What sorcery is this?

I wouldn’t call it unlucky, it constantly happens to me… I’m b4 on my account and my QM I’m playing against gold/plat players. It’s very common.

Doesn’t seem like it to me.

  • You tend to refuse to accept certain possibilities you haven’t considered
  • You process information mostly in the line you see it, and disregard it in a later section of my post. Reading isn’t one thing at a time, it’s a processed whole. That’s part of why ‘reading is hard’.
  • You prioritize what you want, and act like that has more priority (common sense) then stuff that doesn’t agree with your take.
  • Your reliance on heroesprofile means you trust that more than anything else when HP doesn’t actually affect the matching.

“it thinks i’m a master”. That doesn’t matter. While you can see mmr discrepancy between matched players on the site, that doesn’t actually mean they don’t have the same mmr range for qm. Players that have a wide mmr distance don’t have any other games uploaded to the site, so it has to assume their mmr is low. So gap in matching isn’t due to the match history from the game, but the amount of available data on HP. While players can say “I know their ‘skill’” mmr does not match ‘skill’, it’s just a number to approximate average estimated performance.

If you look at the mmr over time, the evaluation of the game pretty much skews mmr upward just because people uploaded, and not really a reflection of their winrate. While you could claim that at “35% winrate at masters” is improbable, that going to be the case for QM evaluations for people that upload enough to the site. There’s also a similar case with “60% winrate bronze” on HP if they don’t upload enough. And none of those really indicate the rating on the actual game because data-availability does skew what HP posts.

HP stormleague used to post that I was silver – when my latest season at that time was diamond – and I haven’t played in a storm league season since that time, but my ranking has gotten to plat. Since the system hasn’t had games from me to change that ranking, when it gets more information on others around it, then it has to bump up the rest of the mmr and push the numbers up compared to players that don’t upload.

If those level 35 players had every replay uploaded, then it would increase their mmr even if they have a 35% winrate. That is part of the corrective assumption it makes in that it sees a match was made, so it has to assume it is a ‘valid’ match and not a outlier, level game, long-queue instance etc etc.

You capacity to call something ‘impossible’ doesn’t mean it isn’ possible, but rather than you’re not willing to consider that things you don’t understand, don’t work like how you want to claim. And that should be ‘common sense’, but people need to be willing to accept something isn’t want they want if they actually want to learn something about it.

No. Matching is like a sudoku: there are slots to fill, and it’s looking for numbers to meet the goal of an estimated average mmr for the players it has already picked. If it already has 4 players on one side (a stack) then it’s running math variations on mmr ranges it can use for the other side based on the available players in the matching queue. The goal is to meet a projected average as it matches more players in a particular mmr range, and not to grab a bunch of players and then average the mmr between them.

what you keep asserting ‘doesn’t explain why they can’t swap’ is because you keep rejecting the process to be different than your expectations. The ability for something to ‘explain’ isn’t that the explanation isn’t wrong, but that someone isn’t willing to accept things they hasn’t considered.

Just as you said “I’m prioritizing queue times” and use that to reject some of what I’m posting – not wanting to consider that – the game does use matching times as an influence on the matching parameters it’s using to complete the match. That’s part of why other topics tend to mention ‘expanded search’ function on the queue waiting screen. IF a match ‘expands’ then it’s increased the range of mmr it’ll use for the average, it will match games without ‘essential’ roles, and a number of things I keep trying to convey to you, that you don’t think are ‘common sense’ because you have a particular assumption of how you think the matching works, and don’t seem to realize that it isn’t doing what you claim.

3 Likes

Not at all. It’s your assumption that it should be the case, but it’s actually far from truth in ranked. And this is confirmed by players, who made statistics with the results that the difference in winrate between stacked teams and solo differs only by 5%.

In fact teamchat is usually not even important ingame to be successful, because anything can be done with pings (camp, retreat, push, attack). Teamchat might only important for the best players to make the difference.

EDIT:

But then how can we explain ranked differences in ranked, when for example enemy team had all rank 1 players from certain league and the other team has only 1, but the other 4 are below rank 1, but they are stacked? Especially if MMR = Rank counts for ranked play. I experienced this inequalities in ranked way too often and I wondered, if it was a lie by Blizzard that Rank = MMR, because it sometimes feel that there is still a hidden MMR.

