Ranked Mode Woes

(wasdman) #1

Long queue times, aptitude all over the place, and a rank that only cares whether you win or lose.

Player performance in game, needs to be factored into rank progression. As it stands, If your going to lose, there is no reason to stay. It should be a show of an individual skill, not games won with total strangers.

It is incredibly frustrating when matchmaking pairs me up with people who just started playing the game, I’ll top score, 80+ frag trying to contribute only to lose because they don’t know what they’re doing. Adding insult to injury, after an average wait time of 20+ minutes for a game, rank just goes down after a loss.

(Mustang) #2

That’s why penalties for leaving exist, I guess you’re saying they aren’t harsh enough if they’re going unnoticed.

(Glottis-3D) #3

i still think that win/lose rank is not correct for team-games. there SHOULD be a correlation between player performance in a match and his rank change. i.e. players in a team get different ELO change, depending on their XP or XP/min

(Szakalot) #4

win loss should be the primary value that affects rank but sone factoring of individual perfomance would be nice too :

+20% for top (more win. if won, less rank lost if lost)
+10 for 2nd etc until -20% for lowest

score might seem at first glance like a bad measure but:

  • good way to get lots of xp is to play a medic
  • good way to get xp is to play objective (engineer/ev destruction)
  • good way to get xp is to make swinging multikills.

in the end its hard to imagine such a factoring would make the team fall apart cause everyone tries to xp whore. losing still results in the loss of rank, so best thing to do is still play as a team.

another great benefit of such factoring is a faster rank adjustment for individual players : great players that carry teams will get promoted faster; bad players that drag the team down will lose rank quickly. increase xp penalty (when punished) for teamkill to -500xp and suddenly you gave a good deterrent against griefers.

(Smooth) #5

I do very much agree with Szakalot’s thinking.

While Win/Loss should always be the primary factor, something like scoreboard position or XP/Min for that match could be included in the future.

(Glottis-3D) #6

Even with clear danger of XP whoring, we cannot argue that Top player Is better than the Bottom player.
I cannot remember any game, where a bottom player could be considered as MVP of the match.
That is more than enough for statistic altering of the rank bonuses.

So as i see this:
Glottis with 2000 ELO Won and placed 1st in his “Team B” (Everage team ELO 2000) in a match Against “Team A” with Average team elo 2000,
Szakalot with 2200 ELO from “Team B” placed last in the same match.

Glottis gets +50 w/l elo and +20% from score placement = +60 elo
Szaka gets +40(because he is higher in rank than :Team A" thus less default delta-elo) w/l elo and -20% from score placement = +32 elo

so after this match
Glottis 2060 ELO (MVP of the match, and applause from teamspeak)
Szakalot 2232 ELO (Says something about still being better than Glottis, lets go 1on1 fletcha vs proxy, what the matter, scared, haa??)

:>

am i right in my calculations?

(Szakalot) #7

[QUOTE=Glottis-3D;548127]Even with clear danger of XP whoring, we cannot argue that Top player Is better than the Bottom player.
I cannot remember any game, where a bottom player could be considered as MVP of the match.
That is more than enough for statistic altering of the rank bonuses.

So as i see this:
Glottis with 2000 ELO Won and placed 1st in his “Team B” (Everage team ELO 2000) in a match Against “Team A” with Average team elo 2000,
Szakalot with 2200 ELO from “Team B” placed last in the same match.

Glottis gets +50 w/l elo and +20% from score placement = +60 elo
Szaka gets +40(because he is higher in rank than :Team A" thus less default delta-elo) w/l elo and -20% from score placement = +32 elo

so after this match
Glottis 2060 ELO (MVP of the match, and applause from teamspeak)
Szakalot 2232 ELO (Says something about still being better than Glottis, lets go 1on1 fletcha vs proxy, what the matter, scared, haa??)

:>

am i right in my calculations?[/QUOTE]

disagree, the base elo adjustment should be the same for both you and me. Just because I am individually higher doesn’t mean that I didn’t have to try just as hard to compensate for that other unnamed 1800 elo player.

So both of us get +50 win base, that later gets boosted to 60 for you, and 40 for me

(MoonOnAStick) #8

[QUOTE=Glottis-3D;548127]Glottis gets +50 w/l elo and +20% from score placement = +60 elo
Szaka gets +40(because he is higher in rank than :Team A" thus less default delta-elo) w/l elo and -20% from score placement = +32 elo[/QUOTE]

Different individual Elo changes based on pre-game rank seems counter-intuitive. Assuming evenly rated teams, if you top frag and carry 4 other (lower ranked) players to victory why should you get a lower payout? Conversely, if those 4 players drag you to defeat surely you wouldn’t expect a bigger rank decrease than them.

Rank changes scaled by in-game xp could incentivise poor play/merc choice, particularly if you’re already resigned to losing. Poorly timed revives, focus on side-objectives might help reduce your rank loss. I like the idea that you win/lose as a team but I guess only the match predictions and player rank distribution would tell whether it works or not.

