(edited by Delweyn.1309)
A more accurate rank system
Those stats are an utterly useless way to judge skill at conquest, unless you want to identify afkers, and you don’t need those stats to identify afkers.
The stats are usefull because it will take into account everything :
- Number of decap – Damages – Finishing kills (for thief)
- Time spend at defending point – Damages guarded – Heals – Revives (tank)
- Damages – kills – Damages near defending points (for damages dealers)
- Heals – Revives – number of condi clean – Defend points (for supports)
- If someone hase good place in numerous categories then it means that he caried the game. So more points for him.
- Reduce by number of deaths -
And it’s just exemples. So there is ways to evaluate more accurately the work done by every players in the match.
(edited by Delweyn.1309)
Maybe it’s because i’m on the EU servers, maybe i’m a weird player but i think the matchmaking works with the current system.
Quite honestly the biggest problem i could see is too small a playerbase to draw matches out of and that’s not something you’ll fix by introducing more and more filters into the search.
Sure, your perceived result will be more accurate but in this case that is not something you want to that level.
Your changes would lead to an incredible increase in queue times as the system would have to look all that over and find people exactly in your range, the possible sample to look into though would get smaller and smaller with each round, and, worst case, different characters played by the same person.
If anything A-Net should do a per character rating and, as sadly was voted out, lock the character relog at the start of a round.
The problem I experience is that this system is fluctuating too much. At one point in time you are top 100, then you lose 3 matches in a row and you can be nearly out of the leaderboards. In this way, the leaderboard does not accurately reflect someones skill level, which is actually the whole point of a leaderboard. Besides, get rid of DuoQ.
Maybe it’s because i’m on the EU servers, maybe i’m a weird player but i think the matchmaking works with the current system.
Quite honestly the biggest problem i could see is too small a playerbase to draw matches out of and that’s not something you’ll fix by introducing more and more filters into the search.
Sure, your perceived result will be more accurate but in this case that is not something you want to that level.Your changes would lead to an incredible increase in queue times as the system would have to look all that over and find people exactly in your range, the possible sample to look into though would get smaller and smaller with each round, and, worst case, different characters played by the same person.
If anything A-Net should do a per character rating and, as sadly was voted out, lock the character relog at the start of a round.
Not at all because if the game match us with ppl much higher or lower of our level, because there is no other players, it’s not a problem. In all case him and you will gain points depending on what u do in the game.
With this system, people will less care to be with bad teammates because they will know that if they are doing well, they will gain more points (or lose less).
Maybe it’s because i’m on the EU servers, maybe i’m a weird player but i think the matchmaking works with the current system.
Quite honestly the biggest problem i could see is too small a playerbase to draw matches out of and that’s not something you’ll fix by introducing more and more filters into the search.
Sure, your perceived result will be more accurate but in this case that is not something you want to that level.Your changes would lead to an incredible increase in queue times as the system would have to look all that over and find people exactly in your range, the possible sample to look into though would get smaller and smaller with each round, and, worst case, different characters played by the same person.
If anything A-Net should do a per character rating and, as sadly was voted out, lock the character relog at the start of a round.
Not at all because if the game match us with ppl much higher or lower of our level, because there is no other players, it’s not a problem. In all case him and you will gain points depending on what u do in the game.
With this system, people will less care to be with bad teammates because they will know that if they are doing well, they will gain more points (or lose less).
I don’t think that is a good idea. Simply using these stats is not enough. You never take into account that sometimes moa and portal change the fate of the game. Or stealing a beast, rather than just killing it by yourself.
You can’t make a matchmaking system like that, it has way too many factors, otherwise other more popular and games with more money in the competitive PvP scene would have made it by now. And still everyone uses MMR.
We need more incentive to play PvP → more players → better matchmaking. It’s the same with the esports. You need people who will watch it first, before you can improve the whole scene.
This isn’t a good system because the game will have to know what is considered " a good performance ".
I already pointed the example multiple times : if we assume i’m playing thief and opponents are constantly chasing me, there’s a big chance i won’t fight on points, kill many players, while my teammates will be 4v2-4v3 for the rest of the game. Does that mean i will get less points than them ? Alternatively, is following players the whole game with some OP 0 risk build, leading to defending and fighting a lot, a good play?
We can find many other situational examples, but there will be objective and subjective analysis on this, and the system won’t work.
To me, to make it easier, there are 2 solutions :
- Keep the actual system, but remove seasons : we will avoid luck on placement match, and on the long term, everyone will be at his true rank
- Keep the actual system, keep seasons, but make everyone start on the same point.
The numbers are 100% completely useless, they tell you nothing. I’ve literally never once looked at my numbers after a match, and I’ll happily argue anyone who truly cares about their numbers probably isn’t a very good player.
If you have 2 menders staff druids auotattacking each other for 10 minutes off point their damage and healing numbers are going to be totally off the charts, but they accomplished nothing.
- I don’t care how much damage I did, I care that the guy who needed to be killed was killed. Perhaps my contribution was a clutch interrupt or lockdown.
- I don’t care how much healing I did, I care that I saved my teammate when he needed help (think smoke field, AoE stun break, other types of peels.)
- I don’t care how often I fought on a point, I care that we held points. If we’re holding 2 points I will intentionally fight off point the entire game.
the sistem proposed is useles, S1 system in a pure solo q environement(without smurf acount manipulation kitten whith mixed q) is more suitable:
1 evaluate team
2 system guess a expected outcome
3 reward team who performs better than expected not only winer