1 Like

www heroesprofile com

Upload all your matches to get the best results. I was amazed at how many games I was in that someone else had uploaded already.

3 Likes

Im both a master and a bronze accourding to heroesprofile. Very accurate

Master has mmr of 2811 and bronze a mmr of 2311
My winrate of muradin is positive rather then 46%

There are ways that mmr and rank can still diverge:

  1. If players leave games and take leaver penalties, they’ll be docked points – and those can be earned back over time – but the effect on how their influences mmr isn’t exactly known (to my knowledge).

  2. Similar deal with ranked decay; a player can decay out of a league, but there’s uncertainty on how that influences mmr. I don’t know how leaver & decay bonuses play out since players can have both.

  3. Groups have a handicap on how that effects the mmr average for a match, but the effect and extent of that changes depending on the ranking for the game. Iirc, there was a blog/blue post on which mmr is used for leagues at high end that isn’t used at lower ranks. So depending on where the match is make, the influence of the mmr for that match is going to be different, esp at the border where group, decay, and leaver points can all happen across several players in the same match.

Some players have expressed that the rank gap allowed by parties is too generous and that allows for more variance – or effectual ‘hidden’ mmr – than should be allowed because it does mess with how the game matches.

  1. As mentioned in my previous post, queue times can have an influence on matching and that’s usually the factor on topics where someone posts a rainbow match on lower population servers that has a range of bronze, silver, gold, plat, diamond in a game. The matcher tries to have averages be close, but after so much time, it just tries to fill, and like ELO, the severity of rank change then shifts depending on the gap between the teams.

Those are the generalized/speculative parts. Different games will have different influences, so when people post specific examples, then it’s usually easier to identify more particulars. Part of the issue of my complaint topics is they’re looking for a ‘why this’ on a specific match and are overlooking other details that require looking at more than just that one example to find.

1 Like

i already said this many times lol, QM is the only mode that is still populated in my server, 5 stacks against a bunch of solos in QM always mean better teamcomp and better teamwork especially in a low populated server.

took a look at your profile, apparently the bronze one are those mode which you invest very little time in, maybe that’s why.

  • possibilities that are so low that is not even suppose to be consider as norm, this is my stand point. bronze level player with 65% or even 80% winrate, do you even know how bad this sounds?

  • maybe you are the one who didnt process my circumstances and only use your input to by what you can see and feel on your side to predict what my circumstances is i guess. i post things here with data, on the other side, you only speculate.

  • then again, throw back at you, because i pretty sure my words of “QM is the only mode that is still alive” has anything to do with how you think my server works.

  • i rely on heroes profiles because SL rank etc are DEAD in my server. QM is the only mode that people still play. im not sure how many times do i need to repeat this.

i win game, i gain mmr, i lose game, i lose mmr, how hard is this to be understood? lol
this is basic 101 game mechanics that is being used across all kind of game across all genre.

you are trying to justify a bad example here using a really rare case that shouldnt even be the norm.

just imagine all game has people with low skills has much higher winrate than player with high skill. does this even make sense to you?

then why should anyone get good at the game?

pretty sure it wont, even in heroes profile, if you lose game, especially if you lose games against players with worse mmr, you lose more point.

your capacity of calling something really rare to be the norm is what that shouldnt even be consider.

you are trying to justify extremely rare cases, but me on the other hand is taking what is the norm and what should be the norm.

i dont deny happens of rare cases, but that shouldnt be the norm.

and the goal should have put the lowest mmr guy in the 4 stacks, which didnt happen like what you said to be average.

because i dont consider a player with 35% and 37% winrate that is lvl 37 and lvl 29 to be good. if they are smurf, they are bad at the game. if they are new, they dont even play enough game to understand the game, the more reason why they shouldnt be place in the same team




you need to understand some basic 101 simple game design here, you win game, you gain mmr, you lose game you drop mmr. then system will adjust the game difficulty according to your winrate and mmr.

not the other way around of “you lose game, but you got skill, so im going to give you harder opponent so that you can lose more game” OR “you win game, but you are bad at the game, so im going to give you weaker opponent so that you can win more game”.

and as i said, i consider the possibility of extremely rare game breaking cases, but that shouldnt be take in account as the norm.

literally no game designer will create game around what you said. because it will drive away the players that care about the game.

