Diamond was supposed to be top 1%, while Very High is something like top 2.5%. So most of Very High is actually Platinum on the Dotabuff catagories.
And Platinum extends down all the way across High almost into Normal. So Gold is basically the top of Normal and the very bottom of High bracket.
The vast majority of players are in Gold and Silver, which is Normal bracket...and of course Bronze is the real nubs at the bottom of Normal....mostly people who did not play dota 1 and so their 50 games or whatever really is (or was) the full extent of their experience when DBR froze.
Keep in mind that these are set for everyone who was playing 8 months ago when there were 2 million players active per month. Now its 6.5 million. One effect of this has been that "bronze" players have probably almost all moved up at least into the Silver and Gold catagories for real by now...but their DBR is stuck in bronze. I think this leads to artificially high stats for Bronze, which should be reflecting truly nub play.
That is... surprising. Only 2.5% of the player base is extremely secluded, but I guess it is sort of understandable if the player pool is 6.5 million per month now.
You get a better handle on the size of very high if you think about the numbers for people online at a given time.
Right now, its low 150k so very high is less than 4,000 players worldwide...maybe only a few hundred on your favorite server.
If half of them are actually playing...instead of afk like my friends list usually is.... its enough to make 180 games...30 pages on the "Live Games" tab.
At peak times with 450k online very high is more like 12,000 players worldwide...so there might easily be over 1,000 on your favorite server. With half playing...
Enough to make 600 games, 100 pages on the "Live games" tab.
That gives you an idea also of the difference between "page 1" and Very High. Page 1 is the top 1-3% of Very High.
Many people will tell you that they play on page 1. Usually this really means "once I stacked with these great players and we got on page 1". People who actually show up on page 1 regularly have winrates that are at least 55% and often 60%+ on over 1,000 games. I've been on page 1 too...because I stacked with some pro player or some other amazing pubstar.
brackets mean absolutely nothing, your elo is what matters.
just get 20+ kills without dying once and then queue for europe west if you want to play with pros on the front page
this mmr is so fucking predictable since the smurf detection, sometimes i wonder if valve even hire people associated with these kind of tasks. would be so much better if they could just revert mmr before the damn smurf detection or hire somebody with phd in math to do their awful elo ranking
I don't expect that most smurfs will be able to stay there. But apparently with the new anti-smurf thing you really an do that on a new account.
Does anyone know if you can do it on an old account? For instance...your little brother has a low Normal account...you play a couple games and it goes to Very High?
I remember 6 months ago it seemed to take about a 10 game win streak to go from Normal to Very High.
there is a way to end up on the front page with a smurf after your first game, i do it for the lulz to make sure people get real mad about the fact i have one win and i'm playing in such a very high match so they can bitch on valve
#trufax
just saw your edit so here goes mine, you need a new account because the old account is already in the "elo" sort of speaking
Yes, I would expect that to be true...but do we know it is? It's hard to be sure of anything with this matching system unless someone actually does it.
create a smurf and make sure you don't die once with 20+ kills advantage and then queue on europe west
watch the magic happen
about the old account, the reason why it works with a new account is because it's fresh and hasn't been established in the elo ranking that valve is keeping hidden from the public, thats why you're so high in the elo ranking when you don't die and have tons of kills.
Yes... I know this is how it should work (despite the fact that its not an elo ranking really)...The question is does that actually happen or can an old, low MMR account also jump to Very High in 1 or 2 games.
People have in the past reported doing this...taking over a low MMR account and almost immediately reaching Very High. It suggests that they must really be using KDA to adjust the rate of change in MMR.
no it is actually elo ranking, it counts your kills, death and assists.
less deaths = higher elo ranking
before the smurf detection or whatever you'd call it, wins were taken into consideration instead of KDA
valve will probably deny this but that's how awful their mmr is.
No if it counts KDA...that's not elo. The entire point of elo systems is that they are based only on strength of schedule and really they are supposed to be for 1 v 1 games.
And valve claims it is an elo system...but I think that's just because their admin doesn't know what the term means.
the redeeming factors are wins and loses usually but they changed it with kill+loses divided with assists
? what? that must be a typo. "killls+loses divided with assists"...that doesn't even mean anything
That's nonsense. I can only play AM for example achieving that sick KDAs and you say that'll give me my precious mmr/elo? What if i often win despite having bad KDAs?
