Match Making

[graceful scaffold]

right

so that's this

ranked assumes you already know how to play Predecessor

you're not a brand new player

you can be assumed to be average until you prove otherwise

[thick ether]

But, if you were a good otb player, and started on chess.com, went 7/3

[graceful scaffold]

you'd be like 1300

[thick ether]

You would still be playing against 1300s

[graceful scaffold]

and player 1300 players with thousands of games

[thick ether]

Not 1600s

[graceful scaffold]

I know, because I'm a 1300 with thousands of games

[thick ether]

This guy's playing against 1600s

[graceful scaffold]

no he's not

[thick ether]

How? He had 4 games played, won lost won lost

[graceful scaffold]

he's gone +4?

not won lost won lost

[thick ether]

After 4 games he got matched with gold 2

And he won lost won lost

[graceful scaffold]

because he started around average MMR, and you're able to party with folks that are above or below you, and MMR does not take a strict rating of +/- 10 MMR points to put together a match

If a gold and a bronze player party together, they average out to about silver, right?

Gold players are around average, the player here started at the average, and they maintained that average rating through their first few games

[thick ether]

And that's why he was in our lobby, because he was given an average mmr

[graceful scaffold]

Well by definition average IS the starting location

because math and stuff

You take two 1300s and put them in a game. One wins, one loses. What's the average MMR after the match?

[thick ether]

Because he wud need to have won like 20 odd ranked games in a row to get gold 1 or plat 3 right?

[graceful scaffold]

1300 because it's a closed system

Do you mean by MMR, or by actual achieved rank?

[thick ether]

Either

[graceful scaffold]

If you go 20-0 I think you could maybe get up to gold 1 by rank?

By MMR you'd probably be a lot higher honestly

[thick ether]

19 MMR a game on average?

You think so?

[graceful scaffold]

Where do you get 19?

[thick ether]

Vp sorry

[graceful scaffold]

Yeah I mean your first win is 35 I think, or something like that

The standard "even" VP gain and loss is +/- 18

[thick ether]

How many vp are in a level?

[graceful scaffold]

100

[thick ether]

So, from bronze 2 to gold 1

[graceful scaffold]

You start bronze 3

[thick ether]

I mean, this guy was bronze 2

For him to get to gold 1, he would need 600vp ish?

[graceful scaffold]

Something like that, yeah

[thick ether]

Divide by 18

[graceful scaffold]

You get more VP when you're underrated

18 is when you're stable

[thick ether]

It's an average

That's how many wins that guy would need

To be on my level

30 wins or so?

[vague saffron]

Sorry, but Predecssor match making has been horrible. Yall need role que. No excuse, when deadlock easily allowed 4 hero selects in a sub 10 minute que. Throw fill after X time.

[graceful scaffold]

20-30

[graceful scaffold]

Not now papaya

[thick ether]

Do you see my point now @graceful scaffold

[graceful scaffold]

What point?

[thick ether]

You still don't see where I'm coming from?

[graceful scaffold]

I'm not getting it

[thick ether]

Fair enough

[graceful scaffold]

Try again?

[thick ether]

Thanks again for going through it

Nah DW man

[graceful scaffold]

No I mean try your point again

[thick ether]

No it's ok

[graceful scaffold]

If I'm not getting it don't just give up

What am I missing?

[thick ether]

To get to gold 1, you need about 30 wins with no losses to get there. Predecessor is matching guys like that up with guys with 7 wins and 3 loses total

[graceful scaffold]

You're equating rank and MMR again

They're not the same

MMR is a measure of strength of play

Rank is a measure of sustained historical success

Think of it this way: You're a chess IM at 2400 strength playing a talented young kid that's been flying up the ranks and is also 2400, but has no norms yet.

You're the same rating, even if the kid hasn't done the "achievement" part of the title for that level.

[thick ether]

How does he get to 2400 with 10 games played?

[graceful scaffold]

He doesn't

It's a metaphor

[thick ether]

I know

[graceful scaffold]

about the difference between strength and rank

[thick ether]

I understand

The 2400 would have to have played a lot to be 2400, no?

[graceful scaffold]

Ok let's go back to the 1300s because that metaphor didn't land

I am a 1300 with a thousand games played

Your friend is a 1300 with 10 games played

Chess.com will put us together for a match, no?

That is this

[thick ether]

How did he get to 1300 with 10 games played?

[graceful scaffold]

Because he started at 1200

and went 7-3

[thick ether]

So what MMR are people given if they have never played in ranked?

[graceful scaffold]

in our system, 1300

[thick ether]

And what MMR is gold 1?

[graceful scaffold]

I don't know exactly

[thick ether]

Roughly?

[graceful scaffold]

something nearish but above that

No I literally don't know

I didn't design the system or the matchmaking

[thick ether]

Well, it must be 135/1360

[graceful scaffold]

I just understand the moving pieces and am willing to hang out at 11pm talkin with folks about it

[slim wave]

You're a goat for that icl

[eager saffron]

I am convinced now, that your main job is Moderating MM posts xD

[normal rose]

1350

[graceful scaffold]

Heh, that's certainly how I spend an unreasonable amount time after I finish work

[eager saffron]

Thats Gold 3, Gold 1 is 1450

[graceful scaffold]

I do real work during the day and hang out with you folks during lunch and after work

[normal rose]

my editor trolled me

[eager saffron]

You should be paid per Message

[thick ether]

Is it plus or minus 10mmr?

Or 100?

What are the allowances for matchmaking?

[graceful scaffold]

It's that thing I described above

your own range expands and so does everyone else, until 10 people have overlapping ranges

[thick ether]

I couldn't find it sorry

[graceful scaffold]

The total range of a match depends on the region, day, time of day, MMR range, etc.

[thick ether]

Ahhhhh

[graceful scaffold]

usually it's probably about 20-30 top to bottom for the center of the distribution, and 100ish for the top MMR games?

[thick ether]

The truth is coming out!

[graceful scaffold]

What do you mean?

The statistics of the available population are going to depend on the population queueing at that time? And more people play on a saturday evening than a wednesday afternoon?

And also be variable relative to the z-score of that subgroup relative to the total distribution?

[thick ether]

So, the reason a bronze 2 and gold 1 get matched is because there was no one online, and it's squeezes him in, because he was just outside usual tolerances

That's what happened

[graceful scaffold]

Nope

[thick ether]

It is

[graceful scaffold]

...it's not?

[thick ether]

But you just said he's about 1300/1350

Playing with 1450#

[graceful scaffold]

I did not, no

[thick ether]

No?

[graceful scaffold]

they started at 1300 and are somewhere above that

What's your account name? I can pull MMR right quick if you're ok with it

If not that's fine, but it'll help illustrate the point I think

[thick ether]

They started at 1300 and are somewhere above that, so 1350

He's plus 4

No?

[graceful scaffold]

they are likely higher than that

they were on a 4-game winning streak

[thick ether]

But if they were higher gold 1 wouldn't be 1450

They would be 1350 ish

Playing with 1450

[graceful scaffold]

Are you 1450?

[thick ether]

The game squeezes them in

[graceful scaffold]

They were 1382 at the time of that match

if you get me your username I can pull your MMR real quick and we can confirm the theory

you can DM if you'd like

I don't have to post the number publicly

[thick ether]

What number?

[graceful scaffold]

your MMR

[thick ether]

Your a comedian

I'll give you that

[graceful scaffold]

What?

[thick ether]

Goodnight

[graceful scaffold]

No I'm being serious

I'm trying to help

[viscid condor]

💀

[graceful scaffold]

what

mans just dipped 🤷‍♂️

[viscid condor]

what mine

nahnah

i didnt understand what the conversation was aobut

was it how 2 1300s got matched up

[graceful scaffold]

Whelp, that's what I get for trying to spend two hours answering someone's questions

[viscid condor]

answer mine pretty please

wif a lil cherry on top

[graceful scaffold]

Good thing I get to go back to being yelled at by the community for a matchmaker I didn't build because I'm trying to help teach people how all this stuff works

💀

[thick ether]

Your not exactly lakenator broski

[graceful scaffold]

What?

[thick ether]

Just your tone I think

[graceful scaffold]

No I'm really trying to be helpful here

[thick ether]

But you know that already

[graceful scaffold]

I've spent an hour and a half on a wednesday night in my free time, I'm really trying to help or I wouldn't be here

[compact geode]

MMR wise how much it can change after a win or a loss? Like if you get a winstreak it will go up quickly or depends also with the people inside the lobby you played? For example a person is stuck at a rank and gaining 15 vp and losing 15 vp, overtime he improves and starts winning and for example his WR goes up to 55/60 %, will mmr (and vp) reflect this?

[graceful scaffold]

I think average is like...12 to 15 per game? Something like that once you've stabilized

First few games it's way more

but to your point, you can totally go up or down by some amount of "ranks" (according to MMR target) with a strong win or loss streak

[vague saffron]

just had paragon game, this one i gained less for the first time

https://omeda.city/matches/eb9a3ab7-8e90-4295-b83b-09cfd4897aa1

wild game youutbe vid tomorow

but like in that game

if we just use the average VP

its a pretty big disparity , wish i got 21 😦

could you see what averfage mmr was for that game?

