The Stacking problem paradox

[k3endo]

So lets imagine its the new ranked season. Huzzah, two ways to play dirty bomb in a party. You have 4 friends online, skill ranges all the way from silver 3 to cobalt 1. You aren’t a pro team, and the extent of your communication is simple call outs (most of it is general chatter and banter) What do you do?

Do you:

1. Play competetive. You get put against a team with an average ranked 2-3 subranked higher than you, usually with at least 2 cobalts/1 elite to ‘match’ your apparent skill being in a stack of friends. You lose most games unless you go full cheese strat like stacking rhino’s or javelins, and even then its a gamble on when the much better players on the enemy team break your defences or swat away your assaults.

or do you:

2. Play casual. You put your self up the mercy of the CMM system, and there are only two outcomes. You either get put against complete randoms, winning around 75% of the time (leading to them complaining about the stacks since its the average skill thats probably out of wack). Or you get put up against the same stack of opponents, over and over and over again until ones side yields and either leaves the game or decides to wait 10 minutes for a different game, to which you go back to playing against randoms.

Ive kinda guess where these problems seem to stem from.

1. This problem stems from the balance needed to counter the full stack pro teams that can decimate the opposition. Yeah, we can try and be all tryhard etc, but you cant do much against a guy that can wipe the floor with you just by sneezing at you.

2. I assume this comes from the way CMM works. Solo playing CMM, the usual composition is one or two high level squaddies, a couple mid range ones with a few hundred hours, and 1-2 complete noobs with default guns, no map knowledge or skill whatsoever. HOWEVER, playing in a party removes the last category and replaces it with the high skilled or mid skilled players (of your choice), meaning the enemy has an unnecessary skill deficit. Or in the case of the party vs party thing, its the matchmaking system putting you against any big stack thats online (usually there is only one other queuing at the same time sadly.)

So…what am i meant to do? Should i suck it up, queue for ranked and have a bad time against elite operatives? Or should i queue in casual, have easy games against bad teams, ruining their experience of the game, and just HOPE that another, similarly skilled stack is on the other team so we can FINALLY have some fun?

[teflonlove]

@k3endo said:
So…what am i meant to do?

What would the matchmaking meant to do? How would an opposing team look like that could result in an interesting game?

[henki000]

Play ranked. You should learn from better players and avoid casual, if you really want to improve. If you keep losing, you lose rank and get easier opponents.

[Ptiloui]

As for me, it’s not really a 5-players group if you don’t use vocal to communicate. Any players that queue solo with minimal thinking about team composition and strategy could perform as well as your 5-players group.

So don’t bother about encountering strong opponents and go for ranked. As @henki000 said, after some games, you should face people of the same rank skill as yours. And use vocals ^^ there’s nothing better to improve together as a team, especially if some of you have their role preference (more supportive or more assaulting).

[teflonlove]

Ranked only ensures that everyone is at least level 7. Most of the time people also tend to take it a little more serious and try to play the mercs they are most experienced with. Also in CMM everyone can participate and (by design) play mercs for practice.

In Ranked you still have insane skill spread across players and a naive matchmaking algorithm based on a role independent single Elo rating that thinks it’s ok to match 5 lower golds with 4 mid golds and a bronze (which of course it’s not because the match turns into a 5 vs 4).

[kopyright]

@k3endo said:
So lets imagine its the new ranked season. Huzzah, two ways to play dirty bomb in a party. You have 4 friends online, skill ranges all the way from silver 3 to cobalt 1. You aren’t a pro team, and the extent of your communication is simple call outs (most of it is general chatter and banter) What do you do?

Do you:

1. Play competetive. You get put against a team with an average ranked 2-3 subranked higher than you, usually with at least 2 cobalts/1 elite to ‘match’ your apparent skill being in a stack of friends. You lose most games unless you go full cheese strat like stacking rhino’s or javelins, and even then its a gamble on when the much better players on the enemy team break your defences or swat away your assaults.

THANK YOU! I have brought this up several times myself in the past. For reasons beyond my understanding the idea behind this decision seems to be that (a) your aim magically improves when playing with friends, and (b) the opposing team cannot communicate via text or voice chat. Neither of this is true, and therefore playing with friends in Ranked is pretty much the worst experience you can have in DB.

[k3endo]

@henki000 said:
Play ranked. You should learn from better players and avoid casual, if you really want to improve. If you keep losing, you lose rank and get easier opponents.

