Jump to content

Photo

Matchaker Updates v3

- - - - -

  • Please log in to reply
83 replies to this topic

#1
capnjosh

capnjosh

    Advanced Member

  • Administrators
  • 260 posts

We've just deployed another matchmaker update.  Only 1 change on this one:

 

Since the matchmaker has routinely been putting new players (below _segregationLow) into matches 100+MMR points higher than their own skill, we've put in a hard threshold at the _segregationLow number, which is currenly 1340.  The goal is to get new players into matches that are super-close to their skill level, which, as we all remember, 100MMR points is a huge deal when you're brand new.   Just 3 matches is a big deal.

 

We also added more resolution into time spent waiting in queue, so it'll be useful a bit later on.

 

 

Here's what we were seeing that made it seem wise to make an effort to improve our brand new players' first few matches.  High skill players typically get matched to lower-skill matches, and everyone else typically gets into harder matches, with our poor newest players getting it most.  Hopefully this change today compresses the MMR spread within matches a bit more. Btw, that low point right at "Fri 06" is about where we launched the previous update.

 

kNMjX6M.png?1

 


  • eth0 and Charcoal like this

#2
Acguy

Acguy

    Advanced Member

  • Members
  • PipPipPip
  • 170 posts

We've just deployed another matchmaker update.  Only 1 change on this one:

 

Since the matchmaker has routinely been putting new players (below _segregationLow) into matches 100+MMR points higher than their own skill, we've put in a hard threshold at the _segregationLow number, which is currenly 1340.  The goal is to get new players into matches that are super-close to their skill level, which, as we all remember, 100MMR points is a huge deal when you're brand new.   Just 3 matches is a big deal.

 

It sounds like reworking how mmr is calculated might be a greater concern.

 

Might also be good to just keep new players from even having an mmr until they complete a certain number of matches, with an exception being made if they do absurdly well as a means of filtering smurfs out. Even if you do get a bunch of 100 mmr players in one room, a new players mmr is far less meaningful of a measurement than a higher player's goes considering how few samples it's comparatively going to be working off of.


Edited by Acguy, 06 January 2017 - 02:28 PM.

  • Charcoal likes this

#3
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

Even if you do get a bunch of 100 mmr players in one room, a new players mmr is far less meaningful of a measurement than a higher player's goes considering how few samples it's comparatively going to be working off of.

 

There shouldn't be any 100 mmr players unless there is something VERY VERY wrong somewhere.

 

People start out at 1250 MMR, and go from there.

 

The variables pertaining to SkillRatingSpread are deltas, or differences.

 

 

 

Thanks Josh, more pretty graphs please, next time I recommend using Pink instead of blue though.


[DELETED]

 

fuzzy bunny you CZeroFive 


#4
capnjosh

capnjosh

    Advanced Member

  • Administrators
  • 260 posts

It sounds like reworking how mmr is calculated might be a greater concern.

 

Might also be good to just keep new players from even having an mmr until they complete a certain number of matches, with an exception being made if they do absurdly well as a means of filtering smurfs out. Even if you do get a bunch of 100 mmr players in one room, a new players mmr is far less meaningful of a measurement than a higher player's goes considering how few samples it's comparatively going to be working off of.

 

Agreed on changing how much MMR is adjusted after a match to be based on the extremity of outcome.  Right now, it's a flat change based on the binary "win" or "lose", which is sometimes a bit shoehorned.  This work will start when we begin more of the PvP ranking-related systems.



#5
wischatesjesus

wischatesjesus

    Advanced Member

  • Members
  • PipPipPip
  • 180 posts

Right now, it's a flat change based on the binary "win" or "lose", which is sometimes a bit shoehorned.

 

Wait, what?

 

We've always been told that MMR is a glicko system, and it has been observed to change differently based on how well you score vs the rest of the server.


oFFOtRH.jpg

 

High Tier Cabal (Noun): A group of people who used to play Hawken.  