[graceful scaffold]

I know this one without checking, since DizKraze was in a game someone had me look at a few days ago

They're super underrated by rank and haven't played a ton of games sinze we fixed VP gains a few weeks ago

If you subtract them out, the average is 1638 to 1625 by VP

But also counting by VP in Paragon games gets silly, since cold is 2100 and you're 1500 but you're the same rank, it just gets impossible to keep things any kind of comparable when there's a bigger gap between 1 (coLd) and 26 (you) than between you and someone in gold 1

Plus if anything, y'all were "underrated" by VP but won pretty handily

[sacred ice]

i feel like kinda bad for J ... idk how that all is so hard to grasp for a lot of people ... i mean you could say they need to work in performance somehow but other then that i dont seee an argument heree

[pallid sable]

It’s very very hard to work in performance without having some sort of role based MMR

[sacred ice]

i know

but you kinda could bring in the argument, idk about how hard it is im the consumer you need to provide

thats not my standpoint to be clear

but thats what i would consider a logicly sound point

[pallid sable]

Yeah I agree

Tbh there’s a lot we don’t know about ranked

[sacred ice]

mybe you could also ask for some more transperency or increased or decreased gains in MMR

[pallid sable]

Would’ve been nice to get some clearer guidelines and how it works. Who you can queue up with etc.

[sacred ice]

or thighter bracketing

[pallid sable]

Yeah agreed

[sacred ice]

but in my case, i dont play a lot of ranked till now and if i do i play with a friend so i know im in lobbys where i dont belong most of the time, so i proberbly skew matchmaking a lot too

and it often sucks ass, cause my games when i duo feel crazy unbalanced at times

but hey i see why that is the case

and id hope that is what more people would see

[rare quarry]

that's what I meant by asking "what ever MMR means"

The value alone has no information included of any if they are good or bad in anything. MMR alone shouldn't be used to select players for a match.

[sacred ice]

why not?

it tells you vs wich players he usually winns or loses

if he did play enough games that is

[iron depot]

So then 40 people get express priority to potentially ruin peoples experiences, because checks notes those 40 people are the only ones who exist in their rank?

I feel like there is definitely a better way to balance the system out than shoving them in with players of a much lower skill bracket.

[rare quarry]

you can read that someone is a good Kira carry just by looking at MMR 1742?

How bout this guy: MMR 1741? Is he a good Serath jungler?

[sacred ice]

nah i dont

1. id hope everyone who is out of silver has some knowledge of the hero they are picking

2. id hope in draft they are asking for roles they are self asssured in

[iron depot]

To me, it sounds like this is what the system was meant to do in the first place.

It sounds like this system is intented to get people to spend as much time playing as possible in order to get actual fair games.

[sacred ice]

how else could it get the data required to do so

i feel like its obvious that results would improve with games played

[iron depot]

I mean yeah, obviously. The results would improve with games played. However, I feel like if I have to spend a ridiculous amount of time in order to get games that properly respect my skill/time - I'm probably just going to find another way to spend my time.

Its fair that they need time + games played in order to improve how fair the matchmaking can be. At what point does that curve of time + games played become a bit too much though?

Why would I spend my time playing 200+ games just to get to games I can finally enjoy, when I could just spend my time actually doing something I enjoy?

[plain plaza]

Shouldnt both MMR and VP be based not on just win/loss but on PS in match? If a team wins a game the guy with 40 kills and 400 PS will have same gain as guy with 0 kills and 30 PS.

[graceful scaffold]

That statement really only applies to the top of the top of the top folks trying to separate themselves into Paragon-only lobbies, which is happening already.

Creating a gap from the rest of the community

[plain plaza]

I remember long time ago i played Bloodline Champions and there was score in rating games. You perform better - you gain more rating when you win and you lose less when you lose. It was simple and effective, sadly the game died.

[graceful scaffold]

Answered a bunch of times in a bunch of places (not your fault this thread is huge), but tl;dr any system that tries to fudge things like that is going to be less accurate than just seeing whether you won or lost. Too many invisible variables way beyond kills and deaths.

You can't "score better" because statistics don't really measure contribution

[iron depot]

I mean I've been avoiding playing the game for nearly a month now. I only played a single game like a week after 1.0 drop just to see how everything was going. What Sei described, is basically what I've had to deal with in many of my games.

I will pop on for a couple games, have to wait 10 minutes or so for one - only for that lobby to disband because some child didn't get their desired role, wait another 10 - 15 minutes for a game, only to have that game be an absolute nightmare that I did not enjoy at all. Either because someone on my team threw or dc'ed, or because the other team had a solid 500+ mmr up on us + our worst guy could not stop actively throwing himself into the meat grinder.

I am at gold and playing in casuals, if these are my experiences and I'm hearing someone in Paragon having that same experience - What is my incentive to even bother with 1. Ranked and 2. the game in general?

[graceful scaffold]

I promise you that you weren't in a game with a 500 MMR gap

[plain plaza]

Well in bloodline champions the final score was calculated of you stats during match. You do more damage, do more CC or healing - you have higher score. Based on that you gain more rating than dude who has done less.

[graceful scaffold]

Totally valid complaints about queue times or players disconnecting or whatnot and that sucks.

[iron depot]

I mean I don't feel like going through the folder rn, but I have plenty of screenshots of my games where the enemy team was rated 500 mmr above us. I know OC isn't the internal MMR and not what you guys use, but if that system is a bit more accurate compared to Glicko's - why shouldn't I be using that as my metric?

[graceful scaffold]

Those systems aren't predictive of success though. Which is more valuable to their team, a support riktor that hits his hooks but always dies in teamfights or a jungle rampage that's excellent at timing objectives but bad at farming?

The number of play styles and variables and heroes and roles are immense, and no human can come up with a formula for "percentage of win added" that will be more accurate at predicting future games than just looking at actual wins and losses, that by definition are the sum total of all of those invisible contributions along the way.

[graceful scaffold]

Because OC is not internal MMR. I still would be very surprised to see a 500 OC score difference unless you were playing with "GM" players, at which point if we're being honest OC's rating system becomes impossible to compare because the gap between #1 and #2 in their system was like 1000 points.

Or if those screenshots are from before march of this year

[iron depot]

No, not GM, but the one game I do remember in particular these were like mid-high teir diamond players while I'm in gold I. Like...huh? The one guy could've been close to GM - his PR was like 1749 or something.

[graceful scaffold]

Casuals don't have queue restrictions for parties

1750 on OC's system was diamond 2

I was there, and partied with friends in gold and silver all the time because they were my friends

There's not the same queue freedom in ranked, for just that reason

1 duo partner, and you need to be within a certain range of each other to queue

[iron depot]

I mean in casuals it should still be stricter imo. I don't feel like I should be playing with anyone beyond plat - regardless of how partying and queue restrictions work. Having to fight someone two skill brackets above you in your lane is just insane.

[graceful scaffold]

But then if I'm "D2" in casuals and one of my best buds picks up the game, how do we play?

A smurf would be worse because then I was hiding my actual strength from the matchmaker

[iron depot]

I mean its not like the game keeps you from smurfing, not that I condone it - but isn't that what most people usually do in every other game?

[graceful scaffold]

Strict matches are what ranked is for

[graceful scaffold]

But it's directly bypassing the matchmaker to make an objectively worse match. If you'd rather pretend a diamond player was silver that's fine I guess, but it doesn't change the actual skill of that person and puts them in a less balanced game

Casuals make the best-balanced match it can with unrestricted parties, while ranked is there with stricter bands and limited parties.

[iron depot]

I mean, I'm not saying casuals should be exactly like ranked. However, I can imagine how frustrating it would be for someone newer to the game to have to deal with some of the matches I've had to go through. Tbh, if I wasn't as much of a fan of Paragon and wanted to support Omeda, some of those matches absolutely would've made me put down the game. It'd be discouraging to find out that I can go up against someone THAT much better than me in casuals. At that point I would be incentivized to play ranked, but ranked is limited.

[graceful scaffold]

Ranked currently covers 90% of the played games a day, when you look at the population curve. It certainly sucks for those that work odd hours or are otherwise limited when it's off and I sympathize, but for the vast majority of the player base it's available for the vast majority of their usual play time.

[iron depot]

How long does ranked run again now, I haven't been keeping up. Is it 12 hours a day now?

[graceful scaffold]

14 on weekdays I think, 16 on weekends? Something like that

Every day

[iron depot]

I guess that is a good majority of the day then, but that kinda worries me to hear that 90% of all matches in a day are taking place in a 14-16 hour window and that window is only when ranked is available.

Its a stretch, but it could be a sign that something is wrong with casuals and that perhaps ranked is the more enjoyable experience because of its stricter queue.

But then again, I don't know what the statistics are for similar games, so that could be the right percentage for that sort of thing.

[graceful scaffold]

Nah, just the reality of sleep and stuff

You could say 90% of people are asleep between midnight and 8am, right?

If that's true, they're also not trying to play pred

[iron depot]

Well, yeah but doesn't ranked only happen on 1 server rn? Or did they change that.

[graceful scaffold]

EU and NAEast

Separate hours

[iron depot]

Ah, so it got open to two thats good. That makes more sense then.

[graceful scaffold]

Nope always open with two

From the get-go, but yes the hours are different in each to match the population

Ranked is closed when peeps are asleep

[iron depot]

I thought it started out with only one? I thought I remembered complaints about peeps not being able to play ranked due to the server times elsewhere or something.

Maybe they mightve been SA.

[graceful scaffold]

And ultimately it expands to cover the night owls too, but that's a very small group of people

[graceful scaffold]

Not currently ranked in SEA or NAWest, yep. NAWest can queue into NAEast without too much trouble, but we don't have a great option yet for Asia-Pacific.

[iron depot]

I might've been remembering someone who was NAWest having an issue with that and just misremembering it as only one server or something.

[graceful scaffold]

Yep if you're in California the hours will be a few off alignment relative to NAE in that it'll start a little early and end a little early relative to most people's work/school/sleep schedules, but even now there's still tons of primetime overlap for those folks.

Ok I have the day off and need to go make breakfast so I'm going to dip for now, if you've got other questions leave them here and I'll cycle back later.

[iron depot]

Still, I always appreciate the insight and conversation you can give into the internal workings of the system. It's always fascinating. I hope you enjoy the breakfast and have a safe day off.

