Exaggerating scenarios

(Amerika) #21

Heh, you’d probably agree with my viewpoint or at least part of it here. I like pubs but I think it’s one of the major sources for a lot of different issues that is nearly impossible to overcome given the limitations of the systems involved and the mindset of players playing today.

(CCP115) #22

I don’t know bout you kiddos, but this seems like that “hacker” scenario all over again.

It’s either blown WAY out of proportion, or you Americans just have problems we Oceanians don’t.

We have balanced matches at least 70% of the time, and only steamrolls around 10% of the time. That 10% is just slightly stacked.

It’s really hard to get stomped in pubs as well, they’d need at least 2 high level players, and even then, frequently you get low levels but with 1000 + hours in CSGO, and it really shows.

(LifeupOmega) #23

Heh, you’d probably agree with my viewpoint or at least part of it here. I like pubs but I think it’s one of the major sources for a lot of different issues that is nearly impossible to overcome given the limitations of the systems involved and the mindset of players playing today.[/quote]

I like pubs to a degree. But until MM gets the attention it sorely needs this game offers nothing to me, which is sad, because I love this game. I don’t even come from a background of matchmaking, I’ve always done PuG lobbies and such, but this game just drove home how useless the majority of players are, which leaves a very bad taste in my mouth when I play public. I’m tired of carrying, and I just want good games that are enjoyable and fun to play again.

I don’t even see why its a huge issue, this game would do so much better as a 5v5 standard. People might even learn to play as a team and communicate once in a while. You’re putting the same investment into a pub anyway if you want credits and whatever, only difference is that with everyone pooled into MM you’d have to wait a minute or two for a game to start, and you’ll eventually get games that don’t give us more “balance plx fix this autoshuffle plz sd” threads after people get proper rankings.

(Amerika) #24

I think SD was afraid of really alienating their core audience and their initial backers by forcing DB to be an MM only experience with no pubs. I mean, it’s hard to sell people on being OK with the idea of creating an old school shooter that isn’t being too hampered by modern console expectations and then immediately restrict it from old school style pubs.

But I think the best and most balanced I have ever seen this game was back during the closed beta after the Steam launch where SD turned off all pub servers and you could only play matchmaking. It was great because there was a large pool of skilled players that could be matched up together as opposed to a tiny pool like there is now.

If we had a pub system where we could rent/own/admin our own servers I most definitely wouldn’t want to see pubs go. But since the servers are just “random region server X 7v7” I wouldn’t mind seeing them go since there isn’t any sense of community outside of forums in DB. So if the game launched and dropped pubs all-together it really wouldn’t bother me. I think some people would definitely be upset and quit but retention rates might be much higher because it would fix a literal mountain of issues that cause players to quit IMO.

(LifeupOmega) #25

Even from my first week of playing Dirty Bomb I was asking why we didn’t have community driven servers. Right now it feels like, say, TF2 does. You join a game, you know no one, you go do your own thing and it doesn’t matter in the slightest what happens because there’s no driving force to get better.

I’m in the same boat though, if pubs disappeared I wouldn’t really care because it’d probably mean we’d get better games in the long run.

(Rjsto) #26

For me it’s usually pretty rare to end up in a game that’s a COMPLETE stomp one way or other. Sure, sometimes the game ends pretty quick because the other side plays better, but rarely it’s just one sided. So I don’t really mind, maybe I’m such an average player that any situation feels pretty basic

(N8o) #27

I can’t really say what the true numbers are. =
never been spawn camped.

(Carniege) #28

@Watsyurdeal I was agreeing

@Lumi That’s to do with skill and time played and stuff. What I’m talking about is simply the number of players in the game.

@CCP115 Exactly! I play on EU though and those are the same proportions for my games whereas for other EU players, they get spawn camped 60% of the time apparently.

@Amerika so basically you said what I’ve been trying to avoid saying in a very nice way - people are shit at the game and blame the match making. Funny that in another game World Of Tanks the exact same thing happens.

(Lumi) #29

@Carniege

I agree, but they’re indirectly related. Have a slightly skill unbalanced game and you will have people leaving on the losing team, leading to further imbalance.

(Carniege) #30

@Lumi yes I know but that’s not my issue with these forums. My issues are people saying that most of their games are unbalanced player count wise, with numbers such as 3 vs 7. And I’m calling bullshit because for 15 minutes it would not be 3 vs 7.

And the skill wise inbalance just has to be fought. The match making as far as I know tries to balance the teams in the minute it gets before the match starts based on skill ( @Amerika help pls idk how it works) but you’ll just have to fight the better players. If players on the losing team leave, then new players will join within a minute, and that will keep and keep happening.