(wasdman) #9

I’m talking in terms of rank progression. If you’re going to lose, you will be going back in rank. I’ve never abandoned a game, but I starting to feel like why I even bother. This is regular enough of an occurance that I decided to find my old SD forum account and make this post :confused:
I dont think more punishment is the awnser here, I think incentives for staying in the current match, and doing the best you can in it is.

As it currently stands in Dirty Bomb, matchmaking rank is simply a measure of how fourtunate you are to be paired with adept teammates, and how many wins you can manage. Pre-season, was a measure of how well you hand picked 5 man premades, letting you stomp anyone foolish enough to not do the same.

So whats the solution?

For a game loss that I worked hard for, I’d at the very least, like my rank loss penalty reduced. Being able to keep my current progression, (when I’m so far ahead of my teammates its not even funny) would also be a welcome addition. I also like what Szakalot had to say. :slight_smile:

Matching players of simular skill with and against each other would help greatly. Though I suspect thats not really possible at the moment, due to so few people queueing for ranked. (At least I can only speculate since I dont have acutal player numbers, but the 20+ minute wait times and the “maximum skill range achieved” lead me to think that’s why.)

(Szakalot) #10

[QUOTE=MoonOnAStick;548155]Different individual Elo changes based on pre-game rank seems counter-intuitive. Assuming evenly rated teams, if you top frag and carry 4 other (lower ranked) players to victory why should you get a lower payout? Conversely, if those 4 players drag you to defeat surely you wouldn’t expect a bigger rank decrease than them.

Rank changes scaled by in-game xp could incentivise poor play/merc choice, particularly if you’re already resigned to losing. Poorly timed revives, focus on side-objectives might help reduce your rank loss. I like the idea that you win/lose as a team but I guess only the match predictions and player rank distribution would tell whether it works or not.[/QUOTE]

an argumentcan be made that if you are resigned to losing it doesnt matteranymore - you will likely not care either way and do silly stuff.

i also find it hard to come up with examplea of gameplaychoices that would give you mpre xp but sabotage your team: sure you could repair that now redundant generator, but once you are done - if you want more xp - you will likely return to playing the map as intended. in all likelihood, the teammates that have not given up will score more xp

(montheponies) #11

Hopefully SD will come up with a 5 v 5 ranked system, so you can challenge others or just go for the luck of draw against someone else looking. The previous scheme was terrible for the vast majority, and to be honest I can’t believe anyone really enjoys a match where you stomp the other team (unless you’re 10).

My very rough and full of holes suggestion for improving things would be something like the following;

Every match has a set ‘Value’. At the start of the match a perfectly balanced set of teams would result in 50% of the value being used as the loser ‘pot’ and 50% added to the winning team pot.

Within each team I would just go for a set amount of the value being added to each members individual ELO (or whatever we’re calling it).

      • 20%
    • +10%
      • 10%
    • +5%
    • +5%

I wouldnt try to do anything too fancy, such as proportionally allocating it. Same story except in reverse for the losing team ie. worst player in the losing team loses the most. Either incentives them to play harder or forces them to drop rank quicker to a level where they are in the middle of the ranking more often than not.

If the teams on paper aren’t perfectly even then the inital game ‘Value’ is simply skewed so the weaker team would get say 60% of the value to share if they won, or 40% of the value if they lost (the stronger team getting the inverse, 40% for a win and 60% for a loss).

All that being said, it does feel like the current scheme has something to it when everyone shares the same risk and reward and as such acts as a team. Perhaps we really just need to let the current system bed in for a couple of months so ranks are better settled.

(Glottis-3D) #12

[QUOTE=Szakalot;548147]disagree, the base elo adjustment should be the same for both you and me. Just because I am individually higher doesn’t mean that I didn’t have to try just as hard to compensate for that other unnamed 1800 elo player.

So both of us get +50 win base, that later gets boosted to 60 for you, and 40 for me[/QUOTE]

you are wrong. ELO should be person-dependant.
Imagine
3000-elo guys getting wins in a team of 1000-elo nubs vs team of 1400-elo low skillers, and getting his ELO go sky high after every game. But he shouldnt get any elo bonus ofater winning with this kind of skill difference.

In duel 3000elo will get ZERO elo bonus after win vs 1400. and this is how it should be.

this is self-consistent system. you get many, when its hard, and you get none, when its easy.

(ailmanki) #13

This is a core aspect of the game, I wrote that already early in alpha.
WIN/LOSS alone is a indicator - Its a symptom of something else. It alone is useless, unless you have many many more matches and players. And even then its slow at what it does. TBH I expected more from fireteam appartement, and from the heatmaps … and all those statistics.
Recognize how important it is SD, and distribute some brains thinking how to do player statistics proper, so that you can match them up against each other.
Sure a one dimensional skill rating is cool - so everyone can be directly compared, but for match making you might want a multidimensional skill representation. Movement, Aiming, Positioning, Team work, Situation awareness; I really hoped fireteam would get a step forward in statistics.

But seems like you guys are just throwing buzzwords around.

https://www.fireteam.net/fireline/

Smart Matchmaking

Configure low-latency, asynchronous matchmaking using buckets and fuzzy numeric ranges – all adjustable on the fly

edit:
just to proof my point, it can be done in Basketball - where they even need to translate the recording into digital data first…

(FireWorks) #14

Bonus depending on the position is as good or bad as the current system in my eyes.