1 Like

Just out of curiosity, in ranked how far from Diamond 2 are you?

No, I’ve processed it. That’s why I gave particulars as I have, and why you keep repeating yourself. I’ve already spelled out a large part of the issues here, and you simply refuse to believe it.

You keep relying on unreliable information in ways it cannot ‘prove’ what you try to assert – what I keep trying to explain – because heroes profile is not the system that actually matches. It only reacts to the information it does get, and it has patterns on how it reacts to that information.

“I win games my mmr goes up” No duh
“I lost games the mmr goes down”

What you are missing is “by how much”

If a player wins a game and only gets 5 mmr, but loses 15 mmr on a loss, then that’s going to change the end result beyond just an evaluation of the winrate percentages. That’s part of why I’m trying to explain particulars you keep neglecting: details matter, and you want to ignore those.

It’s a written format, you can stop skipping bits for repeating yourself and actually go back and read through all that stuff you don’t like. Just because you don’t like it, or who wrote it, doesn’t magically make it less applicable.

1 Like

yes, and game system is not suppose to throw you in unbalanced games CONSISTENLY in the long run. that’s why a bronze with 80% winrate or a master with 35% winrate doesnt exist in any other moba or competitive gaes i have played.




same like other guy in other part of the forum, you are just telling me hots match making system is horrendously bad that this is acceptable.

and you still dont understand the magnitude of how games should be for a master player to have 35% winrate or bronze player to have 80% winrate. it might happen for a short while when they create new account, i dont disagree. but in the long run everything will drop back to what it should be.

you can just try to simulate the circumstances of how many games it should be or how bad that master player has to lose game against abunch of bronze player while winning games against much better to have that winrate.

losing game to all worse players and winning games against player that have much better skills CONSISTENLY. logically speaking this is not even something that should happen consistently.

this is ridiculous to begin with that this is even something to debate about tbh. in every other games i played, especially moba like dota, LoL or even Hon, winrates dictates everything and which rank you should be.

Fully decayed I dont know if that matters. Ive reached master 700 points in. Didnt play half the season and decayed
This is my smurf so not up to date. but it says dia 2

MaestrPotato here, made an alt just to test out your theories. lvl 27, 70 games, 60.9% winrate, all games solo queue.

now im more sure that it is impossible for people around my skills or around high diamond to masters players to have around 35% winrate even if it is their alt account. at least not in this server even though it is one of the least populated server.

for comparison:

hotslog:
mini-3078
maestr-3065

heroesprofiles:
mini-2748
maestr-2901

Let me in on a magic secret for you: it’s not ‘theories’ that I’ve been presenting.

Grand master ranking has its own scoring system. With enough games played in games at that level, the point system can essentially make it so some players can’t ‘fall out’ of the rank, so even with increasing lose to win ratios (but not hard steaks) they can still sit in that rank depending on how much they play, who’re they’re ranked against, and how active the rest of the server is at that rank.

Or in simpler terms, if you look at HP’s “mmr” evaluation, games get to a point people are only changing at 5 points, at which point it’s hard to see much shifts at high rank.

That’s part of why there are reddit topics that talk about low-winrate leaderboard entries. While there probably isn’t an example of ‘exactly’ “35%”, my point hasn’t been about that specific number, (rather it gets repeated because the example you used from qm was a low winrate) but the differences in how the systems work at different mmr. Some people have noted players have sub 45% winrates, but still gaining ‘points’ in the master system, so they can still have ‘negative’ winrates and not lose the rank.

Similarly, the climb from bronze 1 and up is very different from bronze 5 where accounts at really low ranks do need to have more than 60% winrates to effectively climb out of that. The rates and time it takes to climb from those spots have been noted by select players that do bronze-to-gm challenges.

While many known examples of the disparities of ‘skill’ and ‘winrate’ were better pronounced years earlier of the game (when it was more active) a hefty bit of my replies have been trying to get you to understand there are things you are refusing to even consider.

So to rephrase that: what I’ve been talking about is stuff that is ‘known’, it’s not speculation. The goal of a matching is to get people to a level where their mmr is going to settle, however, the math on how that works is very different at the extreme ends of the bell curve (bronze vs grandmaster) The type of ‘proof’ you’re trying to demonstrate with the alt account is just more misunderstanding of what I’ve been trying to convey to you because you don’t want to think it through. And the functional concern is that isn’t not ‘speculation’ it’s how ELO-based systems end up doing cuz math.

