I’ve been able to do much worse or much better based on the account I’m playing because of old history being calculated in into rank and anyone with old accounts knows what I’m talking about. That’s why so many of us older players bought new accounts.
So does anyone know if this will be a fresh start like a new account?
That would be nice. I could play my old accounts with more skins again! (woohoo?)
The answer lies with the question: which do you think would make more money?
No, stats and MMR will stay. There’s a bit of a “soft MMR reset” happening. Basically, MMR is actually two different values: predicted skill and confidence. It seems that the predicted skill is the same, but the confidence is dropped significantly. This means that match win/loss SR gains are far more extreme, in order to move you to your proper rank as quickly as possible.
Old stats are not affecting your actual rank…
I am really interested which excuse you people will use after they reset mmr.
Great, thanks for the info, DaSquyd, almost as good as a reset, where’d you find out?
Old stats are not affecting your actual rank…
First, that’s just ridiculous. It’s a necessity of ranking they’ve talked about a million times, but giving you the benefit of the doubt, you would be somewhat accurate if the player you’re talking about is playing hundreds of games each season.
Tons of data is the only way a ranking system attempting to judge the impact of one out of six players is having on the outcome of a game can come anywhere close to being correct. It’s one of the reasons they’ve refused to allow people to reset MMR. It’s not -only- greed.
The truth of the matter is that regular players of the game resetting from scratch would wildly throw off the stability and therefore accuracy of the system. It’s true that the more regular players will be generally where they belong and that data is very important for keeping the system from fluctuating wildly.
Even this soft reset is going to cause a little instability. It will be good for more casual players or players with alternate accounts, but hard core players might be affected adversely right at first.
Because the matchmaker hasn’t throw the rankings in chaos years ago right…?
No, old stats are not affecting your rank. Your MMR is dynamic and the system expects people to get better over time which makes stats from when you started the game rather irrelevant if you since then have climbed and consistently maintained a higher MMR.
I will flag every single post of yours for being sarcastic and toxic. Byew
Do not fall for his sneer, he just wants you to feel bad.
Blizzard is deliberately putting losers and winners into one team in soloq. I saw this yesterday as I played against the same people over and over again and they lost every game.
It’s actually not a excuse, it’s happening all the time right now. The game seems to set new accounts higher when they perform as good as the old ones, which is pretty good tested and many you tubers showed that
I wasnt talking about new accounts. New accounts simply start at 2350 first placement game and then every win or loss will go ± 150 SR until you are placed. Because initial placement is decided mostly by wins or losses, its actualy placing people way too high then they should be. And they are dropping back to their elos after they play enough games. Atleast like 80% people. Some are able to keep new elo because they learn fast new stuff from higher elo. But these people are usualy able to raise up their olds account after that too.
MMR is pretty dynamic and its changing as you play games. Based on your skill its moving up or down. There is no such thing as stats from old seasons keeping you lower. Only actual MMR is important. Why should your old history affect you when you get better? It doesnt make any sense.
Jeff was on Seagull’s stream and gave a lot of detailed info on how the change affects MMR.
the stats from 1-2 years ago will be removed but your new sr will be calculated about how u play+last few seasons
Bruh giving each account 3 SRs kills smurfing (a little). If i actually felt like they were making these decisions based on money i wouldn’t even play the game.
We dont pay or play to unlock new heroes, they dont sell “map packs”, dident get charged 5$ to use the editing tool. Loot box changes to not give dupes. The things that you basically get taxed for in other games.
The incentive to buy a new account is purely based on speculation… ive never broughten a second account and i wont because there is no reason too.
Because initial placement is decided mostly by wins or losses, its actualy placing people way too high then they should be.
I really don’t understand how you aren’t following me. It doesn’t see like we even disagree. If it places some people too high it also places some people too low. Regardless, the wins and losses calculation YOU are talking about is in opposition to them using old data. That’s entirely the point.
And they are dropping back to their elos after they play enough games. Atleast like 80% people.
Yes, I agree that the system works for a majority of players and is totally broken for the exceptions. When you say 80% you feel like you’re done with the discussion and forget that the 20% are real people with a real problem. (but, yes, a good 80% of the people complaining are just whining like you believe)
Stop forgetting tat 20% is still a lot of real people and just because there’s a useful rule of thumb doesn’t mean you should use it for every circumstance.
There is no such thing as stats from old seasons keeping you lower. Only actual MMR is important. Why should your old history affect you when you get better?
|||>>> Warning incoming storytime attack! <<<|||
Okay now you’ve asked the right question and I suppose it’s hard for you to imagine a life different from your own so let me lay it out for you. Yes, old MMR does hold you back if you don’t play many games.
And if you have other special circumstances you are REALLY boned. For instance, I play with my fiancee, nearly every single game in ranked and most of them in QP. She’s a solid mid silver player who could make it to gold if she tried. Unfortunately she smokes WAY too much weed and throws games when she’s too high. All The Time.
Let me repeat, we almost ALWAYS play ranked together. If we were to never play a game separately, it’s literally impossible for rank system based on win/loss alone to tell us apart. However, I’ve been roughly top-of-middle to bottom-of-high ranked in every shooter I’ve played for 20 years depending on my frequency of play. I’m consistently the dominant gamer in every group of friends I’ve ever had. (local hero, not a global one)
I’ve repeatedly pulled her account and mine out of bronze into mid silver by playing solo for about 20ish games and even decided to put my account into gold for once and played a few more on my account.
It’s not “hard” to increase the rank of the accounts, it’s tedious and time consuming as an adult doing this for entertainment. If I put another solo 8-16 hours into my account maybe the system would be able to get it right but I have to overcome all the “bad play” that the system has logged far more consistently. They have to account for cheaters who get their accounts boosted so they make it tedious and slow to overcome old data when there’s a sudden improvement.
It’s discouraging, slow and annoying to have to dig an account out of silver repeatedly knowing what I’ll have to deal with. You know what those players are like! So I just don’t do it. I only have so much entertainment time and grinding is NEVER part of my allocated game time. (and there’s no reason to go beyond gold since that’s about the top of my fiancee’s play ability when we go back to playing together again)
I’ve played with my son and his friends who easily got into platinum and diamond and none of them have any doubt I’m on their level even though they have the twitchy teenager advantage. One of my son’s friends mains Junkrat like I do so we had a little competition when he got a new account and trading Junk back and forth we were neck and neck on personal stats but, of course, he was gaining DOUBLE the amount of SR that I was PER GAME.
Now. Do you understand that repeatedly playing with someone who isn’t that great of a player and that also isn’t trying hard enough, in season after season where we only play about 30ish games (not enough to overcome the throws) would make certain the system has old crap data that is consistently bad?
Now that you understand how an exception works, imagine there are other weird circumstances that make people repeatedly have too few games for the system to judge an actual skill level appropriately. Like maybe they live way out in the boonies and get DCd over and over. Maybe their mother will randomly shut off the internet because she just doesn’t care. Maybe they have to share an account with their little brother.
So sure, fine, the system works pretty well for people who play a lot of games in a consistent way under consistent circumstances. Some people don’t have that luxury and so crap data follows us around and weighs down our account.
I’m not playing 400 games a season for 4 season to convince the system I’ve finally “improved,”
…and I’m also not playing every quick play with only my mains at peak performance. I play this as a game for entertainment. I love playing silly stupid trash in QP
That’s why an account reset is so valuable. The system uses old data and that data can be consistently wrong.
If that is how you play and continue to play, why does it matter if the system resets or not? What I mean is, unless you are going to take the game more seriously and not queue with your gf, your rank will end up the same regardless, so why does it matter whether or not you ever climb to gold and beyond?
I have had similar circumstances stemming from sharing my account with a 9 year for several months. The system is pretty quick to get a proper rank back, even after an extended break.
Good point Veeshkav. I was planning on having my own separate account anyway just for competitive and I think we’ve gotten to the point that it won’t hurt her feelings, but…
I’ve been recently been able to get her to stop smoking nearly as much and feel like I could climb with her a bit. I want her to have that fun and we do some streaming together as well. I know she sometimes cares more about the skins and things like that than competitive rank but she also gets disappointed about her rank too. So she’s stubborn about letting go of the skins to get a better rank.
I really want the best of both worlds for her: An opportunity to play at her full potential with me as well as her getting to keep all the hard earned skins etc.
So therefore, if the variation in rank was faster (lower confidence level like someone mentioned above) we could climb together faster and have more fun without all the grind required. I’d love to see her finally make it up to gold! If she did I bet she’d apply herself more to the competitive part.
Truthfully she’s more easily discouraged by grind than I am and just wont do any of it and she’s also more easily discouraged, so we have to have more extreme changes for her to be happy and to fulfill our time constraints. If we dipped down low quickly, that would be fine if we could also do the same in the other direction too.
We’re much more okay with rapid fluctuation than some people would be, I suppose. We love the thrill of victory AND the agony of defeat. The cold grey middle of trying to hold some ground or slowly grind and eek out tiny progress is just a totally different mindset. (yuck!)
We don’t play WoW or really any grindy game and maybe a lot of people who play shooters are like us while people who play MMOs like it the other way.
Old stats are not why you can’t climb. I’ve taken my day 1 account to within 50 SR of where I’ve taken multiple alt accounts to. If old stats even really matter, it’s a VERY low factor. The biggest factor in the opponents/teammates you get is your SR vs your MMR–if your SR is higher than your MMR, you get bad teammates and better opponents and vice versa.
You didn’t read my posts pyah.
Given a large enough consistent data set in a short enough time, the system will eventually work and even correct.
I don’t have the time to grind it out so I keep feeding it short crappy datasets. One season I couldn’t even do all my qualifiers. Other seasons I could put roughly 100 SR per hour on my account. The system uses the previous data to a very LARGE extent, with very good technical reasons as well as good rational reasons to do so.
If you were a consistent and frequent player before, then it will work nearly perfectly for you and even match your progress as a player. Your experience makes perfect sense from a specific situation and perspective.
Is my inconsistent life and play style everyone else’s problem? Should the system have to change everything to try to fit me?
No. I understand the decision to disallow even the resets per season that I crave that would be useful to the few like me. Doesn’t mean I don’t still want that option though. I’m still of the opinion that the reset ability wouldn’t be as disruptive as they believe.
But trust me, they have repeatedly denied requests for clean slate resets of account data. If you don’t understand the technical and mathematical methods (including regression analysis) required to maintain this sort of ranking system it might not make sense to you. It makes sense to me but I think they have been too careful about maintaining stability and the new relaxation of “confidence level” on stats will be great for me.
Look, guys, I’m sorry but this just isn’t up for debate. We’re talking about hard technical requirements of maintaining a stable and accurate system.
There are statistical requirements at play. They absolutely CAN NOT flush all the play data for every season and create a reasonably fair and accurate ranking system when all things are considered.
The term “ELO” is a reference to an early games ranking system (chess I believe was the start) that is only for 1v1 and cannot be used for rank of individuals within a team with ANY accuracy. It can rank teams like single individual but CAN NOT accurately rank individuals within those teams without absolutely gargantuan data sets that grow exponentially based on team size.
This is a fertile field of academic study, not just a bunch of people speculating online.
There is a data requirement and considerations of stability in doing what they are attempting to do in Overwatch and there are various alternative ranking systems that have been developed to overcome the problems of ELO being only suited to 1v1,
These are just long-standing hard mathematical facts: (dunno why I can’t post a link)
[add http:/ here]/tht.fangraphs.com/elo-vs-regression-to-the-mean-a-theoretical-comparison/
Yeah sure, system is placing new accounts too low too because initial placement is mostly about wins and not about performance.
Not sure if we understand each other. Sorry, but my english is not that good to understand how people sometimes explain something. I just want to flarify, that I was talking about 80% as people who will drop back where their main is. 20% are people who actualy open they eyes and get better. I know some people like this, they had no issue to climb on their old main accounts after.
Sorry but it doesnt. Even if you will not be playing a month and then suddendly come back and play one game and outperform others, you will be rewarded by higher amount of SR. I had a friend who done some boosting, he used lower ranked acc and climed like in a day. If older bad stats/MMR would hold you back, boosters would not be able to climb that fast.
Dude you made me laugh 
. Its funny but I dont see what you were trying to say with that. Yes some factors can affect your ability to climb. But as you said, weed is affecting her ability to play effectively? not MMR .
It really depends about which ranks are you talking about and too how skilled are you or how good are your ractinons, deduction, social and learning skills? I would say climbing will start to be tedious for an adult ±30 around platinum.
Its not the system and “bad play” logged you are supposed to outplay. Its people and your performace. MMR and SR are very close to each other, unless you are smurf. So even if you will have bad game full of losing and terrible stats, you can log in next game and demolish enemy team and gain like 35 SR in one game. BUt its about you not the system. You think its system affecting your rank, while its you.
Maybe you just reached you limits. Its nothing bad to admit it if you have fun in your rank. I will tell you that I am high plat/low diamond player and I have one gold acc. Account actualy was silver but now its mid gold. And even when i play on it only occasinaly when I dont want to play at 100% or play completely different heroes I am slowly climbing into platinum. Even when I dont need or want that. When I took torb in attack, I am doing it maninly to have fun, but I want to win like that. Its that challenge I like. Game is usualy starting to be tedious at 3000, thats why I never took it more serious above that line. Its starting to be less fun and I have to fully concentrate there. I am 35 btw. So I dont have problem if someone say to me I am hard stuck plat or anything, I dont care as I have fun in my rank and I can pop off in it. I guess I invested more time into it too because I dont have wife or kids so i gained more knowledge about game and knowledge is power. You could probably climb higher too, you just dont want to invest more time in it and i understand it.
Again you are simply talking about outside circumstainces. And they are on the side of the player in most cases. And in my experience, if you are leaving games or DCed, it will actualy not affect your MMR. I kind of experiemented with it when I left 4 games in row. Suddendly I was able to gain a lot SR for my games. I think it kind of works like Decay for accounts above 3000. System will not lower your MMR, only SR, but want to help you get back to your MMR. And if someone is sharing account with little brother, they are lowering their SR on purpose basicly, if they know he will not play good. It has nothing to do with comp system.
I would say one thing. If you want to be better at anything, not only at game, you have to invest time into it. If you will have math test and fail horribly, its not old test data what affected it. It was your actual performace. System wont affect you in a bad way if you perform well ni your games. Its not system, its your performace and it doesnt matter how often you play for the system. Jeff Kaplan was talking about it in the past how system will know better where you belong if you play more often. But I dont think he was saying anything about how system will affect you in a bad way. It was only that system will know better where you belong. Its mostly important in placement(not initial ones)
Actualy placement is only way how to use your MMR to your advantage or disadvantage. If you have played very bad in like 10 games before placement games and then do placement, you will most likely lose a lot of SR in placement. But that only oocasion where MMR can affect you a lot.