Kat XD (NA)
: Yeah because Sivir+Vi kept trying to dive and he just counterganked everytime, then post-6 TF joined and they all got got super fed, impossible to deal with. I had a solid lead on everyone on the enemy team in the early game, but I can't keep up vs semi-global roams...
I don't like your team synergy at all. Temmo wants folks to push and hit mushrooms. Vi and Leo want to go ham. You want to have a window to R and blast their back line. Your Dive is all single target besides leo whos super small aoe with short duration and their comp can peel off targets. Their comp... Protects the koggles. Looks more like the issue was 1 team picked who every they like and did not give a shit about comp. The other team might have done the exact same but their comp was way more synergy. Woulda been interesting to see if your team had a different comp. Going off the towers looks like you guys couldn't spit... but to try and Split vs a TF and WW with squishy champs seems unlikely. Going from damage and team kills+ gold gap. I'd put this on decisions made in game and lobby more then I would for the system that built the lobby.
Kat XD (NA)
: Just another complaint about matchmaking that I really hope is addressed in the preseason.
https://imgur.com/a/zu0Pv75 Looks more like WW was the issue.
pwc2016 (NA)
: How TF are people already saying the changes are awful?
Every change will upset some one. Upset folks are consistently the most vocal. Wait for the next patch, keep an eye out. I expect some ones gonna show up and Bitch about {{champion:555}} getting nerfed and with in a few threads some one else will bitch that he did not get enough of a nerf and some one else is going to shout that Riots stupid because they nerfed the wrong things. Just got to filter out the emotional folks and look for people who are logical if you wanna have a decent conversation around here.
Cdore (NA)
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=MgNGMl3h,comment-id=000100010000,timestamp=2019-10-16T17:03:57.942+0000) > > A lot of the stats from those sports don't have as much noise and some one "padding" Their stats in a live game infront of several hundred folks would be terrible for your career, much in the same way "throwing" will destroy careers. Leauge players get away with it in game pretty easy. Folks give up and soft int, Folks can pad with out much fear of consequence. If in say Baseball Some one who wanted to maintain a high hitting average and does so at the cost of the other runners consistently. A Running back that wont take the ball and lets his OB get sacked because he did not like the look of the defensive line and wanted to ensure his rushing average was high. All done on national TV? > > League provides a high level of control over stats with low oversight. Abilities on CD are not a threat. Constantly spamming something provides far higher Stats. {{champion:30}} is my go to example. If a Karth player just hits R every time he has 5 targets the moment its off cd. That player will have far higher damage from R stat at the end of the game. This is sub optimal usage. > > If the enemy team has {{champion:55}}. You will have higher CC as {{champion:12}} if you spam headbutt Q any time you have a target.... but the right play is to reserve CC to stop {{champion:55}} R. > > Players can control ability usage far better then athletes can control their stats. With out incentive to win its always going to be less effort to pad stats. Here is a Quote from the publication for Microsoft Trueskill 2 which a MMR system with Performance adjustments. > > "Aligned incentives. The skill rating system should create incentives that align with the spirit of the game. For example, consider a team game where players cooperate to achieve a goal. An improperly-designed skill rating system could encourage players to impede their teammates. As another example, consider a system that increased skill rating according to the number of times a player healed themselves. This creates an incentive for a player to repeatedly injure themselves so that they could be healed. **In general, the more control that a player has over a quantity, the less useful it is for skill rating.**" > > When you sit down and ask, what stats objectively can be used for League that don't have easily generated noise? > What is fair to all classes? > What works for all roles? > Wins and losses is the only universal measurement that is not easily faked.The lack of a stat in some kits should make it very clear that you have to build for individual champions. > > If you cant sort noise on a game to game then it just serves as a feel good reduction of LP loss and I frankly don't trust this community to not aggressively be selfish the moment they think its GG. From conversations on the boards I know that for some folks that's literally 1 death in a lane. > > To sort the data game to game is difficult, requires what's best described as an AI judge. Computational requirements for that are high, just not realistic with current tech. You made a fair assessment on the emphasis of presenting skill. But if you're going to talk about noise, being 1 member out of a team of five, along with the variance of five other randoms, screams noise. The problem with the current system is that climbing is very misleading as a term. To climb, you have to win. But to win, you have to play the game to a winning state. There's certain things that make winning "easier," and that's when you get into stuff like tier lists. Some champs or playstyles or classes make winning easier. So solo ranked is more of a test of "how well can you carry in a game where you are queued with four other random players vs five random players that you have no relation with." I can agree that winning and losing a game is the best measurement, but you also have to agree, then, that the best way to show skill is not necessarily how good you are at your champ, or your playing, or anything concrete. All climbing proves is "How good are you at uprooting the status quo of a game and making it winnable by exploiting the fact that everyone in the game is random." Boosters use carries that can 1v5 or 1v9. Smurfs do the same, or duo with another high elo player. On all occassions of climbing, the measurement is not on "skill of playing league of legends". The skill is "how much can I cheat the system"
There's undeniably noise in MMR systems. However Balanced Comps > unbalanced comps. The "Easy" Win champions only function if you have the skill at them. Many players sacrifice a good comp to play a "OP" Champion who they already underperform on and harm their ability to climb. Its a law of large numbers style system that uses multiple data points to remove noise. Folks can spam OP broken... but if they don't achieve any results its idiotic for them to keep doing it. Skill gaps in Mobas are pretty clear and the impact of mismatches are frusturating as hell. This makes being Stuck annoying. Thing is, your not expected to climb when your in the right class interval, Ideally the # of mismatches is lower but for folks in the Center of the curve ( Looks to be built around silver 2 far as i can tell.) your most likely to see a under or overskilled player show up. So, Id put it at S4 to Gold 3 for the highest Mismatch possibilities. There are steps you can take for MM to lower # but it will never solve the frustration of a trashed game, Each win from a smurf/good teammate feels like shit for the other side much the same way a loss from a weak teammate sucks. Nature of the beast, one of the reasons I keep trying to promote 5 man premades as a viable competitive scean. MMR is just a modification on Elo, I think everybody agrees that MM is needed for League. Removing MMR/Elo makes it 100% RNG team distribution and that's fking awfully. Far as i can tell a lot of people don't take time to learn about even the basics for Elo. This kind of system is built around math theory not game theory. Its around the simple principle that in a group of humans you can find a natural average. Average age, hight, weight are pretty easy. IQ and Skill require a bit of testing. Why I constantly call the ladder a curve. Ladder is misleading for a lot of folks. Start at bottom. Go up rung by rung to top. "Climbing". End of the day everybody starts at the middle rungs of the ladder. Because you build off Averages, Forced 50% clearly is a weak argument. Cant accurately rig games with a metric that does not have any relation to skill if its rigged. No naturally occurring averages prevent you from using naturally occurring averages. The, 1 teammate holds me back Argument also fails from a math side. 4 slots on your team. 5 on the enemy. 