@relentless
that means everything, you just switch out names for the variables... same shit basically. you don't need a phd in math to figure it out
@small
your winrate is 50% and you have over 2.3k games, if you could read properly then you wouldn't assume such a thing when i already said you need a new account to reach the highest point of the elo ranking with just one game, you're already in the elo ranking and your rank has already been determined sort of speaking.
if you remember before the patch in june or whenever it was in the summer, you'd figure this out by now because of how low win scrubs can end up on the front page with no real effort...
also damn all those treant games, do you solo queue? treant is such a boring hero
Ok, clearly you don't understand or don't care how elo works...and you just made up gibberish and are calling it an equation.
As bad as Valves MMR equations may be...if they were that stupid matching for every game would be totally meaningless. You are just a troll. There is no way you believe what you wrote.
then care to explain how somebody can end up on the front page due to having a high KDA after his first match on a smurf account?
calling someone a troll just proves you've lost the argument and by no means do i take myself seriously after what i've said, it's only a hypothesis of something that could be true which i highly believe it is.
there is technically no other explanation and valve are known liars so they can't really prove it unless they provide the variables and equations for it
No...you are mad because you were caught. Everyone, even people who are bad at math know what you wrote is meaningless.
i'm not mad at all, what i said is a hypothesis of something that could be true and i highly doubt i'm wrong about it.
but if you care to give me your theory of how a player with only one match and high KDA in his latest match can play on the front page with actual pro players... please do tell me as i'm genuinely interested.
That has nothing to do with the regular calculation of MMR scores. If it did any bad player who happened to get fed would jump immediately into very high.
The way a system like this normally works it has a two part score for each player. One part is the MMR, the other part is an uncertainty. When you win or lose the difference in the MMRs between the teams and the uncertainty levels of the each player impact how large the change in MMR is going to be for that game for each player. As you accumulate a history your MMR becomes more certain if you have consistent results (ie you have a win-lose-win-lose-win-lose...pattern). When your results are inconsistent with the prediction of 50% win (ie you have a winning streak or losing streak) the uncertainty for your account gets larger the longer the streak goes on and each successive game can change the MMR by a larger amount.
But Valve did not want to wait for the 5-10 games it would normally take a new smurf account to reach Very High by the ordinary equation. So they stuck in an extra bit where, right at the start your KDA impacts the uncertainty in some absurdly hyperbolic fashion and 1 win seems to be able to add 2k MMR or even more. There is some evidence that KDA may also have a small impact on the uncertainty in the regular equation also....which would be stupid...but I suspect they may have actually done this.
I do not know what code has been written for the MMR. I merely know what sort of change to the equation has the kind of behavior the system displays. Using KDA or other in-game stats rather than sticking strictly to strength of schedule introduces extra uncertainty into the system that keeps match quality persistently "off" the mark....The impact of jumping someone in one game into Very High is a much bigger problem. That will cause gross distortion of the MMR values and also a lot of inflation of scores since so much extra MMR is being artificially added....that extra MMR they give to the smurfs will end up inflating the scores of everyone at the top...the impact to matchmaking is that it gives all the best players even worse feeders on their teams because their score is too high.
interesting reading, it sounds better than my hypothesis and i also suspect valve has done something rad like that, everything is hidden so its a high possibility after all.
chimney
"create a smurf and make sure you don't die once with 20+ kills advantage and then queue on europe west
watch the magic happen"
First step is done . . .
Can i follow u guys on twitter or something? i just dont wanna miss anything from this e-penis fight, its amazing, rly, it, is.
@chimney
So assuming i am intrested playing on very high and right now i am in high bracket, i only need to to finnish a game with 20 and more kills without any deaths? or imy account is already determinded and it cannot be achived?
Valve devs has already explained on their own forums that before approximately 200 games, they use a different system for approximating elo values. Since it tends to stabilize after that many games, it most probably reverts back to changing values purely based on quality of wins.
Which probably implies that for accounts with too little games played to have a proper ranking, a different system is used to push them into their proper skill bracket earlier, possibly taking into account factors like KDA.
well its weird though! Using 2 differnet systems and yet those players from 2 different systems can meet. For example I have less than 40 games and i still up end up playing vs. AA or qojqva which have a ridicilous amount of games compared 2 mine
http://www.twitch.tv/iwin4poland/c/2864374
http://www.twitch.tv/iwin4poland/c/2860665
Either way my first game was already High bracket and the rest was very high. I think they consider the physical address of a computer too otherwise i wouldnt jump to high right on the spot.
http://dotabuff.com/matches/266950608
http://pubstats.me/133212901/
thats bullshit zenoth, you're being tricked by valve devs because they love to talk bullshit without providing fact
i think it has more to do with the fact that you teamed up with a player who has a ridiculous winrate of 64% and another one of your friends who is consistently in high. they probably pulled you up to their tier.
