439 replies to this topic

[Tarl Cabot]

Krucilatoz, on 04 May 2025 - 01:50 AM, said:

if any in match maker able to fix, probably by using recent W/L ratio (such as last one month) after taking player bucket queue from tier level.
This to ensure groups / player who keeps on winning (coz they are soo good) to meet each other and balance the match.
It's just my simple observation, who at tier 1, being pitted against some of the best player, and having losing streaks.
I don't know why I still playing this game.. old man habits I guess

A few ways that would work out.

• Grab players
  
• Review W/L
  
• Groups - MM would likely take their W/L avg. Currently the MM takes the Group's PSR (Tier) avg.
  
• Great Group Team A and Okay Group Team B
  
• MM assigns worse W/L into Team A with that respective group while the rest are assigned to Team B.
  
  • Take into consideration mech weights would likely be ignored. This is taking into consideration previous steps PGI had taken.
    
  • Team A with the best group and the bottom of the heap players, most who are piloting assaults.
• The side that will usually win will be the side that tries to communicate and work together and plays aggressive but usually not extremely aggressive.

Have you watched any of the previous Tournament matches? Most of those rarely ended in a "close" match, i.e. Last mech standing. And matching a major change to the MM now would be out of the question, monetary and manpower-wise.

[pbiggz]

its worth repeating that players' perception of stomps is highly influenced by MWO not presenting match performance very effectively or clearly. If the total amount of armour each team had was shown as bars, and the end match screen showed how much armour each team had left, in total, there's probably even 12/0 "stomps" that were actually close matches, just where one team's damage was more efficient in scoring kills than the other team.

Much of the discontent over the matchmaker comes from the perception that matches are more often than not, uneven, and that's not really born out by anything except anecdote. It sucks to lose. People get mad. They remember what made them mad. The negative impression lasts alot longer than positive ones.

[Jelan]

pbiggz, on 05 May 2025 - 08:10 PM, said:

its worth repeating that players' perception of stomps is highly influenced by MWO not presenting match performance very effectively or clearly. If the total amount of armour each team had was shown as bars, and the end match screen showed how much armour each team had left, in total, there's probably even 12/0 "stomps" that were actually close matches, just where one team's damage was more efficient in scoring kills than the other team.

Much of the discontent over the matchmaker comes from the perception that matches are more often than not, uneven, and that's not really born out by anything except anecdote. It sucks to lose. People get mad. They remember what made them mad. The negative impression lasts alot longer than positive ones.

Thats a really good suggestion, I'm quite vocal about the MM and how it rarely creates even matches. You know, the ones we all enjoy that go down to the wire, win or lose, but actually having something like that as a metric might actually show that even the 12-0 stomps are closer than it shows. Some still clearly won't be as you can see the damage from each team at the end.

[nanashi0110]

pbiggz, on 05 May 2025 - 08:10 PM, said:

its worth repeating that players' perception of stomps is highly influenced by MWO not presenting match performance very effectively or clearly. If the total amount of armour each team had was shown as bars, and the end match screen showed how much armour each team had left, in total, there's probably even 12/0 "stomps" that were actually close matches, just where one team's damage was more efficient in scoring kills than the other team.

Much of the discontent over the matchmaker comes from the perception that matches are more often than not, uneven, and that's not really born out by anything except anecdote. It sucks to lose. People get mad. They remember what made them mad. The negative impression lasts alot longer than positive ones.

Very good idea! I would love to see it implemented.


Just now I experienced a terrible match. It was a match where I doubt if I was matched with a player...
I even have a conspiracy theory that BOTs are actually being implemented as a way of matching numbers, and are making their allies suffer with their terrible, low-quality AI. This is a conspiracy theory that comes to mind.
I have experienced in other games that players get frustrated with matchmakers due to the small number of incredibly bad matches that occur.
(In fact, that game turned out to be a terrible matchmaker...)
I wish there was something to mitigate these awful experiences... I wonder if it is difficult.

[Void Angel]

The matchmaker does not "rarely" create even matches. You are misunderstanding emergent player behaviors, based on MWO/Mechwarrior's combat damage system and information the end-of-match screen does not relay to players. Pbiggz literally just explained that, and it's very correct.