100% of your games your on your team, so if your not it then logically the enemy team gets the weakest player in a game more then you. Thing is this is also true for if your not the strongest player. Balances out when your a representation for usual level of skill in a MMR. Otherwise you climb or sink respectively, as your always a MVP or your always dead weight. Thing is, despite abuse MMR is a reliably functional system. Boosted players simply cant compete. Just because some one buys a Master level account, they don't magical gain the skill of a master level player. This FK the measurement up, but it will self correct if left alone. With out the ability to back it up that players going to get demolished every game they play. If I had to make a RL example I suppose Its like Buying a Trophy at goodwill, You can show folks and Lie about placing 1st in some tournament, but if you ever had to back it up your lack of skill is apparent. I having helped a lot of friends over the years I would say that its painfully clear when some ones dead weight beyond any ability to carry. You can "cheat" but only to an extent. Smurfs have to expend effort to sink. They have to throw games. Have to Lower K variables, Have to make an effort to not pubstomp 100% of the time. Or they need to break TOS conditions and play on a purchased low MMR account or use some one elses. Imo, far as pub stomping goes, Its easier to just Tank MMR in normals by learning new champions and occasionally play a Main Tryhard style to pub stomp. Especially if you abuse premade. Some folks want to just be toxic and rub it in so they do it in ranked. Not a fan of that behavior myself. PvP… No opponents no game so be a good sport. You have the right concept to an extent, There's no universal equality in ratings. Like Looking at say a Gold 2 account. Some folks raiting is only valid for if they have a specific duo, only time they are gonna be around G2. Some for only 1-2 champions they thrive with. Some players have multiple roles and champions. Personaly I would say the person who can do the most champions and positions is more skilled then the other 2 options thou all 3 of them would earn the same Rank. Boards annoys me at times because the perceptions on what a good champion to climb is not remotely as relevant as what is a good champion to climb for an individual. For me? Supports are my best class. I can hit high gold and low Plat MMR with an off meta {{champion:432}} build and most my other picks are mid gold level of skill. Rough Estimate for me is it should usually take about 100 games of time invested to grind to that. Riots Implementation of Duoing is something You will see me be critical on, Mostly because its off titles not MMR. The Range of skill allowed is to large imo, going off the Player % by divisions and also because MMR moves faster then LP. Riots demotion protection at division 4 is massive. so S4 can be Bronze MMR and G1 can be Plat, the two players can que. And MM has to build a game with a gap of almost 80% of the player base between player A and B... out of who ever is online and in the que. Outlier scenario but its gonna strain MM. I use the phrase Class intervals because of Elo systems, its basically how you sort the players into groups where games are at reasonably fair % for each party to win. They are built off statistical standard deviation expectations for the probability models used (Wordy but... well that's the terms for them. No "layman" for them.) It sorts players to reflect the distribution. When they don't match it, that's how and why they move up or down. It never rests, it never stops readjusting. Long as folks don't break out of the expected deviation they are "stuck" as they will be hovering in that area getting more wins then losses if they slip to far, vice versa. Game quality is a non factor as gameplay is irrelevant to this kind of system. End results are used not the steps that get there. Elo/MMR being used for a massive range of unrelated topics, Online Radios, Dating websites, FPS, RTS, Mobas, Scrabble, Chess, Baseball. All because it works off mathematical concepts. Binominal. 1 or 0. Win or loss. Hot or not. Like or Dislike. Trinomial. Win, Loss, Draw | Yes no maybe | Bad, Good, Ok. End of the day. MMR is the best kind of MM system for a game like league. Riots MMR system has features that are unnecessary and there's improvement that could be made for quality at the sacrifice of popularity. MM is absolutely required for a Moba. The majority of shitty things in gameplay are the fault of shitty behaviors by players... PvP gameplay is driven by players. I am open to being wrong, that's never off the table. Yet when folks present emotional responses over logical reasoning I usually dig in my heels and demand logical steps. Do we need MM? Yes. Is there a better system for it then MMR? No. (open invitation for folks to correct me if this is untrue.) Can Riot improve theirs? I would say yes. Do players have unrealistic expectations? Oh hell to the yes. No idea how folks think Skill can be known with out proof of it first. Magic crystal ball? /end rant. If you make it this far thanks for your time and hopefully it was interesting.
Áery (NA)
: I really HOPE for S10 the ladder gets a HARD MMR reset
: How many times have you guys went through the same promo series?
My worst was like 9 attempts, Promo helper was not a thing thou... I think/hope
pwc2016 (NA)
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=LE45KWAX,comment-id=0003,timestamp=2019-10-16T06:34:29.335+0000) > > Mtg and LOL are built around perfect imbalance as a concept. Its fairly impressive how well both accomplish it. Boards will flip their shit over 3% win rate changes. But 54% is not at all a terrifying number. > > The concept of balance has many variations by individuals. I would love if collectively we just started to approach it as game health. To many folks want 50% WR with no accounting for skill floor or ceeling. Mobas and mrg need friendly low cap options. The learning curve is brutal. > > Homeland was one of the most "balanced"sets. 0 banned cards if my memory serves but its a boring fucking block. > > Rock paper scissor balance is bad for draft games. Team synergy to add viability? Relys on players knowing how to patch counters. Thats not this games playerbase. Its stupidly selfish for a team game. > > 5 man premades are the most competitive format. Harder to run. > > Just my take. Yeah you can give the balance team shit but the fact that we don't have champs with 60 or 70 percent win rates across all elos is fairly impressive considering some decks in MTG become so powerful they overtake everything else.
One thing i see as an issue with running balance in cycles. The expectation is that players will attempt to deal with it, adapt to meta shifts, and your job as a dev for this balance is to provide enough tools that there is always a solution/counter to anything you throw at them. High # of League players don't seem to understand this. MTG has a luck aspect, always. This can be brutal and require cards to get banned when something overperforms. Leauge has thousands of players who literally do the same things over and over and over. OTP who is hard countered by majority of meta? To bad...gonna OTP that every game. A champion who requires a item counter? {{champion:36}} for example? People will Cookie Cutter their build the same for each and every game... better hope the support takes ignite. People are unwilling to learn and invest into the skillsets that would give them tools to deal with an ever chaining meta. To be Fair, The advice people get is not good advice. "play hard carrys" Ok.... Welcome to the damage meta. How many folks spamming Feast or Famine hypercarrys and ignoring a balanced party composition this year? "just 1 trick something" Great as long as your good at it to a level your above your peers. Just takes less effort to learn matchups faster way to improve and climb but not ideal once you hit your cap. One tricking is a great way to learn a champion, once you have their mechanics you should look to widen your pool. How many League Players in the Pro level can be hard countered by the enemy using 1 ban?