Ever since match making was reset(after ti3) I am finally playing in the bracket I am meant to be in. I am winning much more than I am losing and I have been solo Q pretty much the entire time. I was constantly sitting on dead even win/loss ratio, within the match making being reset I am now 30 wins ahead of my losses. I have a lot of faith in match making now. The games I have lost since MM being reset has been 100% communication/ego issues. Getting a team of bipolar dota players is always good fun.
Heres how I think it goes
Very high +
Very high
Very high -
High +
high
High -
Normal +
normal
normal -
I'm pretty sure Dota 2 uses the True Skill system.
I'm not about to go look up the links to it, but a quick google search should get you all the info you need. There is a significant difference between True Skill and ELO however.
Yes, I'm certain dota 2 has a TSR based system for MMR. But they seem to have added a couple distortions to achieve the anti-smurf effect.
So is it easy for a smurf account to stay in normal bracket permanently/very long time (which is what a smurf would want to do I suppose - to own shitty players) if he feeds crazily the first ten games on purpose?
That might be true. I have not seen an example yet to test it.
However, I think what you would need to do is lose the first few games...not really feed, just lose. Then you would need to not win more than 1 or 2 games in a row to stay in Normal. I believe there are a group of people doing something like this.
They keep being sent into Low priority que as people notice they are throwing...but that does not bother them. They just want to play easy games and don't care about winning or teamwork.
TrueSkill is patented by Microsoft, but I would not be surprised if it was based on TS. It sure is a hell of a lot better than elo
Right, main characteristic that TSR systems have and ELO systems do not is that accelerating change in score with win and loss streaks. Of course its hard to say when all the scores are hidden. But everyone knows that win/loss streaks are what dramatically changes your MMR because they have experienced it. That means its a TSR type system not an ELO type system.
If it were an ELO system smurfs would start with enormous win streaks all the time taking many games...maybe 30 or 40 before they were rated high enough to be in a fair match.
@relentless
I know you enjoy testing things like this, over time I have come to notice certain things in Solo Q. When I was in high tier it was taking me 5:30 - 6:00 on average to find games, As I have progressed into very high tier I find games now in 4:15 - 4:45. When I was playing really well for a certain period of time trying to get into teams nerding dota really hard I found myself finding games as fast 3:30 solo Q very high. Try this out for yourself and see what you find.
Valve has said the system is supposed to prioritize finding games for the highest rated players faster.
Basically what I suspect they have done to accomplish this is that the #1 vs #2 groups are set first. Then, #3 vs #4, then #5 vs #6 etc.
This does find matches for the top ranked players quickly and that works pretty well for speeding up the solo queing...but it can cause problems with the regular matching and team matching.
If they set had it to require a certain range of MMR instead it might take a much longer time to find a match for the top players since a fair match may not exist in the que. But the way they have done it (again I am speculating based on system behavior) they have to pull up someone with a low enough score to match the #1 and #2 teams. This results in your "5th" when you que with 4 often being a relatively weak player.
You see this #1 vs #2 technique most clearly in team matchmaking. I have seen the top rated team game be something like 5.3k vs 3.9k...that game is absurd. There is absolutely no chance for the 3.9k team to win and the game is no change in points when its over.
In fact differences much more than 400 are often zero gain for the top rated team if they win...but a big change if its an upset.
I posted a suggestion in the Development Forum to give us the option to wait longer for a better match. Let us choose to set a range of MMR requirement instead of just playing #1 vs #2. That resulted in the thread being locked within minutes and admins pming me with a threat to ban me if I posted anything related to matching again....apparently they were very emotional at the time from hundreds of crude critical posts about the matching. So recently I tried another way of explaining it. I suggested that they decide some lower limit on Very High high games played and MMR required and it just be a selectable option to que. So far I have not received any response.
