OP feels entitled to effortless easy runs at whatever xp level he is playing at. Sorry, with humans involved, that will never consistently happen. End of story.
You want consistency and control, don’t pug keys. It’s always been that simple.
2 Likes
Or just make a better game.
@Everyone in this thread. Guys, it is X-mas, take a break from these forums, enjoy the game, or spend some time with your family. 
1 Like
In other words, OP wants to log-in and hit play game, play with like minded and skilled players.
It’s honestly such a terrible idea, it’s far superior to hunt discords, troll forums, lounge around on the internet to make “friends” to get the chance to play the game the player pays for…
Clearly didn’t read the topic of discussion. Topic of discussion is that there is a severe disconnect in game design and community.
Stop defending broken systems because it doesn’t apply to you.
Right, you used “correct” in quotes. You absolutely can quantity “correct” play, there are tons of binary metrics that assist in doing so. If you are unable to come to those conclusions, then that is a critical thinking error on your part. I can’t help you there.
And that’s what is wrong with the average player. They think my DPS OMG, or MY HPS UR DOG PLAYER. It’s quite easy to understand these players do not grasp the game outside of their small mindset. HPS, DPS, interrupts, all of these metrics alone do not dictate how well you play/played/or will play, but combined with multiple runs of data and something we call a “trend” shows quite a bit about how skilled an individual is.
Sure. But I’m just going to use the original comment you made to me for… reasons that will most likely become obvious.
For the exact same reason “Gearscore” wasn’t a good measure. As has been pointed out an endless number of times in this thread, the types of PvE encounters in WoW that differentiate a good player versus a bad one is usually found in the decisions we don’t make, rather than the ones we do.
For that we require an AI that would be capable of interpreting negative data and separate that from positive data. Or another way to phrase it, you need a measure wherein a wrongful decision can be calculated and measured to be a good one. For this, we could use a chess engine as an example.
In every single match of Chess there’s ALWAYS the ‘best’ move one can make. A machine can calculate it and learn that. However, due to the complexities of chess we can only ever get near to that point, ergo why all decisions has to be weighted. But that means that there are categorically wrong moves and decisions one can make.
How do we classify this in terms of M+? The activity as a whole includes grouping up, selecting players, and picking talents. And this is before we already get into the dungeon. So we have a baseline already that consists of (A) comparable performances (Warcraft logs essentially), (B) performance variants ((B.1) talents, (B.2) gear, (B.3) types of gear, (B.4) ilvl, (B.5) stat weights, (B.6) consumables (which we will get back to again later), (B.7) enchantments, (B.8) enchantment variants ((B.8.1) types, (B.8.2) ranks, (B.8.3) stat weights), (B.9) gems ((B.9.1) types, (B.9.2) ranks, (B.9.3) stat weights) (and this is separate from Warcraft logs because this is real-time simulated calculations comparing against Warcraft logs), (C) group composition (for this one I’m not going to go through it because do everything I just did above but now extrapolate overlaps, combinations of spells, CC, niche utilities, and missing utilities (since if the group is missing something, that is a negative that should be applied to the leader before even starting the M+ in order to be realistic).
And sure. All of this is stuff that AIs can do. After all we can do something like this already between simulated logs, Warcraft logs, raidbots, and so much more. The biggest problem is twofold though:
- This is required for your Mythic+ Social Credit Score (btw, look up what ‘Social Credit Score’ is and you might realize why I used that phrase rather than ELO, because we haven’t even started to extrapolate these things into MMR-like numbers yet, all of this is in preparation for being able to do that) and all of this happens before you have even put the key into the dungeon.
- All of this as much as an AI can do this … people can’t. And this is responsibilities that people would have in order to before even getting started with the activity itself. So the only way to make this work is to also give players the tools required to thoroughly inspect and demand others to play in specific ways in order to even start making the group. After all… you don’t want to lose your Mythic+ Social Credit Score as the group leader, now do you?
And just to make sure this doesn’t drive me utterly insane, I’m going to simplify the hell out of how it would work mid-dungeon!
- Real-time performance check. You do poorly? “MINUS 50 DKP!!!”
- Missed an ability but someone else hit it? “MINUS 50 DKP!!!”
- You didn’t have the ability to deal with an ability that the AI decided was your responsibility to deal with regardless of group composition or whatever happened during the key? “MINUS 50 DKP!!!”
