228 replies to this topic

[dario03]

justcallme A S H, on 01 January 2020 - 04:18 PM, said:

lolwhat?

If you have a high win rate and match score - you are the carryALL, for the most part.

Some seasons my WLR is garbage. It varies wildly from 3.0+ down to 1.8... Kills and Match Score stay similar though and that is the nature of YoloQ. Some games it doesn't matter what you do - you cannot carry 4+ players who consistently output sub 99dmg and thus have absolutely no business being past Tier 4...

But that is the fundamentally broken PSR system for you...



You have no place to be commenting on anyones stats.

It's been proven time and time again you lack the understanding to know what they mean.

Vxheous, on 01 January 2020 - 04:47 PM, said:

This is my best summary of how to tell by stats if someone's W/L is being carried:
They have a W/L of 4+ (that's winning at least 80% of the games you play)
yet:
under 1.0 K/D
under 180 average match score

Guaranteed someone with stats like that would not have a W/L above 1.0 without significant carry.

Edit: BTW, there are lots of Tier 1 PSR players with under 1.0 K/D and under 180 average match score...

See I would agree that is enough info and far enough to say the player is probably getting carried and shouldn't be T1.

IMO The PSR system has been one of the big problems with the game. In most other games if you get to a high rank you need to be very good at the game and you only play other very good players. Or if the population is low then you get mixed skill levels but the MM can actually make a decently balanced match. But in this game you can be the highest rank and end up playing with players that are basically new or the skill level of new.

T1 should have meant very good, not just something that almost everybody can get to but just takes some longer than others. If we had that then teams could be better balanced, even if you don't change who is on them.

Edited by dario03, 01 January 2020 - 05:03 PM.

[justcallme A S H]

Agree.

PSR has led us to the situation the game is in. PSR has also caused players to leave the game because match quality has been on a decline.

I really don't understand why it was used/created when it was just so flawed in terms of skill/match making - and used to, make matches.

[UnofficialOperator]

justcallme A S H, on 01 January 2020 - 05:28 PM, said:

Agree.

PSR has led us to the situation the game is in. PSR has also caused players to leave the game because match quality has been on a decline.

I really don't understand why it was used/created when it was just so flawed in terms of skill/match making - and used to, make matches.

I used to think that ELO is the way to go but I am no longer sure.

Are you aware of the SBMM (skill based match making) controversy in Modern Warfare for the past few months?

A real SBMM matchmaker will be fun for the those with a professional attitude but won't go over well with casuals because of how "sweaty" everything becomes.

Perhaps it can work in MWO because the ttk isn't as low as 3-5 seconds and there is margin for errors.

Anyway the player numbers have dropped to 300-500 per hour?

Pretty pointless now.

[MischiefSC]

justcallme A S H, on 01 January 2020 - 05:28 PM, said:

Agree.

PSR has led us to the situation the game is in. PSR has also caused players to leave the game because match quality has been on a decline.

I really don't understand why it was used/created when it was just so flawed in terms of skill/match making - and used to, make matches.

Matchmaking is and will always be an issue for almost every PvP game. MWO has two big strikes to deal with - low population density at any given time and sweeping variables in mech design and random maps. So even individual players performance can swing significantly just based on circumstance.

You could develop something Elo based (SBMM is functionally an Elo system too) that accounts for map, mech and loadout to some greater or lesser degree and no question that would be more accurate - but it still has to build teams from the people available in a given matchmaking window in a given region and to approximate tonnage.

As bad as PSR is, as open and over-inclusive as it is.... it still can't build matches with any single narrow band of PSR. The analytics behind stuff like a matchmaker is my jam but I really don't see a solution that isn't super time and money intensive.

