@Jacked
According to statistics, you have 50% chance of winning. However if your allies have a fairly negative attitude/behavior you will find it harder to work with them and therefore harder to win and vice-versa.
Behaviour score is irrelevant, I'm 90% sure that it is not involved in matchmaking at all. The only reason people are streaky is because of tilt/random chance, go flip a coin 50 times and you'll probably get a couple of 'streaks'.
speqking about beh score i dropped from 4594 to 4150 pretty fast cquse i abandoned 2 games and was just losing my team picks were awful and no one carried me once those last 16 loses FUCKING 0 TIMES
So you got to 4K by a streak of Good luck due to good teammates and 0 skill too? Or do u attributes your win to your own skill (but your losses to bad teammates and bad luck)?
Did you know why you have these unbelievable teammates? Coz you belong to that category. Your KDA are awful imo. Simple as that.
Dota 2 calibration in the past can be manipulated. There was a Oracle spam healing, Zeus Wraith Spam, and the Spec Haunt Spam. Good players can easily reach 4.5k+ calibration using that hero but after the calibration? They fell short. When they play ranked, their winrate falls to negative 50% till they reach their true mmr.
@CUTNPASTE
Flipping a coin a few times will give you 'streaks' true, but how do you end up getting such vastly different allies is what I have tried to explain.
@Jacked
Not sure what you are trying to say here, but I explained what I have noticed.
@Ayase (╬ Ò v Ó)
Thank you for insulting me and bringing up something that is totally irrelevant.
is interesting and makes sense
i do recall totally shitstomping some games like 7 in a row, and then entire team total dog shit moronic feeders.
but skill remains as the main factor for win/loss
however this surely makes sense because at times you need like 5k mmr level of skill to reliably win some 3k games, for some you dont even need 3k skill and a win comes naturally. which explains why cookie does not have a flat 100% wr on his smurf accounts.
Only agreed with the "win alot get dumbfuck teammates", but then again, it usually happen only for one or two games
Behavior score is just a shitty excuse people use to backlash their incompetence
i dont think it works as you described though. because it woud be harder to climb or boost mmr than it is currently if it was the case.
i think behavior score became irrelevant as soon as people (reddit) found out about its existence
@END
My observation has nothing to do with climbing MMR. I only stated how system tries to queue players, giving them a 50% chance to win. If you play only 1-2 heroes and become mechanically more efficient and better your game sense, you will find it easier to climb up.
@arin
From what I have observed, when the BScore came out, it changed every 10 games and you get similar allies based on what you BScore was. So I would get well behaved allies if I had 9600 for 10 games, then get slightly ill-mannered for another 10 if my score went down to 8500. Now they changed it to a cumulative score where you are matched with similar scores but recent behavior markers are added so you dont have tilted/favorable games for 10 games straight, rather see the impact pretty much in a couple of games.
How do you do that? Do you use the console in-game? I'm a noob regarding those things. I just play
So, imagine you are a hero spammer, winning 20 games in a row with huge impact. what kind of people do you need to get matched with in order to get a "fair game"? does that really make sense?
im talking about the "RWR - 70%" in example
Enable Console in your settings. Then press \ while in-game and then type in the two commands.
@END
Your RWR changes so very slightly and if you are winning every game with spamming one hero, then clearly you are going beyond what the system is expected to achieve. You will not notice a huge disadvantage, unless you pick some other hero and or you reach such a peak point where you have to improve your skill level significantly to win again.
this is true I figure this out long time ago. you will have 2-3 noobs, valve doesn't care.
Getting what you perceive as vastly different allies is subjective and could just be random variance.
Why i ask is because I'm trying to determine if your thinking is affected be any bias. We tend to be extremely bias when it comes to assessing such team based events.
So far u have given me some indication that you are biased
Of course its bias, thats all these threads ever are. Retards that just can't accept that they are where they belong. I mean, just look at the first fucking sentence of this guys thread. Blames his bad team mates for dropping MMR. Couldn't POSSIBLY be that he is the cause.
@Jacked
Skill level is subjective? How about game impact? Tell me more about my bias please.
I have been through super high behavior score and low through abandon sprees due to irl obligations, and honestly I didn't find any real differences
The only thing behavior score is relevant for is to send you to LP more easily if you have low amount of it
No one has a shred of evidence for what behaviour score does. It could literally do nothing, Valve could just be tracking it for testing purposes.
Valve has not said what it does and no one has come up with anything CLOSE to accurate data analysis to form any sort of conclusion as to what it does. I don't even think we have access to the data needed to prove anything.