Eitherway what does that have anything to do with match making?

The match maker as @kAndyREW said in his latest video can not predict whether people would leave or not, so I don’t see how this is it’s fault. This is the fault of there not being more divided servers, for example servers for players level 30+, to separate these better players with newer ones.

I get these games with an alpha player and I just try my best to do what (s)he does and it helps me learn. I once spectated this alpha when (s)he was on my team and I found 2 new long jumps I never knew about. They’re very helpful to learn from, not just be salty about (not that youre being salty, I’m talking about the other players who can’t handle losing).

(Amerika) #31

@Carniege The way the current system works in a pub is it looks at your last 25 matches and assigns a skill rating to each person privately. Most likely this is weighted by your performance which is most likely using your score per minute heavily and other factors like wins and losses. That’s a theory but it makes sense from a database performance perspective (which is part of what I do for a living) and for simplicity. The system is trying to assign players to give each team an equal skill value. Also, one of the more recent changes along with looking at your last 25 games instead of 10 is if teams are even (but slots are open) and a new player joins while the game is in progress that player will be placed on the team with the lower skill rating.

With that said I can say that the system has a number of flaws. Sadly but many of the flaws have less to do with the private ELO system’s own inadequacies and more to do with it not being able to predict players.

The system quickly falls apart when players leave/join or if there is a huge difference in player skill. Players leaving and joining can throw off the skill rating based on when the leave and who joins and the server can auto-shuffle a player who is has an abnormally high rating and offset them with a bunch of players with lower ratings. In the latter example that high rated player might leave and the team gets pounded. The player might stay and the team gets pounded because the less experienced players (players still figuring the game out in their first few hours) simply can’t do enough yet to offset the other team that was filled with a few competent players (to even out the skill rating). Creating a situation where the highly skilled player still does well and might even be able to carry but the lower skilled players simply get destroyed regardless of a win or loss. So you can get stomped and lose (often) or stomped and win (less often but happens). Sometimes just one or two incredibly solid player can carry a server of evenly skilled players which also can produce a stomp. Especially if they are friends and joining the same team when they can.

So between leaving/joins (and the time people have to sit in the lobby), how the ELO system works and assigns teams, players doing their own thing and not caring (because it’s a pub), the system not factoring in if players have a medic/engineer in their squad and if they are inclined to use them (something that could be added potentially) leads to a rather hectic “balance” system that really can’t hope to consistently balance in a pub environment. Things are slightly easier in MM in theory but due to the low population of players who participate in MM it also has tons of issues stemming from the lack of players.

Sorry about being long winded but that’s my perspective of what’s going on and the issues that may or may not be able to be fixed/lessened.

(Ctrix) #32

I think the system also doesn’t take into account if people have less than x games played, which is why we keep getting teams full of noobs vs teams full of MLGpros

(Amerika) #33

I think the system also doesn’t take into account if people have less than x games played, which is why we keep getting teams full of noobs vs teams full of MLGpros

[/quote]

It’s only looking at your last 25 games played and looking at how well you did in those games regardless of your amount of games played total according to what SD has stated. The stats it uses to give you a skill value is what I don’t actually know. It might only use your score per minute. It might weight your wins vs. losses heavily. It might factor in things like your accuracy during those games or KDR and it might even weigh in total amount of time played. That’s all a guess as far as I know unless somebody else has been told differently (I don’t catch everything SD says).

One reason for the system only looking at your last 10 (and now 25) games and not at individual merc scores or re-calculating for currently picked squads might partly be due to performance. A read optimized database with global player stats has to be used to create this skill rating. I laymen terms, once you start adding too many queries and making a request too complicated and making it look at too much information over a long period of time you your result times start slowing down. Now multiply this by thousands of users all hitting it at the same time and the hardware required to do this and you start to see that the rating system might possibly be limited by performance/hardware more than it is anything else. Not everyone has EA/Ubisoft money for hardware. However, that is just a guess on my part but it’s an educated one.

(Carniege) #34

@Amerika thank you for the explanation :slight_smile:

(FalC_16) #35

I agree with the above post. People leaving and joining the match do change the balance and it can influence the result of a match.

What I experienced yesterday, was stomping the opposite team and when I finally decided to switch team, it was not possible. There were equal numbers of players on both teams.

This was in execution, which also means, team shuffle is not available.

All I could do is to continue the stomp until somebody from the opposite team left, enabling me to switch teams.

(Carniege) #36

@FalC_16 yeah exactly. Same thing happened to me recently as well, I couldn’t do anything because the teams were balanced player wise. Since people kept leaving and joining the balanced went into our favour, thus fucking up match making whilst keeping player count balance