buddy, you are an idiot.
how football leagues works? they give you +10 points for beating second placed team 12-0?? no, it doesnt matter, u win u get +3 otherwise u get no points. thats all. simple is the best. f = ma. idc about anything else this simple formula changed everything in our lives.
its so fucking simple, your performance means nothing if u lose. if u are good u win, otherwise u lose. thats the meaning of the competition. idc if u think i performed bad, i have my own way of playing this game, and only things can decide if i am good or bad is winning or losing it. nothing more. gpm/xpm/lpm/tdm can go fuck themselves.
@Kowareta, kid listen in football leagues you are not in a team of random players/strangers! Keep making a fool of yourself like this, it just feeds the whole purpose of this thread.
i just want to give u an example, its about consistency buddy. CONSISTENCY.
and now i read the half bottom of your text. wow. MMR for every hero??? u think dota is about how to play heroes?? says a lot why u are normal skill after many years playing this game.
If it is like you said they should have the same heroes fight each other. I think there was a mirror mode in dota 1. Where both side has the same mid, same safe lane and support heroes. Then if you lose you really have no excuse and is purely skill.
nah
in that case its still not fair, were not same persons, its not my fault my dad had the bad genomes and the other dude had better ones. we should see who worked harder too. i have worse ping :( i have cheaper pc :(
If you wanted to rate someone's skill based off a single match that's when you would need an algorithm. But since people play 100 games before they can even start ranked, the metric of whose ancient died is much more accurate than any algorithm since the sample size of games is huge. It literally - LITERALLY - doesn't matter that people underperform in won matches or play well in lost matches. The average is unaffected.
MMR is already is extremely accurate to player skill, it's just that lower MMR players' performance is very inconsistent from game to game because their game knowledge is patchy.
The only issue that can exist is that as a support spammer I can't play carry at my support MMR because i will just laughably lose, but it's a free game, so I can just smurf to play carry at the MMR appropriate to that.
@Kowareta - "says a lot why u are normal skill after many years playing this game."
Many years of playing this game? LMFAO.
well i thought u said that in the other thread, i dont have a good memory. its really rare to see some new player comes with ideas about matchmaking and etc. like there is lot of stuff u dont know about and all you care is teammates and performance? you think u are the master already?? are u suffering delusion or sth? dont play team games then!
@Kowareta - I won't miss your salt and hate. Good luck, hope you find some meaning to your life.
Changing the way players are being rated would most likely change the way the game is being played. People would no longer have the same need to cooperate, because they are being individually graded. The supports would for instance buy more wards than necessary, because it boosts their score, the carries will probably prioritize farming regardless of game situation, if they are behind on CS. Can you imagine "No I won't join the fight, I need 20 more cs to even my score"?It would completely shift the dynamic of the game, and I can't for the life of me imagine an algorithm or ten that would sufficiently judge each player individually while leaving the game unaffected in terms of dynamic. Atm it's more or less "whatever it takes to win" If that means I'm a brown boots + clarity and stick for the first 30 minutes so my carry is left alone, then so be it. You're proposing a "Winning doesn't matter as long as your score looks good". (I realize winning could be implemented as a factor in your suggestion, I'm just making a point).
The simple fact is, play well, win more. If you have a big enough impact with your consistent performance you WILL have a positive winrate. You're looking at this with a narrow lens, you need to see a bigger picture.
This is the biggest bullshit Ive ever read here.
Mmr based on performance, not the ability to win? Are you high?
fucking hell, for the day i met Kowareta on dotabuff he is toxic as fuck i know, but in this case he is right. IMAGINE ME fucking saying he is RIGHT.
Wanted to say those exact things myself but there was too much to write, I 200% agree with Natalie
MMR score for every hero? Then how are you going to match players with each other? I never play jugg before so I get 0 mmr, then I go in as jugg and stomp the whole game?
Our MMR system is not entirely based on winrate, in case you are unaware. It is based on calibration games where they look at your team impact. Now they dont even look strictly at mmr. They also look at how frequent you are playing the game! This is the trick to scoring high winrate. Dont play for a few days then come back and stomp matches.
Now of course the matchmaking is not perfect, but it has gotten better by removing smurfs. These are new accounts which easily amounts to 1 out of 20 players in crusader and archon. So there is no longer that complain, they got some 7k tinker or invoker in the enemy team thats why I lost. Now is even better, more heroes are banned which makes support life much easier.
Another trick which higher mmr players employ is using analytics to help them decide on the hero they should draft. If you are in very high skill then drafting is very important.
