can be matched against larger low Elo/BV team), defaulting to tight matching at the expense of the queue time with an option to switch it to loose matching with shorter queue times for those who "just want to shoot giant stompy robots". Unfortunately, world is not perfect and we're stuck with PGI's way of doing things.

[Rebas Kradd]

Khobai, on 03 February 2014 - 08:59 AM, said:

another terrible idea in the same vein as ghost heat.

how bout we just balance teams to begin with?

Won't stop stomps.

[BillHones]

Well thought out but as others posted, nerfing the team for doing well is not the way to go. You need...

Dynamic Matchmaking

It needs to take into account hardware on the mechs, level of efficiencies, your personal elo, balance groups against equal groups, etc. We may also need a bigger player base... :/

Edited by BillHones, 04 February 2014 - 09:14 AM.

[Joseph Mallan]

Rebas Kradd, on 04 February 2014 - 09:10 AM, said:

Won't stop stomps.

true words.
Wasn't the Superbowl supposed to be a match up of two strong teams? The Broncos was the 2 point favorites. How did that game turn out?

Edited by Joseph Mallan, 04 February 2014 - 09:55 AM.

[Scratx]

OP's idea is just bad, sorry. This isn't NannyWarrior Online, if you can't deal with the fact you're down 3 mechs from enemy fire and are likely to be roflstomped if you don't get your remaining act together, then that's your problem.

The game, however flawed, does have one thing going for it. The rules are the same for everybody. Deal with it.