#6
Acguy

Acguy

    Advanced Member

  • Members
  • PipPipPip
  • 170 posts

Right now, it's a flat change based on the binary "win" or "lose", which is sometimes a bit shoehorned.

 

I... what?

 

I'm not like, terribly offended, but why on earth have you been dedicating all this effort to sorting people out by MMR when it's that inconsequential of a measurement? You've gone through all this trouble to spruce up your sorting algorithm only to say "guess what; I've been sorting people by the number of eyelashes they have."

 

You said so yourself; "High skill players typically get matched to lower-skill matches, and everyone else typically gets into harder matches". If high mmr players are so scarce that they get thrown into noob lobbies, and mmr is purely based off of winrate, what do you think it going to happen to a high mmr player's mmr? It's going to get bigger, which is going to make mid-level mmr players cut ahead of them when getting into mid level rooms, which is going to cause them to end up in some noob room, which is going to cause their mmr to get bigger which is going to... give you a really bad feedback loop. mmr needs to change based off of the mmr of the people/team you've beaten at least. Someone who stomps a lot of noobs should not have their mmr skyrocket unless they were a noob too.


Edited by Acguy, 06 January 2017 - 03:54 PM.


#7
Call_Me_Ishmael

Call_Me_Ishmael

    Advanced Member

  • Members
  • PipPipPip
  • 1104 posts

Agreed on changing how much MMR is adjusted after a match to be based on the extremity of outcome.  Right now, it's a flat change based on the binary "win" or "lose", which is sometimes a bit shoehorned.  This work will start when we begin more of the PvP ranking-related systems.

 

'Beat median' or 'did not beat median', I hope?


Did I say Call Me Ishmael?

 

You should call me Luna.


#8
Acguy

Acguy

    Advanced Member

  • Members
  • PipPipPip
  • 170 posts

'Beat median' or 'did not beat median', I hope?

 

It needs to be inversely proportional to the difference in mmr between you and the enemy team. Losing to someone whose mmr is higher than you or beating someone who was lower than you means your mmr was accurate, and therefore shouldn't be changed much. Conversely, beating someone above you or losing to someone beneath you means your mmrs weren't accurate, and therefore they should both be changed more significantly.

 

I don't know how you'd make an "opposing MMR" value to compare to though, you could compare to the median on the enemy team, the average, closest player, something else. Depends.



#9
Sriracha_Sauce

Sriracha_Sauce

    Advanced Member

  • Members
  • PipPipPip
  • 55 posts
I lose games and gain MMR
I win games and lose MMR
Therefore it cannot be a win loss MMR system from match results

Perhaps the win and lose Josh refers to is your individual score? Top half score is "win" and bottom half score is "lose"
  • SS396, Miscellaneous, DieselCat and 1 other like this

Can't handle the sauce?


#10
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

Perhaps the win and lose Josh refers to is your individual score? Top half score is "win" and bottom half score is "lose"

 

Ding Ding Ding, we have a winner folks!


[DELETED]

 

fuzzy bunny you CZeroFive 


#11
wischatesjesus

wischatesjesus

    Advanced Member

  • Members
  • PipPipPip
  • 180 posts

Perhaps the win and lose Josh refers to is your individual score? Top half score is "win" and bottom half score is "lose"

 

Right, but he also said:

 

it's a flat change

 

...which was false last time I checked. You can definitely get different amounts of MMR change, else Xacius would have hit 4k during his S3 scrim farming run.

 

IMO, the glicko based MMR system to rate players based on their performance in game is actually very good. The difficulty arises when you have to actually use that rating to create good games despite having only a handful of players to throw into various servers at any given time. Given a group of 12 players it is fairly easy to the most balanced 6v6 match possible for that group, but if that group has a 1000 mmr spread or an outlier the 'most balanced' may still be pretty unpleasant. Getting a group of 12 that can be balanced into a reasonably close match is the hard part, especially when you have so few to work with.