Seems bad. Would favour different playstyles and definitely not mine. Would even influence people to play incorrectly to boost their performance stat. The reason so many stupid VHS posts come about is because for the first few games they do implement a system like this, and how did that work out?
Winning and losing is the only thing that should matter.
I'll give an example:
https://www.dotabuff.com/matches/5332931284
Your metric would say i was most useless on the team, and sure i def didn't play perfect. Yet I would say, AM became that strong because i dedicated my game to him as win condition.. how will your system measure that?
OP, idk where you got this idea from but it won't work. You can actually tweak this 100 times over the course of 5 years and it still won't work, because the base of your system is bad. The reason why mmr works so well is exactly the opposite of what you say, it's because it is a system based on a simple logic: if you win, you get mmr, if you lose, you lose mmr.
That's it, that's the basic formula and that's how every game should behave actually. Your idea is like when math teaches say you need to use only an specific formula to get to the answer, meanwhile everyone knows that it doesn't really matter how you got to the answer, If you got the right answer then you're doing the right thing. The same applies to mmr because there's a lot of ways to win a game, so if you judge someone by some specific way to play, you're breaking the whole game because of the principle of: if you're good, you win a lot, thus you're higher mmr than you should be now.
"But dude, what if the guy got the right answer by luck?" Then, he'll need to have this same luck over 100 games or something, which is pretty unlikely to happen if you ask me. If you wanna see what happens when you try to cheat the system or get lucky, you can search for the threads about people who calibrated higher and are losing tons of mmr now; I'm sure you will find this stories in this forum.
Also, if you want to see what happens when you put a lot of variables instead of making a simple win - lose system, you can search about how great Dead by Daylight matchmaking is in general, but specially in terms of going up and down in ranks.
I am not trying to change the ability to win factor, but rather I am adding performance metrics to it. For me, I shouldn't be losing MMR cause some idiot decided to abandon. Similarly, I shouldn't be gaining MMR when I was a deadweight for my team.
@Natalie Sup'Portman - Like I said everything would be factored in to get you your composite score relative to time. If you avoid any TF i.e. no HD for like say 20 minutes you have already lost a lot of HD points cause every stat is scaled with time.
@Dreadnought325 -
"I never play jugg before so I get 0 mmr, then I go in as jugg and stomp the whole game?"
"It is based on calibration games where they look at your team impact."
- You just answered your own question! Unranked MMR is hidden but it sets up the stage from where your calibration can begin. So in essence, you start this reworked matchmaking the same way you would get matched when you make a new account in Dota. As far as drafting goes, would you consider your skill to be equal in all the heroes in the game?
@Tyski - My system would compare your performance against the current average for that hero along with your win-rate. For the match you shared, your main contribution was being a defensive support i.e healing / wards / smoke / etc. Also, winning as a support would greatly enhance your score cause it is usually harder to win if your carry fails miserably.
@Salty - The reason this system is in place is because it helps retain newer, more inexperienced players. The system helps you win often even if you did poorly. On the flip side, more experienced players improve over time and tend to stick more with the game.
"people who calibrated higher and are losing tons of mmr now" - Brother, I am very familiar with what you are saying but do you know why there's a drastic drop in winrate after they calibrate? I will explain - They spam a couple of heroes to break into VHS bracket, then when their unranked MMR finally drops them off to their newly calibrated MMR, they enter a totally competitive pool i.e. ranked. Now, most players create a new account to enjoy "better games" even though their plays are only of that level on certain heroes which they spammed to get to that level. Therefore, anytime they try a new hero or play a sub-skilled hero, they instantly lose and hence the drastic drop. Why do you think those people created smurfs in the first place?
A system that could do what you want is near impossible with current technology.
I'll give another example my gpm, kda, supporting EVERYTHING is well below average (honestly often is?), yet i find an illusionist cape and bait 3 enemies into committing on to it, and this play won us the game.
Another example is maybe you are talkative on microphone and make good judgement calls, equally so maybe you are toxic on mic and tilts the team.
What measurement will take these into account?
There are literally thousands of scenarios like this. Fairest thing is win/losses.
Play 10,000 games of dota and it's way more likely that the number of abandons of enemy team is higher than your own team (unless you yourself are an abandoner).
For toxic behavior, if you follow Reddit posts you will see since 2013 many doubted the use of behavior system which matches you with similar behavior, but people refuted this idea and now Valve released it themselves. If you are toxic, you will be punished either by the system or by the players or both. If you are helpful, your chances of getting a commend is higher and so are your chances of winning.