[plain plaza]

Here i am. Ranked starts when i have 8pm. So its 3-4 games before i have to sleep. Eu servers.

[vague saffron]

did someone delete my message?

[iron depot]

Which one? I was reading through and catching up so if it was deleted recently, I might remember

[vague saffron]

saying i gapped the enemy team.

i think it got deleted

[iron depot]

Ah yeah that might've been one of the ones deleted, there were a few messages I couldn't find responses to when I was reading

[vague saffron]

moderation in this disc is just so bad now.

[iron depot]

Extremely heavy handed.

[vague saffron]

oh wait bring it bac to topic before we get timed out

match making needs some work

[iron depot]

It really does imo. Especially in casual. I'd like to know the percentage of how many players, during the time where 90% of the games players are active, are playing ranked and playing casual respectively.

Because if the ranked% is exceptionally higher, I feel like that could mean casuals could use some work. And imo, from all my experience I would say it definitely does. I'm by no means a terrible player, nor a great one. I'm very average/slightly-above average.

If I'm getting paired up with people a few positions away from Masters while I'm in mid gold, that's kind of a problem - regardless of the fact that it's casuals. I would still like to have a fair match up, especially if I'm locked into that match for 30+ minutes and will take a growing penalty if I decide to leave.

If casuals is meant to be so casual, then there shouldn't be a penalty for leaving or something. Which is not what I want, but it's an alternative if we just can't get slightly stricter matchmaking. I'm already waiting 10 - 20 minutes for a single match anyway, I'd rather take 5 - 10 more if I have to for a game I truly enjoy.

Its not fun going up against bronze players in the 200's and it's not fun going up against diamonds in the 1750's either. I don't want either of those. I should be with platinums and silvers at worst/best if the system really needs to match me up with someone.

I can understand why it reaches so far out at the higher ranks, but it makes little sense at populations where the most people are.

[tulip trout]

So I kept up to date with the whole thread and thanks to JShreds for answering every question so patiently.
I understand everyone starts at 1300 internal hidden MMR. I kinda don’t understand why.
When u look at this picture u can see two bronze guys who have 6 games between them who got matched with silver/gold players who spent way more time on ranked and my guess is just because they have a winrate of 50%.
This is one example of many what gold players experience right now.
And I can tell from listening to others, that this does not feel fair for people who spent way more time, climbing in ranked to get matched with others who have just started playing ranked.
Just my opinion, maybe worth looking into.

[native pasture]

i Have been reading most of the post, and understand the difference between VP and MMR. But one thing is confusing me, WHY dont you show the MMR as the Ranks??? it would be much better for people to understand and see that the matches are "fair". if you hide the real skills of a player i don't see the point in Rank matches.

Wouldn't it be easier to just start at silver 1 or whatever like any other game, if the MMR is doing exactly this right now?

[graceful scaffold]

Because MMR by its nature is designed to be more flexible and mobile than rank, which is designed to be more "consistent" relative to your total performance.

If you were right on the border and going up and down a rank every match it would feel pretty rough, or if you dropped 3 ranks in a night on a particularly tough loss streak.

Would similarly show someone as a high rank if they started off really well because a friend carried them, even if they didn't "earn" it with consistent play at that level.

[covert prairie]

I think eventually some of these concerns could go away if future seasons have placements assuming at the time ranks are filled up with a decent amount of players so players can accurately be placed

Rather than every season everyone restarts at Bronze 3

Which I assume would only happen if MMR would also reset with it

[graceful scaffold]

I don't want to say those kinds of placement matches are a bad idea, but somewhere 1000 comments above we had a discussion on it. "Strict" placements (first X matches are special) suffer the problem where they HEAVILY bias someone's ranked season and willingness to play based on a very small sample size of matches.

[covert prairie]

If we hard reset I don’t mind but in my opinion it can’t be every 2-3 months

[graceful scaffold]

You mean hard-resetting everyone back to baseline every few months?

Yeah I'd hate that too

[covert prairie]

Yeah

[graceful scaffold]

I don't think/hope that's in the plans for future seasons

[covert prairie]

But with soft resets I imagine you’d just reset Paragon players back to like Diamond 3 for example

[graceful scaffold]

I think that's how most games handle it, as more of a "crunch" mixed with higher initial MMR variance than a full reset back to baseline

[covert prairie]

Yeah

Overwatch does do that, but with placements

Meaning they can’t place lower than that rank

[dim carbon]

Marvel Snap does it

[slim wave]

in ranked, does duo queueing make matchmaking different?

[pallid sable]

this, we just had 3 matches where we faced 3+ plats all game being silver. Would be good to see the internal mmrs for it

https://omeda.city/matches/f3b65c17-e92f-481a-a686-e6b850d06bac

[hushed wave]

This is what I brought up with my example too, except in my example we had to go against tip top players.

How I understood it is that your MMR is super inflated, and your VP hasn't caught up. In your case you don't have that many ranked games so that is understandable that you are still climbing VP.

The "super" part of the inflation part is from them tweaking the MMR algorithm which quickens peoples climbs to their appropriate level. Whatever tweak they did to the aggressiveness to MMR changes threw people with long winstreaks (in your case 11 wins in a row) against much higher MMR players.

Don't lose hope, you'll drop to your correct MMR just as fast as you skyrocketed to that high of MMR.

[pallid sable]

makes sense

i hope i dont drop to "my" MMR tho haha

[hushed wave]

To beat the point to death.

Before MMR increased slower than now, so you were able to win, win, win. They then suddenly made MMR increase faster, which backfilled all your games to make you much higher MMR than you were before they made the change.

[graceful scaffold]

You're only silver because you've played 18 games and won 14 of 'em

Won't be silver for long

nor should you be

don't even have to check MMRs to see that

[graceful scaffold]

Nope, MMR hasn't changed. What changed were VP gains relative to MMR, so you go up in rank faster if you're heavily underrated.

[pallid sable]

makes sense was just a bit confusing as to why we were facing a full team of plats all of a sudden, as when I was solo it felt more "balanced" in terms of the ranks (ik thats not how it works but you get my gist)

[graceful scaffold]

Do you think you should be in silver?

[pallid sable]

no defintely not

[graceful scaffold]

Yeah exactly

The game is trying to place you with your people, and your rank will catch up.

[pallid sable]

gotcha

[graceful scaffold]

You could play against only silvers and destroy them, but then your MMR gains would be restricted and it would take longer to pull you up to where you should be.

The system is "testing" you basically against higher opponents, because it thinks you deserve to be there. With a big gap between where it's placing you (MMR) and your rank (VP), it'll close that gap quicker.

You probably see that you're gaining more VP per win than you lose per loss, by a considerable amount. That's the system moving you up by ranks to where you should be.

[pallid sable]

makes a lot more sense now, thank you

yeah with the 3 matches I played today lost 2 won 1 and still had a net +10VP so makes a lot of sense

[slim wave]

ty jshredz
i was very confused lol
i was blaming onion for rigging my games

[graceful scaffold]

Exactly

[pallid sable]

ty ty ❤️

[graceful scaffold]

The game knows you're not bound for silver for long, so it's trying to get you up out of there. Once your rank matches the MMR target you'll get balanced gains, so it's to your benefit to get in harder and harder lobbies and keep winning so you keep getting imbalanced VP gains.

And also more fun matches that better suit your skill

[pallid sable]

very true

[slim wave]

hm but that feels a bit counterintuitive icl
i understand making someone who should be plat but is in silver, play against silvers would be unfair but they would get out of silver faster if they played against silvers and gained increased vp
valorant for example does double rank ups for some niche cases

but it would lead to worse game quality
and in a ranked system without placements, i can see the current way being the better way

[pallid sable]

nah cos VP gains are based on who you play

so if we're silver and beat plats

we get a super high VP gain

right?

[graceful scaffold]

VP is based on your MMR, and your MMR is based on who you play

there's an extra layer in there, oh oniony one

[pallid sable]

gotcha haha

[graceful scaffold]

MMR says "I think this person is plat, let's put them in plat games"

at the end of the game, the system does the normal MMR adjustments based on that lobby

and then says, "how far away is their rank from their expected rank at this mmr?"

[slim wave]

ahhh right cos that way would take a dynamic vp gain system depending on how you play

[graceful scaffold]

it then adjusts VP gains/losses based on that discrepancy

if you're underrated, it assumes it was an unlucky/bad game and goes easy on you because your MMR still says you should be higher

or if you win, it says "I knew it! Give that person some big points and lets keep climbing."

once you reach the rank that is expected of your MMR, gains/losses are even because you're where your current skill says you should be

[slim wave]

ah i see
thank you

[graceful scaffold]

MMR is essentially saying "I think this person is XXXXX rank, but I want them to prove it"

If you hold your MMR at that level, the VP gains will pull you up to that rank

but you need to actually play the games and prove the MMR you reached wasn't a lucky streak or fluke or whatever

and if you play enough, you'll get there and can say for sure you earned it

[tulip trout]

So why is the system testing someone who had a total of 2 games in ranked?

[graceful scaffold]

It's a slower system than some other ranked methods to get to the high ranks, but it ensures that the people there have proven it

[graceful scaffold]

What do you mean?

[graceful scaffold]

I need context amigo

[tulip trout]

Murdoch 4 Games, fey 2 games…

[graceful scaffold]

Yeah?

Everyone starts somewhere

[viscid condor]

looks like they did pretty good too

[pallid sable]

^

[tulip trout]

Why do gold 1 players with 100+ games have to play with people who are still trying out ranked?

[viscid condor]

if u ignore the vp the game looks mad balanced and even

those bronze players in reality are around that skill level

they just havent played a lot of ranked

[slim wave]

in ranked maybe
how many in standards?