Also, there are no bot players, and it is crazy to think so without proof. Just because an uncommunicative player fights poorly and does things you don't understand, it does not logically follow that bot players are being used to "fill out matches." If that were the case, some of my match wait times would be a dang sight lower...

[Saved By The Bell]

Just now I am in non meta cyclops, cant find good build and have terrible times. If I dropped to Tier 4, it would be easy, but keep working about this build.

[Void Angel]

Eventually, you would drop to Tier 4, if you did not find a build that suited you - but it will take time, as it should.

[Jelan]

Void Angel, on 07 May 2025 - 01:49 AM, said:

The matchmaker does not "rarely" create even matches. You are misunderstanding emergent player behaviors, based on MWO/Mechwarrior's combat damage system and information the end-of-match screen does not relay to players. Pbiggz literally just explained that, and it's very correct.

Also, there are no bot players, and it is crazy to think so without proof. Just because an uncommunicative player fights poorly and does things you don't understand, it does not logically follow that bot players are being used to "fill out matches." If that were the case, some of my match wait times would be a dang sight lower...

I believe it does rarely create even matches, what I perceive as an even match may be different to yours so you cant just dismiss my statement without evidence. I consider an even match any that ends 12-7 or closer win or lose, as I've said before its not about the wins or losses for me, its about how much fun the match is.

However as pbiggz says adding a total damage dealt by each team would show if the 12-0 games are closer than they seem. Its unlikely with a 12-0 match as once it snowballs the OpFor dies pretty fast, but the 12-4 games etc might actually turn out to be closer than the kills suggest

Edited by Jelan, 07 May 2025 - 11:51 AM.

[pbiggz]

Jelan, on 07 May 2025 - 11:50 AM, said:

I believe it does rarely create even matches, what I perceive as an even match may be different to yours so you cant just dismiss my statement without evidence. I consider an even match any that ends 12-7 or closer win or lose, as I've said before its not about the wins or losses for me, its about how much fun the match is.

With respect, what Void Angel is telling you, and what I would agree with, is that your perception of this is anecdotal. That doesn't mean its meaningless, or that it should be ignored, but its not a hard statistical standard. The matchmaker works ok alot more often than not, again, its just that negative impressions last longer than positive ones, so people tend to remember the games that weren't balanced.

Jelan, on 07 May 2025 - 11:50 AM, said:

However as pbiggz says adding a total damage dealt by each team would show if the 12-0 games are closer than they seem. Its unlikely with a 12-0 match as once it snowballs the OpFor dies pretty fast, but the 12-4 games etc might actually turn out to be closer than the kills suggest

The fact that we have issues even agreeing on the definition of a stomp kind of goes to show this is not cut and dry.

[Meep Meep]

T1 mm is just absolutely disgusting now. Long waits for nearly the same lobby over and over and over even during peak hours.

[Void Angel]

Jelan, on 07 May 2025 - 11:50 AM, said:

I believe it does rarely create even matches, what I perceive as an even match may be different to yours so you cant just dismiss my statement without evidence. I consider an even match any that ends 12-7 or closer win or lose, as I've said before its not about the wins or losses for me, its about how much fun the match is.

That's not how it works. People have to present evidence that a claim is true - and relying on private definitions for what is an "even match" isn't appropriate or useful. So yes, I will disagree with your baseless claims and subjective private definitions - because you are presenting claims without evidence. A close match, in the context of discussing the matchmaker (see thread title) means that the skill levels on each side are roughly even. Your subjective definition based on feelings isn't relevant in that context. But even skill on both sides - even with even 'mechs - doesn't guarantee a "close match" by your definition.

Because the first person to sustain severe damage leaves their team at a disadvantage - and since people are not lining up in skirmish lines and playing Roshambo with their Battlemechs, there's no guarantee that the enemy team took equivalent damage - and so, many times, you're going to see that difference snowball. And it's not guaranteed to show up on a scoreboard - I spent the last several minutes of a match in my armless Firemoth last week, squirrelling two other Lights as they tried to finish me off. Damaged/open 'mechs often play more conservatively, which leads to matches where some 'mechs are damaged and hanging back at the end of the match.

These two situations are common in matches, and are integral to the game's damage system. So you're not complaining about the matchmaker - you are complaining about how the game actually works. That's why your complaint - whether you're informed enough to know it or not - is a recycled complaint dating back to the actual public release of the game.