: yeah but how do you improve behaviors when riot gives less and less punishments. back in season 3 if you seen a troll you wont see them again because they get banned for like a week off of that. now if they dont directly run it down its fine of ragequit after dying 2 times oh 5 games of 5 mins longer queue also when enemy team has 2 g4 players with 80% wr each in dia game and you got darius main who has 400 games with barely positive wr on champ its ovbius its gonna be a loss and being support you cant do a damt thing about it just sit there grining your teeth hoping for the best. sometimes thinking about quitting and going for a game on a smurf cuz there is no punishments for leaving anymore.
I consistently see folks who run it down getting Banned, thou folks with awfull KDA are fine. Detection seems to be chat and death locations. I don't mind a bad player losing a game as long as its not intentional. If they make effort to group and be of use I would rather have them on my team then some one whos good but rages then spends the rest of the game splitting or clearing jungle camps. Enemy at hibitor? Better go farm Top lane, that's helpful. I still see folks get punished for AFK or rage quitting. I don't spam reports. Overall I very much want gameplay to be more enjoyable, this involves Improvements to the behavior system. As much shit as the community gave Lyte, having a dedicated Red for behavior was good for the game. If it was up to me I would implement a system where you could spend Honor on reporting a player to escalate it to a human for review. If you spam reports for no reason you lose honor and cant use the escalation system. If you consistently remove Toxic individuals from the game get rewarded for it. Tribunal 2.0. Automation can only go so far, once its metrics are known values its easy to evade. 1337speach for example existed just to get around chat filters.
: I could get behind 5 man premades but they do have that in flex. Sometimes players just don't the friends to climb especially close friends. for example all my friends are plat in 5v5 flex and solo/duo. So i literally have no chance to even get that high which I have climbed to silver 1. Then recently been on a loss streak that sent me back to silver 4. 100% my fault, probably playing tilted and not playing to my strengths. Especially it being the end season climb I got alot of pressure to try and make it. But I have gone through about 15 series in silver and can't seem to get good, dare I even say RNG when it comes to fairly decent teammates, 90% of time im sweating just trying to get my teams to not give up. Series is what ruins the climbing experience. Inter division climbing I don't think there should be series. If you hit 100LP should be automatic promoted like TFT does. Then to get into a new division / tier like silver 1 to gold 4. I feel should be a best of 7 series etc. I feel that would give alot more players motivation/incentive to climb. I feel like there's so many unnecessary games being played. To be honest about the smurfing thing. I encounter smurfs especially in series about 50% of the time. I always seem to get one and idk why its always in my series. Could just be really bad luck. Like awhile back I faced a 2 diamond smurfs. Lucian and Riven. I watched this guy on twitch after he beat us, he literally was like, Yeah I can't climb in diamond anymore so I want to win in low elo. I tried to find the game but can't seem to find it or his twitch but was super annoying.
The best advise I have is ignore promos completely. If your playing well to consistently get to them just play like that. people freak out, "try harder" Aim to make dramatic and big plays or go hard for a snowball win. No idea why, if your climbing how you play normally just play normal. https://na.op.gg/summoner/userName=NotNotkaiguy I have gotten this account from Bronze 1 To silver 1 with 0 promo game wins. I dodge lobby every promo to make it count as a loss. The system plays catchup, if your MMR to Title gap is to large the system skips promos completely.
Cdore (NA)
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=MgNGMl3h,comment-id=0001,timestamp=2019-10-15T19:10:10.295+0000) > > The main issue is most bad stats overlap with good stats. The main metric in a game for succus is win or loss. If you go to a park and join a pick up game of any sport and demanded you get treated like a winner because you played well as an individual folks are going to think your an idiot and self absorbed. > > {{champion:30}} spamming R will have higher R damage then the one who holds it to secure kills or help a lane win a close fight. Same for CC, Heal. ETC. > An ability on CD is not a threat. Spamming results in higher stats, so just a larger value is not a clear indication of skill or proper usage. The context of how its generated matters. > > KDA gets multiplictivly lower each death. 10-1 =10 to 1. 10-2 = 5 to 1. So you have to value each part as an individual to run that as a stat, but as the games not just kills some one whos splitting and controlling map and OBJ loses out on generating relevant numbers. > > Overall, the effort required to pad stats is far less then whats needed to win. > > Many MMR systems do Performance adjustments, I would not trust the league community with this feature for a second. Generating good values for a moba is hard as well because again the weak stats overlap with good ones. It gets harder to build the more advanced you try to make it, like the # of combinations if you try to account for individual matchups? > > NA players need to get their heads out of their asses, Its a team game treat it like such. If its a fair match there's absolutely 0 reason to think your teamamts are all going to be superior and win their roles. Having a hissy fit then soft inting because 1 lane lost is idiotic. > > Treating Silver as Low Elo, for a system that effectively is grading on a curve? So stupid, The majority is "average" not Low. > > Individual statistical systems work best if they can remove noise by identifying context specific to a match. As its insane to think you can have humans grade each game it has to be automated and this requires a huge amount of computational resources and what would best be described as an Ai judge. Unrealistic with current tech. > > Statistical objective systems are easy to automate and have been around for several decades. Arpad Elo was made famous for his work on them. How come Baseball, Football, and Basketball have individual stats that determine the value of a player and determines their marketability, but a game like League can't do that?
