1. #12521
    Deleted
    50% win rate or not, the Matchmaking still sucks dick.

    All you have to do is watch someone like SingSing stream and his team mates are mentally retarded.

  2. #12522
    Retarded team mates don't seem related to MMR though.

    With that said, I think some of the most impressive kind of retarded player is the kind you explicitly tell "enemy hero is headed your way". You glance at minimap again a moment later. The player in question has engaged the enemy duo lane while their support is halfway back to fountain. And then the enemy hero you warned them about well in advance finally appears behind them to take the kill to boot.

  3. #12523
    Deleted
    "Mid is missing guys" - On voice chat

    "Use the chat wheel"

    "Draws a line where the enemy is going"

    "Pings the bot rune / top rune and says middle missing again"

    "Bot / Top gets ganked because they didn't move"

    "Rages at mid"

    Sums up a lot of players.

  4. #12524
    Like I'll be happy to not mind them simply admitting they missed it if they were playing someone who could genuinely fight a duo on their own, but it's always a bloody Spectre or something too.

  5. #12525
    Quote Originally Posted by Ariadne View Post
    He didn't get spoonfed easy games because the account had low win rate but because the account had low MMR. An account with 50% win rate doesn't say anything about the MMR it has.
    I was going to write out a big post but, neither of us have proof of anything and everything we're talking about is just like our opinion. All I know is pizza linked me the screenshot of the 6400 rated player who had a game full of <5k people the other day. How do you think the enemy player at 5k felt? His whole team is 4k rated players and the enemy has a 6400 god tier player ready to stomp him flat wherever he goes so he can't even abuse the enemy team's lower ranked players.

    But, yeah, that was a totally fair game and matchmaking doesn't suck at all.
    Honestly, if I was to take ranked seriously, I would cancel and requeue every 30 seconds so the search range doesn't have a chance to increase more than a little bit. It seems common knowledge to me that if you let the queue go more than a few minutes your games are virtually guaranteed to be a joke.

  6. #12526
    I think the biggest problem is the inflation of 3k +/- 150 after the evaluation games.
    Its a pinnacle of issues that is not easy to solve due to zero sum system.

    That rating have both players who have a lot of games/experience and freshly new players. This is the most unfair rating that can be there punishing both sides, new and experienced players which in the end couses both sides to be stuck at the place they dont really belong.

    I think community and valve should actually do something about it but I dont see an easy solution.

    I just played on my bro account, that has 2000 rating again for few and I must say, people ARE BETTER than at the f**** 3k rating becouse you end up with people who played that game for at least a while, they are not snowflakes and it is seriously easier to carry them and not loose the game in first 10min completly without a chance to solo recover.
    Last edited by Sorcereria; 2014-03-17 at 05:47 PM.

  7. #12527
    Deleted
    @Lysah

    First off, nothing of what you said has anything to do with what was discussed earlier about forced 50% win rate.

    Secondly, no, it's not my opinion. It's the fundamentals of how an ELO system works. The logic behind Arteezy getting matched with those guys was that the system couldn't find players at his MMR. Notice how the overall MMR in that game is still balanced and the system does try to keep the discrepancies between players' MMR small but if it can't it will increase the search range to give you a game in the first place. If it didn't increase the search range Arteezy would never get a game. Ever. Do you think it's reasonable that people at 5.3k+ never can play a game?

    And your example is an extreme case. In 99% of the cases it's much closer than that but it's easy to take that 1% case and use it as a base for why the system is not working. I'd say people between 2k-5k have balanced MMR games and not extremes like that in 99% of their games. Even for Arteezy that game was an extreme case but people noticed that one game and posted on reddit. It usually isn't that large of a difference, trust me I know after watching his stream.

  8. #12528
    I haven't seen a spread larger than 150 rating in any of my games yet, and I play a lot of late night dota

    MMR has been wonky for the topmost of players since beta began, this isn't new

  9. #12529
    I am Murloc! Scummer's Avatar
    10+ Year Old Account
    Join Date
    Jul 2011
    Location
    England
    Posts
    5,262
    The only issue I really have with the current MMR system is that I'd like to see your performance in games have more influence on the rating outcome than it currently does. Not something stupidly big, but I often find that current system of deviating by +1 or -1 to be a little bit disheartening when I've either had a bad or good game.
    I've had games where I feel like I really sucked and games where I feel I deserve a pat on the back, but I don't feel my performance ultimately has any major bearing on the rating change I receive.
    Last edited by Scummer; 2014-03-17 at 06:43 PM.

  10. #12530
    Deleted
    Quote Originally Posted by Longview View Post
    I haven't seen a spread larger than 150 rating in any of my games yet, and I play a lot of late night dota

    MMR has been wonky for the topmost of players since beta began, this isn't new
    There's a really low % of players higher then 5.5k MMR playing at the same time, and the MMR spread only happens in games they're in. Games 2k-4k usually have more then enough players queuing at any hour to allow for a relatively low MMR spread. There are quite a few graphs on reddit about this.