[tulip trout]

So standard games count for ranked? That doesn’t make any sense

Those are two different game modes

[viscid condor]

no they dont

but that murdock and fey were probably undefeated

in ranked

so they put them in a higher tier match to see if they could perform there

[tulip trout]

They have a 50% winrate in their 4 and 2 games they played ranked, not a very big pool of stats

[vague saffron]

U guys fail to grasp though, that some people only know 1 role or character

U can get 3 main mods for example

Mids*

[slim wave]

no
but if they have a lot of games and experience in standards, theyll be decent in ranked too

[vague saffron]

And hell none of them can get mid and they kinda suck

[viscid condor]

murdock won 2 and lost 2 ranked?

[slim wave]

other than adding role queue
or making every person that plays the game ever to learn every role
not much they can do about that

[tulip trout]

I understand that, but many bronze players I meet in gold 1 don’t have that kind of experience, that’s my point, they should have a bigger amount of games before they get ‚tested‘ against higher ranks, if they are so bloody good, they will climb very fast and easy

[tulip trout]

Yes

[viscid condor]

but they did very well that game

they obviously belong

[pallid sable]

well the MMR they start with is higher than bronze thus they get matched with people above bronze, until they drop

[viscid condor]

that is very weird

[graceful scaffold]

That's how it works in standard, but we assume in ranked that you're an average player that already knows the game.

[vague saffron]

Role que would be a good implementation to help bias your pick

[graceful scaffold]

Role queue is so so complicated and a huge can of worms

But we'll still explore it

[vague saffron]

I don't get how y'all lowered draft req level for ranked

[graceful scaffold]

Like we're open to exploring anything

[graceful scaffold]

You mean to 20?

[vague saffron]

Yes

[slim wave]

heroes owned is still the limiting factor unless they spend money

[graceful scaffold]

Shouldn't affect you homie, you're in the big leagues now

[vague saffron]

Don't matter I care about the community

[tulip trout]

But that’s exactly the point, you assume that but that’s not the reality, people playing severing support for example and other weird stuff u wouldn’t do if u have a good understanding of the game

[vague saffron]

Level 20 sounds so low...

And to be placed at 1300 mmr

[graceful scaffold]

It's still like 50 hours or playtime or something like that

[vague saffron]

Lord

[graceful scaffold]

or 100? I forget

[slim wave]

20 heroes is a lot of grinding tho icl
especially cos free rotation heroes dont count

[viscid condor]

its a long time

[vague saffron]

Levels.

[tulip trout]

Thank you

[vague saffron]

Not heros

[slim wave]

heroes is the limiting factor

[graceful scaffold]

I mean I walked into Paragon (the game, not rank) and just kinda smoked off the bat because of Smite

[vague saffron]

Just buy them, makes smurfs to easy as well

[graceful scaffold]

I know a bunch of League folks did that too

[tulip trout]

Or money

Not skill

[slim wave]

i know very few people who will spend money on a smurf

[vague saffron]

I mean I'm hoping y'all's season 2 is better experience

[graceful scaffold]

My account for testing features and stuff has more than 100 games and still isn't close to level 20

[vague saffron]

Content creators do because they make money

[slim wave]

and even fewer people that will spend money on a new game they havent even reached level 20 in

[vague saffron]

Ez clips

But I guess how many will

[slim wave]

content creators are nowhere near the majority of the playerbase

[graceful scaffold]

This mans really got to Paragon and is so mad about it

[vague saffron]

Paragon is to easy. NEXT.

When ai bots

[graceful scaffold]

Relax papaya, lots of stuff getting looked at and improved for the future (including your feedback), I promise

[vague saffron]

Ill be waiting... Don't got much else to do but act like a rampage on stream

[graceful scaffold]

It's why the people love ya

[graceful scaffold]

I mean some people start ranked and crush, some people start ranked and struggle. On average, they're average.

[tulip trout]

U said u assume they have a good understanding of the game, and everyone who can read knows a sevarog support makes no sense and that’s the kind of players u often get when getting matched with new bronze players.
And Someone like that would never land outside of bronze if he plays that way, so why place them with higher ranks?

[graceful scaffold]

Are you sure they wouldn't?

Or were they trying something goofy they saw a streamer do

I mean they don't actually get placed with higher ranks

they get placed at the average

(which, because math and stuff, is wherever people get placed)

If you need 200 or whatever pred matches to even access ranked, you have to assume they're not total newbies

[tulip trout]

Or someone going afk, getting banned a lot.
Let everyone play their first I don’t know 20 games in their rank and test them afterwards.
They wold not climb and would stay down there instead they get matched with gold and silver players

[graceful scaffold]

But that's where they start

Average MMR, and they get tested at the average

if they win more than they lose, they go up above average

If they lose more than they win, they go below

[tulip trout]

But why already after 2 games?

[graceful scaffold]

What do you mean?

They had 2 games at average MMR

[graceful scaffold]

what am I looking for?

[tulip trout]

Fey 2 matches 50%wr, Murdoch 4 matches 50%wr

[graceful scaffold]

Yep

So they're still around average mmr

probably like the rest of the lobby

also one of those players had 17 kills

[tulip trout]

I know, that’s not the point, even trolls start with average mmr

Let them be trolls for 20 games and u won’t see them in higher ranks

[viscid condor]

ok but why is 800 vp and 0 vp average mmr still

[graceful scaffold]

You start with an MMR equal to the average, and if you maintain that you end up with a rank in the silver/gold range I think

[viscid condor]

i see

[tulip trout]

Do u get my point?
Trolls even with a good MMR would get reported and banned and without them this game is so much more fun
So give everyone a certain amount of games before matching them with higher ranks, the players who have a „internally higher hidden MMR“ would be able to carry anyway so no problem there

[graceful scaffold]

I'm sorry, I'm struggling a bit here

If you mean a separate pool of only players with limited ranked games, you can't really test them or let their MMR go up/down, because they're only playing against people that also haven't been placed

[tulip trout]

I thought MMR is based upon stats?
So if they have a above average KDA and winrate over 50f why wouldn’t their MMR go up?
We don’t know what goes into the measurement of MMR because it’s hidden, so I can’t really speak to that

[graceful scaffold]

MMR is based on wins and losses

[tulip trout]

Only?

[graceful scaffold]

Well, relative to the MMR in the match yeah

not just "win + 10" or "lose-10"

[tulip trout]

Ok so they would go up for a win and down for a loss?

[graceful scaffold]

Yep

[tulip trout]

I‘m not talking about a separate pool, let them start in bronze and play with other bronze players who have settled in bronze

[graceful scaffold]

You can't start someone somewhere and not have it be average

Without a super funky biased distribution (which we do in standard, but doesn't work for ranked)

[tulip trout]

Why?

[graceful scaffold]

Two people enter at 1300

One wins, one loses

The amount of MMR gained is equal to the amount lost

What is the average of those two players afterwards?

[tulip trout]

130ö

[graceful scaffold]

rinse and repeat infinite times, what is the average?

still 1300

If we started people at 1000, the average would be 1000 for the same reason

[tulip trout]

But still, good players with an above 50f winrate would climb

[graceful scaffold]

correct

that's what happens

new players start at 1300, and if they win they go up

if they lose they go down

if they equalize at 1300, they stay at 1300 and occasionally get future new players appearing

[maiden relic]

So how come i get bronze players in plat lobbies is it bcs the game can’t find any gold/plat players ?

[graceful scaffold]

Are they really bronze players or do they have like a 80% win rate in their first 20 or whatever games?

[tulip trout]

I know that this is what’s happening, I was just showing the problem with that, getting trolls who have 4 or less games in ranked in my lobby, this shouldn’t be a thing after having made over 100 ranked games and having settled in gold 1
As I mentioned if they would have a longer phase before getting „tested“ u would not find those players in higher ranks than bronze
I guess I just have another understanding of how new players should be placed in ranked.
Or maybe I just suck at this game, haha

Still thank u very much that u have tried to explain this to me, I’m still amazed of your patience 😉

[maiden relic]

Idk haven’t checked it

This is gold tho

[graceful scaffold]

I mean if bronze was average, there wouldn't be anything below average though

[maiden relic]

But i had it happen in plat

But i notices when i just restart mm after 1 min i get higher ranked players instead of the bronze and silver

So every time i hit 1 min of mm i just restart it

[maiden relic]

You see that shinbi for example she was clearly new to the game

I had a bellica same skill

[tulip trout]

While I was reading that I thought, might be a good idea, I’ll try that

[maiden relic]

Not knowing the game

Sitting still not last hitting etc

[tulip trout]

She just had a bad day…

[maiden relic]

Thats not someone that has 80% win lose right ?

[maiden relic]

Idk

Not listening to pings etc

End game she had killed 17 minions

We won but ye

I just called her out like everyone else on the team lol

Got muted again

For like 7 days or longer idk

Thats why i wanna know if its normal

To get bronze players in plat

[woven bay]

im having the same issue i put in a lot of work to solo to gold two. i finally get there and all my matches are with bronze players or people who havent been placed yet. they go double negative and surrender my games, but everyone on the enemy team is gold/plat??? ranked matching is hot dog shit

[graceful scaffold]

Looks like the same thing, that game is "centered" around gold 3 based on the ranks

New player, average MMR

[graceful scaffold]

Same same same, new players start at average MMR and if your current MMR is around there or a bit above you'll get players around there, or players that are new to ranked but on the way up quickly

Reaching gold 2 is certainly an achievement, but in terms of raw MMR you're still on the way up and may be placed with newer players that are also rising

[maiden relic]

How gold 3 when i was gold 1 in that game

[graceful scaffold]

Centered, you were at the high end of the game

Every game is going to be a distribution of some type

[maiden relic]

So how is this possible then

The zarus has 200 vp

Same for the morigesh

[graceful scaffold]

it's the same thing, it always is

they had the two highest KDAs on your team

[maiden relic]

I didn’t play well that match so can’t blame them tho

Ye ik

They where good lol

[graceful scaffold]

I mean, question solved, right?