And here is the other paradox.

Lets imagine i am a silver 3 level team today. When stacking, the game pits a silver 5-gold 1 against me to balance it out. I cant beat them, so i drop in rank. Eventually, i gain experience in the game to get back to silver 3, in which i just get put against the same silver 5-gold 1 people and cant progress. This is fine, its how ranked systems are meant to work.

So i get good, and play and put in the hours, and i progress to the silver 5-gold 1 level of skill.

now the game pits me against gold 3-gold 5 to make up for the fact i am in a stack at the silver 5-gold 1 level and the cycle continues.

So what, am i just expected to play against people far worse than me, then people far better than me, in a metaphorical ping pong just because i want to play with friends? In my personal case i am cobalt 1 with an average ranked team of gold 3 level, and i would like for the same thing to occur on the enemy team. The people in the cobalt 3-elite rank groupings are better than i ever will be, so why does playing in a stack pit me against the gods amongst men? Wouldnt it be wiser to try and match us versus the enemy, in a sort of 1 to 1 thing, pitting me, a cobalt 1, versus an enemy of roughly same skill, with every other player in the stack being balanced versus their equivalent skill?

This isn’t about improving personal skill, this is about stacking in ranked being the worst dirty bomb experience possible. The game sees a stack in ranked, i assume gives a leeway of 1-3 subranks as a balance resort (something i disagree to begin with), but instead of this being spread over the entire team it usually culminates in 1-2 really ****ing good players heading their team, leaving us in the dust.

Edit: Woah, this came off as waaaaay too aggressive. its not aimed at anyone in particular, I’m just a bit

T R I G G E R E D

[B_Montiel]

On public games, rework the shuffle so it shuffles after the map selection.

Either way, running the AIE server, we noticed the lobby shuffle, even if there are minimal player loss during lobby, is doing terribly.

[Deathi]

Stacking doesn’t work, when not playing vs. also stacked teams. Shuffle doesn’t work because using not enough data (or someone uses natural numbers instead of allowing negative or rational numbers). Also shuffling after a map (while people are joining/leaving/switching), in opposit to doing it directly before showing the intro-vid, obviously is an idiotic idea.

[Szakalot]

@B_Montiel i think this is largely due to completely scrambled hidden rating due to previous games. If i play a few rounds of pubs somewhere else first, the AIE shuffle will very likely be wrong about my skill level. Already one player’s being off would be enough to reck balance, now imagine the same for 90% of the server.

The shuffle after lobby might help a lot with overall game quality across the entire game.

But even then, the game will remain difficult to balance, with a large range of skill levels between different mercs (e.g. drag the team down with Vassili, but ill stomp with Redeye) for the same player.

Even still, the AIE server match quality is remarkably better than the crap we see on normal pubs

[kopyright]

@Szakalot said:
@B_Montiel i think this is largely due to completely scrambled hidden rating due to previous games.

Given the number of losses in a row we sometimes had during an evening (due to the stacking problem) we should have been pitted against a group of toddlers at one point.

@Szakalot said:
But even then, the game will remain difficult to balance, with a large range of skill levels between different mercs (e.g. drag the team down with Vassili, but ill stomp with Redeye) for the same player.

Never thought about that, but I don’t see how you could ever find a good balance here.

[OCB-Wildcard]

@kopyright said:

@Szakalot said:
@B_Montiel i think this is largely due to completely scrambled hidden rating due to previous games.

Given the number of losses in a row we sometimes had during an evening (due to the stacking problem) we should have been pitted against a group of toddlers at one point.

@Szakalot said:
But even then, the game will remain difficult to balance, with a large range of skill levels between different mercs (e.g. drag the team down with Vassili, but ill stomp with Redeye) for the same player.

Never thought about that, but I don’t see how you could ever find a good balance here.

Hence the Meme: “Balance” is a myth. There is no “Balance”. So. Git. Gud.

The likely odds of a system ever managing to achieve the kind of balance we all want is highly improbable. This is especially true when you consider the uncontrolled random element that is a human player being thrown into the mix. Any semblance of balance can be easily thrown off kilter by one or two players under-performing, for whatever reason, in that situation. You can’t achieve such a system, one which will always manage to balance out every single factor to maintain fair matches 100% of the time, when you have an influencing factor that it holds no sway over; it’s just impossible from a practical, and realistic, standpoint to achieve such a system. That isn’t to say the system can’t be improved, it definitely can be improved upon immensely. I merely mean to point out that it’ll never be quite what we want from it.