It's not new, it's not the matchmaker, and it's not going away.

[Void Angel]

Meep Meep, on 07 May 2025 - 04:12 PM, said:

T1 mm is just absolutely disgusting now. Long waits for nearly the same lobby over and over and over even during peak hours.

Wait time hasn't been that bad for me, but I've seen the same faces over and over...

[Saved By The Bell]

Void Angel, on 07 May 2025 - 11:26 PM, said:

Wait time hasn't been that bad for me, but I've seen the same faces over and over...

thats why I was talking this:
https://mwomercs.com...-factions-idea/

But people just thinking: "oh, I ll be farmed". You ll be farmed anyway, my Lord.

[Sjorpha]

Void Angel, on 03 May 2025 - 09:48 PM, said:

None. The matchmaker does the best job that can be done with the player population numbers throughout the day. Literally any other method of "improvement" would require more players to implement.

I think this is true up the point where it caps PSR for tier 1 players. If PSR was open ended so the MM could know the difference between a good player and a very very good player it could match them a little better, now they will both have maxed out PSR so the MM is completely blind beyond that point.

It would be a small improvement and only affect the top 10% or so, but it would still be an improvement.

Another way would be to do away with tiers and just match by total team PSR (assuming open ended PSR). This would probably allow for more even matches, at the cost of a much larger skill spread within the teams.

I can imagine bad players would be a bit frustrated by this as their relative performance would go down a lot, but at the same time they would have better teammates to learn from and compete with so it could raise the average skill level.

I personally have never felt matchmaking in MWO is that bad. It's fine, many other games are much worse or don't have skill based MM at all.

I feel like most of the complaints are just venting about recent frustration, you go on a loss streak and blame the matchmaker despite the fact that loss streaks will happen even with perfect matchmaking. Just like flipping a coin a 1000 times will contain long streaks of heads or tails. I think a lot of people don't understand probabilities and thinks losing or winning 5, 10 or 15 times in a row just shouldn't ever happen if the mm is good. But it will if you play enough, it's inevitable.

Edited by Sjorpha, 08 May 2025 - 09:06 AM.

[a 5 year old with an Uzi]

Sjorpha, on 08 May 2025 - 08:57 AM, said:

Another way would be to do away with tiers and just match by total team PSR (assuming open ended PSR). This would probably allow for more even matches, at the cost of a much larger skill spread within the teams.

I can imagine bad players would be a bit frustrated by this as their relative performance would go down a lot, but at the same time they would have better teammates to learn from and compete with so it could raise the average skill level.

Every match becomes this, over and over and over until the dude with the T shirt and pants gives up and a chain reaction starts. De facto reversion to same tier spread as before, possibly with smaller player pool to pull from as lower skilled players flee from game because they get demolished even faster than before.

This change would balance team skill levels better math wise and remove the T1 MM time malus at least temporarily though before possibly making it everyone's problem.

I anticipate fatigue will set in sooner than skill will increase.

If MM changes aren't happening, though... we're kind of sitting at a bar scribbling on napkins.

Edited by a 5 year old with an Uzi, 08 May 2025 - 10:55 AM.

[Sjorpha]

Jelan, on 07 May 2025 - 11:50 AM, said:

I believe it does rarely create even matches, what I perceive as an even match may be different to yours so you cant just dismiss my statement without evidence. I consider an even match any that ends 12-7 or closer win or lose, as I've said before its not about the wins or losses for me, its about how much fun the match is.

However as pbiggz says adding a total damage dealt by each team would show if the 12-0 games are closer than they seem. Its unlikely with a 12-0 match as once it snowballs the OpFor dies pretty fast, but the 12-4 games etc might actually turn out to be closer than the kills suggest

Look at some comp play between teams of close skill levels.

What you'll see in most cases is one team making the better read, getting the upper hand and then snowballing to a 8-2 or something like that. Let's say it's a best of 5, it might end up being a close 3-2 showing the teams are well balanced in terms of skill, but the individual matches will still snowball based on which team get's the advantage or had the better plan.

This just isn't a game designed to generate "close" matches in terms of kills, and the main reason for that is the high TTK.