@Quantum - They don't actually calibrate higher spamming these heroes, they get this hidden mmr higher because of standards that Valve sets just like you, standards that are compared between ranks and if you have X rank standard you'll belong in X rank with that hero / role. The reason they fail is because this standards are not enough to actually measure skill and understanding, and the more variables you put on the equation the harder it gets to actually measure what you can call "skill". So you just go back to the win - lose system.
Also, if they are actually X rank good on Y hero and they only got there spamming it, why not keep spamming and winning? Or why not keep spamming to maintain the 50% winrate? It's because they actually can't win even with their best hero, all of that because they don't belong in the skill bracket they were put at in the first place.
But I get what you're saying, this isn't about calibration, this is about performance to win scenarios, right? Well, yes, it's really trash to perform good and lose, and sometimes being carried is just as trash, since you don't know how you actually won the match; but what you need to understand is that since the system is completely random, this odds tend to equal out, meaning that for every retard teammate you get, you'll also get a retarded enemy. It may not happen in the same game, that's true, but it eventually will. The only way to prevent this is by controlling the only variable that you can, which is yourself. That's why if you're good enough you'll have close to if not 100% winrate while climbing in some mmr that's 3k below yours.
So TL:DR, if you want a system that covers all these variables, you just need to put a system where if you're good enough, this variables won't actually matter to you. How are you able to recognize if this players are good? Well, since the game has an objective (destroy the enemy ancient), you just need to put a system that recognizes the players who are able to do that consistently, and you reward them for winning and punish them for losing. The players with the best winrate (the ones who consistently destroy the enemy ancient) are naturally going to climb and meet people who are also able to destroy the enemy ancient as oftem as them.
Did you just read a backlog of reddit posts since 2013? Earlier in this post you laugh at the fact when somebody said you had been playing for years.
Also that doesn't address the guy making good judgement calls on mic? Or the other 1000 scenarios that could make the difference between a win and loss, and cant be measured with numbers. Thinks its time to hang the gloves up on this idea, any system you implement would be less fair than the one currently in place.
@Salty - If you could give me a rundown of how calibration works from unranked to ranked, you will understand what point I was trying to make.
@Tyski - I have family and friends who have been in the DOTA scene for years and I am very familiar with how it progressed. I work as a game dev at a well-known VGP, so yes I do build similar ranking schemas and I could clearly see how profits clearly triumph logic here.
Profits triumph logic? Doesn't even make any sense in this context.
What does profit have to do with your system, outside of me quitting because it has no logic.
@Tyski - Do you realize that current matchmaking balances wins and losses which gives weak players incentive to keep playing? So the more players, the VGP gets, the more profit they can make.
Lol horrible horrible idea. It's flawed on so many levels I can't even begin to explain. The ENTIRE POINT of the game is to win.
If you just focus on this line, like for 5-10 minutes, you'd understand why your proposition (which has also been proposed before) is flawed. Good players NATURALLY win more than they lose. Performance means NOTHING. Who judges the standard of performance? In a game like Dota where EVERYTHING is so fluid and flexible, there CAN NEVER be a standard good enough to actually change MMR based on that. Last hits? Denies? Kills? Deaths? This is so pathetic, all of these can EASILY be abused by below average skill players and you want the system to reward them with +MMR for that, lol.
Remember this, and remember this well. If a player is GOOD, HE WILL FIND A WAY TO WIN MORE GAMES THAN HE LOSES, irrespective of the performance of his team. The entire POINT of the game is to win. If you can't FIND A WAY TO WIN, it means you are not good enough. There are MANY examples of people, who won 50-100 games in lower bracket smurf accounts in a row.
The point of the game is not to pErfOrM gOoD but to win. Sure, good players do lose and will lose some games as well, even where they perform very good, but if they are truly good, they will find ways to win more games than they lose.
I don't mean to insult you but you obviously don't know how roles are played. Pos4's sacrifice themselves to create space for their cores all the time.
----->>> YOUR BULLSHIT IDEA DE-INCENTIVIZES WINNING ITSELF!! <<<---------
Think over that for a second.
Also these propositions will ALWAYS come from players who are below 3k, have subpar game sense, frustrated and stuck in their bracket for ages (No offence, but this is the how it is in most cases) because good players understand how this idea can not be practical.
Do you realize that current matchmaking balances wins and losses which gives weak players incentive to keep playing? So the more players, the VGP gets, the more profit they can make.
totally disagree, if u know how to check if a binary string is random or not go ahead and check your win/lose stream. i did it for mine, its totally random. matchmaking is random. there is no forced winrate or etc. its just random! i played on smurf and after like 9-10 win streak i got another smurf on my team. do you think that happened because matchmaker tries to force ppl lose/win games???
