21 replies to this topic

[Scout Derek]

Seriously. I take a few matches in a medium mech, team absolutely does nothing. I take a Assault mech, do the same work in my medium, and suddenly 4 times in a row, W O W, Look at that. Wins! And stomps at that! 4 Times in a row! Here, Have a look!









Just do a hard reset - something that will pair me up against similar Skilled opponents. Please....

I hate stomping pugs. Like some here and I think, you don't. Learn. Crap. It's not challenging. It's boring! And each time I do a stomp, it's disgusting. I decided to take screenshots of matches close to or beyond 1k damage. These are all screenshots one after the other.

Just look at the losing side. It's just.... mediocre. Stupidity.

Marauder IIC doing 82 Dmg???

A spider on our team that did 0 dmg??? (he did cap at least, so there's that)


Look, the point is, Matchmaker is still unbalanced, and it's really bugging me that I have to bring the biggest and most boring meta mech I have to make a difference. How sad....

And yeah, there's a good pinch of salt in this post, and it's mostly a gripe with the Matchmaker. It's stupid broken. Fix it PG- you know what, I don't think I'll bother asking that. They've got more priorities like making E-sport championships for the game, rather than finishing their half baked products like Faction Play and other bugs.


(Probably the most salt I have ever had with this game in years)

[Skipmagnet]

Good, good. Let the salt flow through you. Give in to your NaCl feelings and your journey towards the bittervet side will be complete.

[DaZur]

No offence intended but... A hard reset would place everyone at the same tier.

Resulting in... You guessed it, the same sub-par players you're railing against.

[Bombast]

Well... duh.

About 90% of the matches that happen are the result of the MM giving up and just doing whatever. And the few times it succeeds, all it does is prove that PSR is junk.

The only difference the new MM has made, that I can tell, is that I'm (A Tier 1) being put in teams with more people who are willing to use VOIP. And for that reason, I want to go back to how it was before.

[Alcom Isst]

You don't balance by KDR. You balance by Win-Loss difference so players rise and fall in the matchmaker until their Win-Loss ratio is 50%, which is what ELO does in a nutshell.

Though with the gameplay being a relentless snowball, and so many other fluctuating variables in a player's ability beyond skill (skill-trees, mech quality, etc), there's only so much you can do to save MWO's balance via the matchmaker.

[El Bandito]

WLR is far more accurate than KDR. It doesn't shaft contributions towards victory other than killing.

Edited by El Bandito, 15 September 2017 - 08:58 PM.

[Skipmagnet]

It does shaft high performers dealt onto a losing team by the matchmaker, though. That seems uncool.

[Scout Derek]

Alcom Isst, on 15 September 2017 - 08:55 PM, said:

You don't balance by KDR. You balance by Win-Loss difference so players rise and fall in the matchmaker until their Win-Loss ratio is 50%, which is what ELO does in a nutshell.

Though with the gameplay being a relentless snowball, and so many other fluctuating variables in a player's ability beyond skill (skill-trees, mech quality, etc), there's only so much you can do to save MWO's balance via the matchmaker.

El Bandito, on 15 September 2017 - 08:57 PM, said:

WLR is far more accurate than KDR. It doesn't shaft contributions towards victory other than killing.

Here's my question then to this;

What about those who have a low WLR but a higher KDR? What happens then? What happens when those people get matched into games with 11 other winners but their KDR is horrendous, against another team with low WLR but with respectable KDRs of 1.5~?


I remember someone mentioning to balance by Leaderboard, but I'm not sure how that'd work out. Sounds good, but in concept it sounds a bit complex to implement.

[Bombast]

eyeballs, on 15 September 2017 - 09:28 PM, said:

It does shaft high performers dealt onto a losing team by the matchmaker, though. That seems uncool.

The idea is that good players, through their own skills, will be able to win more often. Everyone gets screwed with 11 not-so-good player teams, but once it all averages out and the smoke clears, a players skill should be seen in his WLR.

The same cannot be said for damage dealt or KDR, both of which can be gamed.

Edited by Bombast, 15 September 2017 - 09:34 PM.

[Scout Derek]

Bombast, on 15 September 2017 - 09:33 PM, said:

The idea is that good players, through their own skills, will be able to win more often. Everyone gets screwed with 11 not-so-good player teams, but once it all averages out and the smoke clears, a players skill should be seen in his WLR.

The same cannot be said for damage dealt or KDR, both of which can be gamed.

Should it really though?

My Overall WLR is approximately 1.275, while my KDR Overall is 1.89. Most of my games are Solo, too, about 70% are. the other 30% is group play with units, or Faction Play.

So either this is a player population issue, or a MM issue going on. Maybe even both to be fair.

eyeballs, on 15 September 2017 - 08:44 PM, said:

Good, good. Let the salt flow through you. Give in to your NaCl feelings and your journey towards the bittervet side will be complete.

[Bombast]

Scout Derek, on 15 September 2017 - 09:39 PM, said:

My Overall WLR is approximately 1.275, while my KDR Overall is 1.89. Most of my games are Solo, too, about 70% are. the other 30% is group play with units, or Faction Play.

...I'm not sure what you're trying to say here. Is the presumption that, if WLR is an indicator of skill, it should be higher than a persons KDR?

[Zergling]

Above 1.0 W/L, KDR is normally higher than WLR, see Tarogato's correlation graphs.

Edited by Zergling, 15 September 2017 - 09:53 PM.

[Scout Derek]

Bombast, on 15 September 2017 - 09:44 PM, said:

...I'm not sure what you're trying to say here. Is the presumption that, if WLR is an indicator of skill, it should be higher than a persons KDR?