I’ll close this on a personal thought/ideal of mine that sums this up nicely:

“There is nothing more insatiable, or unyielding, than the expectations of the people one serves”

[B_Montiel]

@Szakalot said:
@B_Montiel i think this is largely due to completely scrambled hidden rating due to previous games. If i play a few rounds of pubs somewhere else first, the AIE shuffle will very likely be wrong about my skill level. Already one player’s being off would be enough to reck balance, now imagine the same for 90% of the server.

The shuffle after lobby might help a lot with overall game quality across the entire game.

But even then, the game will remain difficult to balance, with a large range of skill levels between different mercs (e.g. drag the team down with Vassili, but ill stomp with Redeye) for the same player.

Even still, the AIE server match quality is remarkably better than the crap we see on normal pubs

That’s where the player level you’re playing against should factor in at some point. Yes, a level 7 can be better than most level 60, but it still awfully rare, and someone doing well against 60+ should get better rating in this respect.

Pulling a 1 KDR on our server is somewhat underrated compared to a 3 one on any pub server, taking the average game filled with random Joe.

As it seems, shuffle can do more damage than good in my view. A good old random scramble no matter the levels or skills might do better in this case. Most games I played in the past, even ones with high skill ceiling, had lesser balance issues than DB currently has.

[teflonlove]

@kopyright said:
Never thought about that, but I don’t see how you could ever find a good balance here.

It should be fairly easy to decide if someone is inexperienced with a merc or at least competent by looking at certain behaviors and KPI’s. Then make some servers where you can only pick the mercs you are competent with.

For example for an aura: average HP of revived players (due to charged defibs), average survival time of health station (due to being able to pick good spots and collect it before air strikes etc), percentage of revives after long jump and so on. You might have to balance this numbers somewhat with weak opponents but in the end you should come up with a fairly representative competence score.

This would make more sense than the generic “min level X” servers where I can still be an awful sniper.

[Szakalot]

@teflonlove this isnt gonna work. Charging defibs as aura is very often a bad idea. In combat, asap revive takes precedence, due to how easy it is to finish a downed player.

Average hp station survival time is also pretty bad: a station put in a useless spot might survive the entire map. A critical station that heals a push for hundreds of HP might be promptly destroyed, etc.

[teflonlove]

@Szakalot said:
@teflonlove this isnt gonna work. Charging defibs as aura is very often a bad idea. In combat, asap revive takes precedence, due to how easy it is to finish a downed player.

True, but there’s situations where it makes sense and you should do it. I’m not saying that 80% charged revives is better than 40% charged revives. But there’s a threshold where inexperienced players will stay below, like 5%. And that’s how it should be possible to distinguish them.

Average hp station survival time is also pretty bad: a station put in a useless spot might survive the entire map. A critical station that heals a push for hundreds of HP might be promptly destroyed, etc.

As a single KPI, yes. If you add another like average HP healed per minute then a bad spot will yield a low number there. Again, this is not a “bigger is better” KPI, just a threshold. There’s also other factors like how often do you move the health station to avoid explosives destroying it. Everything also depends a lot on the rest of your team (e.g. a nice Stoker puts a protective ammo station in front of you health station) and the enemy team (e.g. lack of skill causes less damage with in turn need less healing).

The thing is, inexperienced players will fail at pretty much all of these these thresholds at the same time. Consequently they remain on servers where they can practice more with enemies that don’t gank them the entire match.

[bgyoshi]

The answer is you play solo server browser

Most of the players are over level 30 and I don’t think I’ve played a server browser match with fewer than 2 level 100+. And I know I’ve played more than 10 games recently with more than 5 level 100/200+

Such is true even during the ranked seasons. At this point you’ll improve no matter what you choose

If you don’t mind slow imbalanced Ranked matches then sure go for it, the only real difference between ranked and casual from a gameplay perspective is that casual is much faster and you get more chances to do stuff.

Personally, one of the biggest benefits of casual over Ranked is global mute. I don’t have to listen to kids scream, annoying vets cry about how bad the team is, coolguys trying to command people around the map, nothing. I can just play KMFDM and Ride the Lightning and enjoy the game.