  • Bergwein likes this

oFFOtRH.jpg

 

High Tier Cabal (Noun): A group of people who used to play Hawken.  


#12
ArchMech

ArchMech

    Advanced Member

  • Members
  • PipPipPip
  • 395 posts

remember you lot, these guys been on at this for quite some time doing things so...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

who knows what they've gutted


don't mind me, i'm just on a crusade against humanity, by the end of my lifespan earth's population will be 8 billion+ trolls


#13
ReversusRex

ReversusRex

    Advanced Member

  • Members
  • PipPipPip
  • 91 posts
Can we get a graph of the MMR distribution of the current player base? Not sure how useful the data would be given the prevalence of smurfs nowadays, but I'm curious lol.

Edited by ReversusRex, 06 January 2017 - 07:29 PM.


#14
DemitronPrime

DemitronPrime

    Advanced Member

  • Members
  • PipPipPip
  • 879 posts

None of this means diddly fuzzy bunny if you cant join a server due to reservation constantly Failing....


Puck Flinging-Nade Lobbing-Troll Tech

r4GY0PF.jpg

::eoc Pred::Eoc/HC Infil::Vulcanador::Hawkins Tech::EOC Raider::

" I play on north American servers with a 200Ping to give you cry babies a chance"

2XhpJes.png

"Nothing says Good morning Like a well placed Nade"


#15
Onstrava

Onstrava

    Advanced Member

  • Members
  • PipPipPip
  • 283 posts

Sorry ReversusRex, these days I doubt anything that has to do with tracking MMR to make matchmaking work right will do any good. It's worth remembering that besides the obvious smirf accounts, if someone with a smirf account kills a player with a high MMR, that player loses like 5-10 points almost instantly. While only getting 0.5-1 point when he kills the smirf. So at the end of the match that high mmr player will be brought down regardless of how good they do and be put in with the lower mmr players matches, thus creating choas for the new guys and the words fly across the screen. You know the same non-sense that's been happening for what 4 years now. :no:  You have to give more points for multi kills and stuff like that so the mmr Ttier matchmaking can be properly balanced. Nobody should be getting 20+ kills with 2 deaths by him/her self and only go up about 1-3 points at the end of the match. That's just stupid, but w/e nothings going to change, so I'm preaching to the quire.


  • WmMoneyFrmMissouri likes this

Hold on to the things you care about most, even if others see it as insignificant..If you can't be true to yourself, are you really living?

W4znwFO.jpg1sHSjPn.png1sHSjPn.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngbANk8SP.jpgMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngMLIZlEa.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngMLIZlEa.pngMLIZlEa.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.pngZVn3Cxy.png1sHSjPn.pngW4znwFO.jpg


#16
6ixxer

6ixxer

    Advanced Member

  • Members
  • PipPipPip
  • 1104 posts
In regards to the region expansion, could you add some sort of fuzzy estimated MMR factor for out of region matches?

What I mean is, if a player has an mmr of 1800 and there are no matches in their region, you estimate that adding 100ping is somewhat equivalent to giving them a 100MMR handicap so you adjust their evaluated mmr down a bit to 1700 when sending them out of region. You can work out the average ping difference from USA to EU, etc and also USA to Asia, etc to create a matrix of values from region to region and it might make the resulting matches a little more tolerable.

This will help people who suddenly need to deal with leading and lag issues while playing the same ranked players that do not have these issues.

For example, I found that to play games at 300-400 ping, I had to look for one star matches on the low side which is about 300mmr lower.
You might say I was trying to pubstomp, but really I just wanted to play and that was about the only match I could tolerate and it was a fair struggle to maintain a 1:1 kdr for those matches. Rocketard was about all I could play at 400 ping and 300 ping I could get some fair hits with splash damage weapons. Mostly I aimed not to feed the other team.