- Someone accidentally overpulled? “MINUS 50 DKP!!!”
- Adapted during the key because someone else overpulled? Great! Now you need to do all the calculations again. So it is more sensible to assume that an AI would read “deviation from the most logical pathing” as something neutral or negative so… “MINUS 50 DKP!!!” (But someone would probably get like +1 in a score somewhere from this, so realistically just “MINUS 49 DKP!!!”)
- Someone used a combat ress when the AI didn’t like it? “MINUS 50 DKP!!!”
- Someone was slow on doing what the AI wanted them to do? “MINUS 49 DKP!!!”
- You killed a boss at a good time? Good job! No minus DKPs!
8.1. You killed it at a bad time? “MINUS 50 DKP!!!” - You timed the key? YAY! You did it! No negatives for you!
9.1. You timed it but not as fast as the AI wanted you to do it? “MINUS 50 DKP!!!”
Of course I’m being facetious but this is how AIs operate. You either operate on a positive reinforcement which is useful for learning, or you operate on a structured path which means the AI is mostly looking at when things go wrong. But “wrong” is an entirely subjective definition so it would require the AI to be able to interpret it … which is where we go back to the beginning of what I said in regards to this one singular point of yours.
And all of this is just in order to get the baseline … we aren’t even dealing with real fringe cases yet. And I’m not going to because those are quite literally infinitely many, per definition of both how M+ function and because of human error and interpretation.
So that’s the first point. Thankfully this is only the big one I will cover because the other two you mentioned are based on quite easy sociology:
M+ as an infinitely scaling difficulty is a competition yes, but it isn’t the correct type of competition for these types of systems. As I demonstrated above one can realistically create an AI that can provide a random number to people. So… why doesn’t it matter?
Because progress is measured in a few different ways:
- Key completion (with the grading being passed, well passed, and very well passed, depending on +1, +2, or +3). But this already falls short when dealing with high keys because whilst they can be completed, they can’t mathematically be completed within a +2 or +3 the higher and higher you go. So your baseline-progress basically falls back to the simplest form of “Did you time the key or not”. And for that… no MMR is necessary. The key itself showcases your completed MMR.
- Gear. As we progress higher and higher our gear will naturally increase up until it physically cannot increase any higher. When we have reached that point that is when we can pressure ourselves into reaching those astonishingly silly numbers that people try to get at the end of a season. This is another form or progress where our progress pacing is wholly determined by what our gear is. Which again … means no MMR is required.
These are the two I mention (but one can make endless arguments for more) because I focus on the word “progress” here. Progress in a PvP setting and PvE are two VASTLY different things because … there is no clearcut definition of “winning” in PvE. Have you “won” at PvE if you killed normal Fyrakk? A heroic raider would say no, whilst a Normal one would say yes, and someone who only does LFR would say “It is above what I am able to do”. So which is it? Do you win at PvE when YOU succeed at something or not, or do you fail because someone else can do something that you can’t or won’t do?
So we can only focus on progress here. Which, to go back to a PvP setting where we have MMR … you win by winning. That’s why the final goal of every match is to ‘win’, so that isn’t progress it is just what the overall goal is. Progressing in PvP means getting a higher MMR to face against more difficult foes and to win against them.
Who are you trying to win against … in a collaborative effort? In PvE there are no winners, only participants. We have personal goals that determine when we have won but in PvE … there are no winners. There’s plenty of losers but hey, losing a PvP match is what just happens and we do the same thing in PvE so losing isn’t a problem. The problem is that we quite literally cannot define “winning” in PvE without progress, and progress in PvE isn’t determined by winning.
Winning in PvE requires progress of some kind. It can be the ‘final’ progress, and then a personal goal to progress towards after that. But progress in PvE isn’t defined by winning; it only goes one way.
M+ is a competition against ourselves and the environment. It isn’t a competition against other people. Only in tournaments like the MDI is that a thing and we aren’t talking about the MDI. We are talking about the base system itself. So no, M+ is a competition … but not that type of competition that requires numerical progress. It requires practical progress but not a numerical one. Ergo, MMR is not needed. Heck not even our M+ score is a MMR system but just a number that we use personally (we created it ourselves via RaiderIo so yes, it is self-ascribed - not a MMR as you cannot lose your progress) to keep track of our “general” progress across multiple dungeons and affixes.