@Kowareta - "i leave u alone in your retarded life, rest in peace" wHaT HaPpeNeD
@Ace ♡ Ashes - "If a player is GOOD, HE WILL FIND A WAY TO WIN MORE GAMES THAN HE LOSES" - Oh my! Just do a cost-benefit analysis - If you spend 45 minutes of your life on something and do it well, do you not want to get rewarded for that? From what you are saying, you only get rewarded if the outcome is positive, what about the effort? What about the time put in? What about the physical and emotional quotient? Just use this logic anywhere else and you will see how incredibly silly this is. Just ask your boss to not pay you if your daily work-related goals are not met or maybe perhaps not pay you until the project you are working on doesn't finish or is successful?
"----->>> YOUR BULLSHIT IDEA DE-INCENTIVIZES WINNING ITSELF!! <<<---------"
Read the above post again, and re-iterate what I have mentioned previously.
Also, I play for fun. I couldn't care less what arbitrary number the game awards me. I am a dev and I was trying to engage in a constructive discussion. I failed to realize this is a place where kids are afraid of change? Why? The size of their e-p*nis will shrink if the system is altered. I am sorry, this has been a failed expedition. Good Luck!
is that all u can say? fucking miserable manchild. literally everything i said in this thread u just give me bullshit in response.
u claimed you are game dev so i gave u a second chance, but it seems my judgements about ppl is most certainly right.
I understand your life is miserable, but spewing hate only makes it worse. I strongly recommended psychotherapy and SSRI.
oh? as long as i see, im talking to you and u response like an animal. u wanna get respect with your shitty behavior?
its ok tho, now u can live alone in your deluded life where everybody here was wrong and u were right, Sir Galileo!
Mr fun player, can u stop claiming once in your life that you are playing for fun and u are careless while u get triggered about your teammates, thats very stupid and ironic.
@Quantum
Yeah of course theres a system that balances wins and losses, it's called the MMR system. Proof the system works.
However, it seems you're implying that theres some algorithm in place to force you to get 50% WR outside of the mmr system. Just No. There definitely isn't. I don't even need to look at any source codes to see how false that claim is.
Want Proof? Two categories of people who intentionally play outside of the mmr system; Smurfs and Account buyers.
Are their winrates 50%? Over an extended period of time, definetely not.
The third type of person who fits this description is people who calibrated higher/lower than they should be due to a system very much like the one you are proposing.. eventually the mmr system catches them too.
Your example of asking the boss to not pay you if you failed at your job is so bad, because don't worry you won't even have to ask him, you won't be working there for much longer.
You claim you wanted a constructive discussion, but it seems more like you should of changed your topic name from Share YOUR Thoughts to AGREE with ME!
@Kowareta - Kid, check your earlier responses and then decide for yourself. Again, I strongly suggest Xanax and Lexapro for you.
@Tyski - Why are you getting triggered? I know why you love your smurf account so much. And hey, please quote me where I mentioned "forced 50"
"Over an extended period of time, definetely not." - Thank you for proving my point.
"Do you realize that current matchmaking balances wins and losses which gives weak players incentive to keep playing?" This pretty much implies forced 50, unless your understanding of the word balance and mine drastically differ.
Also, how was I proving your point? Pool size of 2 games, they win one and lose one, thats 50%. Over 10 games maybe they get lucky/unlucky and keep this 50%, but over 30+ games, the likelihood is much slimmer.
Only point proven was that you are not open to constructive criticism.
is this tyski pretending to be a nice guy or he is very dumb??
only this miserable guy can do is skimming your wall of text and find something wrong(in his opinion) and mentioning that like a donkey.
obviously the guy suffering issues, probably he never got any acceptance in his life and now he is sad cause nobody is agree with him.
imagine wasting hours thinking about an idea, writing it on a forum and ppl call u dumb, LUL
leave him alone talking with him is pointless, he just wanna win once in his life, give that to the man, he deserves that.
wtf xanax blabla st fu dude with your normal skill dota player ass. u are a nobodyyyyyyyy no one cares your ideas for the matchmaking. besides them being utterly shit.
@kowareta pretending and very dumb. So now that the jig is up, the original comments had more teen angst, even had a get your e-penis stroked at chaturbate comment, which I was particularly proud of.
@Tyski - Do you win every game you perform well? Do you lose every game you perform poorly? Answer it yourself and then reply back.
@Kowareta - "wasting hours thinking about an idea, writing it on a forum " Kid, aren't you doing the same thing?
