33 replies to this topic

[Kuritaclan]

Ok i will intruduce a screenshot to show the bigest issue i have with the Matchmaking-System at this time.




Ok what we see is the most likely win in over-tonnage. More weight means more firepower in the most times and therefore the win. 90 tons more is a full assault mech. But this isn't the point to talk mostly about.

What really annoys that the matchmaking systems fits in 6 CTF's vs 1 CTF on the other side. Yes the CTF is a really popular battlemech. But this doesn't have to translate that such unproportional matchups will happen.

The matchmaking system have to assure that chassis have to be "equaly dropped". In the upper case at least drop 4 vs. 3 CTF's at each side. Secondly is the stacking of same variants. Three CTF 4X and two CTF 3D plus one CTF 1X in one group suck the life out of the enemy group. Plz PGI make sure that only one spawned Battlemech of a special variant is in a team, when puging!

We all are saturated of lances with 3+ Battlemechs of one kind. This is maybee funny for them who drop into this group, but not for the enemys. Plz even if we have to wait a bit longer. But something like the above shouldn't happen.

And to point out what a better drop should look like with said adjustments in the Matchmaking-System i will manipulate the Screenshot, how it could be a better matchup (pic under the text). You could say that the "deafeated"-Team may have a higher ELO, but i doubt it heavly because CFT users with good loadout will know why they are driving this Battle Mech, and are rewarded with good results, in the most cases. Particularly if something like this happens.

I have over 500 games in different CTF's and i guess that I know what i'm talking about. My Elo i think isn't high because i used stock-mechs to get the C-Bill without premium time or any bonus beside cadet. The K/D ratio in the stock-mechs i used to built up the money to get my first owned battlemech are miserable. And after purchase of my first CTF I was killed often, because i had to earn c-bill to just equip loadout in general (weapons) or dhs and endo/ff (internals) to become somewhat competitive. And i'm not done yet beside the fact that i have a CTF-1X, CFT-4X and a CFT-3D in my mechbays, i don't earned enough GXP untill this date, so i only have 2 modules out of 3 unlocked (less than 600 GXP away from my third module to unlock and they arn't updated to second degree). As it stands i have a slight positiv 1,0X+ K/D on all three CTF variants, and those numbers are climbing constantly. I was droped in the "defeated"-Team, and i don't think my ELO is that high to single out, all the other CTF users in the enemy team. Otherwise the matchmaking-systems says I'm the best CTF player out of the 7 CTF drivers and they are bad like hell. And the six who droped in the "winning" are so bad that the "defeated" Team easly would make up the 100 ton gap with skill. This assumption can't be true in my eyes.




TLDR: The matchmaking system worries me. Dropping more than two mechs of a kind should be stoped. And if one team have two mechs form the same kind they should be different variants. If there is one mech in one group in the enemy group should be the same kind of mech as a counterbalance. Weight-differences should be beside of the ELO ratting non existent (in the example it is 10,4%, what adds up to a full assault mech). If the weight would be equal, ELO's would really matter. This mixup in ELO's with compensation on drop-weight suxx. It doesn't show who in general have more skill - It only shows who have more firepower to brawl a mech, which comes around a corner.

[Dan Nashe]

Not sure practical.
Although more balanced teams would be nice.
I don't know what the trade-off is in "failed to find match."
Also, more mechanics like intricate team-comp rules could really turn new players off.

Be nice if it could mix buckets.

I.e., in that first game if it could swap the middle lance from each company after the match is created, much better match.
STill, I note that the points for weight classes in the first match add up to like
38 vs 35? (3 lights, 1 med, 6 heavy, 2 assault vs 10 heavy, 2 assault).
Is 31 vs 38 "weight class" points. (Light 1, Med 2, Hvy 3, Assault 4) which is all the matchmaker considers.

So, this a result of a lack of matching players queued.
They did try to tighten match restrictions 2 weeks ago, but the queues were iffy.

But I think weight class matchup fail is a glaring problem they are working on, so I wouldn't attribute it to multiple of the same chassis.

AS for the other complaint, I agree that ELO gets really painful when you are just trying to elite.
Ask any Raven 3L pilot Ever.
My 3L is in the top 10% of my mechs stat wise.
My raven 2 x and 4x are easily my two worst mechs.
Same elo for all 3 is bad :-).

