What happened to that "skill based" mmr?

Consistently wrong and easily gamed. It was a fiasco when they introduced it.

It was because they have useless interns working on it, it wasn’t manipulated (it wasn’t around enough to even figure out what exactly it did) it was just plain broken. That just shows the lack of skill/care of someone who developed it. Blizzard has data on millions of games, you could backtest the hell out of it… And as far as “easily gamed” it’s just a generic statement with no backing.

If you ask 1000 random players, sure. But if you ask 1,000 grand masters their answers would not be so varied.

Of course it would, there are all sorts of bias we all hold. I bet healer mains would be looking for those malf roots that “changed the game” and tank mains for the setups, assassins for kills, etc. Look at fantasy brackets from sports analytics and pros. There is often some agreement on top 3ish, but then it’s generally all over the place. In addition, since having 1000 gms vote on every game isn’t feasible in implementation then why bring it up at all. The question is between current system (winrate based) and the pbmmr, which one has a better potential?

Actually you are just another useless person with an uninformed opinion and think it is right merely because you thought it.

Dime a dozen personality flaw.

If you think it can work, why dont you explain exactly how do write the algorithm.

Oh let me guess, you don’t know how to even begin, you just magically know that you could do it.

lol

The current one. It is just slow, but it is more accurate. PBMMR has been running in the back ground for a year and there is a reason why they decided not to implement it.

They still could not get it to work even with more data.

The game is too complex for a computer to figure out who was better and who was worse without creating a “game the system” scenario.

The current one. It is just slow, but it is more accurate. PBMMR has been running in the back ground for a year and there is a reason why they decided not to implement it.

Lol, proof please? You can’t run a system in the “background” it’s either one or the other.

If you think it can work, why dont you explain exactly how do write the algorithm.

Couple of reasons:

  • I generally charge for this type of services, if someone is willing to pay, I’d be happy to design one
  • It is based on empirical data, which blizzard doesn’t make public because people would realize how incompetent their MM is. Basically, same reason why they stopped showing number of players online - makes them look bad.

They implemented pbmmr right at the same time that the battleground update and the stealth rework.

It was a worst timing ever, since the set of data they were using to calculate pbmmr was already obsolete.

Valeera mains were very happy, I don’t think it was possible to get a negative pbmmr playing her, unless you were really bad. Probably some more heroes had the same problem to a lesser extend.

Bugs and the following reddit reactions made Blizzard shelve the project.

There’s no such thing as skill based MMR in a game like this?
How in the world is the game going to determine who is more skillful than the others??

Win/loss. Same as Elo type of systems. Except when you apply it to a team game … you get what you get.

Are you trolling or are you this unaware of what you are even trying to discuss?

They have had the data metrics active for a year on PBMMR and the results led them to believe it was not fixable. They have literally stated this.

Im not digging up the source for you. You can look like the idiot idc.

just one reason.

you are bad at the game and possibly an idiot.

If those two were not true you would realize that PBMMR just isnt going to work.

Although if there is going to be an individual rating I would prefer it not based simply on winning/losing since one player can usually not stop a loss if their teammates are throwing or just playing too horribly.

But I think the larger problem is even trying to rate players individually in what is a fundamentally team-based game. As well-discussed on this thread and elsewhere, it is incredibly complicated to try to put a number on just how good a player is when there are so many factors involved as well as so many different opinions on what is really important.

I think the only thing that you can reasonably do for assessment is to judge teams (actual teams of players that always play together, not just randos thrown together) by their wins/losses and that of their opponents like is done for teams in other sports.

Love when clueless people tell me that things I do on a regular basis won’t work, I guess that’s why I can make a decent living on that :smiley:

I think the only thing that you can reasonably do for assessment is to judge teams (actual teams of players that always play together, not just randos thrown together) by their wins/losses and that of their opponents like is done for teams in other sports.

Pretty sure they’ve been doing it since release (maybe even in beta, can’t remember now). They have a separate MMR for groups.

Ok so you are an idiot. Got it.

You can not develop PBMMR systems for team based mobas on a regular basis because this is literally the only team based moba.