And ultimately … the reason why I brought up this to begin with…
Because all of this is solved by doing M+ with friends. Guildmates. Random folks who you said “hey you were good, mind if I add you so I can poke you for any future runs?”. Ultimately … all of this boils down to that simple fact.
None of your ideas are sensible on their own, and the few places where one can safely say you are correct it is a real problem … they are immediately solved by being social and talking with people. If not making outright friends but at least acquaintances with like-minded goals. Joining a Battle.net community dedicated for high-end M+. Creating a community tab for people with those goals.
So … yah’. I ain’t trolling when I said that. Because it genuinely just reads as “I want to turn M+ into a MOBA-styled activity … because I don’t want to socialize with people, and would rather have every single negative that comes with the mentalities that stem from MOBA-mechanics”. So… yah’, I still stand by what I said - it genuinely just sounds like you suck at socializing but really like MOBAs and M+, so you want to combine it into your dream version … be damned of however much it would completely ruin M+.
TL;DR … yah’ there isn’t one for this post. OP wanted a complete deconstruction as to what they have been talking about so… yah’. No TL;DR. Either read it in its entirety, or don’t. The thing is long enough already as it is.
Quick edit: Since I realized I didn’t put any links or the sort as to what I base all of this on… videos, articles, and more regarding game design over the years. There’s not really any one source for any of these things, but I can recommend the book Superintelligence and learning some more as to how AI work both functionally, in-games, and some philosophy behind this. Because in reality this is a conversation about the philosophy behind game design more so than a conversation about AI, Algorithms, and MMR.
2 Likes
We have multiple in game systems to help facilitate that.
Well. Its only a terrible idea because of the severe limitations in having an algorithm decide if you played well in a singular run or not.
First, thank you for making a detailed post, and I greatly value your feedback.
Correct, gearscore is the exactly same thing we have, mythic plus score, or even itemlevel. The SAME problem exists as it did in 2009, A NUMBER THAT DOES NOT REFLECT PLAYER SKILL.
Please provide some details, as I don’t know what you’re trying to say here.
We don’t. You’ve overly complicated this whole discussion.
I am sick and tired of non-technical people using terms they don’t understand, such as AI, or machine learning, or neural networks, in context to what my discussion is about. Please stop.
These things you pointed out are borderline sarcastic but I do appreciate it. Why? Because all of those things are not factors, nor would be factors to the hypothetical vision you have in your mind.
Ok, and what is the “correct type” of competition for these systems?
There is, you just defined them. Someone “winning” at doing LFR has an ELO score of LFR, and so-on. Your SKILL goes up according to the content you do.
Yes, it entirely is. You can’t just state something as fact, because you think so. There is no sane person here who thinks “winning” is attempting a boss 10000x and never killing it for the rest of your life. Please stop trying to play semantics around the word “winning,” it’s very clear in PVE as is in PVP.
Don’t confuse the word “winning” to a personal goal, there is a mixture here. Not a single person doing a single key is going to consider it “winning” to brick 100 keys. Winning IS clearly defined in a PVE setting.
What? Again, just because YOU say so doesn’t make it so. There is a whole leaderboard, io scores, small tournaments, a score within the game that shows how good you are, achievements that give you rewards for hitting a certain number, etc.
None of what you’re writing about progress makes any sense when relation to M+ system. They give you keys, which are numbers, that are numerical. You need a certain HPS, DPS, which are numbers, to time said key. You also get a numerical SCORE that gives you points for how high of a key you can do…
Again, this has been discussed so much that I am begging to think that the people on the forums are incapable of seeing a different respective at all.
The entire post is just about that, a “simple fact” that in THEORY it is very easy to write “go make friends, go on discords, go join guilds, go be social” without addressing the real nature of how time consuming, inconsistent, or even not possible that is for many people who play this game. Especially new players.
I wrote this post because that is the first thing I personally have tried. My friends list has 300+ people, great players, good people, but due to poor game design, I’ve maybe played with them 1-2 times?
There is no team nor group continuity, there are no rewards for playing together with a group, team or community. After KSM, the reward stops, and the only “reward” left is that magical numerical number you talked about.