A lot of the stats from those sports don't have as much noise and some one "padding" Their stats in a live game infront of several hundred folks would be terrible for your career, much in the same way "throwing" will destroy careers. Leauge players get away with it in game pretty easy. Folks give up and soft int, Folks can pad with out much fear of consequence. If in say Baseball Some one who wanted to maintain a high hitting average and does so at the cost of the other runners consistently. A Running back that wont take the ball and lets his OB get sacked because he did not like the look of the defensive line and wanted to ensure his rushing average was high. All done on national TV? League provides a high level of control over stats with low oversight. Abilities on CD are not a threat. Constantly spamming something provides far higher Stats. {{champion:30}} is my go to example. If a Karth player just hits R every time he has 5 targets the moment its off cd. That player will have far higher damage from R stat at the end of the game. This is sub optimal usage. If the enemy team has {{champion:55}}. You will have higher CC as {{champion:12}} if you spam headbutt Q any time you have a target.... but the right play is to reserve CC to stop {{champion:55}} R. Players can control ability usage far better then athletes can control their stats. With out incentive to win its always going to be less effort to pad stats. Here is a Quote from the publication for Microsoft Trueskill 2 which a MMR system with Performance adjustments. "Aligned incentives. The skill rating system should create incentives that align with the spirit of the game. For example, consider a team game where players cooperate to achieve a goal. An improperly-designed skill rating system could encourage players to impede their teammates. As another example, consider a system that increased skill rating according to the number of times a player healed themselves. This creates an incentive for a player to repeatedly injure themselves so that they could be healed. **In general, the more control that a player has over a quantity, the less useful it is for skill rating.**" When you sit down and ask, what stats objectively can be used for League that don't have easily generated noise? What is fair to all classes? What works for all roles? Wins and losses is the only universal measurement that is not easily faked.The lack of a stat in some kits should make it very clear that you have to build for individual champions. If you cant sort noise on a game to game then it just serves as a feel good reduction of LP loss and I frankly don't trust this community to not aggressively be selfish the moment they think its GG. From conversations on the boards I know that for some folks that's literally 1 death in a lane. To sort the data game to game is difficult, requires what's best described as an AI judge. Computational requirements for that are high, just not realistic with current tech.
SIvIaRsH (NA)
: That seems like its the problem tho. Games are literally being given up on just by NA players giving up because something isn't going their way etc. Its depressing and gives me like less of a reason to climb. if that makes sense? I try to play to my best but players seem to give up so easy
Thankfully its more likely to be common on the enemy team as long as you don't tilt. I am grinding a lower level account with just supports right now to show how all roles make impact. Tell folks at the start its a Tilt meta and to mute toxic players. Politely request folks who are soft inting to stop. Silver is rough simply because its how you blend new accounts into the ladder fastest. You can add modifiers to the MM algorithm to protect individuals but end of the day there will be awful mismatches because of high uncertainty accounts. People don't think about the ladder as a curve very often or the impact that MM has to make a game. 10 years now and people still don't expect the massive dip in game quality when Schools are out on breaks. End of the day the best competitive experience is to have a group you play with regularly, sadly the most popular experience for competitive play is Solo/Duo. Clash if Riot can incentivize it stands to improve game quality but trying to nail the calibration is hard. What Id like to see Riot attempt is to upgrade their Club function to reward participating in a community, then to have 5 man premade be the competitive que, then run a "grassroot" style semipro tournament that puts 1-2 winning teams onto the pro stage. Games like this are better with friends, higher end strategies need individuals to have strong synergy. I am consistently vocal and consistently unpopular for my views on the need to improve behavior systems > MM. Sadly, People get mad, I tell them PvP games gameplay is determined by players. This is the logical attitude. That MM ends once a lobby is made. This is a logical view. I don't do a good job accounting for them getting over their emotional response to a shitty match. I'm really bad at dealing with that. End of the day a lot of folks have the same view, we just want the game to be better. For me, anything that people hold up as an excuse to give up in games is awfull. So I throw the math and logic behind Elo styled systems in their face. Curve, Large #, averages, etc. Cant build a forced 50% off averages if your using that same forced 50% for your averages, the math just wont work out. Its not something some one who just lost 5 games to asshole teammates wants to hear. Personally I blame the players who do it, not the MM that ends with a lobby. More teamplay would make this game better imo, so I favor the idea of 5 man premade and tighter groups in clan/club being promoted. Add incentives. Like if a clan buys a "battle pass" All members gain it, but the missions require being in a 5 man group. Up the club size to a high enough # that folks can play consistently. Add some Chat function back to the game to find groups who's free time match up.
Destaice (NA)
: Did Lucian know holding Senna would save her from the lantern’s pull?
Dog, Everybody knows you bodyblock the lantern by now.
pwc2016 (NA)
: Why Is Balancing so Hard?
Mtg and LOL are built around perfect imbalance as a concept. Its fairly impressive how well both accomplish it. Boards will flip their shit over 3% win rate changes. But 54% is not at all a terrifying number. The concept of balance has many variations by individuals. I would love if collectively we just started to approach it as game health. To many folks want 50% WR with no accounting for skill floor or ceeling. Mobas and mrg need friendly low cap options. The learning curve is brutal. Homeland was one of the most "balanced"sets. 0 banned cards if my memory serves but its a boring fucking block. Rock paper scissor balance is bad for draft games. Team synergy to add viability? Relys on players knowing how to patch counters. Thats not this games playerbase. Its stupidly selfish for a team game. 5 man premades are the most competitive format. Harder to run. Just my take.
: i have like 4cs per min
Terozu (NA)
: New Drake effects and other PreSeason changes!
Thanks. At work so I can't stream. If you update reply so I can read up when its slow.
: Apparently you can control your Ashe ult to bend in Mobile LoL
: Here here, to all the who have taken the role of Bronze, come and bare your pride; your Bronze Pride
: I’m so tired of “it’s your fault you’re in low elo”... Maybe give advice instead of having an ego?
Can you Share a replay of a game you feel you did meh and one you did good?
: Do any of the reds aram? It's all I've been playing lately.
I assume yes. Who thou I don't know. I will ask if I end up chatting with one.
Rioter Comments
: I missed you all
Welcome back. I miss your no nonsense replays to folks.
: Remove/gut yasuo, riven, and fiora. They are absolutely unwanted and they just 1v9/destroy low elo
Best thing you can do is play them. Get their mechanics down to understand how to react. Most champions pub stomp if there is a large skill gap.
: First time ever in Gold Promos
Stay calm and treat them like any other game. If you tilt after a regular loss, don't que up after a promo loss. Trying to Play differently from whats gotten you to that point is a mistake. Treat it the same as any other match. Ignore the X out of 5 count completely. Promos serve to slow "lucky" streaks and also give the system extra time to dial in on an accounts MMR range. They also make players feel "pressure" and have a High Stake mentality. MMR is what matters thou. Just play well and consistent. https://na.op.gg/summoner/userName=NotNotkaiguy I have made an alt account and leveled it from Bronze 1 to Silver 1 and did this with out passing a single promotional game. MMR is what matters, Promos are a distraction. Dodging lobby during promo fails it. I have done this every time on this alt account to test what the MMR catchup looks like.