@Cowboy - "normal skill dota player ass. u are a nobodyyyyyyyy no one cares your ideas" Yep, MMR is everything.
tyski, sry buddy but i didnt get what u mean at all, im bad at english
it doesnt matter tho, keep sucking this guy dick for free while he behave like a shit to you
yeah, go tell these secrets to better community with higher avg IQ, we dont understand u here. bye bye!!
Of course I dont? I don't make some long winded thread proposing a stupid change because of it though. In the long run if I consistently perform well/poorly, I will win/lose more.
Even if you could magically make a system that accurately measured players performance accurately (spoiler: you can't), what is the point? You think it would make the playerbase happier?
n0Tail: Ana you sucked in game 2 and 3, and got carried by topson. So taking that into account, your share of the prize pool is now 500,000 dollars less.
Ana: but i won..
n0tail: computer says no
^ if it Was about the prize money ana would take all, and the other 4 dogs had to pay him lmao.
"Make money". If you want that you can have heroes that can only be purchased to be played, and make those heroes op. You can have certain buffs like + atk dmg or + hp to pay to use.
@Quantum. Of course your ideas do work. Other games have blessing points which can be earned and converted to climb rank depending on your match performance. Mobile Legends have protection points that can prevent you from losing a star depending on your performance over the last few games. In fact these ideas are quite old, can be applied to any game and have actually found success in games like Mobile Legends.
However, winrate still accounts for more than 80% of the metric that is used to measure your skill. So at the end of the day, winrate is still important. It is ok to agree to disagree, I mean it is up to you what you want to do to your game.
კომენტარის დასაწერად გაიარეთ ავტორიზაცია.
DOTA is an amazing game and has an amazing community! I cannot imagine my life without it. However, I feel like matchmaking often holds SOME good players back from getting recruited into the PRO scene (I’m a noob and I’m not talking about myself). Now, you would probably argue that if you are good you will be noticed anyway which is true, but what about those players who understood, loved and played the game well, but their MMR dictated what tag they were given and hence they eventually quit.
Winning > Performance
This is the central theme of matchmaking even though you are matched with 9 strangers.
Currently, matchmaking ranks you based on win rate, rather than your individual performance!
For example,
Game 1: Player Performance – Great ..... Result: Won
Game 2: Player Performance – Bad ..... Result: Won
Game 3: Player Performance – Great ..... Result: Lost
Game 4: Player Performance – Bad ..... Result: Lost
Game 5: Player Performance – Great ..... Result: Lost
Overall Performance 3/5 .....Overall Winrate 2/5
Now, this could go in any number of ways depending on individual performance, teammates, enemies, server connection, etc. etc. You could perform outstandingly but still lose cause maybe somebody lost connection or maybe someone fed in your team. On the flip side, you could perform terribly and yet win cause of the same reasons on the enemy team. We tend to ignore this because we enjoy the flip side more! We get frustrated when we play well and still lose, BUT we never complain when we win even though we perform miserably in certain games.
Now, here’s my suggestion to rework matchmaking:
MMR should be hero-specific, where your individual performance is graded rather than the overall outcome. Like, with all the data analytics in place it shouldn’t be too hard to rate your hero-specific performance against all the players on that particular hero (leaning more towards percentile-based scoring metric). To avoid abuse of this method, all areas of the performance metric should be graded. For example,
If you picked AM, how does your performance scale against the average performance on that hero? i.e. GPM, XPM, LH/D, TD, HD, Item Timings, etc. etc. If you AFK farmed you will massively miss out on HD and then if you further lose the game, you will also lose a lot of points. Similarly, if you picked say Omni, you will lose fewer points for LH/D, while earning more points for healing and support. Lastly, your teammates report/commend would also add significantly to your score. So, your overall MMR points would be a summary of your performance/behavior and not a reflection of just the typical win-loss outcome.
As far as queuing is concerned, you could select a pool of heroes with hero-specific MMR points and the average of the points is how you would be matched with other players. Like, you selected AM (4000), Dazzle (2000) and Invoker (3000), then your average MMR search pool will be 4500. This would greatly reduce the randomness of the outcome and provide an incentive for players to perform better. Behavior scores would also be a huge factor in this kind of rework.
Advantages of this kind of rework are very distinct, i.e.
- Reward good performances, while punishing weak/bad performances.
- Balanced Outcome / More Enjoyable matches.
- It would be easier for teams to recruit players skilled on particular heroes.
- A better indication of skill level for solo players.
Please provide insights on what you think and how we could improve matchmaking for solo players.