Although man, I was unstoppable on my 3L after I levelled the 2x and 4x. My ELO must have been "worst player ever."

[IceLom]

The matchmaker takes into account elo as well as tonnage..... so you don't know how well the teams were balanced just from tonnage.

I am not saying that its working perfect, I am just saying that we have no clue how well it actually is working. Even the most balanced teams can snowball into a 12-0 game.

I personally feel like the matchmaker is working well, not perfect but better then just random drops that's for sure.

[Anti ARMD]

Calling for any more then 2 of the same mech to be in any match is like saying you want the rain to become donuts!

You use the excuse that pre-mades using the same chassis skew the system, in most cases the one's with many of the same chassis tend to be out-numbered by the enemy in the same chassis. E.g. 4 Victor's in one lance might mean no other assault on your team but you might find 4 Atlas and 2 Stalkers on the other team... whereas your team might be more leaning into the heavy chassis category whilst the enemy have lights, the teams my only be 100t or less difference. Who wins is based on luck and skills (and if you can hit SPIDERS!).

The problem is matching tonnage AND Elo perfectly everygame... can't happen because at any one time you might not be able to match it together, so you might create a perpetual loop of 'failed to find a game' for those outside the largest player primetimes... in effect isolating international communities.

This game is still VERY much about skills, teamwork and LUCK!

So may Lady Fortune shine on you today!

[OneEyed Jack]

As far as they've said, they haven't even begun instituting an actual tonnage balance. There's some limit on the weight class, but there can be a lot of difference across twelve mechs even if weight classes were evenly matched.

That said, 90 tons across 12 mechs is not even close to a decisive tonnage advantage.

[Devil Fox]

OneEyed Jack, on 26 September 2013 - 03:37 PM, said:

As far as they've said, they haven't even begun instituting an actual tonnage balance. There's some limit on the weight class, but there can be a lot of difference across twelve mechs even if weight classes were evenly matched.

That said, 90 tons across 12 mechs is not even close to a decisive tonnage advantage.

Agree... I think their current system allows for 20t difference per weight class (240t for 12 mechs) so getting anything between 50-140t is the right window, then it becomes skill VS tactics. A bad team can win if they use tactics well, likewise a skilled team can dominate if left to their own devises.

[Chavette]

OneEyed Jack, on 26 September 2013 - 03:37 PM, said:

That said, 90 tons across 12 mechs is not even close to a decisive tonnage advantage.

Says who?

[Kuritaclan]

DanNashe, on 26 September 2013 - 03:06 PM, said:

I.e., in that first game if it could swap the middle lance from each company after the match is created, much better match.
STill, I note that the points for weight classes in the first match add up to like

It is the same game... both pictures. In the last one i single handed made what the matchmaker should have done.

And again, as i said in the text ELO couldn't be the point that in one group are 6 CTFs and in the other is only 1 CTF.

IceLom, on 26 September 2013 - 03:18 PM, said:

The matchmaker takes into account elo as well as tonnage..... so you don't know how well the teams were balanced just from tonnage.

I can argue all day long,... You can't tell me, that 1 CTF player have such an high Elo, that there is not other CTF in the same team but in the opposing theam are 6 CTF's!

IceLom, on 26 September 2013 - 03:18 PM, said:

I am not saying that its working perfect, I am just saying that we have no clue how well it actually is working. Even the most balanced teams can snowball into a 12-0 game.

The snowball effect is something what is a ingame thing and how the "pug" works together. All i'm talking about is the start before the match happens. And a drop with half of 12man team filled up with 6 battle mechs of the same kind isn't a good matchmaking system, when on the other side is only one of this kind.

OneEyed Jack, on 26 September 2013 - 03:37 PM, said:

That said, 90 tons across 12 mechs is not even close to a decisive tonnage advantage.

This is by all respect a major disadvantage - it says that there is a virtual 13th heavy weighted mech on the battlefiel. If PGI has a ELO equality, you can't make up this difference by skill, beside theire are only stock mechs driven by new players who can't aim. And as seen in the first picture. There is no Hero Jenner, Centurion, or Dragon Trial-Mech within. And the both AS7-RS maybee © are equaly fitted on both sides by the matchmaker. And both didn't had a major impact anyhow.

Anti ARMD, on 26 September 2013 - 03:33 PM, said:

Calling for any more then 2 of the same mech to be in any match is like saying you want the rain to become donuts!