I think it would be worth seeing what happens when you send an 1800 US player to EU and put them up against 1700 players.
I know it probably won't create good balance, but if you're going to send players out of region, I think its worth a look.
  • _ckc_ and Onstrava like this

#17
6ixxer

6ixxer

    Advanced Member

  • Members
  • PipPipPip
  • 1104 posts
I saw that there are a lot of servers running right now. still none in oceanic.

Gave queueing a try for any mode.

110sec and it deployed me into an empty DM in oceanic. I was kinda expecting to be sent out of region...
Following attempt it quite optimistically had me trying to seed an oceanic siege match.

Edited by 6ixxer, 07 January 2017 - 03:14 AM.

  • DemitronPrime likes this

#18
MomOw

MomOw

    Advanced Member

  • Members
  • PipPipPip
  • 1038 posts

Here a feedback of the new MM :

 

tried the MM this morning (my MMR is 2300ish, but it's a lie, I should be more like 2100 if good players were still there).

there was ~60 pple playing and one 1700~1800 MMR.

 

The good point is that low MMR players seems to stay in adequate servers.

 

After a longer time (as expected) the MM decided to create a new server.

Fortunately pple came and eventually we had a 2000ish server, but it also could have ended in a no game loophole.

 

 

 

I'd really like that some issues would be solved, such as adding uneven new players in balanced match with even teams.

--> Could you make them wait to add new players only when the teams stay even ?

 

I'm OK with ending in a new server, but when there are OKish players in lower MMR server, can they receive invites to join the new better server ? I do it when I have them in my contact list, but I won't "add as friends" all players with MMR > 2k...


Edited by MomOw, 07 January 2017 - 04:04 AM.

IRZUTYo.png

gXO9Nfd.pngmXasTsY.pngft4VqcE.png

 

KDR Member | Streamer | Priority Target II

Spoiler

#19
ReversusRex

ReversusRex

    Advanced Member

  • Members
  • PipPipPip
  • 91 posts

Sorry ReversusRex, these days I doubt anything that has to do with tracking MMR to make matchmaking work right will do any good. It's worth remembering that besides the obvious smirf accounts, if someone with a smirf account kills a player with a high MMR, that player loses like 5-10 points almost instantly. While only getting 0.5-1 point when he kills the smirf. So at the end of the match that high mmr player will be brought down regardless of how good they do and be put in with the lower mmr players matches, thus creating choas for the new guys and the words fly across the screen. You know the same non-sense that's been happening for what 4 years now. :no: You have to give more points for multi kills and stuff like that so the mmr Ttier matchmaking can be properly balanced. Nobody should be getting 20+ kills with 2 deaths by him/her self and only go up about 1-3 points at the end of the match. That's just stupid, but w/e nothings going to change, so I'm preaching to the quire.

Yeah I doubt it will be of any practical use, but I'm curious to see what the distribution looks like. If smurfs are as damaging as we make them out to be, the graph would be skewed left to some degree. I remember seeing an MMR distribution posted years ago before smurfs became an epidemic and the distribution looked roughly normal.

Posted by Hyginos using data from December 2014:
http://i.imgur.com/MgX9nNm.png
http://i.imgur.com/mgfmnuW.png

Edited by ReversusRex, 07 January 2017 - 05:55 AM.


#20
talon70

talon70

    Advanced Member

  • Members
  • PipPipPip
  • 264 posts

I am all for protecting new players so they have an opportunity to learn and have fun. It is really important. But segregating the new  player population creates other problems.

 

It would be a good idea to tell newer pilots, "beware. you are now moving up into higher level servers!". I feel most people go through feelings they are getting pretty good, then move into higher level  servers and are decimated and discouraged because they don't get how the system works. More communication to the new player would go a long way.


  • ARCH3TYP3 likes this

#21
CrimsonKaim

CrimsonKaim

    Advanced Member

  • Members
  • PipPipPip
  • 1234 posts

Agreed on changing how much MMR is adjusted after a match to be based on the extremity of outcome.  Right now, it's a flat change based on the binary "win" or "lose", which is sometimes a bit shoehorned.  This work will start when we begin more of the PvP ranking-related systems.