[maiden relic]

Just wondering why they ended up in the mm y know

I don’t mind if they can play good

Good coms etc

[graceful scaffold]

That's the whole point of the system

Their MMR was high enough for that lobby, and that indicates they were good enough for that game

[maiden relic]

Would love to see more pings like enemy blink on cd etc but thats a diff topic

Also bellica skin enabled when ?

[graceful scaffold]

they haven't played enough games so they're not just automatically awarded "gold" or "plat" or whatever ranks without more time spent proving they're that good, but their MMR put them in that match for a reason

[maiden relic]

So why don’t you guys do placement

Like apex does

For example idk

[graceful scaffold]

Because everyone started from the same place this season, because it was the first season

There was nothing to place against

[maiden relic]

Oh okay

[graceful scaffold]

Placements implies that there's a bracket you get moved up and down on to test where you should land, but if there's no ladder you have nothing to test

[maiden relic]

I see

All good tho

Its just frustrating some players don’t use coms

Even if plat

[tulip trout]

I thought MMR was only about win/loss?

[fast apex]

Hi, just one question, there's been tons of questions about VP gain and loss affected by the performance of the player during match?
Right now there is no reward for doing your best while you get matched with people who complain that people in ranked "sweat too much" (its supposed to be the top competetive mode, if they dont want to play their best then i think its not good to keep them at the same level)

[dim briar]

You should get a higher rank after some time, if you play better. If you have the same rank with someone always trolling you should have the same skill like him, when he is trolling.
Problematic are players who are sometimes playing their best and sometimes playing meme stuff, because they will be in lobbys with skill level something between both of that, so if they play for real they will somehow be smurfs and if they play some stupid fun builds they will lose the game for the team.
Besides of that even if meme builds are working for people on that elo, it is just not fun for players who are taking the game seriously.

[normal rose]

If you do your best you win more games, if you win more games you gain more vp.

[fast apex]

Ofc youre getting the higher rank after some time, but the grind just to have the grind is pointless. Not without a reason in other MOBAs its a variable and you loose less if someone leave the game or go afk

[normal rose]

if stats have impact in your vp changes it only leads to players trying to maximize stats to gain more/lose less

you are about to lose? quick, run to duo lane and take all their farm so you lose less vp

sup? doesnt matter, take all kills

your stats do indirectly affect your intermal MMR changes IF they lead to you winning more games

and higher MMR means you gain more vp

[fast apex]

From what I saw before ranked when people grinded your rank on OC they cared more in the games to not throw/int/afk

Games were higher quality for OC rank than currently. They just dont care now, they loose the same visible amount no matter what

[rare quarry]

@normal rose you assume the same stats apply to all roles. In this case yes, a support would be incentivized to take kills himself.

Role specific stats could indeed improve overall gameplay. Encourage the jungler to clear enemy jungle.
Encourage the offlaner to kill turrets elsewhere.

Since Pred tracks locations to each event, stats based incentives could be more complex.

Encourage offlaner to kill turrets else after he killed opponents offlane tower.

[normal rose]

that still doesnt change the fact that people will try to maximize their stats

doesnt matter which one you count or how you weigh them

an offlaner prioritizing killing his turret over helping his teammate not die is not something you want

if your stats lead to you winning more games, you gain more mmr

if your mmr is "too far apart" from your VP, you gain more vp

[dim briar]

It' s partly because of people not in their supposed ranks. With everyone getting to gold 1 for free and most people seemingly at the same rank it might be better to treat ranked as casual until it does not work anymore. Less frustrating for themself. So atleast they have fun, even though they have a teammate with 0/6.

[plain plaza]

The question is not about that. If oyu do 15-0 and your teammate have done 0-15 you have same VP gain if won.

And its wrong, it shouldnt be like that.

You shouldnt be equaled to worst teammate.

There are stats of damage and healing done, there should be stats of CC done, objectives and structure damage. PS should be calculated out of that and VP shouold be based on PS.

I think omeda's and many more companies problem is that american-european mindset that "everybody is a winner" when you have prizes for participation. While in reality the better you do the more you have. And if youre doing bad youre going broke. It should be a thing for our VP.

[normal rose]

You only gain the same if you are both close to equal in mmr

If that guy keeps going 0-15, he will lose mmr much faster and his vp will drop too

[plain plaza]

I mean dont you really find it wrong that total value of player is based not on personal performance but on vague win-loss ratio which in half the games doesnt show the correct picture?

I have 2 games here in first ive lost but ive done good and shouldnt lose vp. In second i won but ive done really bad and shouldnt gain any VP.

https://omeda.city/matches/a5cb5f8a-2b8a-43a7-9439-f3e88d03c86b

https://omeda.city/matches/abf3eadc-2358-4930-9bc6-4069eecc0b7b

[strange shard]

We dont care about matchmaking Right NOWWWWWWWWWW we want games improvements instead of skin !!! and then matchmaking will follow

[normal rose]

the whole point of MMR systems is to find the individual skills of players given a large dataset of matches, it doesnt need stats for that

cuz math

[graceful scaffold]

Tell me which is more important, a body-blocking steel support that sets up kills or a muriel that's great at timing their shields?

How about a muriel that's great at timing their shields, or a grux that bullies out their solo laner early?

Or a grux that bullies out their solo lane opponent vs. an iggy mid that doesn't get many kills but is always on time to burn down objectives?

There are SO many variables that add up to helping your team win or lose, and so many different valid playstyles on so many heroes in so many roles, that no formula can accurately predict your "win contribution".

MMR isn't about rewarding you for putting up good stats, it's about trying to assess your probability of contributing to a win at any given level of play so it needs to be as accurate as it can be. The only way to really do that is to boil everything down to whether or not you won or lost, and then over time building up a picture of how often you win or lose with or against expectations.

[iron depot]

Isn't that what PS is in a nutshell though? The various things that happen during a match - compared to the other people in the match - to give a score based upon their performance in relation to those other people?

Couldn't that just be used or would that be an innacurate form of measurement?

[graceful scaffold]

PS is neat, but I think @normal rose would tell you it's not actually a "win contribution" measurement.

It's an arbitrary formula that compares your statistics to other statistics, but how do you measure things like the quality of dekker cages, or shotcalling, or any of that stuff that can't be measured statistically?

PS looks at your damage, kills, farm, etc., but that's only a slice of your actual team contribution

If I'm a jungler that totally mercs enemy players over and over with tons of farm and kills but I just REFUSE to do an objective, I'm going to lose teams the game with a great player score

[iron depot]

Ah I see. So do you think there could be some alteration or form of PS that takes more things into account, that could possibly work? Or would it just be too far fetched?

[iron depot]

I've seen that one personally x.x

[graceful scaffold]

Take the examples above to explain why that's basically impossible. How do you measure not only one player against the statistics of other players in the same role with the same hero, but the relative contributions of different roles?

Also how do you account for things like shotcalling, communication (both in terms of using in-game pings but also keeping morale up and being positive), non-damage skills like dekker cage or good body blocking, etc.

the reality is, you can't, or at least you can't ever do it accurately

[normal rose]

Yup

[graceful scaffold]

the easiest and most reliable way to see whether or not you helped your team win is just to see whether or not your team won, and do that a whole bunch of times in a whole bunch of games

[normal rose]

Its okayish as an avg across many games but actually not great to compare players within a particular match

[graceful scaffold]

It's a super neat heuristic for "did you frag or get bodied?" and I use it at a glance too

it's just not reliable as a predictive method for future matchmaking

[normal rose]

Ofc not

[tawny basalt]

I think there is theoretically a way to analyze all player stats in relation to winrate and extract a few important metrics that are very predictive of your win outcome. So small things like body blocks, shot calling, etc. are magnitudes less influential than say, kills and gold and thus don't need to be directly accounted for. If integrated with current w/l systems, I think you could have a system that predicts your skill faster and more accurately on average.

However, a fundamental issue with this is that it tries to fit a "correct" way of playing onto the entire player base when one of the core aspects of a moba is the vastly different methods of play. So, off-meta players would be punished for not playing "correctly" in the eyes of the system, despite it still being a perfectly viable strategy. Even if this wasn't a problem, this system would be a vastly more complicated and difficult system all for marginal improvements.

[graceful scaffold]

Bruv you just called shot calling "small", I already got beef

Once you reach a certain level of general competency in play, decision making becomes MUCH more important (relatively) than mechanics, because everyone at that level has mechanics

[tawny basalt]

small in comparison to kills and gold as predictors of wr. Also, shotcalling in this game is way weaker than other because of... a lack of a certain feature. Idk what it is but it might start with voice and end with chat

[graceful scaffold]

Oh I totally disagree, in an actual sense of contributing to win rate

[iron depot]

I mean I'd say its relatively small. It's very helpful and can absolutely sway games in comparison to bad comms/none. However, I've won many games where that wasn't a shotcaller.

But yes, at the higher levels it is absolutely more integral than at lower levels.

[graceful scaffold]

I play most of my duo games with a buddy of mine that has no prior MOBA experience and limited hero pool, and he's still coming to terms with a lot of moba mechanics. But he is an EXCELLENT decision-maker on objective timing and contributes to wins far beyond what you'd expect from his damage or kills tats.

He can do 10k damage in 30 minutes but we win with 4 fangs and 2 orbs

[iron depot]

That's how I feel about my brother in lane. He has wave timings and everything down to a T. He is such a carry-brain in that way. I can't wave that same way.
I'm a support through and through T.T

[graceful scaffold]

The awesome thing about MMR is it automatically adapts to the "important" skills at any given level. If mechanics are important in bronze, then if you have good mechanics you're more likely to win games in bronze and your MMR will go up. If mechanics stop being important and shotcalling and positioning are relatively more critical at higher levels and you're good at those, you'll be more likely to win games and your MMR will go up.