SIvIaRsH (NA)
: Possible ranked changes? Based on how TFT ranked is / OP.gg ranking system
The main issue is most bad stats overlap with good stats. The main metric in a game for succus is win or loss. If you go to a park and join a pick up game of any sport and demanded you get treated like a winner because you played well as an individual folks are going to think your an idiot and self absorbed. {{champion:30}} spamming R will have higher R damage then the one who holds it to secure kills or help a lane win a close fight. Same for CC, Heal. ETC. An ability on CD is not a threat. Spamming results in higher stats, so just a larger value is not a clear indication of skill or proper usage. The context of how its generated matters. KDA gets multiplictivly lower each death. 10-1 =10 to 1. 10-2 = 5 to 1. So you have to value each part as an individual to run that as a stat, but as the games not just kills some one whos splitting and controlling map and OBJ loses out on generating relevant numbers. Overall, the effort required to pad stats is far less then whats needed to win. Many MMR systems do Performance adjustments, I would not trust the league community with this feature for a second. Generating good values for a moba is hard as well because again the weak stats overlap with good ones. It gets harder to build the more advanced you try to make it, like the # of combinations if you try to account for individual matchups? NA players need to get their heads out of their asses, Its a team game treat it like such. If its a fair match there's absolutely 0 reason to think your teamamts are all going to be superior and win their roles. Having a hissy fit then soft inting because 1 lane lost is idiotic. Treating Silver as Low Elo, for a system that effectively is grading on a curve? So stupid, The majority is "average" not Low. Individual statistical systems work best if they can remove noise by identifying context specific to a match. As its insane to think you can have humans grade each game it has to be automated and this requires a huge amount of computational resources and what would best be described as an Ai judge. Unrealistic with current tech. Statistical objective systems are easy to automate and have been around for several decades. Arpad Elo was made famous for his work on them.
GilxeN (EUW)
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=Uka8FpyI,comment-id=000000000000000000000000,timestamp=2019-10-15T07:48:57.671+0000) > > I instead went and played another game. Learning some of the basics for the mode resulted with me snagging first. https://imgur.com/a/wLnGNTQ > > Weird how it seems like only the folks who lose constantly and don't know how to play it whine about luck and RNG. > > Meanwhile some one who has by all accounts a fairly basic understanding of the mechanics can consistently place. 4 out of 4 games. Different comps every single time. > > Feel free to respond for a 4th time... Keep on discussing the topic. It really sells how much you don't care.{{sticker:slayer-jinx-wink}} What's your point lol I don't get it, yet u didn't understood simple sentence "I dont care". Why so ignorant?
Your sharing your negative opinion but also totally like you don't even care at all, not in the littlest bit. Which is why you feel the need to respond 5 times on the topic. I am just calling attention to how your putting in a lot of effort for this not caring thing. Kind of counter productive on your end. Prior to the "I don't care" You have some pretty negative emotional reactions. Basic logic here is those two viewpoints are contradictions. Its ok. Not everybody is going to be good at games like TFT. No shame in it. Don't like it don't play it. Don't care don't take the time to proactively whine about it. You should also avoid poker and deck builder games like MTG and Dominion. There is no shame in that, but to look for validation of your view but also not care is silly. Literally no one is demanding that you have to reply. Just ignore me completely and don't care if that's what you want.
GilxeN (EUW)
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=Uka8FpyI,comment-id=0000000000000000,timestamp=2019-10-15T06:48:59.227+0000) > > https://imgur.com/a/Jh6WdnF > > I like to let players Pick my comp. Then aim to place with it. > > I love folks who don't learn the basics and just hope for the best or mindlessly obey tier lists. > > In addition To wild being a hard counter to dodge a high # of ability's function just fine as a counter to yodels, and simply grabbing some Gnars can knee cap the comp. What are u talking about lol, i dont give a flying s'hit about this game mode, delete the post already.
I instead went and played another game. Learning some of the basics for the mode resulted with me snagging first. https://imgur.com/a/wLnGNTQ Weird how it seems like only the folks who lose constantly and don't know how to play it whine about luck and RNG. Meanwhile some one who has by all accounts a fairly basic understanding of the mechanics can consistently place. 4 out of 4 games. Different comps every single time. Feel free to respond for a 4th time... Keep on discussing the topic. It really sells how much you don't care.{{sticker:slayer-jinx-wink}}
NeekOwU (NA)
: If I lose lane as an enchanter, what can I do?
Remain calm, its going to happen don't let it blow your mentality out the window. Imo players who think they need all 4 teammates to be not just equal but superior to the enemy counterparts to win are oblivious to the fact that it strongly implys they are dead weight on a team. For lane try to tell if its cuz your ADC is weak or just gave up after a few deaths. Soft inting behaviors are sadly very common atm with folks just going thru the motion of playing to avoid getting nailed with a leave buster punishment. Look to identify who is useful on your team that's playing for a win. Work around them primarily in fights. Depending on your comp Aim for all around helpful support items like {{item:3107}} or aim to get some damage to compensate if that's needed. I like Sona powercord for damage reduction as it can really destroy some ones damage pattern if they don't expect it, most folks I see pretty much only ever spam the Q powercord for more damage. Aim to Target the teammates who will pull weight, and aim to interfere with the enemy who pulls weight. A lot of folks burn spells just because they see a target. Do you really need to R their front line instantly, or can you hold onto it and use the threat of the ability to keep some of their back line farther away. {{item:3107}} RWQE is a lot stronger then {{item:3504}} W when your most helpful teammates are not auto reliant. Be ready to adjust your build on a game to game to match game to game needs. You don't have to CS, so you can communicate. I try to get ADC's to focus on their next item power spike or coordinate help with mid or jungle. Don't be overly chatty as that's gonna result in a lot of back and forth chatter between teamamtes you don't want.
GilxeN (EUW)
: > [{quoted}](name=Jamaree,realm=NA,application-id=3ErqAdtq,discussion-id=Uka8FpyI,comment-id=00000000,timestamp=2019-10-15T05:57:15.169+0000) > > Sorcerers is literally the counterplay? It was my extremely BIASED opinion, I don't play this garbage gamemode lol. When I tried to play this game and realized it's a lottery RnG, dumped this game mode.
https://imgur.com/a/Jh6WdnF I like to let players Pick my comp. Then aim to place with it. I love folks who don't learn the basics and just hope for the best or mindlessly obey tier lists. In addition To wild being a hard counter to dodge a high # of ability's function just fine as a counter to yodels, and simply grabbing some Gnars can knee cap the comp.