 

Wouldn't it be wise to not adjust the MMR due to a win or loss? Instead, adjust it in relation to personal performence such as damage dealt, kills secured, fully kills done (~90% of dmg done), amount of health restored (Technician), amount of assists combined with damage done, etc.

At the end, check how many MMR points are added or take naway based on how large the skill gap was in that match.

 

This would create a more accurate rating of a player's abilities since a bad player isn't necessarily good just because the team wins or a good player isn't necessarily bad because the team loses.


- Sitting next to the sound box in Last Eco -


#22
wischatesjesus

wischatesjesus

    Advanced Member

  • Members
  • PipPipPip
  • 180 posts

Yeah I doubt it will be of any practical use, but I'm curious to see what the distribution looks like. If smurfs are as damaging as we make them out to be, the graph would be skewed left to some degree. I remember seeing an MMR distribution posted years ago before smurfs became an epidemic and the distribution looked roughly normal.

Posted by Hyginos using data from December 2014:
http://i.imgur.com/MgX9nNm.png
http://i.imgur.com/mgfmnuW.png

 

Smurfing was as much a problem when that data was taken as it ever was. At that point it might even have been worse it if preceded that change that allowed players to simply match into the highest game instead off waiting indefinitely. The thing is a glicko system rates players under the assumption that they are normally distributed, so it will end up looking fairly normal no matter what.

 

The reason you see that weird spike at 1250 is that at some point the starting MMR was changed from 1400 (maybe it was 1500) to 1250, so there are a bunch of accounts in there that are only 1-2 games off of those two starting MMRs. Also the system seems to be set up (anecdotally) to rate around a 1500ish average.

 

 

Wouldn't it be wise to not adjust the MMR due to a win or loss? Instead, adjust it in relation to personal performence such as damage dealt, kills secured, fully kills done (~90% of dmg done), amount of health restored (Technician), amount of assists combined with damage done, etc.

 

That depends. What you adjust based on will determine what skill you are actually rating. If you adjust based only on score, you are rating a players ability to score. If you adjust based on win, you are rating a players ability to win.


Edited by wischatesjesus, 07 January 2017 - 09:56 AM.

  • ReversusRex likes this

oFFOtRH.jpg

 

High Tier Cabal (Noun): A group of people who used to play Hawken.  


#23
ReversusRex

ReversusRex

    Advanced Member

  • Members
  • PipPipPip
  • 91 posts
The presence of smurfs should be more proliferated in the graph given how much smaller the population of the game is now. I don't get what you mean about the glicko system. Even if glicko assumes the sample is distributed normally it will look messed up if it's not normal. Are you saying that in its calculations, glicko compares all players against each other? When I'm asleep glicko will factor my MMR into calculating that of other players?

Edited by ReversusRex, 07 January 2017 - 01:25 PM.


#24
wischatesjesus

wischatesjesus

    Advanced Member

  • Members
  • PipPipPip
  • 180 posts

The presence of smurfs should be more proliferated in the graph given how much smaller the population of the game is now. I don't get what you mean about the glicko system. Even if glicko assumes the sample is distributed normally it will look messed up if it's not normal. Are you saying that in its calculations, glicko compares all players against each other? When I'm asleep glicko will factor my MMR into calculating that of other players?

 

I can't claim to be an expert on Glicko, but the way I understand it every match it looks at where the players land on the normal curve, then based on the result gives them new positions on that normal curve. Because it is always operating under the assumption that players' skill is represented by a normal curve, each time it gives a player a new rating that rating will be somewhere roughly on that curve.

 

I'm having trouble finding a good way to describe this without anthropomorphizing the algorithm.


  • DemitronPrime likes this

oFFOtRH.jpg

 

High Tier Cabal (Noun): A group of people who used to play Hawken.  