You'd have to create something to incentivize smart behaviors and teamwork to make playing more like a 12man 'the norm'. Then you'd need to incentivize running 50-60 ton mechs on average with approximately as many mechs over that as under it. Then you'd need to create a sort of Buy Value system (almost like tabletop, but based off MWO performance) for mechs and gear and it'd need to be both wide and deep - two gauss is way, way stronger than one, laservomit PLUS adequate heating to cool it, SRMs on an LBK but not a Grasshopper, etc. You'd literally need to adjust MM value for every individual player based both on synergy of the loadout and the mech PLUS their historic performance with not just that loadout but similar ones. So a guy who rocks the laservomit HBR is likely going to do well in a gaussvomit Deathstrike.

You can make a strong SBMM system - but you still need a deep enough well of players on at any given moment to fill 24 to a match with similar ping and approximately equal tonnage. Without that all the analytics in the world don't matter.

[justcallme A S H]

MischiefSC, on 01 January 2020 - 10:08 PM, said:

Matchmaking is and will always be an issue for almost every PvP game. MWO has two big strikes to deal with - low population density at any given time and sweeping variables in mech design and random maps. So even individual players performance can swing significantly just based on circumstance.

Indeed this is true.

However putting 3 x EmP pilots on the same team and essentially no-bodies on the other...

It can absolutely do a hell of a lot more than it does now. Which is partly why we are at a point of low population.

[Kroete]

Humble Dexter, on 01 January 2020 - 01:59 PM, said:

If someone has a big win rate they're being carried.

Was he?
Or was he in a group farming smaller groups?
Or did he syncdrop in in solo inviroment?
Or was he just good in pug games?

We dont know,
the stats dont tell enough because we still have group and solo stats mixed
and we dont even see kmdd against kills in the stats ....

But we are in the age of the orange clown, where facts and number dont matter anymore,
so just discuss futher with assumptions and numbers without much meaning.

Edited by Kroete, 02 January 2020 - 12:29 AM.

[FRAGTAST1C]

For almost the whole of December, I was leveling my newly bought Anni's and the Fafnir's from the Collector's Pack in QP. Plus, some Crabs. Compared to November, my MS has increased slightly but my W/R ratio has gone below 1 and yet, my ranking has gone up to 90% compared to 88% from the previous month. If MWO think that I'm to be used as a balance against some of the actual 95%+ players 'cause I, too, am T1 like they are, then you can clearly see where the problem is.

Also, the reason why my ranking has gone up is 'cause even more top players have left the game.

Edited by FRAGTAST1C, 02 January 2020 - 02:41 AM.

[UnofficialOperator]

So I just decided to do 10 34 matches as an assault as a gauge. Have not played for 6 months. 4 17 wins 17 losses with 300 375 MS playing with a 300 ping on NA. I think that's pretty fair. And not once did I ever get nascared as an assault. Got nascared once but that's my own fault.

Its really the player's playstyle that is the problem, not the mech.

Edited by UnofficialOperator, 03 January 2020 - 05:05 AM.

[MischiefSC]

justcallme A S H, on 01 January 2020 - 11:45 PM, said:

Indeed this is true.

However putting 3 x EmP pilots on the same team and essentially no-bodies on the other...

It can absolutely do a hell of a lot more than it does now. Which is partly why we are at a point of low population.

Problem is cost/benefit. All the cheap and easy solutions would make very, very little difference. There's some time/money expensive options that would make at best a moderate difference. PGI has made it clear MWO is nearing end-of-life. Is what it is. Finally reaching the point where even QP is getting the FW treatment. 'We know we said we wanted it to have all these things in it... but nah. Let's just put it all in 1 bukkit.'

Only change you're likely to see literally is a mixed group/pug queue and closing FW.

Don't worry though! You'll always have Solaris VII! Maybe if you had AI from MW5 to run around in stock mechs trying to hold locks for the brave gladiator player hiding behind a hill 800m off trying to keep that wiggly reticle on the red box with his joystick so he can bravely launch his LRMs before falling back again you'd see more people in Solaris VII.

Oh! Wait!

What if Solaris VII took place on a round disc of a map, that automatically just rotated to the right. It'd be friendly and familiar to most players. You could just randomly scatter cover around because 95% of people don't understand positioning anyway there's no need for designing the map around strategic points of interest.