This is ********. If you can manipulate the matching systems to find a match for weight classes, you could easly manipulate the database machting programm it deeper into Chassis and a step above that into variants.

And after finding a opposit chassi for every 1st available chassi in group a, the matchmaker only needs to alter the ELOs so it is on average the same ELO.

If the matchmaker first try to find players with the same ELO and don't look for their actually choosen mechs such useless onesided dropes like the above happens.

Edited by Kuritaclan, 26 September 2013 - 05:27 PM.

[Kaijin]

The matchmaker is beyond worrisome - It's downright useless. I've been on the winning side of some awful steamrolls myself, but none of them were as pathetic as this last drop in which 8 out of 12 of my team got wasted in the first 2 minutes of a match. I doubt I've got a stellar Elo with a 1324/676 win/loss ratio, but still, it's probably high enough that I shouldn't be getting matched up the likes of those idio.ts.

[OneEyed Jack]

Chavette, on 26 September 2013 - 03:45 PM, said:

Says who?

Says anyone who's not looking for some pathetic excuse as to why they lost other than they got beaten, or were stupid and beat themselves.

90 tons isn't even an average of 10 tons per mech. I dunno about you, but I kill mechs more than 10 tons over my weight pretty regularly.

Kuritaclan, on 26 September 2013 - 05:14 PM, said:

This is by all respect a major disadvantage - it says that there is a virtual 13th heavy weighted mech on the battlefiel. If PGI has a ELO equality, you can't make up this difference by skill,

I heartily disagree. It's absolutely nothing like having a 13th mech, because that would be both a separate target and another enemy to be shooting at you from another direction. It is neither. And the tonnage, in and of itself, is not a "major disadvantage" as you claim. A large enough difference can sometimes create a game where the armor and firepower is simply too heavy for the other team to deal with, primarily on maps that allow the heavier team to force a decisive engagement. This is simply not that heavy of a difference.

The most important point, though, is that the larger mechs are not even remotely always the more powerful/better ones. Nor is the larger mech necessarily more effective in the hands of the given player. Elo is a measure of effectiveness. If two players are equally effective in their given mechs, then one being 10 tons heavier makes little to no difference. For instance, I'd pit myself in one of my Jagers against anyone of a similar Elo in an Orion. Likewise, I'd pit my brawler Highlander or Heavy Metal against any Atlas. Or even either of my Stalkers against an Atlas.

In my opinion, while there are matches with a huge difference pre-determining the outcome, 90 tons is well within an acceptable range to be considered reasonably balanced.

[Kaiser Thermidor]

I think most people are frustrated with the matchmaking system for the wrong reasons, namely weight matching.
The thing that worried me, personally, about the matchmaking system the most is the inclusion of the Elo ranking system.

Elo ranks players by their win/loss ratio. The more you win, the better you are (in Elo). In some games, this may in fact be exactly what it takes to judge a player's relative skill and/or ability.
But not in a team-based game like MWO.
The conditions for victory are
1) Destroy all Enemy Mechs, or
2a) Capture the Enemy Base, 2b) Collect 750 Resources.
Of those three things, you can reasonably expect to do zero of them on your own. Yet, your 'skill' rating only takes into account how effective you are at those three things.
If players were dropped as teams, and were locked to teams (i.e. everything was a 12v12 queue all the time, and you could only play with those 11 people on your team) it could maybe make sense. But that isn't how it works.

Making my own personal "skill" rating depend largely on 11 other people makes no right minded sense. Period.
The inclusion of Elo is stupid beyond compare and p****es me off in how it flies in the face of any possible logic.

There will never be any perfect tonnage matches, obviously. And I don't think anyone expects there to be. But on the other hand, at least we're not getting teams full of Atlai anymore.

Edit: Typos, formatting.

Edited by Vadim Krasvanya, 27 September 2013 - 10:55 AM.

[Chavette]

OneEyed Jack, on 27 September 2013 - 02:01 AM, said:

Says anyone who's not looking for some pathetic excuse as to why they lost other than they got beaten, or were stupid and beat themselves.

90 tons isn't even an average of 10 tons per mech. I dunno about you, but I kill mechs more than 10 tons over my weight pretty regularly.