#25
hestoned

hestoned

    Advanced Member

  • Members
  • PipPipPip
  • 251 posts

you cant grind mmr in low mmr servers if you already have a high mmr. your gain or loss of mmr is mostly dependent on the sever average. if i for example join a 1800 rated server with my 2800something mmr then the most gain i can hope for is in the .01 range. if the server is 2500 or more than i can usually net about 5-10 gain. your performance also matters so even if my teams loses the round as long as im in the top 5 or 6 players for points then i wont lose mmr. 


  • DieselCat and popeco like this

#26
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

^

 

Yes because overqualified individuals that are 1000 MMR above the average of the server should be rewarded for playing people much lower than ones in their own skill level.


[DELETED]

 

fuzzy bunny you CZeroFive 


#27
Silverfire

Silverfire

    Advanced Member

  • Members
  • PipPipPip
  • 1395 posts

Can we get a graph of the MMR distribution of the current player base? Not sure how useful the data would be given the prevalence of smurfs nowadays, but I'm curious lol.

 

Not quite as precise in the breakdown as the other graphs provided, but it is *slightly* more up to date, this data is from May -> June 2016, logging a little more than 26,000 unique account log-ins:

 

Sc5JoUR.png

 

Curve still looks the same.  Regular ole normal distribution


Edited by Silverfire, 07 January 2017 - 04:52 PM.

  • ReversusRex likes this

lNM7VnC.png

( ^ click for the EMP song ^ )

 

Come take a look at Hawken guides | Join me on #hawkenscrim IRC

 

 


#28
ReversusRex

ReversusRex

    Advanced Member

  • Members
  • PipPipPip
  • 91 posts

Not quite as precise in the breakdown as the other graphs provided, but it is *slightly* more up to date, this data is from May/June 2016, logging a little more than 26,000 unique account log-ins:
 
Sc5JoUR.png
 
Curve still looks the same.  Regular ole normal distribution


So is this basically including accounts that haven't been accessed in years? That's a lot of people 0.0

#29
Silverfire

Silverfire

    Advanced Member

  • Members
  • PipPipPip
  • 1395 posts

It's 26000 unique account log ins across the month timeframe between May and June 2016.


  • nepacaka likes this

lNM7VnC.png

( ^ click for the EMP song ^ )

 

Come take a look at Hawken guides | Join me on #hawkenscrim IRC

 

 


#30
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

It's 26000 unique account log ins across the month timeframe between May and June 2016.

 

But how many of them are smurfs?


[DELETED]

 

fuzzy bunny you CZeroFive 


#31
wischatesjesus

wischatesjesus

    Advanced Member

  • Members
  • PipPipPip
  • 180 posts

But how many of them are smurfs?

 

25999.

 

7, if you don't include craftydus' accounts.


  • Miscellaneous, -Tj-, eth0 and 3 others like this

oFFOtRH.jpg

 

High Tier Cabal (Noun): A group of people who used to play Hawken.  


#32
ReversusRex

ReversusRex

    Advanced Member

  • Members
  • PipPipPip
  • 91 posts

It's 26000 unique account log ins across the month timeframe between May and June 2016.

welp



#33
hestoned

hestoned

    Advanced Member

  • Members
  • PipPipPip
  • 251 posts

^

 

Yes because overqualified individuals that are 1000 MMR above the average of the server should be rewarded for playing people much lower than ones in their own skill level.

 

yea thats how it works. again you should practice reading comprehension. i was obviously not stating this to be a problem. i was dispelling the notion that its even possible to grind mmr off lower level players. i mean technically you could but it would take months to see any kind of gain. again please think before you speak.


  • DemitronPrime likes this

#34
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

yea thats how it works. again you should practice reading comprehension. i was obviously not stating this to be a problem. i was dispelling the notion that its even possible to grind mmr off lower level players. i mean technically you could but it would take months to see any kind of gain. again please think before you speak.

 

You need a 3rd grade class in sarcasm and how to detect it.

 

I thought that having to repeatedly do something over and over for a long time in online gaming is the definition of grinding.