I just had a game where a bs picked, i had marked mid, the bs then marks mid saying he doesnt know how to jungle, forces me, a dp, into the jungle, then after i buy a talon and shit walks into the jungle and says "im jungling" and even blocks a camp and then jungles without using bloodrage. But ive also had games with extremely positive, understanding, and cooperative players. How can u prove Behavior Score affected this? You cant and its just an excuse for u to pass urself off as good enough for a skill level higher than ur own when ur just not good enough.
omg jacked u rly are stupid
i was playing for months in 4600 4700 4800 avg games very most time was 4500
its not that i got there by luck playing for months isnt it ? plus i never played unranked its shit imo peopls less tryhard
when i play safe carry on 4600 i have 15 0 scors its not that im shit, but losestreaks on 4k is nothing strange
u have players who had 5k and now are very low 4k
also there are co.mon acc buyers mad peoples etc
or u get matched vs 4k parttys with 6k solo mmr lul ( i experienced that shit very bad shit )
also when i play mid i rekt , my worst games are when im playing suport and maybe offlane ( only know play axe offlane and have a good score rest heroes i feed hard )
it just hapens i know i gona get back again but rightnow i dont feel playing dota hard or med at all just chill and pick my voker and dont give a shit
@Yukiko
From what I have noticed the maximum BScore you can get is 10,000.
@ADFGK;LSG;KLDF';KAD;SFNGL...
I used this a/c for testing purposes only.
Alright, I am actually really confused by the whole 'Conduct Summary' thing right now...
A few days back I played a game of Nature's Prophet (3rd pick) after a Rubick and TA, which is immediately followed by a Juggernaut and the classic Sniper last pick. Now I obviously know I'm not gonna get my offlane so I go jungle and prepare for the late game (as all games always go with heroes like this) and buy a midas. After some levels I decide to gank aggressive heroes and get pick-offs. Meanwhile, the solo safelane juggernaut has boots, TP and no regen. He quits the lane after one death and goes roaming. Meanwhile the TA has been feeding mid to a magnus. The Rubick, incapable of helping the offlane Sniper comes and joins me in the jungle, helping me with my last hits of course. Eventually they all give up in 15 minutes and go afk in base while I try to solo split push and delay the game as much as possible in case they changed their minds.
https://www.dotabuff.com/matches/3151769254
The result: I get 4 reports since these are 4 solo players not in a party. Now I don't get LP but my Conuct Summary stands at "1 Abandon (unavoidable due to power cut), 4 reports and 3 commends".
Strangely, this is a time in which I'm winning games playing core heroes with much ease. I put this aside and keep playing to a point where the current Conduct Summary is as such:
http://imgur.com/gallery/taJ9R
Despite this I have been pairing with more toxic players than before and losing more games than winning. I would either get constant rage pinging people or some so new to the game they have no idea what to do or what items to build. They don't know what lane to go to and when to fight. I mean, I've been solo carrying my last few games but you can only do so much with teammates having double-digit deaths under 15 minutes.
@RayZLiENT I didn't exactly understand how you find out your BS (Behavior Score) or what RWR is... Can you expand?
where are you pulling all of this data from???
seems ridiculous without a valid source.
seems like a lot of shit that you just pulled out of your ass.
without somehow proving your hypothesis all that you've written down here is just useless nonsense
if it were correct i'd love to know why so i can abuse it myself, but for that we need credible proof, where is yours?
where's the data, only way you'll getting this data is if you either work at VALVE or if you did thousands of games data analysis for which i request that data to see it for myself.
This is bullshit, my latest game had everybody on my team cursing, yelling at each other for a mistake beyond their control while I dig a mine in an enemy turf then get away with a large bag full of gold scot-free but not before sabotaging one of their base exterior structure while crippling their teamfighting capability at the same time by forcing Tinker to leave the scene and defend a tower which is already too late because I have already take down his tower by the time he materialize to the tower area and I keep repeating this process until I become fully loaded that none of those enemy can touch me without getting bitch slapped in return let alone BS who I used to be afraid of due to his ultimate. 800+ GPM in a game where the team are just being too emotional and I had to count on myself alone should be saying something.
this theory vanishes once a minimum amount of reasoning is applied
Russell's teapot, sometimes called the celestial teapot or cosmic teapot, is an analogy, coined by the philosopher Bertrand Russell (1872–1970), to illustrate that the philosophic burden of proof lies upon a person making scientifically unfalsifiable claims, rather than shifting the burden of disproof to others. Russell specifically applied his analogy in the context of religion.[1] He wrote that if he were to assert, without offering proof, that a teapot orbits the Sun somewhere in space between the Earth and Mars, he could not expect anyone to believe him solely because his assertion could not be proven wrong. Russell's teapot is still invoked in discussions concerning the existence of God, and in various other contexts.
