They get it from people uploading their data. Since you aren’t personally uploading all your games, it’s only going to have a sample from other players that do uploaded and you were matched with them.
It’s likely “you” and it’s not “cherry picked” but simply an incomplete sample. However, your response and limited understanding when presented with information is par for the course: people don’t know, don’t want to know, and think something was “proven” that doesn’t align with their claim, but they just want to be “right”.
He main concern still stands that people that profess massive exceptions (ie, they routinely happen) do not provide information, nor do they want to consider their limitations of their own incomplete information.
If people want fair, competitive, matches they should play ranked. If they want to get into a PvP match fast just to have fun or practice a hero they should play QM but then not complain when the matches are not perfectly fair as that is to be expected.
Free will and skill are pretty much contradictions to each other. Skill limits the choices a player makes towards those that are more likely to have good outcomes based on learning from past experiences. I guess the lower the skill level of the match, the more free will could be considered as the dominate factor driving decision making, but that would do so equally for both parties.
A player could purposely decide to start playing worse using their free will. However, this would effectively reduce their skill which the match maker would try to correct for after enough games.
The main issue with QM is not so much free will and other aspects skewing win rate, but that there are just so many random factors and errors added to the system in order to be “quick”. In extreme cases QM even allows competitively impossible matches, such as mirrored heroes.
You can deflect as much as you want, but the concern you express – years of broken rules – lacks evidence of the claim. Even if I neglect you as the specific example, other accounts – that do upload – lack the extent of what people continue to claim here.
Heroesprofile is just an avenue to try to find experience that either
a) corroborates what people try to claim
b) be something that is different from my own.
QM or “algorithm” complaints generally come from players that don’t know QM rules, make sweeping generalizations, and don’t provide screenshots or other validations to confirm their claims, esp as a means to learn more about something (is it a ‘problem’?)
The typical conduct is people that refuse to believe that something is an option they didn’t consider, and it turn they demand more from anyone else that isn’t them.
You aren’t “biting” you’re missing the point in favor of validating your own ego. I don’t care about the “Average American” I care that you seem to be more clueless than you’re willing to consider because you found something to blame. Far as I can tell, you are the “average American” and won’t realize how ‘average’ you are as that’s the basic loop for rationalizing willful ignorance.
For having a concern about “Cherry picking” as a fallacy, that seems more to be the case of what you’re doing than my efforts to try to find proof of the things people claim are happening. I’m not the one denying possibilities, I want to confirm concerns or provide a means to counter ignorances of people that don’t know how the matching works, but want to fault it regardless.
The capacity for something to be incomplete doesn’t mean it’s “cherry picking”. Might help if you knew the difference between what is being discussed, and what you’re willing to consider. Did your stats 101 not cover bias?
I will agree that Xul might not be as durable compared to other bruiser. However, both in Ranked play, or QM I have often played “tank” Xul to good results.
If you have another frontline hero on your team, with some CC like Kerrigan or even an Uther, Xul can be an effective solo frontline vs a Sonya, or even a tank. I’m far from being an amazing Xul player, but I’ve won many games with him as double bruiser (no tank) in Ranked. You will need to adjust your play style and build to achieve this, but it’s doable.
Well, devs already said what we see on heroes profile is pretty much the same data they have there. It might not be exactly the data from your account because you don’t upload, but for overall heroes winrates, talents etc, it’s close enough.
But just share your account that and we can all check it out. Or upload your matches and we can see everything and explain to your whats happening or see if you’re actually right. Or just upload screenshots.
I’m not an engineer, but I can take a crack at this. The forced 50% win rate discussion is weird and complicated because it’s both true and not true at the same time, and people often inject their own biases very heavily on games to support what they think is happening to them, usually to assign blame for a loss to an external source (in this case the matchmaker) because it’s an easy thing to do (the matchmaker can’t defend itself after all!). First, let’s discuss what forced 50% isn’t.
Forced 50% isn’t the matchmaker putting potatoes on your team because you had a win streak. Frankly it’s hard enough just to find good games that there’s no way we would program such a thing, and there is no incentive for us to. We are incentivized to find matches of equally skilled players in our games because they are the most fun overall, which is what the matchmaker always tries to do.
The way it does this is by assigning a skill level to every player as they play our game called MMR. This skill rating is based on whether or not you win or lose games and nothing else (this is up for debate and some day we may find improvements in the future in how MMR is calculated, but there are many good reasons for this and almost every game uses this kind of system to determine player skill. I’d recommend watching some informational videos on how MMR/ELO works if you’re really interested in the nitty gritty). What the matchmaker does is tries to find games where each team has as close to an average MMR as possible while also having the MMR between each player be in as small a range as possible.
Now, imagine that everyone in the entire game has a MMR of 1, meaning that everyone has the same skill level. If we play an infinite amount of games, then everyone will eventually have a win rate of 50%, since their skill is the same. The 50% win rate of those players isn’t the matchmaker punishing those who have win or loss streaks, it’s just a natural consequence of the matchmaker working properly.
In real life it’s not quite so cut and dry since every player doesn’t have exactly the same MMR, but the fundamentals still hold true. As you win games, your MMR gets higher, which means that you will get into games with higher and higher skilled players until you hit your “ceiling”, where you start to lose games because you are the worst person in the game and bring your team down. Conversely, as you lose lots of games your MMR goes down until you eventually get into games where you are the best person on your team. The matchmaker getting you to a 50% win rate is a natural consequence of the matchmaker finding where you, over a long period of time, actually belong. Your MMR being an accurate reflection of your skill results in the natural consequence of you having something very close to a 50% win rate.