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=gwaoPkzH,comment-id=0000000000000000,timestamp=2019-10-14T23:16:27.037+0000) > > Well I made a comment in the thread as well. Its a race who sinks more. Feel free to call me out if I give bad information btw zero, I will do the same. I really don't get why folks take it so personal so fast. Works for me! Let's see how it goes! So far your advice seems sound. Particularly the last bit, a _lot_ of people do that at my MMR. Some even use Flash to get to that range and then die three seconds later...
My go to advice. If you think something is "broken op". Then go Play it. Learn the kit to learn the counters.
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=gwaoPkzH,comment-id=00000000,timestamp=2019-10-14T20:54:24.875+0000) > > No idea if there's a downvote bot but in my experience there's a good # of users who will actively downvote folks the moment they see a name they dislike. Hell some times folks will chase people around threads using the profile features. > > I certainly piss off a # of folks by constantly point out they are being unrealistic in expectations for MM systems. You tell any one they are wrong about something recently? Frequently! That must be it, haha. :D
Well I made a comment in the thread as well. Its a race who sinks more. Feel free to call me out if I give bad information btw zero, I will do the same. I really don't get why folks take it so personal so fast.
: > [{quoted}](name=Kai Guy,realm=NA,application-id=3ErqAdtq,discussion-id=gwaoPkzH,comment-id=0006,timestamp=2019-10-14T21:04:29.074+0000) > > Never fight {{champion:420}} if you're melee 4Head That sounds balanced too me. {{sticker:slayer-jinx-unamused}} Also, i will now reefer to {{champion:420}} as Mary Jane
Standing there in her R is just stupid. Kite. If you don't know how to do anything besides right click and rush at a target don't play Melee champions. Not a difficult mechanic. Plenty of champions you need to play around specefic ability's high points and this is usualy done by moving away form the threat then behaving around their CD. {{champion:23}} {{champion:122}} {{champion:114}} {{champion:86}} {{champion:24}} {{champion:75}} {{champion:223}} {{champion:150}} When you blow all your mobility to get into point blank range and lack kill potential your giving away kills.
: Its 100% luck. The top 4 will always have the meta comp. either you get it by being lucky or you dont
https://imgur.com/a/Jh6WdnF I let random players call out a comp. Then built it and made top 4 back to back 3 games in ranked. The gap between knowing how to play and not understanding the mechanics in the game is pretty big. Your 100% the kind of person I love to play poker vs. Pirates. 4th. Glacial. 3rd. Nobles. 1st. Relied on simple mechanics of managing my HP for items and Gold for champions. Its easier to play your hand then try to force a comp but most folks still don't get the game and rely on "luck" wasting reroll or not manipulating wins or losses to finish core items. I Literally let people pick the comp.
: Its harder to display a legitimate outplay these days
Players dropped accountability. Its not Skill its lack of skill. No ones BETTER then you its cuz something just Broken or some one else is awful or the Devs are stupid. People are always looking for any "Elo Hell" Excuses to justify any gaps between their opinion of personal skill and a gap in results. I had a recent loss on an Alt primarily to 2 players on the enemy team. A {{champion:89}} {{champion:420}} . Teammates would attempt to 100 to 0 Rush their {{champion:420}} in areas she had established her passive. Get hit by the E. Then she would R. They would then stand and fight to the death in the middle of that R and bitch about how broken she was. As the Leo chain CC to keep folks in place for massive aoe tentacle …. stuff and allow {{champion:420}} to sustain by breaking the burst damage patterns of our team. Rather then listen to Adc and My request of don't rushlike you have all game and please Run from her R. Team just did the same thing over and over and got the same result every time. No accountability for it. Just rage and blame. Blamed teammates. Blamed Top. Blamed Devs. Blamed everything....besides themselves. At any point the decision to not try to rush into and kill a massively ahead champion could have been made. The smart play would be to kite and peel for adc. It never happened. The issue with "Damage" Meta is sheep players are idiots if they don't understand why something is strong. Players drive gameplay in PVP games. If you play constantly feast or famine and your opponents are also constantly playing feast or famine... Your going to see gameplay be feast or famine.
: Agreed. Right _now_, in the current meta... - Illaoi is legitimately tricky to play, especially if they have stuns. - Darius getting two kills isn't that difficult. - The permaslow _is_ pretty stupid. - Irelia is still a legitimate case of outplaying. - Peeling three assassins off as Thresh is no small feat, for sure. So while some of these, I'd view as legitimately excellent play... Others I'd immediately assume it was the benefit of the damage meta. Fortunately for me, I don't get kills anyway. My skill always comes in how I peel, how I tank, how I initiate, how I herd. None of these _ever_ go noticed, but they never did. My viewers notice, and that suits me fine. But when it comes to playing a legitimate fighter? Almost anything can outplay if it builds a shred of damage, and it's making mechanics less impressive. Edit: Now I _know_ there's a downvote bot stuck to me. Literally zero reason to downvote this. It's an opinion, it's not controversial, and it agrees with a generally upvoted OP.
No idea if there's a downvote bot but in my experience there's a good # of users who will actively downvote folks the moment they see a name they dislike. Hell some times folks will chase people around threads using the profile features. I certainly piss off a # of folks by constantly point out they are being unrealistic in expectations for MM systems. You tell any one they are wrong about something recently?
: Below 20% winrate
Out of how many games?
: Understanding mmr and Matchmaking (ranked vs draft)
Short answer. All ques use individual mmr exclusive to the que. Extra information. MMR values only have meaning with the context of a que due to how they are built around population averages. 1200 MMR does not set any expectation for an exact measurement like say 35 pounds. 1200 May be different levels of skill in 3 years but 35 pounds is 35. The core concept is that there are naturally occurring averages in a population. Even For the concept of skill, so you judge off the mathematics for averages. Averages get expressed as a curve. MMR style systems job is not just to "build fair games" but also have to measure out what the current average level of skill is and any single account in relation to that so it can build said fair games. As there's constant fluctuation it's a never ending job and there is an expectation for mismatched low quality games. Protections in place protect ratings not egos or "fun" for the game.
