Psst, hey buddy. This thread is about completely new accounts starting your initial first comp match at a base of 2350. Something which at least 2 people here have backed up with actual data, while you not only missed the entire point but also are trying to argue with “feelings”.
Psst, hey buddy, it’s irrelevant what it’s about, when I disagree with the OP.
New account, old account, or smurf account it’s all applicable and for the exact same reasons.
You want actual data?
Then look at my profile: There’s over 8000 games worth of data, instead of a ballpark guess which is what the OP has done.
Feel free to do the number crunching, but trust me, it doesn’t take a mathematician to figure out how MMR, and MatchMaker works.
Do you also believe that SR gets hard reset every season?
That’s the only explanation one can get why you’re so adamant with your theory.
You’ve played comp games before, we all know the SR doesn’t get reset between seasons. So your next placement is fairly similar to what you had before. Which is exactly what you’re saying.
What Kuwaamba is saying is that a fresh new account is almost certainly 100% going to start at ~2350 for their VERY FIRST COMP GAME regardless of what else that account has done. Which by 2 data sets, has been fairly accurate with teammates near those SR.
I go into QP/Arcade games and I generally get teammates from Diamond -> GM. Does that make me a Masters player? Oh hell no, I’m floating around gold and Plat.
I guess the bottom line is your data is useless and irrelevant to this thread. It could be used in another thread but not this one because you’re missing the point.
As long as the formula is hidden, I will not care about it at all. It’s ridiculous to slap the name ‘Competition’ on anything with a hidden scoreboard.
Besides, low ranks are more fun, it seems.
how can you disagree with a topic you dont even know the subject of? LOL
This topic is about how completely new accounts place for the FIRST TIME in comp. Not how they are after 8000 games or ballpark guesses.
Its been confirmed that no data from quick play is used by multiple players with actual data to back this up on NEW accounts.
You 8000 games are literally meaningless since the topic is about NEW ACCOUNTS. So next time you buy a NEW account and do placements come weigh in on your experiences about this.
As others have said, this post discusses new accounts. For accounts that have already placed, the data is at Season 9 Placements Analysis, High Variance Explained
Put simply, a GM cannot push your established account more than a few hundred SR in 10 games because nothing resets between seasons. For active accounts, placements don’t affect your account any more than normal games. If he kept playing, he would have no problem getting to GM (unless he got banned for boosting).
im pretty sure it does
You can become comp-eligible playing Arcade only (Twitch streamer Slayergramps did that with his one-trick Tracer account https://www.overbuff.com/players/pc/Sojun-11663). Since Practice-vs-AI also grants XP, you can also get to Level 25 by playing that, although I’m pretty sure no one has ever tested that to see if they became Comp-eligible.
JAKE has also made the following fresh account and was taking on GMs in QP, but his very first placement with that account went as follows: https://clips.twitch.tv/SecretiveHeadstrongWaspOSfrog
1 Like
Source?
Because I’ve never heard it confirmed, and it doesn’t ring true to my experience.
I have numerous accounts… And what doesn’t make sense…
I’m not saying the OP isn’t on the right track… I just dislike when people create a theory out of thin air, or based upon their own individual sample size, which is that 1 ten-millionth of of 1%, and act like it’s confirmed facts.
I have 5 accounts, and all of the variables vary WILDY, and as a result I have placed everywhere between high diamond and mid-silver… All were fresh new accounts.
It’s the absolute certainty that is putting people off. Theory passed off as confirmed facts never go over well.
I don’t know if you read the original post, or didn’t understand it, but it is only the initial game that is at 2350. Depending on how you do during your initial placements, your placement after 10 games does vary widely.
Since I wrote the original post, I have gotten four more data points that agree with the original conclusion of the first game being around 2350. But I have also gotten data that the maximum placement is in low master, not low diamond like I originally believed. Some of the new data takes more than 18 games to settle as well. I’ll update the original post, by this weekend at the latest.
Three new data points are at Streamer Data - Google Sheets → PlatChat 11, Rainbow 10, and Dmum 11 (there has been a rash of streamers leveling up alt accounts lately). The fourth is above in this thread: M0rchPonkey.
As promised, I updated the original post.
Yeh, but nothing you said or showed, made anything you’re saying more than a well researched theory.
Until i hear it from someone is Blizzard, in black and white, Thai is one of hundreds of good and possible theories… But it’s still a theory.
Okay… You can say that if you like, and ignore the data I collect and everything I write except my references to official Blizzard statements. But your first post in this thread is claiming a theory (that your first 25 levels establishes your competitive MMR) that has never been supported by Blizzard. You’ll need to be more consistent about what kinds of information you accept.
3 Likes
Agreed. But my point is that both are theories… And both are valid and in the realm of possible, or even likely.
My theory is based off of data and experience the same as yours. But I am not saying my theory is a fact. And while I may not be wrong, it doesn’t mean I’m right.
Your theory has a couple of very large holes in it. When you’re creating a theory, you’re not supposed to only look for the things and the data that support your theory, you also have to look for the things that poke holes in your theories, and you didn’t do that.
Show me believable data or Blizzard statements that contradicts my model, and I will change it. You have not done that.