    Its fair to say your possible chance to win in a fresh new ranked matchmaking match is NOT 50%, and that can always be frustrating. Its not correct saying the matchmaking system is deliberately making you lose.

  11. #12531
    Quote Originally Posted by Aladya View Post
    Its fair to say your possible chance to win in a fresh new ranked matchmaking match is NOT 50%, and that can always be frustrating.
    there are a bunch of things that mmr cannot account for

    like the lowest rated player on your team going middle lane vs the highest rated player in opposing team, people instalocking heroes regardless of situation, people trying to learn new heroes, people with shitty attitudes, people that are not sober, etc

    also, just because games are one-sided doesn't mean the game wasn't statistically probable to go either way

  12. #12532
    Deleted
    Quote Originally Posted by Longview View Post
    there are a bunch of things that mmr cannot account for

    like the lowest rated player on your team going middle lane vs the highest rated player in opposing team, people instalocking heroes regardless of situation, people trying to learn new heroes, people with shitty attitudes, people that are not sober, etc

    also, just because games are one-sided doesn't mean the game wasn't statistically probable to go either way
    Yeah, exactly.

    Anyone else saw the Dreamleague Sigma-Rox.Kis incident this evening? FATA leveled -armor on SF at level 1, instantly abandoned the game, Sigma asked for remake while Rox.Kis were even doing a level 1 Roshan? No idea why the admins even allowed this, should of been instant win for Rox.Kis. And some people actually think Sigma has a chance for a direct TI invite

  13. #12533
    I think the remake was fine but rox.kiss should have been given a free rosh at start of the remake. FATA missclick would have insane impact on the game considering he basicly fuck up his lane with armor red.

  14. #12534
    Deleted
    I think Sigma should be disqualified for their disgusting behaviour and how they reacted after the incident and how they called Rox.KiS out.

  15. #12535
    Quote Originally Posted by Ariadne View Post
    And your example is an extreme case.
    I have yet to get a game faster than 4 minutes and max or close search range. Though, I haven't played since they started showing MMR so I can't say more yet.

    Also can you link to Valve's release on the specifics of how their system works, I must have missed it. Last I heard we have no idea exactly how the system matches up players outside of the facts that it wants to keep the average MMR the same and the MMR delta is small as possible. If we were talking about 3k rated games where there a thousands of people online at once to pool from I would agree the system is probably close to as good as it's ever going to get. But the system breaks down at the extremes, like all systems.

    I have no data for this outside of looking at LoL stats, but it seems plain to me that carry players have a higher MMR than support players, simply because they tend to player higher impact heroes that can more easily change the outcome of a game on their own. Therefore, the system is fundamentally flawed because of this inherent MMR bias. If you take two teams of 5000 MMR players, team A is 5 carry players and team B is 5 support players, team B is way more likely to win. The players are better, regardless of what their MMR says. Maybe their carry will be worse, but as a team they will do a much better job and when you get to 5000 rating even bad players can still farm with space and get items, which is all a carry needs to do honestly. The game will end up looking like a stomp because you will have 3 players on A who never play support trying to ward and roam and they will get stomped by the natural supporters of B, which will turn the game into a 25 minute rax game. A might have better carrys, but after they lose map control by 10 minutes they won't be able to farm so it won't matter.

    When I played League (though I admit is has been quite a while) the top 100 in ranked were almost exclusively ADC players, and for good reason. The ADC gets all the glory and he also wins the most games which inevitably results in the highest MMR. The same is likely true of DotA, just go look at any professional team and see who the shining star of the group is; 9/10 times it's the carry player that everyone considers the best. You watch Arteezy and from what you tell me he seems to pretty much play mid exclusively. Surprise? I also have plenty of friends around the 4500 bracket, the carry players are all higher than the supporters bar none. I actually made friends with a support player @5k during my placement matches and I can easily see that he is a far better player than the 5k carry players I know. But he can't win games as a support, not by himself, not like the carry players can.

    All I'm saying is that MMR means a lot, but it doesn't mean everything. As longview said, it can't account for people with bad days or who get their least favorite heroes. And it also can't account for who plays what role. Pubs always yell at me for last picking a carry instead of a support (because support never gets picked in ranked - go figure, everyone already knows their winrate will be higher as a carry than support) but I can promise you I will have a better game as a carry and will be more helpful to the team. Manni says I suck at supporting, maybe I do, I don't know I only do it once every 500 games. But I can definitely do better with no gold trilane support Timbersaw than I can with CM.
    Last edited by Lysah; 2014-03-18 at 12:11 AM.

  16. #12536
    The Patient
    10+ Year Old Account
    Join Date
    Jul 2010
    Location
    Tromsø, Norway
    Posts
    238
    So, I've recently attained level 13 and started looking at ranked matches. But I find myself with an issue when it comes to my playstyle, it's rather lacking in impact. This regardless of whether I'm playing support or semi-carry. Usually ending the game something like 1/10/14 in a game where I was supporting as Ogre Magi: http://dotabuff.com/matches/565792992
    Most of my deaths come from getting focused hard, by either initiating too early or being in the wrong place at the wrong time. Both which I'm trying to work on, but playing it more defensively will reduce my impact even further.