Again, this is terrible game design. I’d rather spend 10 hrs a week playing this game than 10 hrs a week looking for people to play with.
Again, moot point. It’s not how you imagine it to be in your head, because I can bet that you, along with all of the other people touting the “just make friends” either already have a group from years and years, or just don’t participate in the content or level of content I am in.
This is one thing I do agree with you on. Blizzard created by definition, a MOBA-styledactivity, without all of the supporting infrastructure. Then you have people defending them like it’s a great system.
Interesting, because not only did I never say that, but my solution would allow solo gamers the ability to interact with people by design, rather than spend 80% of their time looking for people to play with.
I am greatly disappointed in those who lack the fundamental understanding of what “AI” is, and how algorithms (you know IF STATEMENTS) existed before marketing got a hold of buzz words.
Agreed. I want the game that is designed around giving players what they paid for, not logging on to chat with people for an hour, then searching off game to find people to play with.
Of which I pointed out the critical failures and their uselessness. Also, how the game systems do not encourage nor reward the players for using them.
That would be a terrible idea if it was based on a singular run or not. You are absolutely correct. If you read my post, it’s clear that it’s not based on an outlier, one time performance, nor one particular “metric”, rather it’s based on a longitudinal series of metrics and performance over time.
This is all forms of non queued content, not just m+. If it was queued, you’d have a point, but as is, make friends, find a guild, or deal with pugs.
There is no reward for grouping up to kill Mythic Fyrakk, and I wonder how many people pugged it.
Infinitely easier to find a mythic raiding guild than an m+ team.
Sure. I’ll do so with questions though to address my concerns. I’m genuinely curious how you would make this work.
What’s to stop players from abusing this? Scenario: Player is bullied the entire run by 4/5 the group. Then they can rate the others low, but will also get rated low 4x the amount.
How do you envision your rating system being implemented, especially to factor in abusive rating behavior? What if you don’t want to rate people and vice versa?
How do you intend to set standards for the KPI’s of individual players?
Some scenarios to think through:
-
How do you attribute missed interrupts? Player 1 and Player 2 are going to have an interrupt available for an upcoming ability. Player 2 is supposed to interrupt it but doesn’t, how do you know to only penalize Player 2 when it looks like both failed, because like you should in any team, Player 1 trusted Player 2 to execute their job.
-
It’s bursting and you’re supposed to stop damage at X amount of stacks. Players 1-4 do everything in their power not to roll the stack, but player 5 doesn’t and rolls it. The group wipes. How do you convert that scenario to a tangible number that represents the individual performance with Bursting?
What if someone is on your team and you want them gone? Can you remove them? Who holds that power? What’s to prevent abuse of that power? What if it’s the other way around? How do you think players going to perceive a recently reset team rating?
How is losing all of your team rating not a significant penalty?
How can this be tracked at the individual level when everything the rest of the group does impacts their performance in all of these areas?
A good tank is so much more than this though. The initial pull, the gathering of mobs(though not just a tank job), routing, pacing, understanding the dps in your group and pulling around cooldowns to maximize throughput. How would you measure those? They’re critical.
How do you measure these? What entails contributing to team strategy? How do you score adaptability?
1 Like
Pretty much all of your suggestions would ruin M+. The appeal of M+ is not needing a team and still being able to push decently high score. Also, taking away score would fundamentally damage group creation since nobody would want to risk taking people they don’t know or who aren’t massively overqualified.
Some of your systems have been tried in PvP and were removed because they are so bad.
1 Like
Would allow exploitation. Would never be implemented
Would that be messed up or what?
I would HATE it if they did that. Absolutely, positively, HATE if my rating goes down because I join a group with someone who shouldn’t have been invited or is deliberately throwing my key. It’s already bad enough that my time (and repair bills and consumables) are wasted.
M+ dungeons are too long, and too low a percentage of total gameplay, to join enough groups for this to even out.
You’re talking about a queue for M+? Can’t wait to try that during incorp/afflicted weeks.
The irony.
The complete lack of self awareness is amazing.
You have yet to give any “technical” explanation on how this would work. You just use vague words and if anyone questions you, you belittle them and tell them to google it.
In fact, you outright have said you dont want to talk about the technical aspect. This combined with how you just tell people to do your home work for you makes everyone think you have no clue what you are talking about. So once again, ironic.