The fact you are clueless as to why developing one for this game would be so hard it is next to impossible makes me doubt you even actually do anything related to it.

But go ahead and keep saying they couldn’t do it because they “were interns” without an idea who they had on the team trying to develop it.

You are just another clueless idiot in a sea of clueless idiots.

Good day.

1 Like

If you want we make out Blizzard devs to start work on my ideas…please support my best ideas…
I either think about personal ranked points system based of these 20 achievements…

We still wait you start work on personal ranked system.
My guess is 2020 year would be to finaly add.

Bullsh*t…exist such system wich can work so well and reward players wich are really high game skilled.
I either play different genre games!

What about such game version?!..

Read all theae over 20 ideas…and think about them,a tleast 1 week! And then say what you think!
Do you know why peoples must think so long time until start think or rate such ideas with his own gamer xp and dependence from his own high intellect?!
Because our human race always think with worng way and do mistakes and because our human intellect or brains,had false illusione do not think with worng way!.

It didnt’ fit in with overall MO to give really bad players with no intelligence … a good time and a carry so they buy gems/skins/mounts. Doesn’t fit in with the overall mind games Blizzard likes to give players so they feel good about themselves even if they are the bane of everyone else’s existence… E. g making them think rooting 12x and dying 40x is deserving of some sort of metal /recognition at the end of the match.

Johnny-Feed-32 times wouldn’t have anyone to play with if they actually implemented such a system that recognizes a players general IQ, motor skills and grasp of the game.

The often found rainbow matches existed for this reason.

Alright guys, get your thinking caps on. I’m coming at you with E=mc style. What I would’ve done is try to figure out the factors and the variables and turn it into a consistent formula based for rewarding MMR. Yes, delicious MMR. The good thing about the stats indicate where you are and what you’ve been trying to do a whole lot of time during the game. Because shared assists is shown, it indicates that you were there in a teamfights, and not always trying to stat pad. Also personal MMR shouldn’t be huge to the point that it impacts your MMR alone, but rather a notable incentive over time. They need to make MMR and personal MMR separate, but will be combined in total in the end match.

MMR points in this post are used for scale only:

For example:

FACTORS. There should be 2 TYPES of factors for your personal MMR. Teamfights and Siege laning:

kills:
This is usually shown in synergy with assists BUT the higher the kills away from assists, the more bonus of MMR points is rewarded, as it is harder to get kills without assists in the first place, or if not, you were always that remaining factor in the teamfights.

assists:
This indicates the very reason if you were helping the team or not. So this clears your evidence of just solo or laning all the time.

deaths: Deaths detract points from your personal MMR.


VARIABLES: Whatever their hero roles are, then that’s what their personal points attributes too. Hero damage, healing, taken damage, sustain. Take in count, even if a healer gets the highest damage, he’ll get a reduction for that MMR variable as his personal role was healing in the first place and will get those points. Same thing applies for tank while the Assasins get the obvious dps. Otherwise healers and tanks will just get a major boost in everything.

So for the variable bonus, for every 10’000 damage an assassin might get, gets a xx amount of bonus added to his personal factor. Pure example is, he might get 5 personal mmr points for his factors, but his variables boosts it by xx.

FOR EXAMPLE:

Exact same hero, same situation and same time.
Let’s pretend the variable hero damage is 10’000 = 10 MMR points.

Valla: 18 kills + 80’000 damage = 80 points

Valla: 18 kills + 40’000 = 40 points

Then those variable bonuses get’s added into the factors. Keep in mind, those numbers are just examples. Those could be 2-3 points per variable gain.
Also The second Valla was obviously trying to jip the system and kill steal, but the systems got it figured out.


The second factor: If you’re a support type, then they need to be fine tuned for what they are. Otherwise a hero like TLV will get nothing, while Abathur will get everything lol. With the removal of most of the lane specialists, there’s usually not a whole lot of reason to push alone most of the time.