The aim is to filter out exceptional players (on both ends), right?

On a scoreboard like this:
18k
18k
12k
9k
3k

I assume both top players did very well. I dont see a point in rewarding them differently by a wide margin (5% in the posts above).
On the other hand, I feel the last guy should see only a low bonus on his rank.

Id suggest to work with this model to avoid the nasty things people could come to their mind:
Build the Average team score and only look at deviations of a certain threshold. Personally Id be for 30% which leaves plenty of room for being the guy not spawncamp on Bridge while defending the EV for that ninja eng. Or not ammo whoring, or mine whoring etc etc.

People above get a percentage bonus for every percent above this threshold. So 5% if 35%.
Same malus for people below.

For the example above:
18k+18k+12k+9k+3k=60k and gives an average of 12k.
30% of 12k would be ±3.6k
The threshold for the bonus to kick in would be 15k6. For the 18k players this means that they are 2.4k in the green. In percentages this means: 18k/12k=150% means so they are 20% above our 30% threshold and get 20% extra progress/Elo.

Our poor 3k guy would have needed 8.4k to get the full points.
3k/12k=25%
He is short of the 70% by a full 45%… His progress would be almost halved… 45%!

Sounds good to me :slight_smile:

(Szakalot) #15

good system too, imo it doesnt matter TOO MUCH, as long as the crappy and excelling players are distinguished

(Szakalot) #16

[QUOTE=Glottis-3D;548249]you are wrong. ELO should be person-dependant.
Imagine
3000-elo guys getting wins in a team of 1000-elo nubs vs team of 1400-elo low skillers, and getting his ELO go sky high after every game. But he shouldnt get any elo bonus ofater winning with this kind of skill difference.

In duel 3000elo will get ZERO elo bonus after win vs 1400. and this is how it should be.

this is self-consistent system. you get many, when its hard, and you get none, when its easy.[/QUOTE]

in your example its hardly easy for the 3k dudes to carry 3noobs and basically playing 2v5.

the measure of how easy it is is expressed in the average elo. while an amazing player might have an easy time winning 1on1s, it might not be easy to actually win the map - which is what counts in the end.

you WANT to have players who are obviously out of their current rank to get promoted/demoted asap, dont see what the problem is. if two 3000 dudes can carry 2v5 they obviously should be elevated to a level where its not possible for them to get matched against such newbies.

there is also the idea of team-building : if i see that im twice as good as anyone else on the server, i dont want to be demotivated because my elo gains are super low. play as a tea - win as a team : which is why win/loss should always remain the primary elo boost

(Glottis-3D) #17

Szaka, my main point is that uber-ELo players will not get any bonuses, when facing noobs.
+plus i do not think that it should be impossible for 30000guy to face and play with 1000 guys.
visa versa, coming from quake i want to any kind of skill low to be able to face higher skills, if both sides are willing to do so

(Szakalot) #18

and my point is: if uber players are facing noobs it menas their rank is underrated - they should get their bonus, get out of noobs rank-range and play people their level.

Imagine a pro CS player starting DB, stomping through placement matches. Are you saying such a player shouldnt get any bonuses cause its ‘not fair’?

Primary point of any rank is to guarantee good quality games, not an ego boost to the respective players

(Glottis-3D) #19

[QUOTE=Szakalot;548308]and my point is: if uber players are facing noobs it menas their rank is underrated - they should get their bonus, get out of noobs rank-range and play people their level.

Imagine a pro CS player starting DB, stomping through placement matches. Are you saying such a player shouldnt get any bonuses cause its ‘not fair’?

Primary point of any rank is to guarantee good quality games, not an ego boost to the respective players[/QUOTE]

  1. they still will face eachother on max range conditions. you cannot avoid that (see axample below), and you should not avoid that imho.
  2. your ranks would only work for match making, while my system will work for any ranked matches, including ranked publics.

Example:
1 uver-player and his 9 noob friends Take the DOme as the Map for MM (prbbly least played map) at NOT Peak times.
These guys can playes dozens of matches every day and that guy will boost his elo with your system UNlimitely!

With my system this and any analogue of this would be waste of time.

(Szakalot) #20

no they won’t and yes you should! :stuck_out_tongue: Max range conditions is something like 3-5 ranks at the moment. Bronze players will never get matched with cobalts, even if the only players available are 4 cobalts and 6 bronzes (the only people queuing at that time). Nor should ever bronze players get matched against cobalts.

  1. your ranks would only work for match making, while my system will work for any ranked matches, including ranked publics.

why would you ever want one system for both? ranked skillset can be vastly different from 8v8 pubs, e.g. proxy can be amazing on 8v8 pubs spamming mines everywhere, whereas she will get destroyed in rank. Trying to balance said proxy-main players for both modes is a waste of time. Not to mention that some people tryhard pubs and some are mucking about, some are great at spamming but cause too much friendly fire damage in ranked, some are great at objective mode but suck terribly at SW