Examples of #1 vs #2 queing (top 2 team in the que) remember anything over ~400 points the high ranked team can't win any points.
http://dotabuff.com/matches/234926679 5600 vs 3900
http://dotabuff.com/matches/234953649 5600 vs 4500
http://dotabuff.com/matches/234972161 5600 vs 4500
http://dotabuff.com/matches/226233520 5500 vs 4300
http://dotabuff.com/matches/226271567 5500 vs 4500
http://dotabuff.com/matches/226364193 5500 vs 4100
http://dotabuff.com/matches/265656056 5400 vs 4800
http://dotabuff.com/matches/266575963 5400 vs 3900
http://dotabuff.com/matches/213335438 5400 vs 4200
http://dotabuff.com/matches/213344730 5400 vs 4500
http://dotabuff.com/matches/213356088 5400 vs 4500
The top rated team won all of these games so easily there was really no point in playing the game....but the trouble becomes when the top team in the que is over 1000 points ahead of #2...well who do you put against them? There should be a challenge system instead of an automatic match. The 4500 team should have the chance to know they are challenging a 5500 team and the 5500 team should have the choice to decide to accept a no-points stomp...or wait for a real challenge.
Then when you screwed the #2 team in the que...often #3 vs #4 will be an easy win for #3 team...and then #5 vs #6 easy win for #5...and so on. A ways down the list the matches start getting fair, but they can easily be unfair all the way into the teams around 4k points. My latest team was about 4k points and we played teams ranging from 4.7k down to 3.3k...very unfair games.
A couple months ago I checked through every game played by the top 20 teams on the highscore and only 1 game had been played between 2 of the top teams. All the other games were one sided stomps of teams ranked far below them.
----------------------------
The approach...if you can call it that...of Valve to all these problems has been to say, "more players and it will fix itself".
While it is true that a larger number of players will minimize these errors it does not fix them. Also it has been fixable for well over a year now and 6.5 million players is not enough. In fact 65 million players might really start to get these problem down to a tolerable level of error if they leave everything as it is...but how long will that take? 5 more years if dota 2 population keeps an exponential growrth rate?
Understanding there is no hot-fix to this issue is paramount in this game setup is rather important. While it is true that a lot of their implementation can be improved, it really comes down to the evolution of the game and how the community responds to these situations.
For months now there has been shouts and cries over the matchmaking system, yet the general consensus is that while in it's essence it is still very much a work in progress, there has been a lot of general ideas being thrown about to make this work closer to a more enjoyable game.
Here we have a dota community largely dominated by children and teenagers with the tendency to follow suit with trend, and general populace of the internet community. I remember community gaming back then, probably in the late 1990s being cordial and helpful in their own start-up, allowing game evolution to occur as with the assistance of the community, feedback as such.
Moving on from the prologue, I think this has some ways hinder the progress of the match making and fairness of the game. DoTA seems to still be the game where so many variant factors could change the outcome of the game. The early game phase, to the mid meta phase and towards the end phase, whereby the variables of all the hero skillset, laning phase brings about a an end result which is nary a certainty. Sure there are certain team composition which are geared towards certain phase and if remain unscatched will bring about a win for them.
As with the challenges of providing a fair team game as mentioned above, I think it really comes down to how far the system is willing to go for that to happen. We've got 6.5mill unique players, scattered across multiple regions, playing on different times, all with different skill set, and not all runs on 5 stack. There's really not much potentially could be done unless team games schedules are affixed, like tournaments. The fact that public matchmaking is used, is really a tool provided by Valve and at it's essence provide a game which comes closest to the level of ones team. If real challenge is what team wants, there's where scrim comes in, and private in house game is organized by these players.
A challenge system is good, more so for teams in the 4000 region where there are still a bigger pool of team in that skill level. If using the existing algorithm valve has, teams in the 5000 level cannot find games closest to them, they get pitted forcibly by the system and really, in its essence a waste of time. But in order for the challenge system to be implemented, Valve has to provide a visual indicative of the team rating before the play, and that itself, is another pitfall, a pandora's box if you would call it which they are not willing to open.
Sifting through most of the forums for the dota 2 community, I've read through a lot of complaints, suggestions, on the matchmaking system. Largely condemning it's implementation. I think the general rule of thumb, for most of the people complaining that the supposedly ELO system used is ineffective, they just have to look at the amount of variables having to take into place before a game is considered "fair" by all accounts. Consider that each individual may not play the same role each game, and that their skill set for each game play differs, then you'll have a game potentially where it does not reflect the true game play of the system.