 

You completely contradicted your own statement, bravo well done.  

 

In both of your examples, you ARE grinding MMR off of lower level players (lower relative to you, the average of the lobby tells us so), you just have determined its faster and more efficient in a 2500 lobby than in a 1800 lobby, and NOT THAT IT IS IMPOSSIBLE.

 

Your example is only valid for the few players near your own MMR (what 1% of the population according to the graph provided by SilverFire maybe less).  A player with a lower MMR (say 200 lower than yours) and using your provided example of a 1800 average server should they end the game at the top of the leaderboard would experience a higher result than your declared gain of only .01 because their MMR is closer to the servers average than yours was.  Now if we take another player with a lower MMR (say 400 lower than yours), in another 1800 server they should experience a higher result than both you and the player 200 above them, again because the servers average is closer to theirs than yours and the player 200 above them was.  We can continue repeating this and in every case until we match the players MMR and the sever average MMR, the amount gained gets bigger and bigger up to the imposed limit of 10.  This is the reason why you saw greater results as the server average became closer to your own MMR.  Its simple 3rd grade math.


Edited by SS396, 08 January 2017 - 02:09 AM.

[DELETED]

 

fuzzy bunny you CZeroFive 


#35
DemitronPrime

DemitronPrime

    Advanced Member

  • Members
  • PipPipPip
  • 879 posts
--blah blah condescension and ego -

 

you need to get off your high horse and take a lession in how to communicate without being a complete douche bag....

 

may not have been your intention, but of all the posts i have seen of yours today all i see is " i think im better than you with no justified reason "

 

maybe work on them social skills...

 

yes coming from ME that is ammusing


Edited by DemitronPrime, 08 January 2017 - 05:06 AM.

Puck Flinging-Nade Lobbing-Troll Tech

r4GY0PF.jpg

::eoc Pred::Eoc/HC Infil::Vulcanador::Hawkins Tech::EOC Raider::

" I play on north American servers with a 200Ping to give you cry babies a chance"

2XhpJes.png

"Nothing says Good morning Like a well placed Nade"


#36
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

you need to get off your high horse and take a lession in how to communicate without being a complete douche bag....

 

may not have been your intention, but of all the posts i have seen of yours today all i see is " i think im better than you with no justified reason "

 

maybe work on them social skills...

 

yes coming from ME that is ammusing

 

What I find amusing is why have you felt the need to only criticize me in this situation?  Why is it acceptable for one individual to repeatedly use insults such as "you need to practice reading comprehension" and "think before you speak" when responding to recent posts (in more than just this thread too) of another specific individual, and yet when you read it and decide to respond to it you completely ignore the behavior of the other party and only identify the issues with the second?  Were they being just as antisocial and unfriendly in their post or not?  Did the offending phrases have any meaningful purpose in their post other than to personally attack and insult the other user and elicit an angry response i.e. trolling?

 

When he can manage to respond without including phrases that are strategically placed to insult someones intelligence, he will get responses that are are free from the return fire.

 

Its simple, treat others as you wish to be treated.


[DELETED]

 

fuzzy bunny you CZeroFive 


#37
talon70

talon70

    Advanced Member

  • Members
  • PipPipPip
  • 264 posts

(Some of you spew so much crap at each other I just skip over all of your posts. There might even be something useful in one but I will never know. )

 

A bit ago I was in a match started 1250-2150, and after a few minutes it was horribly stacked and 1350-2600. It was around 12-2 and I said  something like ' dont pound the noobs' and stopped playing pretty much hoping for some balance to set in. It got worse. I really didn't play, except for droppin enemies orbs, and it still ended 40-xx blowout. Smurfhearding mentality might have led to the unbalanced pounding but it was still 13-2600 spread and that should be unacceptable.

 

Split the servers, and force 3v3 or 4v4 matches for now to make them even, and maybe also penalize leaving players until the end of the left match to encourage them to stick it out.

