|
|
| Author |
Thread Statistics | Show CCP posts - 0 post(s) |
lrian Locust
Chatelain Rapid Response
Gallente Federation
382
   |
Posted - 2014.04.18 15:32:00 -
[1] - Quote
Bad matchmaking leads to pubstomping and is one of Dust's biggest problems. More pubstomping, less new players that stay, leading to worse matchmaking and less income for CCP, and less conten, until the game dies.
So, how to fix this? I assume that for most players, a great match is an even match. The ones where the battle sways from one team to the next, when one's actions really determine the outcome between a glorious win or a bitter defeat. So, let's take win/loss ratio as a Key Performance Indicator and use it for pub matches.
Here's how:
- a player's value is calculated by the total win/loss ratio of the last, say, 30 matches, divided by 30.
- random players in a squad have an advantage, so they get a 5% boost to their ratio.
- random players in a squad with microphones have more of an advantage, so they get boosted by 8%
- players in a squad from a corps have more of an advantage, so they get a boost of 10%
- Scotty should place squads on opposite sides and fill the team's ranks according to their win/loss ratio. A team with the lower cumulative win/loss ratio should get the next available merc.
This might lead to uneven matches with one synched squad against a team of newbies that's twice as big, but still the odds will be more even.
In the end, this will lead to the following:
- Matches will be more closely matched, making fights more worthwhile and exciting.
- Proto gear and SP will make less of an advantage, but skill, experience and teamwork will pay off.
- Newbies get a boost for the first 30 games, which decreases gradually over the 30 matches.
- Newbies will be immersed into the pits of hell more gradually.
- Vet players with a new alt will 'level up' quickly, to reflect their skills
- Long-time players with low skill or low value (e.g. K/D padding redliners) will have less negative impact on their teammates
- Implementation and real-time calculation should be fairly easy and fast.
- CCP delivers on its promise that one shot can really make an empire fall, as actions will have more impact
Of course, this system needs tweaking. Boost percentages can be tweaked as we go, and there should be separate win/loss ratios for pub and faction warfare.
So, what do you guys/gals think? |
lrian Locust
Chatelain Rapid Response
Gallente Federation
383
|
Posted - 2014.04.18 15:54:00 -
[2] - Quote
IceShifter Childhaspawn wrote:Need players to have matchmaking. No barrier between die-hards and die-oftens is the problem. Your idea is impractical because it is too easily gamed and that is the core issue.
All these people want is to win. They don't care about it being fair and they don't care if their actions kill the game. They'll find another. Thanks for your input. But how can this mechanic be gamed? The only way to game it, is to lose. And it'll decrease the gap between diehards and newbies if there's a lot more newbies posed against a squad of vets.
|
lrian Locust
Chatelain Rapid Response
Gallente Federation
383
|
Posted - 2014.04.18 15:57:00 -
[3] - Quote
Forever ETC wrote:This would have worked a year ago when we had 9,000 players. But right now the only solution is through gear. I don't think gear is that important. Good gun-game, knowing the maps, and teamwork are way more important.
|
lrian Locust
Chatelain Rapid Response
Gallente Federation
385
|
Posted - 2014.04.18 16:16:00 -
[4] - Quote
Django Quik wrote:No matchmaking algorithm will work with ~3000 concurrent players worldwide; not within reasonable waiting times anyway.
Here's an idea to help matchmaking - get more people playing the game and sort out the damn NPE so they don't all quit after 3 matches. I agree, NPE is very important, but that alone won't solve the pubstomping problem. Even with 3,000 players this system should make quite a difference, especially for newbies. I don't think that it should increase waiting time considerably. Every merc or squad has a KPI value for win/loss ratio and Scotty only has to add a merc or squad to the team with the lowest total KPI value. It might even lower waiting time, as calculations are very basic and it would lead to player retention.
|
lrian Locust
Chatelain Rapid Response
Gallente Federation
385
|
Posted - 2014.04.18 19:29:00 -
[5] - Quote
The Robot Devil wrote:There are two things that can fix matchmaking. #1 Create multiple different areas that are security status based and then only allow a suit to be used that has a certain total meta number. The total meta would be all the mods, equipment, weapons and the suit added together. the higher the sec stat the lower the total meta allowed.