: Matchmaking
I really do not understand posts like this. " every game with a jungler that afk farms and then shit-talks cause they say everyone else sucks." "But there is literally people that want to FF after 5 mins in the game, and then just ends up trolling till 15 mins." Why do you think MM is responsible and not Weak Behavior systems? How can MM control anything after a lobby is made? Riot needs to start being more proactive in removeing and punishing the Negative game destroying behavior. Automated systems don't function great once their perimeters are known. Folks just don't rage quit or afk to long anymore to get flagged by leave buster as they run to side lanes and CS jungle. MMR MM wont Fix this. Elo MM wont fix this. They are systems built from math principles of averages and large numbers. They are used to build a skill curve. MM then pulls from a range on that curve to lower the # of mismatched games. Gameplay itself is never a factor. That's why Elo is used for both Dating sites and games. Why it works for chess or scrabble. Its objective and functions very well with Binominal or Trinomial Results. Target the correct issue. Players who ruin games. I am sick of them in my games as well. folks who literally give up after 1 death are awful, on my team or the enemy. Its just not fun to play snowballed to hell games because of toxic soft inting behaviors. League is not a battle royal people need to play to win. MM systems Will not ever be the best solution to those behaviors. At best they sink folks who do it far to often down the ladder. Folks who only throw games that they look to lose and are playing to win when they are ahead can hover at any point they maintain roughly 50% WR.
: For Season 10 - Riot needs to be honest with themselves about the state of matchmaking.
Imo alot of the issues are player related as personally most the mismatches I play are cuz idiots give up and soft afk games, so I want better behavior systems to fix it. Hence why you constantly see me pop up in MM threads. I attribute gameplay issues to gameplay as MMR systems do not use gameplay data besides the end result of a match. For leauge thats Win or Loss. You dislike me for this stance. Your not wrong that better MM makes for a better game. Mm side. The main Improvements I'd like are: 1,Tighter ristrictions on what duo range allowed. It should be MMR not titles. 2,Better accounting for uncertainty per team. Protection in place to prevent established players from back to back high uncertainty games. I hope they do a dev post on positional mm. I want more information on positional MM. Is there a cap for the range? Alot of folks think it got gutted but the info I saw was positional mm went live preseason for s9 and positional ques went live only for Korea and Na. Positional ques got gutted but sapmagic only has nice things to say for positional mm. Have seen no comments that its been removed.
: > [{quoted}](name=Kneelbeforezoth,realm=NA,application-id=3ErqAdtq,discussion-id=nYW68tEu,comment-id=00000000,timestamp=2019-10-13T02:05:54.115+0000) > > that isnt what he was saying....The question basically was when i new champ comes out like senna is do they have to altar coding for the other champs to have her interact correctly....for example Sivirs E, If seena has a skillshot and you E it via sivirs ability have they altared sivirs coding for the senna skillshot to be blocked.. Oh. No, that's where game engines come into play. Engines contain underlying code that allows them to more easily design the game. In such a case, all they have to do is classify Senna's skillshot as an ability and Sivir's code will recognize that it can be blocked.
DalekZec (EUNE)
: If turrets can be damaged by abilities...
Play Hero's of the Storm and find out first hand.
Incognitó (EUNE)
: 58% WR in 140 games = always 18lp per win
TLDR. Your LP gains are in relation to a Gap between MMR and Title. MMR below Title. Slanted Gains on loss. Above. Slanted Gains on Wins. In the right range, more or less equal Gains and losses. Far as I can tell the LP system is entirely reactive to MMR. Wall of text. In Elo systems There is a K variable in the Rating adjustment formula. Which is Rn=Ro+K(W-We). With the Context of an Elo MM system this means that Ro is used to build games and RN is your points after a match. W is the result of a game, win loss draw and is what controls the if you have positive or negative gains. We is system expectations and its used to impact uneven games, beating some one lower then you rewards less points, beating some one higher rewards you with more. K is uncertainty, its the idea that low data = high inaccuracy on a correct skill rating. A player with 2 games vs 200 should be treated differently. K is lowered in 2 ways usually. # of games and Position on ladder curve. K multiplies the result of the (w-we) which makes early games have a more rapid change in your rating. This moves player away from their starting Elo who are outclassed or Outclass the average player. As it slows, its reduces rubber banding. MMR systems Are a Tweak on Elo systems, designed to track an individual on a team rather then treat the team as a singular value. That way each player has a individualized Stake in a match and consistent trends in play result in a skill rating specific to a single member for RNG "pick up" style games. Riot has added extra layers to their MMR system, Mostly for Player perception. They do a lot of stuff to keep people queuing up and playing rank, not all of which adds any mathematical quality or positives to a standard MMR system, thou some features do help prevent abuse despite some being abuseable. Elo is well documented, very easy to study. MMR is a bit harder, but understanding the needs of the system give you a good idea what changes are required and again as its mathematical systems your able to run Simulations and test for accuracy. Microsoft Also publishes a good deal of information on their MMR system called TrueSkill so you can look at the concept, math, and logic behind a MMR system. Its also a pretty widely used MMR system. Riots very proprietary with their systems but studying the foundations gives you a pretty good idea on some of their systems A Title is a concept that goes all the way back to Elo. You have a Player in a Range of MMR that earns them the Title. Say 2200-2600 = Masters and everything 2601 and above = GM. The idea of a range of Rankings = a Title is not new. What Riots MMR to Titles are is hidden, They also have extra steps as titles are built of LP not MMR. Resulting in many moments where some ones Rank does not accurately reflect their Title. Like I mentioned before Riot has extra features for player perception. Division 4 is stupidly difficult to demote from. The demotion protection on Division4 is higher then any other rank. To keep people from rage quitting if they drop a Tier. The LP system consistently looks the be playing catchup for your Title to the MMR range. Something I am actively [Testing ](https://na.op.gg/summoner/userName=NotNotkaiguy) I took advantage of the fact Remakes and Lobby dodges Drop LP. I have Flopped every promo via dodges. Yet I still promoted the account from B1 to S1 despite winning 0 promotional games.
: My smurfs are higher ranked than my main
: Give us more emote slots!!
9 would be good but 18 is better.
: I thought ranked required lvl 30
https://matchhistory.na.leagueoflegends.com/en/#match-history/NA1/2250767457069856 You talking about TFT?
: the LoL community seems to be unable to identify the problem and solve it
Imo. Make the competitive scean pre made 5 man teams. "Pick up" style mm ques will be less competitive.
Kai Guy (NA)
: Distributions off the probability model. It's not even. Its not zero sum. If irs normal or logistical it's a bell curve. The center is designed and built around the starting mmr value. Changing the title or value does not impact how it aims to distribute and it will correct over time. Core of Elo theory is built from natural averages. If you have 100 people in a room there is an average age, weight, hight. Similar concept. The room is the que and it looks to mesure in comparison to the average from win rate and context of where on the curve you stand.