1 Like
Oh, two people you say? What a data set! Two people placed gold, so that must be true for everyone.
1 Like
True, not a large sample size.
I’m sorry bro, let me just clarify by saying I don’t know for a fact that the theory you proposed is wrong.
In fact it’s one of the more probable theories that I have heard. But if you look back, you’ll see that there have been several different theories regarding how your first placement matches work, and the ultimate impact that they have on your final rank. And many of them had data/research to support their position. The issue with the “data” that we have access to is that it’s limited, and the vagueness of how that data is quantified leaves reaching definitive answers impossible in some cases.
And while your theory may be likely, or even probable, I can’t say that it’s correct.
Part of the problem that a lot of people have when creating and presenting these theories on placements and ranking is that no one truly knows exactly how the MMR functions, and its role in your placements. I’m not even convinced Blizzard fully understands it, otherwise they would understand that it’s findamentall flawed, and they’d change it.
One major issue that I personally take with your theory, is that totally negates the first 25 levels that you have played to reach your placement matches. And if the first 25 levels have absolutely nothing to do with the teammates and enemies you get matched with in your placement matches, then that could lead to some incredibly unbalanced initial placement matches. Even if the system works to balance it out from that point forward, that would still skew thing dramatically.
For example, if I consistently put up GM/top 500 level stats for the first 25 levels, and in my first placement match they put me on a team with other people who happened to be in the diamond or above range, and we went against an enemy team that happened to be composed of platinum, gold, and Silver level players, they are going to get dominated at a rate and pace that would make it impossible for any type of system or algorithm to accurately judge their ability and efficacy, because they would be getting outplayed in every facet to such significant degree that any type of system in place to measure a persons actual ability would be skewed by the sheer un-evenness of the teams.
That might lead to them being placed lower in their following match, and the sheer inability and ineffectiveness of their teammates could prevent them from displaying their true capability, especially if you main a non-dps healer, for example… As you can’t heal people who are just asking to be killed. Synergy wins matches, and while extremely strong players can carry a couple of weaker ones to victory, you can’t consistently synergize with people who aren’t as skilled as, or immensely more skilled than you.
And likewise, if a gold-level player plays his first match with a couple of GM and masters-level players, they might inadvertently boost their perceived rank. And even if you say the system will “work it out” in the following matches, that’s not assuming that gold-level player doesn’t group up with those masters and diamond-level players. Which would further skew their stats and perceived ability.
Your theory also doesn’t account for things like leavers, throwers, and smurfs… And the odds are, that 9 out of 10 of your fist placements matches will have at least 1 (if not more) of these people on either, or both teams. How does the system quantify that?
That’s why it wouldn’t make any sense to put players in a pool where everyone is more or less evenly ranked in their first placement match, because it only makes sense to assume that blizzard would understand that everyone definitely won’t be anywhere near evenly matched come their first placement match/matches, and that placements might be scaled based upon the first 25 levels of data that the game has gathered on you , and your efficacy.
If the game isn’t using the first 25 levels you play through to gather data, build your MMR, and match you against evenly-skilled players to create more fun and competitive games,it just wouldn’t make sense to me. It makes more sense that your rank isn’t solely, or even prImarily based upon your performance in your initial placements.
If you really go in depth with how severely putting every player on an even keel for their first match could impact their final rank, or goving too much priority to a persons performance in their placements, then it just wouldn’t make sense.
It seems more likely that your MMR is being built during your first 25 levels of play, and your placement matches are used to see if you belong in the rank that the game thinks you are, based upon your MMR, or if you should be promoted or demoted appropriately. Especially if you factor in how placements can often be a crap-storm because of leavers, thrower, and smurfs.
The problem with playing with people who are not of an equal rank, weather on the same team as, or against is that all of the stats becomes skewed, as the strong players will appear even stronger, and the weaker players will appear even weaker.
Again, I’m not saying that you’re wrong, and I’m not saying that you didn’t present what is possibly the most valid theory that I have heard to-date, because you definitely did.
I’m saying that without knowing exactly what the MMR measures, how with measures it, and what role that could play in the matchmaking, and your placements ultimately, I can’t immediately go along with your theory as a fact.
I like you two. (QTQ and Ka.)
Have you already noticed that you are basically 2 sides of the same coin? Taking observations, making predictions with your dataset and forming a Theory, then defending it against an opposed view until a consensus is formed… the core cycle of the scientific method. It’s rare to see something like that here and no “get gud mate” in sight yet. Keep up the good work.
That said, I have one problem with your approach QTooQuiet. (besides the missing data ofc). It seems to me as if your whole view on this matter depends strongly on the premise that it would make “more sense” for Bliz to use the QP/Arcade Data for your MMR.
Thats a very very highly subjective approach. Of course it would make sense from from a players view on the matter. But that doesn’t make it more or less likely to be fact.
Let’s take the cosmetics as an example. It would make sense for me as customer to be able to buy Credits with real money, so that I can buy the skins that I truly want. THats the most sensible thing I could think of to enable me as informed customer. Instead we have a randomized lootbox system. Just beacause something is sensible from an outsiders perspective doesn’t mean the System owner sees it likewise or that he has other goals/priorities in mind.
2 Likes