I heartily disagree. It's absolutely nothing like having a 13th mech, because that would be both a separate target and another enemy to be shooting at you from another direction. It is neither. And the tonnage, in and of itself, is not a "major disadvantage" as you claim. A large enough difference can sometimes create a game where the armor and firepower is simply too heavy for the other team to deal with, primarily on maps that allow the heavier team to force a decisive engagement. This is simply not that heavy of a difference.

The most important point, though, is that the larger mechs are not even remotely always the more powerful/better ones. Nor is the larger mech necessarily more effective in the hands of the given player. Elo is a measure of effectiveness. If two players are equally effective in their given mechs, then one being 10 tons heavier makes little to no difference. For instance, I'd pit myself in one of my Jagers against anyone of a similar Elo in an Orion. Likewise, I'd pit my brawler Highlander or Heavy Metal against any Atlas. Or even either of my Stalkers against an Atlas.

In my opinion, while there are matches with a huge difference pre-determining the outcome, 90 tons is well within an acceptable range to be considered reasonably balanced.

My bad for posting anything, your initial reply said everything about your competence with the subject.

[Joseph Mallan]

I realize we are playing a game, but it is a combat game.

So if you are trying to take a planet (if we ever get CW) and you drop in all lights and mediums and the defenders are all heavy Assault... Are you are going to cry foul? Really? If you are trying to take my stuff, well, All is fair in Love and War! See my objective is to kick your ash, Just like yours is to try to kick mine. Best team wins. So frankly if you bring a knife to a gun fight, you only have yourself to blame.

I have had this attitude long before I got into this game. If you cannot accept that others will not play a game the way you want, maybe you shouldn't play.

I played a 10,000 BV game vs some friends. When my opponent fielded 4 Stone Rhinos with 1/1 Pilots I called foul. Asked to see his sheets. After looking at his builds he asked me if I was mad. I said I was... Cause I didn't think of doing what he did. Of course I lost, but I learned something AND I earned a higher level of respect cause I didn't whine like a baby when I was faced with a superior force.

[Joseph Mallan]

Chavette, on 26 September 2013 - 03:45 PM, said:

Says who?

I agree with Jack, Chavette. A 100 ton ton advantage is not huge. That is just over an 8 ton advantage per Mech. I know you remember the time when a 4 man of Ravens were the terror of the game. 4 Ravens could eat the average assault lance for dinner. 140 tons could... and often did ravage 400 tons.

[FrDrake]

Chavette, on 27 September 2013 - 03:14 AM, said:

My bad for posting anything, your initial reply said everything about your competence with the subject.

90 tons is not a problem, for someone who is usually talking about how good he is, I'm surprised you're going to claim that the enemy having an average of 7.5 more tons per mech was what caused a team to lose.

You know skill is worth an extra 50 tons on any mech.

[Joseph Mallan]

FrDrake, on 27 September 2013 - 05:24 AM, said:

90 tons is not a problem, for someone who is usually talking about how good he is, I'm surprised you're going to claim that the enemy having an average of 7.5 more tons per mech was what caused a team to lose.

You know skill is worth an extra 50 tons on any mech.

For some its worth 65 Tons. I know a few Jenner pilots...

[Chavette]

Joseph Mallan, on 27 September 2013 - 03:45 AM, said:

I agree with Jack, Chavette. A 100 ton ton advantage is not huge. That is just over an 8 ton advantage per Mech. I know you remember the time when a 4 man of Ravens were the terror of the game. 4 Ravens could eat the average assault lance for dinner. 140 tons could... and often did ravage 400 tons.

FrDrake, on 27 September 2013 - 05:24 AM, said:

90 tons is not a problem, for someone who is usually talking about how good he is, I'm surprised you're going to claim that the enemy having an average of 7.5 more tons per mech was what caused a team to lose.

You know skill is worth an extra 50 tons on any mech.

Look at it my way. What are your chances of convincing the organizer of a tournament that your teams' extra 90 tonnage over the agreed on limit doesn't matter because its reasonably possible to win against it? Lets say the organizer doesn't care about the rules too much and is willing to talk.

They won't give a damn what you want, because everyone who has half a brain knows tonnage is an advantage to a degree, and 90 is a LOT in those circumstances.

90t will give you a class up on ~5 mechs. It will turn 5 mediums into 5 heavies, or 5 heavies into 5 assaults, or any mix of those.

Its basically one whole stage heavier comp. To say it doesn't matter alot, cmon man... cmon.

Edited by Chavette, 27 September 2013 - 07:33 PM.