I guess moreover this, but mainly it's the fact that for years of playing, I've managed to somehow keep a rising and positive values of both ratios even though MM is supposed to, from what you're telling me, is to balance the two out.

I'm just saying the MM, is still bad..... For all that it's worth. From the time that Group Que and Solo Que were together, to Group weights being gamed, to MM balancing back before tiers were a thing, up to now...

[Bombast]

Scout Derek, on 15 September 2017 - 09:53 PM, said:

I guess moreover this, but mainly it's the fact that for years of playing, I've managed to somehow keep a rising and positive values of both ratios even though MM is supposed to, from what you're telling me, is to balance the two out.

Whoa. I never said the MM balances them out. I said that, in theory, a good players WLR will rise to the top, showing their skill.

Naturally, as one's skill level goes up, KDR will go up as well.

Scout Derek, on 15 September 2017 - 09:53 PM, said:

I'm just saying the MM, is still bad.....

I wont argue against that. The MM consistently fails to do 'whats its supposed to do' (Sort by tier), and even when it does, the metrics that tier run on are so jacked it doesn't matter.

[Nightbird]

MM only looks at Tier, Tier only looks at Match Score, 75% of your Match Score is damage. People who use LRMs aren't effective in winning a match can pad damage i.e. match score, easily. This is why Tiers don't seem to work. Switching to KDR is fine, a less drastic change would be to reduce the amount damage contributes to match score significantly.

Another problem is not all mechs are equal. Just because you're a good pilot doesn't mean you'll do well in a brand new mech, but MM assumes your effectiveness does not change whether in a meta mech or noob mech. This is not true. Stats need to be tracked by chassis, so that piloting new or under-performing mechs gives a +1, +2, +3 modifier to tier for the purpose of match-making when piloting bad mechs. An alternative to this is make Tier swings much faster. Want to pilot a vindicator? Sure. Lose 3 matches and drop a tier to get better matches in the future. Want to pilot a meta mech? Get 3 wins and raise a tier.

Edited by NlGHTBlRD, 15 September 2017 - 10:08 PM.

[Alcom Isst]

Scout Derek, on 15 September 2017 - 09:30 PM, said:

Here's my question then to this;

What about those who have a low WLR but a higher KDR? What happens then? What happens when those people get matched into games with 11 other winners but their KDR is horrendous, against another team with low WLR but with respectable KDRs of 1.5~?

I remember someone mentioning to balance by Leaderboard, but I'm not sure how that'd work out. Sounds good, but in concept it sounds a bit complex to implement.

Teams will be matched against teams with the same average WLD or ELO (along with average weight), as the game attempts to form matchups with a 50% chance of victory. You won't have a team with 1 low WLD and 11 high WLD's match against a team with 12 low WLD's.

As far as KDR goes, the kill objective is by far the most commonly used win-condition in MechWarrior Online, so KDR and WLR strongly weakly correlate. Any player with a low KDR and high WLR, or vice-versa is an anomaly. I can imagine a player with poor piloting or damage output but with strong communication having a low KDR and high WLR, but that'll still be pretty rare. Even with anomalies, the matchmaker should focus on Win/Loss for balance, as the ultimate goal is to have matches with a 50% chance of victory.

Alcom Isst said:

Teams will be matched against teams with the same average WLD or ELO

Though I'm pretty sure this will fail too, as different weight classes have different values for... value. An assault mech is more valuable than a medium mech, so a team with its better players in assault mechs should have an advantage over a team with its better players in medium mechs... so you'd have to like add weights to the WLD/ELO balancing.

Again, MechWarrior Online is a stupidly difficult game to balance the matchmaker for.

Edit: Apparently the correlation between KDR and WLR is weaker than I thought, but the point still stands. The goal of the game is to win, not to kill and live.

Edited by Alcom Isst, 15 September 2017 - 10:41 PM.

[El Bandito]

Scout Derek, on 15 September 2017 - 09:30 PM, said:

Here's my question then to this;

What about those who have a low WLR but a higher KDR? What happens then? What happens when those people get matched into games with 11 other winners but their KDR is horrendous, against another team with low WLR but with respectable KDRs of 1.5~?

If someone has high KDR but low WLR, then they are mere cowardly kill stealers who likely do not share armor with the team. They are not good players, cause good players will have better than average WLR, overall. I have not seen a contradicting evidence anywhere, in my MWO life.

As for your second question, WLR based MM will not match high amount of good WLR players against high amount of bad WLR players, in the first place. It will mix them up.

Edited by El Bandito, 15 September 2017 - 10:51 PM.

[Valhallan]

Alcom Isst, on 15 September 2017 - 10:24 PM, said:

so you'd have to like add weights to the WLD/ELO balancing.

You mean like BV? . BV+Elo = gold.

[Thorqemada]

Zergling, on 15 September 2017 - 09:49 PM, said:

Above 1.0 W/L, KDR is normally higher than WLR, see Tarogato's correlation graphs.

Cant say that, my W/L 1.39 is higher than my K/D 1.15 since ages though i dont play that much so i advance only slowly toward T1 - maybe it changes there or grouplay changes it bcs i only do QP.

[Zergling]

Thorqemada, on 16 September 2017 - 12:09 AM, said:

Cant say that, my W/L 1.39 is higher than my K/D 1.15 since ages though i dont play that much so i advance only slowly toward T1 - maybe it changes there or grouplay changes it bcs i only do QP.

Which is why I said 'normally'; there are exceptions.