Of course the best course of action is if everyone has their own team, and a definite role in their team to have a true reflection of the ELO implementation, but let's face it, this is supposedly done in a casual environment and on a "CLICK NOW PLAY NOW" basis.
I personally started dota 8 years back, when all I had was this client called "Hamachi", and to really have what we call a fair game, I had to always play in a team. It'd usually be one sided games, until we get to play against another equally skilled team, but these usually required us to play at a certain time, and arranging a game could take hours.
I have no qualms these days with the system as I've taken most games with a grain of salt. Some things I try to defend Valve's implementation, but It'd be good if they have some transparency in their implementation to a selected pool of individuals to further tighten the screws in this game.
http://dotabuff.com/matches/295053668
I just ran out of superlatives on how they could potentially matchmake this game to me couple days back. The dire with a complete 5 stack with at least 1k games, and 60% win rate at the very least.
Other times, they have been rather consistent. Understanding the current pitfalls of the game allows me to take the game with stride and enjoy it as I play it. Though as a developer personally, I am really keen on helping the community with this game as most of us had with IceFrog at the early stages of Dota.
@chimney
u may be right, i just made a smurf and went 41-0-3 with nightstalker.
next game and i was on the 3rd page of US West
"the best course of action is if everyone has their own team"
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This is the key. Integral support for the team matching is lacking. 98% of accounts are not currently on a team. And even fewer of the teams have ratings. People need to be able to find teams and join them much more easily.
Currently it is very difficult to have a group of 5 players of similar skill to go into the que. This is because people can only draw on those they have encountered in the game or real life. So the apparently vast millions becomes effectively less than 20 people...most of which are in a different game, or offline, or asleep etc.
I realize that there are legitimate concerns with allowing people to know their ratings. But not knowing them is in some ways even worse...because it makes impossible the kind of solutions that are really needed to make the gaming experience team based and excellent at all levels.
It needs to be possible, ultimately for people to go to a large pool of players of similar skill and put together a team easily. Many of the problems with matching come because of the half-freedom for people to stack up with anyone. That freedom would be good if it were real...but it is artificial. People do not really have the freedom to make their teams because they do not have the information to do it.
So there are two things missing.
[1] the ability to go to an abundant pool of rated players and say "I want 4 people with rating between X and X"
and also
[2] the incentive to play in that team environment...the connection to the rest of the chain from nubs to pros. Team dota will massively trump all individual play if it is allowed to occur. No one will be playing for individual stats because that WILL LOSE games.
------------------------------------------------------------------------
As I explained in my post on the dev forum. The reason there are not enough players for the top is that they are quickly stuffed into lower games instead of accumulating until an appropriate match exists. If the option to wait longer until a match that is really on equal MMR shows up...people will wait for it. In ThronIT tier 3 there were only 3 games a day and the top players waited hours to play them.
They are willing to wait because it is worth doing for the higher quality game. Everyone at the 4k level would be very happy to wait 10 min for a good game.
If you do let the Very High players accumulate this will also improve the matching for the top of High because the maximum difference in skill level of the players will be much smaller.
And of course the current smurf system is overkill...I think everyone knows that. Just send smurfs to High in 1 game, not Very High. They can work through a few High bracket games to get up to Very High at the ordinary rate without causing problems. If they deserve to be in Very High they will be soon enough. In the meantime. Adding +1000 to MMR adds far less uncertainty and instability to the system than adding +2000.
---------------------------------------------------------
So many things have been improved already over the dota 1 system, but there is still a lot of work left to do to get it right.
Pandora's box.
Whether it'd be apt to metaphorically give definition to publicly release rating to pandora's box, would be hypothetically at this stage. One can only speculate.
At some point you'll have to give the developers some benefit of doubt here. Ratings are not revolutionary. Rankings and points has been incepted by a lot of sports, organizations, e-games et cetera. I think the concern here pertains to the community, and to the majority playing crowd as a whole.
A lot of the dilemmas and pitfalls can be too verbose to be listed down, but as I have not seen anyone in developers forum or community forum attempt to address these concerns in relative to the community, bear with me here.
Again, this is put as a context to a NON-ideal world. In an ideal world, we'd have a more supportive community embracing the game positively and Dota2 developers allocating time to communicate with the community.