 

It seems to me 3v3 or 4v4 is really all the current server/backend situation seems to be able to handle right now anyhow. (ping limit might help also as a couple  200 ping player seems to kill it for everyone even at 4v4 imho). Population increase will solve almost all of hwkn's problems but server performance is hurting the population, therefore the $. There won't be better pipelines without more $. So  there's the vicious fuzzy buzzy cycle aint it folks...

 

edit'd for spellin errs


Edited by talon70, 08 January 2017 - 10:54 AM.

  • eth0, DieselCat, JackVandal and 2 others like this

#38
DieselCat

DieselCat

    Advanced Member

  • Members
  • PipPipPip
  • 683 posts

(Some of you spew so much crap at each other I just skip over all of your posts. There might even be something useful in one but I will never know. )

 

A bit ago I was in a match started 1250-2150, and after a few minutes it was horribly stacked and 1350-2600. It was around 12-2 and I said  something like ' dont pound the noobs' and stopped playing pretty much hoping for some balance to set in. It got worse. I really didn't play, except for droppin enemies orbs, and it still ended 40-xx blowout. Smurfhearding mentality might have led to the unbalanced pounding but it was still 13-2600 spread and that should be unacceptable.

 

Split the servers, and force 3v3 or 4v4 matches for now to make them even, and maybe also penalize leaving players until the end of the left match to encourage them to stick it out.

 

It seems to me 3v3 or 4v4 is really all the current server/backend situation seems to be able to handle right now anyhow. (ping limit might help also as a couple  200 ping player seems to kill it for everyone even at 4v4 imho). Population increase will solve almost all of hwkn's problems but server performance is hurting the population, therefore the $. There won't be better pipelines without more $. So  there's the vicious fuzzy buzzy cycle aint it folks...

 

edit'd for spellin errs

 

Talon..U hit on some great points and I agree...Most people understand that the only real impactful solution is many more players on a consistent daily basis. But if and when that happens, your above suggestion make a lot of sense....I would like to see those things implemented [Toot-sweet] by RLD for the interim until the game starts to see more improvement and signs of becoming successful for the long (longer) term.      


  • DemitronPrime likes this

Just Relax....and take life one game at a time....

Don't run to your death....walk

 

th_Duckman.jpg   th_82c0a97c-98de-4aac-be47-05e5e099be80.

 

*+

 


#39
SS396

SS396

    Advanced Member

  • Members
  • PipPipPip
  • 691 posts

Talon..U hit on some great points and I agree...Most people understand that the only real impactful solution is many more players on a consistent daily basis. But if and when that happens, your above suggestion make a lot of sense....I would like to see those things implemented [Toot-sweet] by RLD for the interim until the game starts to see more improvement and signs of becoming successful for the long (longer) term.      

 

The only suggestion of his that I see that they could implement right this second is lowering the match player limit on the servers to help create more lobbies.

 

The rest of them are just not all that possible in the interim, because the current system doesn't have the ability to kick people off of servers and move them to new ones, or punish quitters (mostly because they are already exploiting a console command (at least on the PC) and avoiding the team abandonment stat penalty).

 

Right now is the perfect time to create an entirely new matchmaker from scratch, and test it.


[DELETED]

 

fuzzy bunny you CZeroFive 


#40
Silverfire

Silverfire

    Advanced Member

  • Members
  • PipPipPip
  • 1395 posts

 

Right now is the perfect time to create an entirely new matchmaker from scratch, and test it.

 

Time is not on RLD's side, unfortunately. Is it a wise move, being that the word on the street is that RLD is going to release a PC patch Q1 of 2017? RLD would be the broken-promises company and Hawken would yet again prove itself to be the broken promises game.


Edited by Silverfire, 08 January 2017 - 04:08 PM.

  • SS396 and Miscellaneous like this

lNM7VnC.png

( ^ click for the EMP song ^ )

 

Come take a look at Hawken guides | Join me on #hawkenscrim IRC

 

 





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users