23 replies to this topic

[Zen555]

Is it just me, or are a ton of games ending with a 12-2 score (or worse)?

The MM seemed to be doing pretty good for a while there. By "pretty good" I mean I had a 50/50 chance of winning and the score was 12-6 or closer. That makes sense: even if I'm a top-tier player, the other team should have a top-tier player to cancel me out, resulting in about a 50/50 win/loss ratio no matter how good or bad I am personally.

And I know games tend to "avalanche" so that when things go pear-shaped for one team, the other will just steamroll them with few additional losses. But that's why I'm happy with a 12-6 score. A score of 12-2 is just pathetic.

If there were changes in the last month to the MM I think they were for the worse.

[Reaper3015]

Not just you. Been seeing it more and more since CW Beta dropped. My guess is that the MM is pulling from a wider pool to try and create "fair" teams. Problem is, 1-2 high ranked people are going to be hard pressed to carry a team in MWO against a group of all midranged rank people.

[Basskicker]

Weekend + Easy Event means the underhive has come out to derp.

[Dahzer]

I understand the thinking that the event and the weekend would lead you too think it is the cause but not solely. Yes it makes a bigger pool of players but if you look at it objectively you can see it is just how it goes. You can have 2 identical teams and have the exact same result. If team A loses a medium level player due to focus fire then that is 1 less to fire on team B. If team A doesnt focus for as well it becomes a waterfall of lost mechs. It is based on firepower on single targets.

Team A: firepower = 100
Team B: firepower = 100
If team A loses 5% of its firepower then it has less firepower to take team B on.
Team A: firepower = 95
Team B: firepower = 100
Thus team A can only focus 95 towards team B where team B can bring 100 to bare.
This leads to a over abundance of firepower.
This way the gap continues to grow. As team A can only fight back with less and less.
This doesn't factor in bad mech builds, running that leave the group and die or Afks.

Better example if you have a fight between 2 lances. Assume each mech does 5 damage and has 15 hp. If team A focus fires on team B and team B spreads the damage.

Team A. :: Team B
[A15] -- [B15]
[A15] -- [B15]
[A15] -- [B15]
[A15] -- [B15]

[A10] -- [Dies]
[A10] -- [B10]
[A10] -- [B15]
[A10] -- [B15]

[A10] -- [Dead]
[A5] -- [Dies]
[A5] -- [B5]
[A5] -- [B15]

[A5] -- [Dead]
[A5] -- [Dead]
[A5] -- [Dies]
[Dies] -- [Dies]

Also this is in the lore: combat loss grouping
The principle that a group of 'Mechs (or other vehicles) engaged in combat will accumulate damage over time, leading to a point where several or all of the group will succumb to damage at the same time. Can also be used to describe the time a unit sustains enough losses that it is no longer a viable fightin

Which is also based on a power law relating to two quantities or groups

Edited by Dahzer, 15 February 2015 - 05:53 PM.

[Ironwithin]

Solo-dropping was just downright awful for me over weekend while group-queue was surprisingly nice to play in for a change.

[Insects]

If a murderball meets a spread out team then there is going to be a stomping no matter how skilled the team is.
Of course you could say that a team of skilled players wouldn't let themselves get scattered... But skill can easily lead to arrogance with everyone believing their way is best. Been in games where that exact thing has happened, 3 players had their idea of the best tactics, they were right everyone else was wrong, 3 went in different directions fragmenting the team and of course we got stomped.
I'm sure there was plenty of skill there, but when the perfect storm of personalities happens it all falls apart.

So much relies on the team concentrating fire, so easy to go wrong.

[Catra Lanis]

Do we need a matchmaker? Suppose teams were just matched as far as possible with weightclasses and ecm, what would happen?

Sure there would be stomps but that already happens. In the long run you'd think it would even out just as it did back in the days where games just threw two random teams togheter. In fact I believe that after 200 games with the current matchmaker and 200 games with teams matched weightclass for weightclasss and as far as possible equal amounts of ecm mechs in each team you'd be hard pressed to see any difference when looking at the stats.

[Insects]

I think it does separate the extreme ends of the skill scale and there will be a lot of tears when those two ends meet.

I'm all for a trial week though where matchmaker is disabled. If for no other reason than it might stop all the complaints about matchmaker being bad. We could use the knowledge of what life without Elo would be like so we have more than just speculation.

I do enjoy how civilized my light Elo is though compared to my assault ranking (where i suck) where every single match is someone abusing the team for being bad etc (hint for those people, you are in the same Elo bracket as the "bad" people, therefore must be bad too).

[PhoenixFire55]

We've talked about it over and over and over. The most recent. MM isn't there to give you balanced matches. Working as Intended^TM.

[Tarogato]

Also check out this Reddit thread where I started drilling about MM and got a dev response (yes, they actually exist!)

http://www.reddit.co...ore_the/co68o9u

[PhoenixFire55]

Tarogato, on 16 February 2015 - 03:57 AM, said:

Also check out this Reddit thread where I started drilling about MM and got a dev response (yes, they actually exist!)

http://www.reddit.co...ore_the/co68o9u

Yeah, we've heard that here as well. Blah blah blah statistical proof blah blah blah. Only stats I see every day is 12-0 ggclose.