--------------------------------------------------------------------------------------------------------------------------------
I was lucky enough to be hired by Yahoo Games back in the late 90's to spearhead their short game system. At that time, it was a very fresh idea on the internet. Started off with board games, Chess, Checkers (Draughts for you English lads), Go and Reversi. Simple concept, create a lobby, two players play with any individual connected to the lobby. Casual gaming, providing interface for players to engage in games online.
One of the team member then had an idea to implement a rating system. We explored using the ELO ranking system with modified Bayesian based. These days Glicko's rating system is a closer equivalent to what we had back then. All players start with a base rating of 1200. The rating system was implemented and we exponent on it by classifying them by different groups of skill set.
~ - 1199 players will have green colour badge to their ID, and is the base rank
1200- 1499 - Second tiers with light blue I believe.
1500 - 1799 - Third tiers, again, if memory served me right was dark green
1800-2199 - Fourth tier, Orange
2200 above - Red
We create multiple lobbies, tagged them as Casual, Beginner, Intermediate, Advanced, with the colors indicatively providing users a rough gauge on which room to join. We had one of our programmers play through the beta release in a closed environment for couple weeks, as he improved his game play, he moves up and plays with players of the higher level.
As this was released to the public, we noticed that the gauge of games were pretty accurate for the players skill set. Playing against better rating players provided a capped increment based on the differences, and they will always be around their skill set even if they are on a losing streak. The variance were low for these games. Form is calculated but their skill level will rarely diminish.
*Do bear in mind this is still a single form of specialization of the game, in Dota, we are dealing with multiple different heroes, with different roles in team.
It all went good for the first year until we noticed a few trends cropping up. They were what we called boosters. Creating multiple accounts, creating accounts to achieve high ratings. This distorted the entire gaming concept of playing the game for what it is, and basically have high rating accounts because it simple was aesthetically better. We got that influx these days in the game called Smurfs.
Cutting to the end, the community of these games started spawning tools to cheat, hack and crack into the system because it simply gotten "too competitive" to put it mildly. The indicative of rating was too substantial that we had majority of people playing it for all the wrong reasons. These days, the game system has evolved greatly that it's masked, but the amount effort put into it was enormous to make it work correctly.
--------------------------------------------------------------------------------------------------------------------------------
In saying that, I am not against displaying rating. But as a project co-ordinator, I will be greatly concerned with the social impact of it. They'll be a portion of it which will respect it. But as we have seen during Garena, this community is just not mature enough at this stage to be provided with this feature.
I am sure there will be some of you that remembered Garena implementing a ladder/win-loss system during one phase, and it was magically removed. Part of it was because people were spending more time looking for games to casually play because the community got too engrossed in "numbers". Protecting that streak at all cost, or the percentage.
It is one of the reason why Valve, provided an option for players to display their match history but by default hides it. They must have had an influx of cases during load-outs players checking for their teammates, and if they reckon that it's going to be a bad game, disconnect.
This community, as it stands, for every 1 decent player, there will be 10 irrational ones (or maybe more)
For that exact reason we are seeing a very conservative approach taken by Valve to make these radical changes (or displaying rating). It may happen, but I wouldn't count on it happening anytime soon.
--------------------------------------------------------------------------------------------------------------------------------
I agree whole heartedly on the wait times. Challenge system is a great addition. Again, limitations.
Same goes to the concede (give up) function. Great addition, but again, community, limitations. (seen it happen in competitive/team games which is great)
--------------------------------------------------------------------------------------------------------------------------------
Smurf system is terrible at this pace. Necessary, but too radical.
--------------------------------------------------------------------------------------------------------------------------------
They have provided a dev forum, but let's face it. It's rubbish. Rubbish because Valve will never be able to filter appropriate suggestions to pure, obnoxious threads. This really hinders growth, and it's really unfortunate that this has happened because what made this game what it is, was the input of the community.
Zaloguj się aby dodawać komentarze.
Normal.
High.
Very high.
I'm sort of confused over the distribution of players into these brackets. Back when DBR was still active, the bell curve showed the distribution of the five ranks. Diamond was top what, like 3-5%? Then Platinum was 15%-ish. I'm not sure, I'm just stating the numbers off my head.
So with only three brackets to classify players, you can have a very wide range of skill in each bracket. So I'm asking more about the Very High bracket, I get that some people mention the low-tier Very High and upper-tier Very High, but can anyone elaborate more on this?
I assume upper-tier would mean a person who would be considered a Diamond or among the top in Platinum if DBR was used, how low is the lower-tier?