It's self-correcting

[normal rose]

I was trying to improve the PS system using logistic regression to wheigh the different stats depending on their correlation with the outcome of the match. Some (positive) stats actually correlate more with losing if they get too high, for example cs. If you farm too much you will be less present in teamfights and you are more likely too lose. Its really hard to give a meaningful measurement of a players impacz on the match just using his stats. And again, even if it was possible you would have people just trying to maximize stats instead of trying to win.

@tawny basalt

[tawny basalt]

thats me!!!

[iron depot]

So then, if it's pretty solid all around - what would you say are the areas/things it could use some improvement in?

[graceful scaffold]

The most most most important thing to remember is that MMR is designed and built to be reflective over time

You will carry games, be the best player, totally dominate, and lose

Messi and LeBron and Tom Brady have all lost more than their fair share of games while being the best player on the planet, because team games are like that

But that doesn't change the overall perception of them, because the important thing is longevity and consistency over time

[maiden relic]

I had so many games that i did so good but my team wasn’t it

But i had games where i was trash and my team did the work

[graceful scaffold]

And plenty where you played poorly and won anyways

Yep

Same for me too, and everyone

Just part of the unpredictability of human performance in small slices

[maiden relic]

Last games i did good but my jungler did no objectives etc

So easy L

Every role is important

[iron depot]

That's most of my L's tbh. Just a jungler ignoring fang while the enemy jungler is on it from spawn.

[maiden relic]

And its not all about 1 person being trash

I carried games that had 2 ppl going so bad

And i lost games where we went awesome

1 objective or team play cam change a lot

[tawny basalt]

I imagine there are many other similar problems that would just complicate the situation a bunch on top of that. The weight of each stat wouldn't be the same for every elo, it'll change per patch, and it'll probably be different per region. I still think with enough dedication, time, and skill there would be a way to solve these issues and make it fairly accurate. It would still have to be accessory to current mmr systems and even then, it would still have some fundamental problems that I don't think you could ever get rid of.

[maiden relic]

The mm isn’t to bad but its not the best but even if you get a full plat team

It could still screw you up

Its not always the bronze

[graceful scaffold]

The matchmaker itself as a tool that takes groups of the 10 closest-matched players by MMR and puts them in a game together is really, really, really damn good.

The areas of improvement are in making that MMR more predictive of expected performance when 10 players are all close together.

[maiden relic]

The only thing i really need is a swap role option

[graceful scaffold]

We can put 10 people together with identical ratings and have a lopsided game. Some of that is in MMR itself, a lot of it is also in game design to prevent snowballing.

It's a multifaceted problem

[graceful scaffold]

It's coming

[maiden relic]

Any eta ?

[fast apex]

Waaaaaaay to many people dont get that ranked is supposed to be top competetive mode in this game, but they just treat it like standard mode and complain that people sweat too much. Well we're sweating hard to get away from them

[graceful scaffold]

Not tomorrow

SoonTM

[maiden relic]

Yes but still

We can all help

By saying ill swap

And it might win the game

[tawny basalt]

Is this how it works? I was under the impression that the mm tried to get 2 teams to the closest difference in average mmr (and thus closest to a 50/50 match outcome) and that individual mmr differences were secondary.

[graceful scaffold]

Nope, it's the opposite order

[tawny basalt]

really? was that changed recently?

[graceful scaffold]

Making teams even by average is just a math sorting problem that's trivial for modern systems

Nope, always been that way

[fast apex]

The care too little to even write "i cant play this role", while everyone else would take it from him

[normal rose]

There is absolutly no issue to do this per rank/role etc, in fact thats what i did. It doesnt solve the actual issur though.

[graceful scaffold]

Even on the old system we didn't build, it was just a lot worse and slower at it (and we had less people)

[normal rose]

Those stats dont paint the whole picture and they can easily be manipulated

[fast apex]

My biggest wish atm would be 24/7 rank availability, cause right now many of us cant grind out the rank due to working hours irl

And we stopped getting rank changes already

[graceful scaffold]

When you enter queue, you have a rating. Let's say it's 1500 for the sake of argument. You initially look around for games where every player is between let's say 1480 and 1520. After a minute or so, you expand and say you'll take a game where everyone is between 1460 and 1540, and so on and so on. Actual numbers and rate of expansion here is not linear and I don't have them, but this is to illustrated the idea.

[fast apex]

No increments at all

[graceful scaffold]

Once you get 10 players that have all agree to be "in range" of each other, they're grouped together in a match

Only at that point are they sorted onto two teams, which is a max of only 252 combinations if everyone is solo and so it's trivial to just try them all and see what the closest average is. Those are then the two teams

#1 prio is find people that are close and agree to be in a match with each other
#2 which we can do a complete test every time and get the mathematical best average is the balancing step

[graceful scaffold]

The challenge is that the curve of population is not smooth, it's an absolute cliff

Because pretty much everyone is asleep at the same time

(in a given region)

[rare quarry]

you say it's better to have a single value match making than to figure which of the minions killed, was the one that enabled the win? Why?

Why is it important to figure if one contribution was the one that enabled the win?

All contributions pave the way. You suck all game long, but the only Dekker stun you hit, that kills one hero, leads to surrender vote the enemy is now worth all the gold in the universe. Why?

That argument has many holes.

But I guess, we'll hear the same statements again why probably is not random, right?

The match quality is bad. We can debate if a single value is enough to improve match quality. If the answer is "yes, everyone just needs to play more to be placed at true rank" then we always say "you haven't player enough yet" or "the community hasn't played enough yet". Nothing will improve.

It's getting rolled hard or rofl stomping. 3 out of last 20 games were close. The other 17 were slap in the face.

I think, PS, or trend of last games, or role based values must be included. It's currently no fun, neither winning nor losing.

[normal rose]

if you win more due to your great stats, you gain more mmr

if your stats dont lead to you winning games that other players of your mmr dont win, you dont gain mmr for them

your stats are baked into the match result, does it give an accurate representation for a single match? no, but it does over time given enough matches

[fast apex]

Almost all of my friends stopped playing in any other time than the ranked availabile times(standard games stopped having meaning for them), so if we're to look at the given population during different times I guess it's only gotten worse with time and will be worse if we limit the ranked time.

I used to play for couple hours daily in the mornings, but now I just play couple hours weekly cause of timegate ranked.

Looking at almost no player retention we'll see less of them and if we stick to this scenario we will never have 24h up ranked cause naturally there wont be anyone to play in other times than 15-02am.

Why not just give us opportunity to even wait these 30-60min in queue if you worry about the player matching, I bet we all would gladly take even that, just to have a possibilty

[graceful scaffold]

Ranked's available windows cover 90% of games played. It's not perfect, but it's available for times when the overwhelming majority of players are playing.

Adding in overnights certainly helps the small % of people that work odd shifts and those folks are definitely important, but it's not as easy as saying "just open it" when the population drops off so drastically (and with it match quality and queue times) overnight.

[fast apex]

Is the 90% currently or from before the ranked was out, cause like I said many of us just stopped playing standard at all and only launch game to play ranked at the times we can

[graceful scaffold]

Both

It was true before ranked, and continues to be true with ranked open every day

[fast apex]

If it hasnt changed then well, i guess we gotta start at the roots with gaining players back/stick

Thanks for the replies though

For once someone talk with us 🤌

[graceful scaffold]

I mean, fundamentally people are asleep for basically the same hours every day, right?

That'll always be true, and always be a challenge for overnight ranked

Even at a point that it's available, match quality and queue times will always be proportionally MUCH worse overnight even if we had 10x or 100x or 1000x our population, because it's all relative.

[graceful scaffold]

I'm talkin all the time as much as I can, it's just impossible fro everyone to see every conversation

Not your fault, just reality

[fast apex]

I realise that, it's the factor we take into account in any other moba, did it in league and Dota as well.

Its just that its decided for us if we want to risk playing it. Hundreds of my games were fine with no PS players and 30% of current steam player count(30% of 24h player peak)at these times you talked about. Thats why for me, while we have way more players, taking the opportunity away seem "unlogical"

[graceful scaffold]

It's not 30%, the difference between high and low population in one region is up to 20x across a 24 hour period, and this is true across any online multiplayer game.

Designing a system that works well for both peak AND for 5% of that peak is really tricky

You gotta remember any online trackers you find are both imperfect AND amalgamating every region together at once, regardless of their current time of day.

[plain plaza]

So much words defending current system here. I hope you all understand that current MM is BAD and should be reworked?

[slim wave]

If you read any of the words you'd understand that mm isn't the issue
There maybe other issues but mm isn't it

[tawny basalt]

How do you know, for certain, that MM is bad?

[dusk parrot]

I dont think bronze should be able to match up with golds. If you look at a game with a great ranking system. lets say rocket league. Bronze and gold are light years apart in the skill gap. I understand that a lot of bronze have over 1000+ games and some barely play ranked or just started. but the bronze that have around 80 games. shouldnt get ranked up with someone in gold. it makes it totally uneven and not even fun. I'm fine to wait in a longer Q for an even matchup.