So if there’s not a whole lot of kills or kills assists shown on the pad, then the second dominant factor kicks in. Which will reward lane specialists or players that soaked hard. They might not turn up in tf but they can rewarded for pushing with siege, but the second factor should be capped so players wont go the easy route and stat pad and flick minions for stat pad gains. Especially the fact now that most specialists have been… neutered.


If not, why not just do some bubble coding.

The common factors are forts, objs and MVP. It’s obvious when you can see forts are taken down when a flag comes up with the players name, so they get MMR bonus for it. If you cap the obj, again you see the players name and they get that bonus for it. So it’s a win/win situation. You can push forts or you can turn up to the obj and still get personal gain. If you get MVP, then that’s a boosted personal MMR. Don’t allow voting to adjust MMR as people won’t vote anymore :stuck_out_tongue:

1 Like

Apparently it wasn’t mean-adjusted. So in high rank games where people mostly scouted and soaked the whole time, and only got into team fights when there was advantage to it, it gave everyone a low performance rating. But instead of offsetting the performance adjustments such that the average adjustment of every game is the average MMR difference of the two teams, they just turned off the whole system… weird, but w/e.

1 Like

I can see this topic getting in the spotlight somewhat once again, which is good but here’s the problem

IMO, the very way they approached the solution was kinda too much of a “universal dreamwork” at best… It was too much of an overstretch and would’ve consumed (or at least require invested) immense amount of time and money (or at least that’s what I deem to be true)

What they wanted to implement is/was a one and only FULLY CENTRALIZED quality-of-play evaluation system where gameplay of your current game would've been compared to what the best-in-the-business-ATM people on the ladder do, and practically every gameplay, every game would get a grade (% sucess or otherwise) compared to the current/best players on the leaderboard

Why I dislike and why do I think was a bad idea to begin with ?

1 = the game is too complex in order for it to justify doing such a thing like that, like, put one different player with different preferance/playstyle and you might get a completely different per-gameplay success ratio overall
2 = the mechanic implemented that way would furthermore increase toxicity as it would force newcomers (and overall players in general) to “follow the rules”, i.e. = pick IDK Falstad, and then get compared to Mene’s Falstad and get success-measured by amount of things you’d do the same (or whatever)

SURE, fine, let it be that way in Master league or GM, but already not gonna work for lower ranks as there are different “rules” for each level, it’s not just “mimic X” and you’ll succeed like “X”… Certainly doesn’t resemble real life IMO, if that was the case we’d all be self-expert surgeon physicians and do operation interventions to ourselves on our own :P…

Yes, the current system of PRA is even worse = adjusting MMR exclusively on WR%, but the “leap of faith” that should’ve been done otherwise in order to implement that-way imagined PBMMR system was collosal in every shape or form tbh

IF ANYTHING I think the problem isn't the MMR-evaluation system (although it is), IMO the biggest problem is the fact that the Matchmaker is affected (and therefore fully dependent) by the MMR system itself

IF ANYTHING = would've liked to see (before the game died somewhat) an attempt to "revolutionize" the Ranked gameplay without changing such a fundamental (for other gameplay types ?) and attempting at creation of the "biggest leviathan of leviathans centralized systems ever" approach

Here’s what I think should’ve been attempted at a point when the game’s “peak” passed and the course of going down was already established/inevitable

Make ranked different, more real-life team-sports resembling whilest not necessarily making it a team-only thing and HERE’s HOW:

Split the whole gamebase in divisions, 200 (or 400, whatever suits better) players per division, one season lasts 1 month (for start, might wanna adjust later on), at the end of the season top 15% go up a division, bottom 15% go down a division

THAT is what the Ranked system should’ve been like, every rank resemble like a mini Master/GM “leaderboard race” tournament of it’s own

WHY would’ve this approach performed better ? (or at least hope would’ve done so), well, mainly 2 reasons:

1 = No more MMR-based bull, you can literally climb a notch higher than you’ve ever been and you have like a month in order to adapt (instead of just 2 games after a bunch exhausting to-the-end “super extra effort” of subtle winning series), it’s like = noone wants to play 50 games (about 10 of which will resemble a total smackdown and forced loss after failure to stay higher in the first 2)