Player A thinks bursting should stop at 5, so stops dpsing.
Player B thinks that 5 is minimal issue, should go to 6.
Player C thinks going to fifteen is fine, just die and move on.
Healer thinks just don’t roll the stacks.
Tank is moving on to the next pack to minimize downtime.
End result: 3 mobs left with full group wipe.
2 Likes
Great post. This is one of those times I wish I could give an award instead of just a like as I think you explained the issues more concisely than I did in a way someone not familiar with math or programming would understand.
Your examples made me think of another one:
Its bolstering week and you are running a +22 EB. The group makes a decision to pull 5 packs, hero to kill all of the petal and than wipe to reset bolstering. This adds 25 seconds to the timer but will allow a second pull with no threat of bolstering on all of the hard hitting mobs thus saving a ton of time a ton of time having to pull each pack individually. “MINUS 50 DKP!!!”
Why because the group failed to take a druid to vortex mobs when they get bolstered, a pally to provide freedom and a evoker for rescue all of which would allow the tank to kite until the bolstering fell off and prevent the wipe in the first place.
This is the kind of issue that is impossible for an AI to determine because there are to many weighted variables. Its like trying to simulate the exact position of every particle in the universe (thats of course hyperbolic but in practice its the same issue because you need more storage and processing power than exists.)
I think the best way to illiterate how complicated this is, is to show that we can sim dps, we can to some extent run logs and process them to determine if you used your rotation incorrectly. But we can’t sim heals properly, we can’t sim tanking properly and there exists no sim that players run to determine the best route, pull pattern, what to soak by who and what CC to use by whom in a party let alone party comp the day a season starts.
If it was easy we would know all of this the very first day with a simulation and the goal would be to try to play that simulation as closely as possible.
So after smoking I realized we haven’t even asked the question about how to score this. In the above example everything was -50 DKP but in reality mistakes need to be plotted to determine how bad they are.
A simple exmaple is a game were you have two buttons. The best possible outcome would be:
AABABABAAA
However player 1 does:
BABABABAAA
We could score this as a 90% but what if player 2 does:
AABABABAAB
Is this also a 90% score after all its a single letter of difference. On the other hand perhaps the Letters at the beginning have more importance because they have a compounding effect on later letters. In that case its clear that player 1 played worse than player 2 so should we score it differently or do we score them the same. For this its going to be an answer that is relative to the observer as both ways are objectively correct. And this is the simplest you can make this start adding thousand of variables and the ways to score this increase tremendously.
Ok I think that point was made clearly but I tend to have difficulty in explaining complex systems so I probably just muddied the waters.
Moo!
This has been the case for all content (M+, raiding and pvp) and all systems from talents, to legos, to legendary paths, to Azrite traits etc. If you want to push as high of content as you can you need to use what is the mathematical best.
That doesn’t mean you can’t pug your way into +20’s using whatever you want though. I was doing +20’s as a Necrolord VDH in the end of SL’s as a pug. But it will make it more difficult, as in my case I did a whole lot less damage and it was random instead of on demand like a Kyrian.
This was said in legion as well. Its also been said about finding guilds, getting into raids, finding dungeons in vanilla etc. Its not a new thing.
See above about BiS as far as playing with friends again this has been all content since the games inception. If you wanted to heroic raid in ICC you wouldn’t join a pvp guild and you would waste a ton of time attempting to pug it each week when you could just log in and have a set time with friends.
Pugging isn’t dead, there are tons of people here on the forums that pug to KSH (or even better) each season. What people do suggest is if you pug and find a great player ask them to add you as a friend to run more keys. Why? Because you will fill your groups faster and you already know that they are good players so you have one less person you take a risk on.
Personally I puged all of season 1 and 2 and even in season 3 half the groups or more are full pugs and even the ones ran with friends have included some pugs.
The goal posts have never shifted, we hear exactly the same advise we ever have. Perhaps its the way it is said or your perception of what is said that has changed.
Moo!
2 Likes
I feel like this is one of those cases where I should just walk away but watching the train wreck is absolutely fascinating.
Moo!
2 Likes
This happens 4 more times during the key, fail the timer, and they don’t finish. Despite no one saying anything harmful or even anything at all, everyone receives a 1/5 social rating out of frustration.