[sinful apex]

I have been playing Pred almost every day since release and I just uninstalled after yet another completely one-sided match. The match making is the most abysmal I have seen in any game I have played in my life. it feels like 4 out of 5 games one team gets completely stomped, someone goes AFK, or some other BS happens. It's absolutely baffling to me that the devs try to balance the game by giving some abilities 5 more damage or whatever and call this an update, yet actually balancing the teams in a match seems to have no priority at all. For more than a year since release the excuse has been that the external match making service had some bugs and as soon as their own backend is finished everything will be better. Now we have reached full release and still most matches feel just as bad as ever in terms of team balance. I know that there have been a lot of posts about this topic and the answer by the devs is always that the average internal MMR numbers are even. Yet, if hundreds of people feel like the match making is bad maybe the devs should consider the possibility that these numbers do not properly reflect the actually skill level of players. There must be so much data to analyse by now that anyone who knows a bit about statistics could extract the actual factors that correlate with someone's skill and use this to improve team balance instead of just taking raw winrates or whatever you use to match people. Saying that the game is too complex to properly predict a players performance seems ridiculous to me and basically invalidates the whole field of psychology. I have been quite invested in this game and I'm actually a bit sad that the devs seem to be unable to realise its full potential. Maybe I will check if anything changed in a couple of months but until then I will look for something else to play.

[graceful scaffold]

consider the possibility that these numbers do not properly reflect the actually skill level of players
Yes, this is definitely true and something we've covered extensively above. You cannot perfectly predict human performance with a single number, MMR is just a best-current-industry-standard to try to do as good a job as possible in aggregate.

There must be so much data to analyse by now that anyone who knows a bit about statistics could extract the actual factors that correlate with someone's skill and use this to improve team balance instead of just taking raw winrates or whatever you use to match people.
Nope. We just had the same conversation earlier today, I'd recommend reading back through it above.

Saying that the game is too complex to properly predict a players performance seems ridiculous to me and basically invalidates the whole field of psychology.
I don't really know how to respond to this one. Use as much psychology as you want and let me know who to bet on in next week's NFL games

[dusk parrot]

how bout we stop letting one rank play with another rank? as in stop letting bronze, play with silver, and silver play with gold, and giving both gold teams a silver? it dont make sense i think we would all be okay will a longer matchmaking. if your not the same rank you shouldnt be able to play with each other...? i think that would benefit everyone playing ranked?

i get it paragon should be able to play with diamond? thats fine but gold shouldnt play with bronze or silver.

totally different.

i think that would be fair each rank has three steps? so yeah.

[warped imp]

This isn’t needed in the slightest, no point in typing this out.

Please refrain from doing stuff like this again in threads so that they stay constructive and aren’t argumentative for 0 reason.

[dusk parrot]

sorry silver ecli

ill try not to be rude to people.

-.-

[sacred ice]

bro did you even read the things J said above??

if the internal MMR reflects yours they should be in your lobbys

is it realy that hard to understand

if i win 15 of my first games in ranked and im bronze 2 i hope i dont play with bronze 2 players

[marsh oak]

You'd comfortably be in silver if you win your first 15

but i agree with your point lol

[sacred ice]

bro it was a figure of speech

oh okey 😄

[marsh oak]

Excluding adding a role queue, i'm not sure how much more they can do to improve the current matchmaking for ranked

[sacred ice]

if they add a role Q now im quitting pred

no way that helps in the current situation

[marsh oak]

It would

[sacred ice]

to get 45+ min Q times in Plat and diamond

sure

it will help with that

[marsh oak]

that's a playerbase issue

[sacred ice]

true

so we cant do that now

[marsh oak]

matchmaking issue can be reduced with role queue, but if the playerbase isnt big enough to support role queue then you cant implement it

[sacred ice]

yes thats what i was playing for at in my first comment

only thing that would fix MM is if you would see your own internal MMR

then people would stop crying

but others would say : why is there Intenral mmr and Rank blabla bla

so i guess thats all fucked

[marsh oak]

I don't think showing internall MMR would fix people crying, it would just give them something numerical to cry about

it's the same dozen or so people, constantly moaning

[sacred ice]

i mean it would show people that the players who are in the lobby all have an alomst identical winrate

[marsh oak]

unfortunately most people don't understand math

[sacred ice]

true ... as in game its mostly brain or mental diff

[dim briar]

You can look at the winrate and often it is quite different. Besides of that winrate doesn't mean that much. You can have a 50% winrate on a 1800 mmr lobby and a 50% winrate on a 900 mmr lobby. People who are playing in their supposed elo for a long time are more closer to 50%, people who are climbing the ranks do have a higher winrate.
Another thing is if you can climb with a 50% winrate or not. On internal mmr maybe not, on the showed rank you can do that for sure.

[graceful scaffold]

Only up to the "target" rank for your MMR. If you've reached the rank corresponding to your MMR, you can't climb at 50%.

[vague saffron]

in paragon , i get like 17 up 15 down, is there a limit on paragon?

was getting like 21+ 17 down

after this season, mm might start getting better

but

player base will still be a issue

also

do u have any data that shows the most selected role in que?

in order

[graceful scaffold]

You mean selected in draft vs. assigned? That's super interested any maybe we already have telemetry on it but I don't know offhand

Instinct and experience suggest mid first support last, but closer than I'm used to in other games

[graceful scaffold]

🤷‍♂️
I poke around and get involved in stuff that is concerning from a player experience standpoint but haven't thought about/checked into how things are handled once you already reach paragon

[vague saffron]

Ah you think mid is the most selected? Interesting. Looken at paragon n players, lots of carrys

[vague saffron]

Soon you will experience Paragon ranked .. sign up for my 69 dollar coaching sessions and I will reach you . /Sarcasm so I don't get banned

[graceful scaffold]

Certainly lots of carries, but I think mid is the one I anecdotally see most selected. It's also the most common in both Smite and League so it would make sense to carry over to Pred too

[vague saffron]

thats true

[graceful scaffold]

Jokes on you I'm still scrapping to finish the climb to diamond

[vague saffron]

hehe

[graceful scaffold]

I got a ways to go for Paragon

[vague saffron]

steel op now baby, you can climb it now

[vague saffron]

now we need steel twerking

[pallid sable]

this thread isnt for ranting

[maiden relic]

Imo it wasn’t but ok ig

[indigo ocean]

@graceful scaffold this might be the first negative winrate paragon player that you said was impossible. Unless omeda city has incorrect numbers

[slim wave]

Theyre dia 1?

[indigo ocean]

yea

omeda city could be incorrect

but I have no idea

[graceful scaffold]

Those numbers don't seem right

Their OC account says they have a 52.6% win rate and 2.71 KDA

If you sort matches by "ranked", they have 51 pages and 510 matches, but the whole account page supposedly has 584 matches.

[slim wave]

The ss shows 74 matches which might be where the missing ones are

[graceful scaffold]

47% of 510 is 240 wins in ranked

52.6% of 584 is 307 wins

So in order for those numbers to line up, they must have won 67 out of the other 74 matches, ya?

No more than 7 losses total

There are 6 losses in their last 10 pvp games

7 more on the next page, which means the numbers can't add up

[indigo ocean]

Ill have to sort mannually but ill do it if he hits paragon to see how it adds up

[graceful scaffold]

Even putting aside the math not working, they just had a 6 game win streak in diamond 1 with an average KDA in that stretch of better than 4, and I'd be highly surprised for them to be that good with a 47% win rate

@indigo ocean I can do it with math by pulling game records tomorrow, but either way I just proved above that at least one set of Omeda.City numbers are not adding up correctly because the math doesn't line up

[indigo ocean]

yea thats fair. but still curious

[graceful scaffold]

For sure, I'll report back

eh screw it it only took a few seconds to modify an exist query

52.9% win rate in ranked @indigo ocean

OC display is incorrect. If you go to the "statistics" page of the player profile, it's only counting the 74 non-ranked games, in which they do have 35 wins in 74 non-ranked games for a win rate of 47%. I won't tag C0re this late, but it appears to only be showing non-ranked win rates on that page.

https://tenor.com/view/there-you-have-it-boys-case-closed-the-simpsons-gif-13616870

[normal rose]

I think it currently uses ranked & pvp matches

I wanted to add filters for that in the future

[graceful scaffold]

It's just PvP right now

[normal rose]

Ah ok

Yea dont remember tbh 😄

[graceful scaffold]

Heh, no big deal

I just like detective-ing

[normal rose]

😄

I will eventually add filters there & also on profiles so you can see stats for each gamemode seperatly

[dense monolith]

they should start punishing ppl for trolling

[pallid sable]

they already do, this isnt the thread to talk about "punishments"

[dense monolith]

i mean deranking or resetting stats would be awesome

[vague saffron]

would possibly make smurfs more easy to make i guess.

[jagged flume]

match making is stup* as fuc*

[jagged flume]

Now i play with whoever, bronze, silver, gold, plat, and often times bronze is better than plat. I don't understand what the point of the rank is

[slim wave]

Matchmaking is based on mmr not vp

[jolly cairn]

It’s amazing how much the population DOESNT KNOW and can’t seem to UNDERSTAND😭. I’ve stopped bothering tbh

1000+ messages with tons of great insider context and buddy couldn’t be bothered to even glance at it considering he cares enough to make a comment about it.

[slim wave]

It is kinda wild

[jagged flume]

match making is stup* and nobody will ever understand it except for a few people you can count on one hand to whom the dev explained it, and always everyone will come here to complain. when is there mmr why i cant see it. i dont get it.. why is there vp? why can't it be easily understood without having to read the manual?.. i have no time and energy read 1000+ messages

[jolly cairn]

Read the thread

[outer bloom]

Explain this why Dimonds players can't play duo rank with gold players but they get gold players in their team

[pallid sable]

🤷‍♂️

[outer bloom]

Predecessor game: "Oh, sorry, you're a Diamond player. Clearly, you're way too fancy to play Duo Rank with a lowly Gold peasant."

Matchmaking: "But hey, no worries! I’ll just throw three Gold players on your team anyway. Good luck with that!"

[pallid sable]

And you’re tagging me why?