Eh? Silver is only about 30% of the population. Ah. Its my fault. I fked up. "it represents roughly the 50% marker" I should have just said Percentile. The middle of the bell curve Is going to be the Near the starting MMR. There are more titles on the higher side of the curve then lower, That does not impact the "Skill" curve however. Titles are arbitrary Ranges Riot sets. The distribution will want to build the Bell from the starting MMR as that's going to end up representing the average level of skill. If you turned Bronze into the new starting MMR the entire ladder will redistribute to make Bronze the Middle of the curve over time. Look at Various site that Graph the ladder or share ladder position. Silver tends to be visibly the 50 Percentile. I don't think of Silver as "low" I think of it as " Standard level of skill" What's Normal is by definition not sub average. Do not Think of a Bell curve as built from Titles. Cuz that wont match any distribution very well. Titles are distractions. If you treat them as equal then your looking at I,B,S, G,**P**,D M,GM,C For Tiers only Or go by divisions as well to get i4,i3,i2,i1,b4,b3,b2,b1,s4,s3,s2,s1,g4,**g3**,g2,g1,p4,p3,p2,p1,d4,d3,d2,d1,m,gm,c that creates a unrealistic idea of what the ladders distribution looks like. Think of The curve as the more abstract Player Skill. Then law of averages, A majority value exists in any group of humans. This value we are building the curve with is skill at league of legends. By putting them at the same starting point, Above and below average players stand out fairly clearly. Once the curve is established by blending new accounts into the curve at the majority level of skill you get information on if some one is above or below majority level very quickly. Because there is a finite # of players and set starting rating, the probability model used for rating helps smooth it to resemble a curve. Accounts cant earn a rating with out achieving the require results. (You can how ever earn an inaccurate title in riots system.) Man I hope I am explaining this clearly. I had to see it graphed out to start to get the concept personally.
Saezio (EUNE)
: Maybe you don't belong in your current tier and you should be lower. Have you ever considered that? How many games have you played and what is your winratio? If you have played over 100-150 games and you are bellow 50% it means your MMR is terrible and that is why you can't climb. Because in reality you should be lower than what you are rn, so the game will slowly pull you to your actual elo given enough games.
https://na.op.gg/summoner/userName=5050BS 482W 519L Win Ratio 48%. I have tried to point out the 1 reliable Variable in every game is yourself but OP strongly feels its the other 9,009 Or so accounts in his games and/or Riot's Fault.
: I think your primary deficiency is in thinking you have more to offer in this topic than you actually do, to the point of ignoring the actual facts as they come in and are presented. Youve literally dis missed hundreds of new matchmaking complaints incident to just this season "because mmr", which is beyond nonsense. Over and out.
If **you **did any reading on the topic By all means then. > Gameplay quality is the only accountability mechanic that matters Please provide how an Elo system would be influenced by "Gameplay quality." Its not a metric inside an Elo or MMR system in any of the variables I know about. MM impacts gameplay but you can not use Gameplay to cite MM issues. Players have constant unrealistic expectations for MM, this is not just league its a common frustration in team games. Understanding how and why things work is my approach. I study. Folks who Rant about Feeders afk Any issue that happens after a lobby is made. I don't understand their logic in attributing that to MM not the People behind keyboards. Once a lobby is built MM stops, that really should be self evident. Edit. Heads up Mods Deleted one of my earlier posts for tone. I just want you to understand MMR systems. Getting our conversation deleted or us banned would be counter productive imo. Look, I don't think the Perceived drop in quality is to surprising. I expect that given Positional MM. I also do not think S9 has the best gameplay and that there is room to improve. How many times have you seen me call out Duo implementation as a negative value? My opinion on quality loss is direct related to player behavior systems because soft inting/# of folks who just give up instantly is far to often, but that's simply an opinion I have built from observation. However. If some one says 4+4=44 they are wrong. I don't understand how you would call this a "critical reasoning flaw" The Stats pulled from game are W and L for this type of system, so when folks site stats or events that are not W or L they are providing bad information. I don't mind folks who show they actually looked into the system. You don't seem to be one. **By all means if this is not correct then simply prove me wrong**. Folks who have the ability to discuss the terms involved, the concepts. More then welcome to be of a different opinion. When its people who call me pretentious because I am literally just using the Terms inherent to a conversation about Probability and Statistic theory do not get that benefit.
: I think your primary deficiency is in thinking you have more to offer in this topic than you actually do, to the point of ignoring the actual facts as they come in and are presented. Youve literally dis missed hundreds of new matchmaking complaints incident to just this season "because mmr", which is beyond nonsense. Over and out.
And I think your primary Deficiency is you have not done a single search on the topic. Think for a moment. Why can Elo ratings be used for Scrabble, Baseball, Chess, Tinder, Facebook, Pandora, Poker, and MTG? It should be pretty obvious that gameplay is a non factor. Its a mathematic concept. > Elo's central assumption was that the chess performance of each player in each game is a normally distributed random variable. Although a player might perform significantly better or worse from one game to the next, Elo assumed that the mean value of the performances of any given player changes only slowly over time. Elo thought of a player's true skill as the mean of that player's performance random variable. A further assumption is necessary because chess performance in the above sense is still not measurable. One cannot look at a sequence of moves and say, "That performance is 2039." Performance can only be inferred from** wins, draws and losses**. Therefore, if a player wins a game, they are assumed to have performed at a higher level than their opponent for that game. Conversely, if the player loses, they are assumed to have performed at a lower level. If the game is a draw, the two players are assumed to have performed at nearly the same level
: MMR
1. Its not, you just have not escaped the Silver class interval yet. Keep playing. 2. Because 16 games is nothing. 3. No, ques are separately. 4. Because 16 games is nothing, even with a high Uncertainty variable. The Starting MMR is silver. Your looking at -8 games right now when you balance your W to L. Seriously, there are players with 48% WR that have a bigger MMR debit then you do. 5. MMR is a Balance. WR is not as relevant so much as the gap in the # of games. If you don't fix WR and maintain 25%. My guess is that around 30 games you should only be seeing the occasional S4-S3 player. Last time Riot gave a value they listed it as 150-300 Games before their system hit low uncertainty, if my memory is correct. Think of it as a balance, WR is not going to build great expectations. Starting value then subtract a bit for each loss over Win. Add a bit for each Win over loss. Its more complex then this but its a better way to view MMR then Flat WR.
Show more

Kai Guy

Level 174 (NA)
Lifetime Upvotes
Create a Discussion