Multiple arena's would make matchmaking harder, as there's less players available for each arena. Also, the game would need to calculate all meta levels for all different suits (leading to a longer waiting time), and players will have to deal with certain suits being disabled for certain game types.
The Robot Devil wrote:I have been here for two years, I have 21M SP and a running KDR of 0.64 and a monthly of 0.43 and a W/L of 1.5. I will get owned by any noob whom can shoot halfway straight. I am a support player and a force multiplier. My team does better because I am there making them better. Where would a metric based system place me or me in a squad?
If you are a good support player, your team will benefit and you'll end up with a higher win/loss ratio. Then you'll be matched against teams with a higher win/loss ratio as well, no matter your KDR.
KDR shouldn't play a role in matchmaking. That would make redline snipers the most valuable players on the field, although most of them don't contribute much to winning. The opposite would go for eWar scouts.
|
lrian Locust
Chatelain Rapid Response
Gallente Federation
385
|
Posted - 2014.04.18 19:32:00 -
[6] - Quote
Delta 749 wrote:Or heres a much simpler idea, let people queue up for different tiers of battle IE high sec allows up to standard, low sec up to adv and null sec being a free for all with after battle rewards adjusted to grant more/less in which tier you fight in That won't remove protostomping. Synched squads with standard gear and all their skills maxed out, will still beat the crap out of any team of newbies and random blueberries. The differences are too big - matchmaking by meta level isn't enough anymore.
|
lrian Locust
Chatelain Rapid Response
Gallente Federation
385
|
Posted - 2014.04.18 19:34:00 -
[7] - Quote
KILLER EI ITE16 wrote:My opinion to getting new players to stay is to add PVE so that then they have something to have fun in, it builds up their skills and isk so that when they want to play PVE they have a better chance.
I agree, PVE is a necessity to let newbies learn the ropes. But they'll still get stomped as soon as they get out of PVE. Better matchmaking is another necessity, that won't just benefit newbies but advanced and vet players as well. |
lrian Locust
Chatelain Rapid Response
Gallente Federation
387
|
Posted - 2014.04.18 22:20:00 -
[8] - Quote
A'Real Fury wrote:short term base it off monthly kdr. the elite players are less likely to ruin this stat and keep it low just so they can play the less skilled players. even if they do they won't be able to go 30/1 often without moving up the board.
Pre-formed squads go into battles based on the 2 highest kdr players in the squad if there are 4 or more people in the squad. A squad of 3 or less then it is based on the kdr of the single highest player.
KDR doesn't say anything of how well a merc performs. A redline sniper can have a huge KDR, but hardly contribute to winning a match. That's why win/loss ratio or even average WP/match would be a better metric. |
lrian Locust
Chatelain Rapid Response
Gallente Federation
387
|
Posted - 2014.04.18 22:28:00 -
[9] - Quote
Delta 749 wrote:Theres a world of difference between a guy having a 15% bonus over you and a guy having a 15% bonus along with mods that give more health and a suit with more slots letting him stack even more above you
That's right, but if he consistently performs better, he'll be matched against higher level mercs, or more lower-level mercs. |
lrian Locust
Chatelain Rapid Response
Gallente Federation
388
|
Posted - 2014.04.18 22:31:00 -
[10] - Quote
IceShifter Childhaspawn wrote:Biggest problem I can see with your fix is that a queue synced team on comms in a private chat will beat it.
Secondly we have never seen Scotty actually work. So we dont know if its broken. There are not enough players to support matchmaking.
Even players consistently outperform others because of a privatce chat, they'll get a higher win/lose ratio, and they'll be put against teams with an equal win/lose ratio. You can only game the system for a limited time. |
|
|
lrian Locust
Chatelain Rapid Response
Gallente Federation
388
|
Posted - 2014.04.18 22:32:00 -
[11] - Quote
IceShifter Childhaspawn wrote:Secondly we have never seen Scotty actually work. So we dont know if its broken. There are not enough players to support matchmaking.
As long as Scotty puts 2 synched teams against a bunch of randoms, or 2 against 16 players, I'm pretty sure it's broken.
|
|
|
| |
|