I don’t work for Omeda broski

[outer bloom]

[pallid sable]

I don’t know the ins and outs of the game other than what’s been shared here

[outer bloom]

I know I'm just trying to be sarcastic

[slim wave]

Short answer is
Mmr is used to make matches more fair and actually have sbmm
Almost no games show internal mmr, despite almost all game having it.
Vp gives you something to grind for
It can be easily understood by those who want to understand
Those who want something to complain about and something to explain why they lost games refuse to understand

The short short answer is - mm is based on mmr not vp

[plain plaza]

There are no points in rank vp because it doesn't reflect a skill. Its just grind like level of account.

[jolly cairn]

That’s a bit reductive but you are partially right

[rare quarry]

The rank in ranked mode is just cosmetic for a profile website and match loading screen. Otherwise ranked doesn't differ from PVP except the party size is limited to two.

[plain plaza]

Also it was said here that with influx of players the mm will be better and vp will come closer but its wrong since vp is not reflecting a skill level. With current system there always will be bronzes playing like plats and golds who should be in wood league.

[dense monolith]

and he will always suffer from that fate XD

[vague saffron]

And no mirrors and bans

[vague saffron]

Yes. But u get bans and no mirror

[dense monolith]

the only good thing ahahah

[jagged flume]

mmr doesn't exist nobody has ever seen it

[plain plaza]

Can anyone explain me how is that happening https://omeda.city/matches/76298ad8-8ebd-497f-b812-1a07bd4d6453

Im the only non-gold in our team placed against only gold in enemies team. and this only gold enemy is actually gold while all golds on my side are below all bronze enemis?

Is that the matchmaking we are worth of?

[vague saffron]

lets ismplify it.

everyone starts ~ 1300 mmr

lets say 1300 mr is gold

your rank, doesnt mean shit

its just a color, a image.

[thick ether]

it is, gold starts at 1300

[vague saffron]

wen u first start

and get a win

u gain ~ 50 mmr right? for example

as you play more that mmr drops, as it averages your games.

if that guy gets lucky

and wins 3, he now is 1650, that might be plat , who knows.

so thats how the current mmr system works

so those people in ur game

[thick ether]

if he wins 3 its puts him to 1450 rouhly, which puts him its thr 100 + or - catergory therfore he cud play against plats after winning 3 games

[vague saffron]

are near the 1300 mmr

we generally dont know how much mmr icnrease there is though

[thick ether]

yeh i understand just roughly

[vague saffron]

i have a plat 3 smurf, and i get paired with paragons and diamond 1s constantly.

[thick ether]

it seems like it wud be nicer if the MMRs were displayed on omeda, and the rank also like bronze silver etc

[vague saffron]

also, im sure the game doesnt have enough players to truly place people in deserved rank

[thick ether]

In online chess, all you see is the MMR and their name. If your a 1500 playing against an 1800, you can see at a glance how good or bad they are. If you win, against someone higher rater you get a bit more MMR each win, and vice versa. There is no confusion that way and it works nicely I think

[vague saffron]

so

one thing is

alot of people do complaina bout not being able to climb but

thats kinda how it works.

i got into paragon almost solo exclusivley

and have a smurf where i showed on youtube how i climb out of turbo inting/ afk teamamtes

i mean

you literally gotta play against them.

[slim wave]

Thats in a strictly 1v1 format though
Its harder to judge an individuals impact on a game in a 5v5 game, especially a moba
And that's similar to how the mmr pred had works, you just dont see it, and we don't have placements

[vague saffron]

in realtiy, think of it as darksouls style game.

[plain plaza]

I think of it as shitdevs game. Darksouls is supposed to be darksouls while pred isnt.

Unfortunately my whole pvp gaming experience is ruined by devs all the time. Age of Conan from Funcom, then Bloodline Champions of Stunlock, then Paragon, then Battlerite. Now Pred is going exactly the same way.

Every pvp game i start is fcked by its own devs.

Looks like a curse on me 🙂

[thick ether]

But that is exactly my point 'you just don't see it'. Well, show it then so we can see it. Instead of VP, no?

With chess, you can see the MMR up front, and there has never in the hisotry of the game been a discussion about the matchmaking. I think thats because its pretty simple and easy to see

[dawn bolt]

ranked mode need to use VP for matchmaking

mmr use in normal game and brawl mode

If they make these changes it will improve a lot

[sacred ice]

no it wont

why would that be the case ?

imagine pople who come over from lol or other mobas start out in bronze and always play according to rank... they will steamroll lobbys till they reach plat

also what negatives would an internal MMR have other then no transperency

[viscid condor]

but thats exactly how it is now

what are the downsides to vp based matchmaking

[marsh oak]

Smurfs would be a massive downside to VP based matchmaking. You'll have good players smashing noobs for 25+ games to even get to gold 3, whereas in the current system they will start being matched with platinums within 10 or so games

[viscid condor]

if your smashing your current vp give the same high gains they give now

and again thats still how it currently is

i smashed 50 games in row

because the game put me at newbie mmr

[marsh oak]

no, you start at the same mmr that everyone starts at

[viscid condor]

which is newbie mmr

[marsh oak]

that mmr is reflective of gold 3

[viscid condor]

right

[marsh oak]

it's meant to be the expected average skill

you start at the expeced average skill level

[viscid condor]

what you are saying is not what happens

it took me 50+ games to start getting real matches

in this system

and even when I started to get real games the players are in diamond and paragon

and im still plat for now

i dont see how the smurf argument holds up at all

[marsh oak]

let me guess, you started when ranked mode was released?

[viscid condor]

no

I started when omeda city mmr was removed

right before actually

the whole plat in 10 games thing does not happen

its still the same 50+ games

[marsh oak]

if you win your first 10 games, you'll be in silver 2 which I believe can be matched with platinum 3

[viscid condor]

oh well okay then yeah maybe you could match with plats

but again

what is the point of that

right now ur rank does not mean anything

[marsh oak]

because its based on the mmr

[viscid condor]

it doesnt equal your skill

or real rank

right now all it shows is time invested

[marsh oak]

the more you play, the closer your mmr gets

[viscid condor]

because the game matches off internal not vp

what is the point of vp

if you arent going to matchmake off it

[marsh oak]

vp is just a measuring stick, who cares.... you'll be roughly in the right division for your skill set and the more games you play the higher likelihood you're in the right division

[viscid condor]

??

im saying vp should mean something

what are the downsides to matchmaking off of it

tell me

[marsh oak]

because smurfs

[viscid condor]

because smurfing is not one of them

what you predict will happen is already happening

with this system?

[marsh oak]

dude there are only 2000 platinum players and only 300 players in diamond/paragon

[viscid condor]

right

[marsh oak]

that took a couple months of playing time to normalise out

[viscid condor]

what does that have to do with anything

[marsh oak]

but now they have a normal distribution to match people against

so mmr becomes more refelxtive of your skill against others and becomes more accurate over time

[viscid condor]

u are just talking and you dont know what you are saying

[marsh oak]

more like you failed math bro

[viscid condor]

mmr is reflective of skill yeah bro everyone knows wtf are you talking about why are you saying this

for no reason

u are just saying it to say it

you arent talking about what i am talking about

clown

im asking vp based matchmaking system wouldnt work. you keep saying your clown answer of smurfs 💀.

currently there is 0 point to vp

a measuring stick?

fr

?

could half of these players have made it to their current rank if they had to match people in their tank before

[marsh oak]

if you cant understand that using vp to match would lead to smurfs playing more games against lower skilled opponents then i don't know what to tell you mate

[viscid condor]

How could that possibly be the case if you gave the same gains you give now for dominating your current vp

?

if im bronze and i win 10 games in a row

boost my gains so i get to silver quicker

?

[marsh oak]

because you spend 10 games smashing bronzes, 15 games smashing silvers, 20-25 games smashing golds etc

[viscid condor]

but you would actually belong in your rank

[marsh oak]

no, because they are smurfs you know their rank is higher by the fact they are using smurgs

[viscid condor]

and it would just be 10 -15 in each

dunno why you went to 25

[viscid condor]

wdym by this

hypothetically lets say a player is a diamond skill player

everyone starts in bronze

he wins 10 in a row

his gains are massively boosted

he still wins 10 in a row

his gains are still massively boosted

10 more he is now in plat

playing against people of plat skill level

here is where his gains slow down

eventually he will make it to diamond

and now he is in well balanced games

what is the problem with this

i think the whole entire system is fucked for not using placements anyways

btw

[marsh oak]

it's far more than 30 wins to get to plat, you have someone smurfing that could be potentially ruining the experience for 5 other players over a long period of time

[viscid condor]

?

[marsh oak]

they couldn't use placements, there was nothing to place people against

[viscid condor]

it took me maybe 50 games to get to plat

50 games

of stomping

was that fun for anyone?

in this system?

look at my history if u want

30 stomps in a row

30 boring

no fun for anyone games

this system is terrible

[viscid condor]

wdym they couldnt use placements

just have the max rank you could be placed be like gold 1 or something

and place based off wr in 20 games

newer players could settle quickly at the bottom and not have their games ruined

better players could quickly skip new player ranks

the system they have rn is not good

[marsh oak]

https://tenor.com/view/facepalm-really-stressed-mad-angry-gif-16109475

[viscid condor]

i take back my dm

ur a clown

[jagged flume]

this will kill the game, ranked is a joke, mmr doesn't exist

every new player who will play ranked for first time will wonder why he playing against plat when he is bronze

[slim wave]

Only if that bronze has rhe siill of a plat
Mmr doesn't just randomly give a terrible player plat level mmr lol

[marsh oak]

that's who you can queue as a duo with and the matches dont match bronzes with platinums, only up to golds. Silvers can be matched with platinums

[warped imp]

I advise you to please simply read the thread