That may be counterintuitive but in low TTK shooters you can turn a match around because you can plausibly win say a 3 vs 1 scenario with a few shots even if you're not that much better than your opponents, but in high TTK shooters like MWO this isn't plausible. You need a much higher skill advantage in MWO to pull off even a 1 vs 2.

The tier system increases this effect even further. When teams are somewhat even not just vs each other but also internally, the odds of someone having the skill advantage to pull a team back from a numerical disadvantage is lower.

If you had larger skill spread within the teams, you would see the bad players on both teams dying first leading to "closer" games and much fewer so called "stomps". But this would make bad players on both teams feel like cannon fodder so it's probably not a good idea.

So basically "stomps" in MWO are by design and have little or nothing to do with matchmaking.

Edited by Sjorpha, 08 May 2025 - 11:32 AM.

[1453 R]

Someone earlier mentioned that "Stomps" would drastically diminish overnight if PGI switched from displaying kills/still-alives at the end of a match to displaying total team durability remaining. A 12-3 "stomp" in which the 12 team had less than thirty percent total durability remaining on the field is not a stomp, but nobody sees that because remaining durability is never shown. You see how much damage people did, but not how much they took.

A simple display change like that could fix half the pointless griping about the matchmaker. Not all of it, because some folks aren't griping about anything real in the first place and are just venting their own inadequacies, but a good half of it at least.

[a 5 year old with an Uzi]

1453 R, on 08 May 2025 - 11:43 AM, said:

Someone earlier mentioned that "Stomps" would drastically diminish overnight if PGI switched from displaying kills/still-alives at the end of a match to displaying total team durability remaining. A 12-3 "stomp" in which the 12 team had less than thirty percent total durability remaining on the field is not a stomp, but nobody sees that because remaining durability is never shown. You see how much damage people did, but not how much they took.

A simple display change like that could fix half the pointless griping about the matchmaker. Not all of it, because some folks aren't griping about anything real in the first place and are just venting their own inadequacies, but a good half of it at least.

End of match screen is designed to get you high/make you feel like s#!t in a lot of online games so not surprised MWO opted for it if the e-sports angle was what they were going for

It's s#!tty and manipulative to poke at lizard-brain level emotion in order to keep people playing but I'm not a dev so what do I know

Edited by a 5 year old with an Uzi, 08 May 2025 - 12:29 PM.

[Quicksilver Aberration]

Sjorpha, on 08 May 2025 - 11:23 AM, said:

That may be counterintuitive but in low TTK shooters you can turn a match around because you can plausibly win say a 3 vs 1 scenario with a few shots even if you're not that much better than your opponents, but in high TTK shooters like MWO this isn't plausible. You need a much higher skill advantage in MWO to pull off even a 1 vs 2.

It's more than that though, since in other games you can't lose firepower like you can in Mechwarrior. Getting legged, stripped, etc. It all can contribute to a snowball that is hard to recover from outside the enemy just making mistakes.

a 5 year old with an Uzi, on 08 May 2025 - 12:24 PM, said:

End of match screen is designed to get you high/make you feel like s#!t in a lot of online games so not surprised MWO opted for it if the e-sports angle was what they were going for

To be clear, even from the e-sports angle hiding it doesn't necessarily make a whole lot of sense. Knowing that info is useful at the end without having to look at a stream or have a spectator. Hell I'd love more metrics for example knowing how much damage you've done per weapon would be insightful into build optimization. There were matches way back in MRBC where I used the UAC5/UAC10 Whale and didn't notice just how much I missed with the UAC10s due to velocity desync until I looked at a cast of that match. Stuff like that is super useful. If they had honestly gone for the esports angle being able to save matches and replay them would also be huge.

However they never were going for the e-sports angle, a lot of their early focus was more towards casual players. Just think how long it took us to get spectators, lobbies, etc.

Edited by Quicksilver Aberration, 08 May 2025 - 05:08 PM.

[Void Angel]

Sjorpha, on 08 May 2025 - 08:57 AM, said:

I think this is true up the point where it caps PSR for tier 1 players. If PSR was open ended so the MM could know the difference between a good player and a very very good player it could match them a little better, now they will both have maxed out PSR so the MM is completely blind beyond that point.

This is essentially just tier matching with an extra step - but it might be better. Trouble is that they're just not resourced for software engineering any more. =(