[Mad Cow Jenkins]

I´ve also seen the issues being a mediocre player hitting 2+ kills on average and sometimes soloing 2 players is just not normal for me, it really seems like the teams are consiting of very different skill level. I see very few team mates targeting anything and sometimes they start firing at each other because they cannot seem to see the difference between enemies and friends, its almost comical.
When i compare the current stomp though with what was in the past this is a breeze, there you could have whole night where you ran into the same 4 man pro lance every 2 games.
For me personally i think the biggest issue is the 3/3/3/3 when my team loose bad, which is often when I drive an AWS and that just doesn´t cope when the opponent team is made out of heaviest mech in each category.

[SethAbercromby]

We've had the discussion time and time over again. No, 12-2 is not a fair representation of skill balance between the players. Just after 3 kills, a match can be decided for one team. The les players one team has, the less targets there are to select, meaining that more players, chosing randomly, will select the same target to shoot at.

This effect quickly cascades into the losing tem losing 'Mechs much quicker than they can kill to even the field again until there are not enough 'Mechs to have a large effect on the end result any more.

[Martin Oberhofer]

true, the outcoming score has nothing to do with how well the MM worked

in fact last night I had a great game
long fire Exchanges, 1-2 failed tries to flank and after 10mins there were still only 2-3 Kills per side
then they rushed us - greatly done I have to say - and we lost 12-3
was it a bad game? no
was MM playing against me? no

MM is not your garantuee to a 50% Win ratio

[PhoenixFire55]

SethAbercromby, on 16 February 2015 - 05:49 AM, said:

We've had the discussion time and time over again. No, 12-2 is not a fair representation of skill balance between the players. Just after 3 kills, a match can be decided for one team. The les players one team has, the less targets there are to select, meaining that more players, chosing randomly, will select the same target to shoot at.

This effect quickly cascades into the losing tem losing 'Mechs much quicker than they can kill to even the field again until there are not enough 'Mechs to have a large effect on the end result any more.

So, if two teams were actually balanced why did one team "happen" to lose 2-3 players early on while other didn't?

[SethAbercromby]

PhoenixFire55, on 16 February 2015 - 06:57 AM, said:

So, if two teams were actually balanced why did one team "happen" to lose 2-3 players early on while other didn't?

Because even on two perfectly equally skilled teams asymmetrical warfare makes a 1 for 1 trade near impossible. Also, I actually wanted to code a little application that would simulate Dahzer's model. 12 targets with 15 hp, each doing 5 damage to a randomly selected target. I'm very interested in how many of those "Matches" turn out to be close calls.

[PhoenixFire55]

SethAbercromby, on 16 February 2015 - 08:06 AM, said:

Because even on two perfectly equally skilled teams asymmetrical warfare makes a 1 for 1 trade near impossible. Also, I actually wanted to code a little application that would simulate Dahzer's model. 12 targets with 15 hp, each doing 5 damage to a randomly selected target. I'm very interested in how many of those "Matches" turn out to be close calls.

I'd say about 40% (within 12-8), however, the number of 12-0s will be far less than what we have in MWO. Not the best model, but its actually pretty interesting to see what happens. Let me know if my guess was close.

[Icebergdx]

Working as intended, its a feature.

[SethAbercromby]

PhoenixFire55, on 16 February 2015 - 08:12 AM, said:

I'd say about 40% (within 12-8), however, the number of 12-0s will be far less than what we have in MWO. Not the best model, but its actually pretty interesting to see what happens. Let me know if my guess was close.

My first test will be that each turn the targets will be selected randomly for damage. Once I've seen how that turned out, I might extend the code to make the targets remain focused on a single target until it's dead and see how how that alters the results.

[Tarl Cabot]

Martin Oberhofer, on 16 February 2015 - 06:53 AM, said:

true, the outcoming score has nothing to do with how well the MM worked in fact last night I had a great game long fire Exchanges, 1-2 failed tries to flank and after 10mins there were still only 2-3 Kills per side then they rushed us - greatly done I have to say - and we lost 12-3 was it a bad game? no was MM playing against me? no MM is not your garantuee to a 50% Win ratio

Just quoting this, as I had a similar match but was on the side of the victor. The match had been going on for 8/10 minutes, we had the crests while they had the valley but they were setup to where they had better fire lanes. One of theirs came around a butt, took it out, hit the jump jets as the next mech, a phract I think, was next in line, almost covered up but not aiming at me. Landed on top of it, it moved back as I continued to fire at it. It went down and the next one was a dire or warhawk, stripped half of its weapons before buying it. At that point it was 9vs5 but it took about 45 secs before my team started to move in, about when I chat that it what they had left, a cat and 2 ravens, ended up being a raven and locust but still.

The other is communicate or lack thereof can change things. Announcing when the assault + light lance just rolled 2 mechs in a far sector is very informative. Or if you are one of the last standing dies to the last enemy mech, what are the others facing? A firestarter or a lightly damaged dire? Communication and information, even in a pug situation, can change how those still going will approach it.