SO THAT is one problem = the “skepticism” of MMR of letting people try-out games higher

2 = HIGH TRANSPARENCY (or at least the ability to do so) = when having a thing/mechanic like leaderboard/s, you can have a surplus of “extra information” at grasp, like, every player within a division would have had super easy data to access to, therefore when paired with a “Murky main” you’d sort of already know in advance if that’s gonna happen to you

Having each of these pieces of data attached to a player’s name within a division would help IMMENSELY for everyone

Leaderboard data (example)
1 = (was already kinda implemented but would go further) = Primary and Secondary roles.. Should probably even force people pick from those 2 roles alone during a season, players pick/adjust these at the start of each season (simply prevents "unpleasant surprises" and dumb losses happen, or at least draft-wise it should help so)

2 = Winning/Losing streaks, along with W/L of last X games (somewhere between 5 and 10 would do).. Just enough so you know the "momentum" of your teammates (and/or opponents), further more can "spice it up" with a bit of extra mechanics such as streak achievements/rewards

3 = Player's best (highest WR% * games played)  hero.. Perhaps best hero not just lifetime but also last season (and even during the last couple of weeks within a season, best hero per player for current season)

4 = Best/Worst map WR% per player (both lifetime, last, and in 2nd half during season current season), and so on

Ofcourse = this would require high/er activity of gameplay from basically everyone involved, but, BUT if there were indeed well-implemented achievement/rewarding system = might’ve worked

Some extra hypotheses I based that opinion upon
Overall = had a 25 page word document talking about these things before but it comes down to "secularisation is a must" (at least in my view) of MMR and Ranked gameplay

The game's simply not designed in a way where MMR-based matchmaker would've worked (maybe in high leagues, like Master/GM, but in lower leagues often the winner is determined by the team that does less mistakes overall).. And again - instead of forcing everyone else do less mistakes even further or die trying (by ab/using a "X player trainer" "invisible bot" that you might "catch a grip of how it works" if you watch what those do somewhat), there should be more/higher transparency and more of a "family membership" mechanic/system like the "leaderboard" mentioned above

SURE = it may be boring to get stuck more than a few seasons in the same spot but again, if they made it "interesting enough" (and they could've) for those "native members" to feel a new season like something new, would've worked

It would've had MUCH LESS toxicity, cause players would've had higher control of the situation (or at least those that would be aware of what's going on), HIGHER TRANSPARENCY (therefore making matching teams much less of a shock) and lastly but not the least = would've provided really nice chance, not just awarding system but CHANCE to "practise higher" for a good amount of time instead of having a "universal-evaulation" system decide it having another try was a thing for you or not

No ofense, but after all these 4+ (almost 5 ?) years of literally almost nothing but struggle if you defend the current/MMR-system, then something’s really wrong…

CONCLUSION:

(IMO) the only, THE ONLY really revolutionizing thing they could've done/implemented/tried-out IMO was to forego the MMR system completely and try make a system that would've worked without.. After all these years of "defending" it and having it produce worse and worse by the time goes on, about time to get rid of it and "start all over"

Ofcourse, might be a bit too little too late (considering/provided-that the playerbase is relatively small for that particular "division-based" system to feel "alive" overall), but point being:

Think that dividing the playerbase into divisions, each having a number of 200 (400 if/when low activity) players would've (much better and fairer) worked, if anything the toxicity would've/should've been lower by having a more sustainable/predictable matchmaker instead of "you've been trhown into fire, survive" (i.e. survival of the fitting) kind of approach

Yeah…very good try to help and sayed.
Yeah…thats all these factors wich are so important,how must be this PBMM system.

1 Like

Yeah…man…i really like you and what you sayed and explain very well ,what other different ranked system need.
I hope Blizzard devs read your reply or re read again whole this topic.
But i doubt a little they try different type ranked system.
Lets see what will happen in future.
I either did try out to help with better ideas to now.
Btw i stoped play(reasons is this half trash player base,my job) and now i return to best MMO RPG " RIFT". Honestly i log in so rare times.

2 Likes