[OneEyed Jack]

Chavette, on 27 September 2013 - 09:44 AM, said:

Look at it my way. What are your chances of convincing the organizer of a tournament that your teams' extra 90 tonnage over the agreed on limit doesn't matter because its reasonably possible to win against it? Lets say he organizer doesn't care about the rules too much and is willing to talk.

They won't give a damn what you want, because everyone who has half a brain knows tonnage is an advantage to a degree, and 90 is a LOT in those circumstances.

Now, I can't claim to be an expert on the subject, since I've never played in a BT tourney in my life, but I'd expect them to be using BV, rather than tonnage. Why BV? Because tonnage alone is about meaningless, while BV takes into account the relative effectiveness of the mech, of which tonnage is only one factor. Sound familiar?

[Kuritaclan]

The statements in the article got a drive what is only a small thing of what i put into it. You are arguing about Tonnage. What is one little passage. I might be wrong in this Forum if the users only can read the first line of an article and start a prolong statement about ONE point.

I wrote not only about tonnage. I wrote about proportionately drops of certain chassis. If the Matchmaker can't handle a balanced quantity on both sides, the matchmaker is a mess.

The matchmaker is a Database what only have to match two Datas Chassis (including Variants)/Weight and ELO and have to track time in Queue. With a certain number of people who get into this database by queue to the register, the matchmaker only have compare are theire two equal Chassis and if so are these ELOs of the players nearby, if so then put them in opposing groups. After done this interrogation, do it with the next, and alter the higher ELO player in the 1x1 match to opposing Team to eaven out the ELO's between bove teams. After have 4 vs 4 (a lance) matched, compare the dropweight to a given value like 300 tonnes (what could alternate to make some heigher weighted allover the Team matchup and lower ones). If the value is below setup for the next lance higher weighted Chassis. If the value is above the set value than search for lower weighted. After filling the second 4x4 match compare the overall weight with a tonnage like 500 tonnes. And make the process again for the next 8 players.

To match near even ELO's in the process you only need to make a request after the first pair of players is matched together, that the next pair shouldnt be far away from this couple and so on.

Now you make rotation bettween ELO equations. The first one is a search routine to find the highest ELO and set this player's for the 1st 1x1 matchup and building a group out of this. It will build up a 12vs12 to match up the highest ELOplayer and drop those in a game. Next rotation is a search routine for the lowest ELO, what matches the "bad player's". The 3rd routine will interpolate the the best vs the worst ELO and make an average to matchup an average Match. After those routines. You will make a search for the player who is was the longest time in queue, and set him as the startpoint to build up matches. If now within this time nobody is over a certain value of time w8ting to be queued start the rotation again, if the value of the w8ting player is above a given point, match build up again a match with the startpoint of the player who is the longest time in queue.

This isn't hard to program and nobody could tell me that if the players waiting to be queued is great enough that this model wouldn't work. Also if there arn't enough players to match with each other, the model could use the "old model" instead of matching certain Chassis and or a point deeper Variants, it will match Weight-Classes and or Chassis in a 5 tonnage range to the 1st picked player.

Lances would fit a special database, and are added up into the 4x4 comparision as 1st part if the w8time has reached a certain value and or ar altered in as last part with a messure of a avg elo and a avg weight to setup a 4 equal 4 between those. Lances in this case don't have to w8 much longer to be droped into matches.

It isn't that much of requested computing power, cause it only uses if/then requests an some loops to beginn the equation process again for the next match to matchmake.

I rather would w8 a couple of seconds longer to be matched in an allmost evening uped match, than set vs drops with unproportional arrangement of chassis and "weights" like in the case of the screenshot.

Edited by Kuritaclan, 27 September 2013 - 02:40 PM.

[C E Dwyer]

Kind of surprised your lights didn't cap win the game, which is where, in this case, there is a weight imbalance, its not so much that there are 6v1 CTF, but your team had 3 lights while they had none, so your team was more nimble, but not able to take punishment, also the atlas was either a disco, or afk, so you lost another 100 tonnes.

This isn't the worse imbalance by far, I've been in games where the weight difference was over 250, as much as I do think the MM still needs work, and it isn't fun if the mis match is to far apart, you'll also find complants about teams being OP because it was full of lights, that the assaults couldn't hit and made that match unfair, if you care to look far enough in these forums, so being light of tonnage isn't allways a disadvantage