    Another example would be my last game as Juggernaut, where I lost the lane completely (not unexpectedly, as it was me and Bristleback vs Tinker and KotL), but the top lane (Viper/Mirana vs Spectre and Crystal Maiden I think) managed to win theirs and turned the game around, still I ended up with 7/8/10 in that game as well.
    http://dotabuff.com/matches/566972242

    TL;DR How I can make more impact on games when I die too often and how can I reduce my deaths as much as possible?

  17. #12537
    General rule taken from WoW. "Dead people do no dmg" applies in dota aswell. Many people dont understand that.
    I think the game as jugger while I didnt watch, I feel you probably lost the lane becouse you didnt play agressive from level 2. Bristle with 1/1/0/0 on level 2 with you should win that lane but you cant be passive and wait for kotl to supply tinker with mana as a battery. You do needed the kills early on to stay ahead and be able to sustain the lane itself with healing ward.

    EDIT. Bristle level goo first time at level 10. 9 level too late. You also did a little mistake, picking ward at level 2 instead of stats. 2/1/0/0 + 1 stat at level 2 would be much better. Also tinker went march/rocket build that should grant you easy kills since he had no laser for -100% hit debuff.

    The 3k MMR has its big flaw I said earlier that basicly, regardless of whether you play a carry or not slows your progression.

    For once I think the system at this mmr should account for how many real games you played prior to level 13 untill you reach around 3500 for those that just started MMR.
    I also dont get the way people enter MMR with base 3k without some other system in place. Its completly stupid, my bro that is at 2.1k right now is much better than most at 3k due to starters getting matched with experienced.
    Last edited by Sorcereria; 2014-03-18 at 01:56 AM.

  18. #12538
    The Patient
    10+ Year Old Account
    Join Date
    Jul 2010
    Location
    Tromsø, Norway
    Posts
    238
    Yeah, I felt that we could've stood a chance by as you said the Bristleback didn't really do what he is supposed to do, while I was probably giving them too much control. I would normally skill stats at level 2, but due to the heavy hitting harass from Kotl and tinker I felt that the ward would let me stick around to at least get some xp.

    And I certainly don't really feel at home at the rating level there, I haven't really been looking at the rating level I was playing with/against. I need a lot more than 200 some games (with ability draft, which dotabuff doesn't track) to get the experience needed to climb brackets. Probably doesn't help that when I queue with my friends I tend to either get opponents at their level or my level, I'd say they are around 4-5k rating at least.

    Thanks for the tips and information, much appreciated.
    Last edited by Keldarin; 2014-03-18 at 02:11 AM. Reason: Forgot to say thanks.

  19. #12539
    Quote Originally Posted by Keldarin View Post
    TL;DR How I can make more impact on games when I die too often and how can I reduce my deaths as much as possible?
    buy wards
    look at minimap
    buy bkb sometimes

    also a general tip of winning games at low-midskill pubs: learn to farm. you shouldn't miss last-hits if you are uncontested. go lobby train that shit.

    Quote Originally Posted by Sorcereria View Post
    For once I think the system at this mmr should account for how many real games you played prior to level 13 untill you reach around 3500 for those that just started MMR.
    I also dont get the way people enter MMR with base 3k without some other system in place. Its completly stupid, my bro that is at 2.1k right now is much better than most at 3k due to starters getting matched with experienced.
    my alt account was seeded 4.1k+ at level 13 and im not especially good at dota.. mmr isnt random, you end up roughly where you belong

  20. #12540
    Deleted
    @Lysah

    http://blog.dota2.com/2013/12/matchmaking/

    Yes, pure support players will probably have lower MMR because they have less impact. But a good player, no matter support or not can rise in MMR by picking mid/carry heroes until they reach close to their true MMR and then play the role they want to play.

    I disagree about the 5v5 support vs carry players. It completely depends on the players and no one is ever a pure. I think I got a firm grasp on how to play support, how to ward and general map rotations. The reason for why I'm quite confident in my supporting skills is because by playing mid and carry a lot I know what I want of my supports. I know where to place wards to make it easy for my mid and carries and general places enemy supports place theirs because when you are playing a ganking mid you have to be aware of their vision to make most use of your ganking potential. I also realize the importance of carrying dust/sentries. I also got a good understanding of support rotations because they actually rotate on me when I mid and I need to understand their rotations to be successful mid.

    And what you're forgetting is that the mid/carry players will win their lanes because they are generally stronger laners and as you've stated, mid/carry heroes have more impact. While a pure support player may know about lane equilibrium, creep aggro tricks and last hitting it doesn't matter if they don't do those things on a regular basis. The guy won't be on the same level as a full time mid or carry. My friend who mostly play offlane, support and jungle gets destroyed by me on the mid lane both in last hits and lane control and we are about same rating and I don't even consider myself a good last hitter compared to most full-time mids.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •