They are still using your oldest data against you!

General Discussion
21

I’m sorry. I’m not trying to be a jerk. Maybe the problem is that I’m trying to shorten my posts constantly.

When I say “win/loss” I’m talking about the dominant strategy. I’m simultaneously referring to the weighting system that was created from previous win/loss.

So, yes I understand that you had an upside down win/loss rate and went up. There are two scales/sizes of consideration here. The big picture over time win/loss and the quicker win/loss.

When they have a large pool of data from you like all your games from quick play and competitive over years, they will favor that history over a recent dip in your ranking. It will pull you up or down faster based on that large set of data and if you are very far from the higher place they believe you should be then even losing a lot will still rank you up.

There are things like changes in skill and player improvement and new players that they have to try to put into the design to decide the interaction of different scales of datasets. My issue is with something called Bayesian priors. AKA “Where they think you should be” and how they create that is what is broken.

My problem is that the weight of their Bayesian priors and the way they created that data was all wrong and perpetually screwing me.

My problem is that I just wanted my skins and instead they screwed me with old wrong data. (big picture. Where you should be. Bayesian priors.) I know they did this with other people too.

I could just stop insisting on having my skins and go play my higher ranked accounts, but I see a clear problem and know how to fix it. So I’m hoping to communicate that.

22

Myabe they do like the game but don’t want to leave cause there’s still more to offer than any other game and the current system in place doesn’t work and want to be vocal about it in hope that it DOES get fixed and they can get back to enjoying it again.

1 Like
23

Sigh. You’re probably right, but even QP is stomp-counterstomp for my low ranked account. I have better ranked accounts I could use that aren’t weighted all wrong but they don’t have my skins and I really want to be able to play with my wife in a less nonsensical environment.

It’s all smurfs mowing down everyone while their teammates are running into walls and falling off cliffs with no-one nearby. Sometimes it’s hilarious but it also makes for a frustrating game that could be sooooo good.

I see others having similar problems and can’t help but want to fix it for everyone instead of just switching over to one of my accounts where it’s not a problem.

I could fix it and it frustrates me to not be allowed to do so.

24

oh boy do i have news for you.
players of every elo are right now complaining about the game being stomp mania in either direction. all things considered thats just the state of the game right now.

ive always been a strong advocate for “players in every elo hate their teammates”. its true: ppl might think the game is inherently more fun in higher elos when all that really changes is every players individual skill. there is no more real team play anywhere, theres still snowball everywhere. theres still ppl dying in a stupid way, ppl flaming each other, people leaving games etc. essentially, its the same game but ppl tend to hit more shots and they tend to do smart things more often.

25

i guess what im aiming to say here is: the game is as fun as you allow it to be. rn i am basically hardstuck in all my roles (as most ppl are for the majority of time). and while i still sometimes am frustrated by my games or my teammates, the game is fun for me. my games are not particularely brilliant, theres ups and downs. but i play the game for the games sake and whatever happens in the game i just try to roll with it. we have no other choice, bc we cant control our teammates all we can do is play the best we can and hope for the best.

26

That’s simply not true. Why are you just making things up still? lol. You say this all like it’s so “matter of fact,” yet you have no actual data or proof or rhyme or reason to actually believe it.

In fact, if you haven’t played on your account for a long time, it basically “soft resets” your account. It will pull you up or down quicker based on your “small set of data.” They don’t care at all about your history. This is what the developers, that have CREATED THE SYSTEM, have said. It is how it works. It’s not some rambling of a madman on the forums, saying, “this is how it is.”

It’s not doing this, my friend.

27

Hmm, yeah that tracks. I’ve only been playing my oldest low rank account this game and have heard the instability crept up into other ranks but I thought it was hyperbole that it was going all the way up the chain.

That’s just a symptom of the problem I’m talking about. They should have never ported over the data from OW1. They thought the data would give their rank system stability because they thought they were doing it right there. They weren’t.

They are committed to a philosophy (of teamwork) instead of trying to really commit to a workable theory of ranking.

Ooof. Yeah, the influx of new players, many more smurf accounts and other factors would drag noobs to the top and trap some good players closer to the bottom.

It’s exactly what’s described in this article I can’t directly link: just google “Elo vs. Regression to the Mean: A Theoretical Comparison”

They basically are no better than random at this point because there is a critical instability threshold in their methodology. (IE you can have some small success below a certain amount of chaos in your data, but almost none above a certain level)

Even my suggestion of only using personal performance metrics only at low levels won’t fix it now. They will have to switch to personal stats (a very poor ranking metric too) for a while to make anything work, now.

I think they had that at the beginning of OW1 and that gave them some data stability for the system to continue to function under the chaotic threshold even after they made it much stronger win/loss later on.

28

Overbuff percentiles compare you with other players within the same SR brackets.

In these areas you are in the top 11% of Bronze players, not the Top 11% of all players.

29

dont get me wrong. i do genuinely believe my teammates and enemies play just as good or bad as i am. but the game itself right now is just very snowbally. thats not a “new players problem” (as i basically dont ever come into contact with true new players, because all my ranks and my qp mmr is way above average).

the only way they could fix how bad the game feels right now in terms of onesidedness is to work out a good hero balance, lower the amount of oneshots or almost oneshots and look at the push mode which is a large portion of why ow2 feels so snowbally compared to 1.

30

it actually kinda compares you to all players but what it doesnt take into account is that players obviously achieved those stats against players of their skill level. i think that is what you mean though you just worded it a little weirdly.

31

Did I not link this video in this thread? Sorry if I didn’t: youtu.be/jyWPuifJg1k

Here I prove that their method for placement is unambigiously based on prior data by playing qualifiers on three different accounts with three different sets of data and getting 1000 point difference in placement accross those accounts.

The girl I played every qualifier round with placed 1000 points higher than the account I was playing as well.

That’s just not true at all. I wish it was. I’ve posted the evidence above so I guess it’s your turn for some evidence.

I’d love a link to that!! Sounds like I have evidence the writer was either lying or was some high level communicator who didn’t understand what the engineers told them. (as a developer I know this happens constantly)

I used to be in a company where the sales people were constantly saying things about the system that weren’t even close to the truth and I was the lead developer of the primary team. The primary system was my baby so I just kept trying to create what the salesmen said already existed. It was simultaneously sad and hilarious.

Please go watch the evidence that it is.

32

Hi, Masters Junk main here. You’re not being held down in bronze due to the system and frankly your stats aren’t impressive for bronze games. He’s a free pass to plat for anyone vaguely competent at the game.

Actually yeah it does. Losing that much in bronze is a sign that you belong at best in silver. It is possible that you deserve to be higher but in that case keep playing and eventually the system will figure it out.

Why am I so sure that the system isn’t using old data to old you back? I recently took a very old decently levelled account that was previously gold and climbed to masters with it. It took two lots of 7 wins to convince the system that I was underranked but after that it was easy. Keep winning, keep stomping and the system will move you up relatively quickly.

1 Like
33

Uhm. Of course it is. No one is saying it isn’t.

You had to test that? lol. That should be obvious…

Go read the developer notes that I linked. I would find quotes from developers in the past, but I am too lazy. Unless they are lying (lol) it absolutely is how it works.

I think it was posted by the lead game designer? Jeff kaplan? Or geoff something? I forget. At any rate, it was legitimate.

Again, it seems like all your paranoia comes from you simply not understanding how things work.

34

i think my work here is done, other ppl are now fighting the good but pointless fight of arguing with ppl who basically have an unchanging opinion anyways of why they will never climb and its the system thats at fault and not that they are actually not that good at the videogame.

35

You’re completely full of BS Mr keeps-his-info-private. We do know that 44 months ago your stats were: (imgur[dot]com/a/MeXv4nv)

Damage: (Bottom 18%) …LOL WOW!
Solo Kills: (Top 24%)
Final Blows: (Top 45%)
Elims: (Bottom 46%)
Objective Kills: (Bottom 29%) ROTFL!!!

Top 3-6% Does seem impressive and you know it.

…but I already said that “it’s artificially inflated”

And maybe you are one of the carried bad players that has skilled friends who WOULD be all about insisting the system works.

You’ve not understood one word of this thread. Let me give you an account to play and you record EVERY game for us and PROVE your point.

Keep pretending you can 1 v 5 every game consistently right up till you decide NOT to post the video and your lesson learned.

Then again I don’t really believe anything you’ve said now that I’ve looked at your account.

Let’s pretend you’re not making this up.

  • Was that in OW1 or 2?
    Gold and above is the point at which the system starts to level off. I’ve said repeatedly that the OW1 system works at plat and above. You have a skewed experience if you have always been plat or higher

  • How many games?
    Pay attention to the words being said and the points being made. The problem is requiring too much data.

  • Was it random queue or playing with friends?
    Win loss problems and bronze is ONLY RELEVANT TO RANDOM QUEUE

Actually, yes if you queue with a bunch of high rank players on a low rank account and play a bunch of games, any account will slowly rise.

Try the same thing with a low account QUEUED WITH RANDOM and there’s no hope. You’re not going to stomp. Only a brain damaged person who can’t play this game at all thinks they can 1 v 5 all the time. (and there are some of those in bronze causing us all to lose)

Stop pretending you can carry all by yourself. It’s inconsistent with what you should know about how dumb it is to go in without your team.

Again. We’re talking about the MMR that creates your team, not just what selects opponents.

36

Watch the ego
He’s masters on every account he plays on and you’re over here stuck in bronze? Really?

2 Likes
37

Blizzard doesnt keep track of mmr across platforms. Its just that new accounts default to around that mmr range. It might be challanging but as long as you can keep up a good performance youll get out eventually

with that said there is prob a major mistake ur making that ur not aware off thats keeping you down in bronze. (Could be anything cd management to target priority). Use the chaotic nature of lower ranked games to ur advantage. Id be happy to review a game if you want me to(Masters on all 3 roles).

1 Like
38

…that doesn’t really prove anything as you’re comparing 3 sets of placements with a wealth of variables such as maps/heroes/throwers/try-hards/etc…? :thinking:

39

Fairly irrelevant for anybody. More so on an alt that is used only for the forums.

Edit: I opened the profile on my forum account and played a couple days of comp, just for you. 30w with 10L. Started gold and finished masters with 75% winrate. Rigged, clearly.

No, not really. Overbuff stats are vs the players at your own rank and also dont really tell you the whole story. The only useful stat is winrate after a lot of games. Which is what my focus was on.

I never said that. Being better than your rank will cause you to win more than you lose which will eventually make you climb. You wont 1v5 every game and wont win all of your matches. The more you win than lose, the faster it will be.

I don’t really care if you do or not.

OW2. The account in question was relatively high leveled but hadnt been used for years.

It took around 65 games with 49 wins.

All solo queue.

I just did that.

Look, instead of lashing out at people ask for replay reviews. You’re doing something fundamentally wrong if you’re stuck in bronze.

Shhh Omni you’re clearly lying too. You can’t have possibly played with me or anything like that…

40

And how was that account weighted? …oh you don’t know what I mean? I know…

Accounts gain momentum like a train. The longer they remain in a place the harder it is to move from that place. It’s how any ranking system attempting stability will work.

Yes, I know specifically what it is. It’s not playing enough games per season. I already said that. Pay attention.

They can only hold game data specifics for so long before they aggregate and purge to save space. When they do the aggregation they lock in a weight for an account.

If for multiple seasons you don’t play enough games to rank an account up and hold it there, it will gain momentum that makes it exponentially harder to get it out.

More games required next season because too few put in previous season. It’s a negative feedback loop. (especially since I play that account even less when there is less progress per game)

You’re just not getting the point.

I have accounts I can easily rank to gold even with the minimal commitment of total games I am willing to put in. And I have bronze accounts that are a nightmare to play. Those accounts have my skins and are holding the skins hostage.

That’s the point. Focus. Please.

Just focus on the point for a second or two. I believe in you. You can do it!

Address my point, not he point you think I must have because others make the point you are thinking of. I have my own point. It’s different from what you know and perhaps even different from what you are capable of understanding. Just stop spitting out the same reply to everything for a second and think.

Weighting. Time input required. NOT Skill required.

You guys really aren’t this dense. I know your minds are better than this!

I CAN and I HAVE REPEATEDLY ranked deep bronze accounts into gold with fewer games than you listed above. And then I stop playing because it exhausts me.

The point: The game is testing endurance more than skill. (if you’re not willing to input gargantuan amounts of time anyway)

I want one of these done:

  1. Fix ranking to work faster.
  2. Allow account rank reset.
  3. Allow unlocks transfer to a different account.

Please. FOCUS ON THE ACTUAL POINT OF THE CONVERSATION.

The only way I can “Git Gud Scrub” in this situation is to alter my real life dramatically since what is required of me is more time input. (not skill since I actually enjoy gold account play just fine which is my ONLY personal goal: Fun)

That’s a secondary point… oh no, I hope I didn’t muddy the waters!