Just a disc, rotating steadily to the right, with AI stock mechs blundering around the middle and random low cover to break LOS but not block LRMs scattered around. Someone tweet a link of this to Russ I feel like I've found a winner. MWO 2 in the bag!

[dario03]

UnofficialOperator, on 02 January 2020 - 04:47 AM, said:

So I just decided to do 10 matches as an assault as a gauge. Have not played for 6 months. 4 wins 6 loses with 300 MS playing with a 300 ping on NA. I think that's pretty fair. And not once did I ever get nascared as an assault.

Its really the player's playstyle that is the problem, not the mech.

Yeah I've never understood where all the complaints about assaults having it rough come from. They have almost always been the highest scoring class in the game. If the enemy doesn't focus an assault it can do absurd damage. And even if you get left behind you can still do decent by just accepting that you won't make it, turning around, and use your armor and firepower to put up at least a few hundred damage before dying. The top assaults are pretty much the best mechs in the game. Sure some assaults are better than others but that is true of all weight classes.

Edited by dario03, 02 January 2020 - 09:38 AM.

[justcallme A S H]

MischiefSC, on 02 January 2020 - 09:18 AM, said:

Problem is cost/benefit. All the cheap and easy solutions would make very, very little difference.

I stopped reading here because that is absolute garbage.

In what world is it fair to have 3 x EmP on one side against a totally lacklustre enemy team?

It is a very VERY simple issue/calc and code to fix to balance it all. It wont be perfect, never will with PSR that is fundamentally flawed - but it will help match quality - sometimes.

[VonBruinwald]

justcallme A S H, on 02 January 2020 - 09:51 AM, said:

In what world is it fair to have 3 x EmP on one side against a totally lacklustre enemy team?

It is a very VERY simple issue/calc and code to fix to balance it all. It wont be perfect, never will with PSR that is fundamentally flawed - but it will help match quality - sometimes.

I'd rather see a snippet of code that excludes players with the same unit tag dropping in the same game unless the matchmaker gates are wide wide open. And then you have it place them on alternating teams so they can't stack.

Now it won't stop the try-hards who leave/rejoin their unit just so they can synch-drop but it wouldn't be hard for PGI to place a tracker listing what units a player joined and when, they could even list it openly on peoples profiles under "Mercenary Unit Employment History" so when we see someone leave/join a unit 10 times a week we can call them out on it.

Now that won't stop the real hardcore scrubs running multiple units but at that point we're dealing with advanced algorithms.

[Prototelis]

Should just let solos opt into GQ just like the FPQ.

Give better elo matching for groups, and a group counter, and I guarantee people will play.

[dario03]

VonBruinwald, on 02 January 2020 - 02:50 PM, said:

I'd rather see a snippet of code that excludes players with the same unit tag dropping in the same game unless the matchmaker gates are wide wide open. And then you have it place them on alternating teams so they can't stack.

Now it won't stop the try-hards who leave/rejoin their unit just so they can synch-drop but it wouldn't be hard for PGI to place a tracker listing what units a player joined and when, they could even list it openly on peoples profiles under "Mercenary Unit Employment History" so when we see someone leave/join a unit 10 times a week we can call them out on it.

Now that won't stop the real hardcore scrubs running multiple units but at that point we're dealing with advanced algorithms.

Don't think that would do much except maybe make it harder for people in large units to get a match. Players that might not even actually be on comms or sync dropping.
Besides the leaving and rejoining units letting you get around it, players could simply sync with other people that don't have the same tags. They could even be comp team mates and just not have the same tags. I played with the same team for years without having the tag. And of course there is the good ol' alt accounts.

A better ranking system is what could actually fix the issues. That plus either mixing queues or allowing solo opt in could actually make a difference.

[JediPanther]

ai bots. problem solved. Return of the atlas vip. With lrms this time.

[MischiefSC]

justcallme A S H, on 02 January 2020 - 09:51 AM, said:

I stopped reading here because that is absolute garbage.