Keep in mind also that even 10+ win and loss streak games are perfectly natural in the realm of statistics, even when the true chance of something is 50%. These win and loss streaks are often what people point to as proof that something’s broken, which is simply not true. If you were to flip a coin 10,000 you would not get heads, then tails, then heads, then tails again over and over. Over those 10,000 times you would get each result half the time overall, but you would see streaks of getting heads or tails. This is also true for players who are at their appropriate MMR but still get streaks. It’s not that the matchmaker is broken, it’s just how things are. Also keep in mind that people have good and bad days, and sometimes decide to throw or to try extra hard to win. These behavior changes on an individual game level are impossible to predict and account for, which is why some games, even when they should be evenly matched, can still easily be blowouts.
This is just scratching the surface of how these things work, but again, for a TL:DR:
The idea that the matchmaker forces you to win or lose after having streaks is completely false
Having a 50% win rate over time does happen and is “true” in that it’s the result of the matchmaker finding your proper place in the game over time
You are almost guaranteed to match against smurfs if you play alone. People that play in a group tend to have around 60% winrate and at higher MMRs these people get up to 90% winrate in QM and i think around 70% in rank.
Solo MMR around 50% simply means you are at your skill level, but considering the low population i wouldnt consider it even remotely accurate.
It isnt the best system. It has a lot of flaws and exploitable features in some game modes. Cant really help it. They wont do anything no matter how much people complain.
It is fairly apparent that the average population has left and we actually have the extremely casual players with really low winrate doing a few games here and there and the more hardcore people with very very high winrates left.
The matchmaker should obviously only match the people with extreme MMR against eachother but it simply doesnt and never will.
Ive had master mmr in QM over multiple accounts according to heroes profile. My solo account stays in the 50-53%ish range. I could probably get higher but i despise some of the no skill heroes that can easily win QM if not cc countered.
My solo smurf account stayed around 70% until around level 120 and then started to ease back into the 60% range as i approached level 200.
GM duo can maintain 75%+ in QM as well. I run into them pretty often. Ones i see usually run aba hat+valla/illdian/gm/sylv/kerri.
5 stack masters in qm can easily keep 70-90%wr. Getting to pre choose your A and S tier hero comp, cheese/ wombo combos, no afk/troll/newbs, and voip versus potentially unaware solo random players with considerable mmr differences.
All this said I can play a days worth of solo QM and will at some point get a bronze 1-5 non smurf ally. I actually got stuck on a 4 stack of them this week and it was brutal. Mass feeding, no map awareness, genji didnt know how to select talents, etc.
Thanks for the stats. I’ve been checking on how heroesprofile numbers work as well.
60+% winrate player will be able to reach QM Diamond mmr in about 100 games. With some easing back on the winrate, I would guess the same player would reach QM Master mmr in about 150 games.
This is my assumption, but judging from how, normally, you get more points a win than from a loss, I presume even 49.5% winrate player would be able to reach QM Master mmr if played thousands of games. A little iffy part is, I would say QM 49.5% winrate player is more a Bronze player in terms of Ladder/Ranked. This discrepancy could screw overall player experience imo.
This is assuming solo play experience. I would guess winning as stacks would award you less points, etc, making the numbers work differently.
Seems like rainbow 5-stack (which may include Bronze/low Silver player) can reach 70% winrate in QM as well if mostly played in off hours.
Even winning in those games is stressful. But I can’t deny that there is that cathartic feeling. Macro
That is mostly because they abuse the system cause they know there are none in thier mmr range to give them a fair match. I have been stomped in QM alot of times against full 3000+mmr teams that run around farming randoms for fun instead of staying in ranked where they belong.
Some of those 3000+ teams also sometimes let a silver smurf join thier team so enemy team also get silvers.
This is just some of those QM abusers I played against that still premade in 5 stacks today cause its fun to farm free wins against randoms with high ranked players. Most of them where either diamond or master in ranked that just cordinated destroyes random QM teams so they can feel good about themself.
The last profile is a premade that usually plays in the late evenings when there are no one so those few who play get matched up against a team with almost 1000 mmr higher then them.
Nothing more fun then to get B-stepped and taunt spammed by toxic 5 stacks that are too afraid to play ranked and only stays in QM cause of easy wins. They also mock people they kill I saw in thier replays.
Just one of the reasons why QM should have been solo mode and if people then want to play in groupes they can play ranked or unranked to make it more fair.
There are no reason why QM should stay like it is now.
This is correct. People of obviously high skill match against people who can barely figure out how the mouse works. Somehow this a large population in many games. How did they not die as a teenager.
Just an incredibly self-centered viewpoint. I don’t even play in groups that often, but I recognize that this is first and foremost a group based pvp game. To not allow friends to play each other in a casual mode that shouldn’t even be really taken seriously is just so like, what is even the point, when the types of players that you talk about are an extreme minority of the population. There can be other guardrails to limit the impact of such teams than just outright mandatory solo queue. Like ridiculous.
Revan on Nazeebo is cringe (90% - 5 stack - 3k matches). Basically in a fair and random high mmr environment Nazeebo shouldn’t be winning any games with zombie build but somehow he is only and always playing with four friends, supported by slows and stuns, making it impossible to fail at landing the wall.
Safely assume all their mmr values are realistically at least 15% lower because they deliberately avoid uploading losses. None of them is anything special and they’re gaming the matchmaker by building comps which are impossible to counter.
Even when they receive strong opponents, they’re almost guaranteed to get a few early picks which enables the snowball effect. I’ve played against in suprisingly equal teams of randoms around 10 times so far and just warning everyone of what to expect was more than enough to deny them any advantage. It’s only when matchmaker gave up on life did I lose in less than 10 minutes and we always had a single player 500 mmr below the average to thank for that.
Yeah, I think a compromise would be to disallow 5 mans from being able to pick their heroes. Give people the ability to queue with three heroes that will be randomly selected.