As another point of comparison, if you look at the blizzard listed leader boards, and compare that to the same region on Heroes Profile, you’re going to get very different results. Some names aren’t listed at all on HP and some of the players matched against literal grand master players are appraised as bronze/silver to heroes profile. One of Kure’s entries is even listed as ‘bronze’ in other modes of the game due to how few games they play in that mode on that account. So if HP were to be taken at face value, it shows estimations with wide mmr disparity because of low mode exposure that doesn’t indicate what ‘rank’ that player is actually at from one mode to the next.

1 Like

and you missed out the “most critical part” of the evidence that will fool prove your theory.

this account is lvl 27, but have 60.9% winrate.

both the players in this topics are:
lvl 37 and lvl 29, which is not far from me.

like how i explained, cases like these are extremely rare to begin with, and not to mention, there are 2 in my team instead of 1, which is theoretically impossible to happen.

again, “consistently” is the keyword here. to reach the states you are talking about is extremely unlikely. since it require extreme match making “consistently” for a long time, and not on accounts that hvnt even hit lvl 50 yet.

to be able to play at super high rank games while they have 35% winrate and only play under 100 games because game system thinks they are “hidden high skill players”? this is too far fetch.

what you are trying to tell me here is that a guy won a lottery’s first prize 3 times in a row. theoretically speaking it is not impossible, BUT the chances of it happening is so slim that it is negligible.

yes, and when you lose game against players with less mmr than you, you lose more mmr when you lost the game. i know this basics game mechanic.

this is me, the reason behind this is because server is filled with smurf and 5 man stacks that farm winrates to even 80% winrate.

maestrPotato this account has 49.2% winrate, but it is a 6 years old accounts, i faced countless smurf and 5 man stacks with this account. at super high rank games, some people will have to lose more games, and old accounts like mine are the victims.

miniPotato on the hand is a smurf account, that’s why it is at 60.9% winrate.

yes, but on a normal healthy server that is not filled by smurf or 5 man stacks winrate farming, that player should never ever have negative winrate.






look at this topic again, there are 2 guy under lvl 50 that have around 35% winrate in my team, and you are literally telling me and trying to explain to me that these 2 guys have the same skills and mmr as me or enemy’s stacked team across all your comments.

and i repeated this quite alot, QM is the only mode that are still being played in my region. if you are talking about rank like SL, these 2 new players havent even started playing rank, not to mention, to queue for high rank SL games, i tried to queue for 45 minutes and cant even get into a game. so their only mmr are within QM.

dodge extended queue (30s i think) if u want better making or queue as a group

Ye I also started to cancel extended ques now. Even tho it was only 1 sec extended que before game popped up matchmaker still managed to give me a 5-10 score Zarya and a 1-10 score Jaina with only 35% winrate and 1.2kda lol.

One of the worst onetrick Jainas I have ever seen. 100 games played only on Jaina and she still manage to have that low winrate and even lower kda.

1 Like

will probably dodge into another 5 stacks or smurf, haha

i dont resent player like this, because it is not their fault that they are not in the proper mmr game tbh.

so true man, sadly nothing u can do about it. in my opinion qm wasnt meant to be a serious mode anyway, just to get familiar with different heroes and map mechanics.
though i can totally understand how depressing it is to get teammates which make u think “wtf are they even doing” as it happens to me a lot too when trying to lvl up a specific hero in qm mode.
so just give it ur best dont get influenced by the negativity. in the worst case just change mode. i found coop vs elite ai to be a very pleasurable experience in regards to player behaviour. also aram can be very fun, note though that the rnd factor has a huge impact as some heroes are super imbalanced there.
if u intent to play ranked seriously, i d recommend reading up map strategies and map specific hero comps and train those while looking for like-minded ppl in draft mode.

1 Like

i acutally envy you guys still able to play aram or vs ai haha. for me in my server, that’s unachievable anymore even coop vs ai, most of the time, you will ended up playing with and against ai all the time, meaning a fully bot filled lobby.

and aram queue was alrd very long few years back, nowadays no one plays it in my server anymore. sadge haha

The people who think there’s a third party site with any reliable statistics xD