In what world is it fair to have 3 x EmP on one side against a totally lacklustre enemy team?

It is a very VERY simple issue/calc and code to fix to balance it all. It wont be perfect, never will with PSR that is fundamentally flawed - but it will help match quality - sometimes.

How many members of EMP are there? What percentage of the player population are they?

That's the math you keep missing. The top 100 players represent 0.67% of the player population. As in 2/3rd of 1% of the population. And that's the TOP 100 players. As in approximately the 12 best teams in comp play, all together.

You're absolutely right that you could make some basic changes to the matchmaker - though don't assume that's simple and inexpensive - and try to weight outliers value regardless of their tonnage, etc.

However that's still only going to impact about 7 out of every 1,000 total matches played (and that's assuming every single match that top 100 players play in is better balanced), to help put how small the top tier population is. Sure, for that top 100 players it feels like every game but for everyone else? Not so much. So from a cost vs reward benefit for PGI, how's that going to look?

So the question then becomes what's a simple change that will matter to the majority of players and that would likely be something related to a mechs value (FFS I'd even take its Solaris VII ranking or whatever if they don't want to create an Elo by chassis, which they should have the data to do already based on metrics they've released before) as a value and average it with the players value.

If you're already opening the box to make a change then you can throw in the changes for the top and bottom 1% because you're already investing in changing the MM, that's just a small piece of what you're changing. A change just for the statistical outliers though? I wouldn't expect to see it as it's likely not (to PGI) worth the cost.

[Khobai]

UnofficialOperator, on 01 January 2020 - 06:59 PM, said:

I used to think that ELO is the way to go but I am no longer sure.

Pretty pointless now.

ELO is not the way to go. Because ELO requires a high enough player population to match players up. Without a high player population youll have to have higher skilled players play lower skilled players anyway which defeats the purpose of ELO. Might as well just keep the tier system since ELO cant possibly work anyway.

Theres nothing obstensibly wrong with the tier system other than the fact tier 1 has become way too saturated and they need more tiers to separate out players. So they need to add more tiers or reset the tiers or a combination of both.

justcallme A S H, on 02 January 2020 - 09:51 AM, said:

In what world is it fair to have 3 x EmP on one side against a totally lacklustre enemy team?

Which is exactly why EMP shouldnt be allowed to stack on one team. Given the lack of good players in the queue, the only viable solution to that problem is for the matchmaker to forcibly break up groups and redistribute the players to create two roughly equal teams. That can be done either by limiting group sizes to 2-3 or having the matchmaker forcibly break up larger teams when deemed necessary for game balance.

Although Group queue should honestly just be dissolved at this point. Its a miserable experience for everyone involved and PGI will never fix it. So they should just get rid of it.

PGI should focus on eliminating as many buckets as they can so the remaining buckets get filled up as much as possible. Because the more filled a bucket is the better matchmaker will work. At this point they should only have two buckets: quickplay and faction play. The current player population cant support more than that.

Edited by Khobai, 03 January 2020 - 12:20 AM.

[UnofficialOperator]

Khobai, on 02 January 2020 - 11:50 PM, said:

ELO is not the way to go. Because ELO requires a high enough play population to match players up. Without a high player population youll have to have higher skilled players play lower skilled players anyway which defeats the purpose of ELO.

Nah that wasn't really my point. In Modern Warfare there are millions of players, so population isn't the problem. And yet, a lot of people just don't like SBMM for casual play because no one really likes to be "sweaty" where every game, you have to be on your toes. And this is how Modern Warfare has implemented SBMM for even casuals in the last few months and there is a bit of an outcry.

ELO works as a mechanic for ranked play and most are fine with it.

[Prototelis]

https://mwomercs.com...eed/page__st__0

[Paul Meyers DEST]

Prototelis, on 02 January 2020 - 04:35 PM, said:

Should just let solos opt into GQ just like the FPQ.

Give better elo matching for groups, and a group counter, and I guarantee people will play.

And make it 8vs8,