I have another theory for the streaks. I recently played several games of ranked that were all below 1500 avg, stomped them all because I was 500 mmr higher than the avg and I play mostly carry. my last four games right after this streak were all at least 2k avg meaning I was below avg mmr and struggled in some of those games and lost all of them. Having this swing of skill avg could lead to these streaks of wins and losses. Not saying this is true just a possibility
it is pointless for us to keep our minds busy for this matter. whatever we cant control = waste of time thinking about
aka omg government is so bad i cant find a job
aka my family members are the reason i feel so shitty in my life
aka mei ememer is dewn bikez ef mei teemmeitz
hey this is only a theory, but Okkam's razor will advice against these speculations. See, if you win mmr u get to play versus higher mmr and eventually u will set back. And sometimes it has to go further a while before it goes down.
Veuillez vous connecter pour publier un commentaire.
I have always defended the idea that matchmaking isn't forcing games on you and you just have to get better to climb up. However after dropping significantly I decided to try and see why the hell am I getting such unbelievable teammates. From what I have found over the course of a last few weeks is as follows -
Every player is tagged as
"Good" or "Bad"
i.e. in terms of game impact which includes XPM, GPM, KDA, TD, HD, Winrate, Experience, Current W/L Streak etc.
"Up" or "Down"
i.e. UP means recent behavior has been good and/or improving whereas DOWN means fairly negative behavior recently.
Behavior Score is just a cumulative score and not an accurate measurement of recent game ally behavior quality, which is why you can get obnoxious retards even if you have 9K score.
Based on how many appropriate players it can find during the queue time (< 150 seconds), it tries to achieve an overall 50% chances of winning for each team. It doesn't necessarily mean the same % for each player. For example,
Radiant
Player 1 - Good (55% RWR)
Player 2 - Good (51% RWR)
Player 3 - Bad (49% RWR)
Player 4 - Bad (48% RWR)
Player 5 - Bad (47% RWR)
Dire
Player 1 - Good (52% RWR)
Player 2 - Good (51% RWR)
Player 3 - Good (50% RWR)
Player 4 - Bad (49% RWR)
Player 5 - Bad (48% RWR)
Even though in theory, they have 50% chance of winning in reality Dire has a slightly better chance of winning especially the skill disparity amongst its players is less which usually leads to better cohesion and sync.
NOTE:
If the matchmaking is able to find a perfectly balanced 50-50 game, then the MMR won/lost is +/- 25. If not, then 1 point is added/reduced for every 1% difference.
Now keep in mind that even though this is like a typical stack, but depending on how well you have performed in your recent games, your position in the stack will change. For e.g.
Radiant Player 1 would usually have been winning or have been having very high game impact in recent games and therefore is now paired with weaker allies, who haven't been able to have significant game impact in recent games. So according to the matchmaking, pairing Player 1 with Player 4 and 5, would lead to more balanced outcome for both parties. So say, if you have performed worse in recent games you will more than likely be paired with stronger allies to build your winrate back up. And vice-versa.
-----------------------------------------------------------------------------------------
Now lets talk about behavior score and its impact.
Regardless of the "Good/Bad" marker, your "Up/Down" markers work independently.
Like, if your recent behavior have been negative you will get similar behavioral allies even though your team still has a 50% chance of winning (according to statistics mentioned above). However it will be very difficult to work with them due to their behavior. So, even if on paper you have a 50% chance of winning you might actually have way less chance.
Also, the chat words used and the type & frequency of chat used is also filtered in tailoring your behavior marker. You might not get reported or commended and still you will see significant changes in your BScore, regardless of wins or losses.
TRIVIA:
1 Report ~ - 500 BScore to -600 Bscore (Depending on type of report)
1 Commend ~ +100 BScore
1 Abandon ~ -2000 BScore
-----------------------------------------------------------------------------------------
Couple of notes to minimize this effect of matchmaking -
- Play only your main heroes (if you are doing ranked)
- Take time-focused Objectives
- Check player profiles before and don't let the weaker ones pick Core (as much as possible)
- Minimal, yet effective Communication (use chat-wheel more often)
- Refrain from getting reported AT ALL COST
- Minimize Fatigue / Tilting by playing only 1-2 games at a time (if ranked)
-----------------------------------------------------------------------------------------
Let me know what you guys think of this. Sorry if the post was too long.
PS:
This was the alt account where I carried out the tests and it shouldn't be used to judge and correlate the skill level.