Loading...
  OR  Zero-K Name:    Password:   

Resigned/exited players shouldn't gain elo

66 posts, 1677 views
Post comment
Filter:    Player:  
Page of 4 (66 records)
sort
Title. This would fix a lot of stuff, including Firepluk's elo problem,

Edit: they still lose elo if team loses
+5 / -1

4 years ago
Totally agree, it doesn't make sense. I'd even say they should lose elo even if their team wins.
+0 / -0
Please don't encourage players that have given up on the game to stay in the game and wasting their resource share just so they are not penalized by leaving.

Firepluk despite resigning often contributes to his teams victory, his rank reflects that his actions significantly contribute towards a teams victory. I would much rather have him on team to do his thing and resign if it doesn't work out, than have his rank not be reflective of his impact on a game.

If you leave out his contributions then his rank will unbalance games.
+2 / -0
4 years ago
If you don't play you shouldn't be rewarded, end of story really.
+0 / -1
4 years ago
You can't determine that by checking resign/exit. Someone can play for 90% of a game and resign at the end.
+3 / -0
4 years ago
maybe get reward proportional to amount of time played, then?
+0 / -0
I expect many players would abuse this to keep their rating arbitrarily lower than it should be, so as to end up with more skilled teammates and less skilled opponents in future games.

quote:
If you don't play you shouldn't be rewarded, end of story really.

Remember that not everybody feels "rewarded" by gaining rating.
+10 / -0
4 years ago
quote:
I expect many players would abuse this to keep their rating arbitrarily lower than it should be, so as to end up with more skilled teammates and less skilled opponents in future games.
If people want to do that, what's stopping them to do it now? (ex: Firepluk resigning after his comm dies, although he could help team even without comm).

I like the suggestion because I feel that now there are players that resign fast/at first sign of problems, which can snowball to an all team resign, and then it gives the wrong information to the WHR system. Extreme example of such game: https://zero-k.info/Battles/Detail/830413 (the game was won by patrician alone, most others resigned after around 13 minutes of a 29 minute game).
+1 / -0
ye he rushes paladin behind me while i we have to hold front alone. getting gang-banged by many players is definitely not fun. and now you come and tell me i give up and let my teammates alone.
Do you really think any change in elo would change our behaviour? i resigned under the assumption that a) we would lose anyway, so why bother, b) that my teammate used me as a meat-shield for their personal shenanigans. If you don`t want us to resign, maybe communicate or play the game.
+0 / -0
Firepluk
Yea, elo is a curse of Mighty Spring Gods that brings you nothing more than suffering and useless half potato teammates :D
We need more ways to pluk those nasty elons away :P
so +1
+3 / -0
4 years ago
quote:
Do you really think any change in elo would change our behaviour?
No, and it is not the purpose of the suggestion.

Do you think is fair that someone who played 29 minutes, gets same credit as someone who played 11 minutes? In my opinion, you are free to resign the first minute, but if the team wins somehow, you should not gain same ELO/WHR, because it would not be fair for the ones that stayed and put the effort.
+1 / -0
4 years ago
What if the dude who stayed 29 minutes played like crap and the dude who stayed 11 minutes played like gold?

What if the dude who stayed 29 minutes was playing Simcity in the corner, and the dude who stayed 11 minutes was holding the frontline by themselves?

Its just not possible to determine who was good and who was crap based on mechanical attributes such as time spent in the game.

I feel the only thing that would work for this purpose would be some kind of post-game rating system in the style of League of Legends or so, where at the end of the game you can rate your teammates, either giving them a thumbs-up or thumbs-down (or just no opinion).
+4 / -0
quote:
What if the dude who stayed 29 minutes played like crap and the dude who stayed 11 minutes played like gold?

So how was the game won? Somebody must have played like gold 29 minutes, more than double like the dude who played like gold 11 minutes. Is it correct both of them to get the same credit?

quote:
What if the dude who stayed 29 minutes was playing Simcity in the corner, and the dude who stayed 11 minutes was holding the frontline by themselves?

So how was the game won? After the guy left at 11 minutes the Simcity won the game? (doubt it - more probable there is another player that is able not only "to hold" but to "win the game"). So whoever wins the game deserves more credit that someone who abandons a winnable game.

It takes effort to play a game which you do not know you will win. I like to make that effort, but lots of people quit at some sign of problems. As soon as they quit I know that if the rest of the team struggles will succeed (even if improbable), everybody gets credit.

+0 / -0
4 years ago
Teamgames are a thing. Also, the objective of the rank system is to reflect skill, not effort.
+3 / -0
4 years ago
My opinion is that the less you play in a game (time wise) the less impact your skill has on winning the game, hence less impact a win should have on your rating. Do you disagree with this statement?
+0 / -1
4 years ago
Yes.

1. Time spent in game does not linearly correlate with contribution. If all your things were just annihilated, you will be useless for a while. Time spent ingame when you have a lot of things is more valuable than time spent ingame when you have nothing.

2. Time spent in game can be measured objectively but skill cannot, meaning a system like this would always be missing a vital component of the equation and thus be arbitrary.

I maintain my position of that only humans can judge someone's contribution to a match, an automated system relying on things such as time spent in game or game stats cannot.
+4 / -0
4 years ago
As evidenced in this thread, often not even humans can judge contribution.
+0 / -0
4 years ago
quote:
1. Time spent in game does not linearly correlate with contribution.

Proportional is not linear, and there are many ways to model a correlation.

quote:
2. Time spent in game can be measured objectively but skill cannot,

Skill is currently approximately measured by WHR based only on win/loose of games in which you participated. Do you think adding more information to the system (length of time played) would make it more or less arbitrary?

quote:
an automated system relying on things such as time spent in game or game stats cannot.

The method would be: make a model in which time is taken into account based on some formula (various options can be tested), and call it WHR-Time. Then you take the existing information about games and feed it, game by game to both WHR and WHR-Time. For each game you record which of the system made a better prediction (based on all the games information at that point in time). A better prediction can mean algorithm was correct about winning team (can be also by how much). At the end you check which algorithm is better.

I am not aware of any analysis of the performance of the WHR system (how many games it predicts accurately), but my impression is that generally is doing an ok job, so would be curious to see actual proofs (if you know about any) that "not even humans can judge contribution" (while I see that the algorithm seems to judge ok the contributions).

+1 / -1
4 years ago
quote:
Proportional is not linear, and there are many ways to model a correlation.
Gonna have to explain this phrase better.

quote:
Skill is currently approximately measured by WHR based only on win/loose of games in which you participated. Do you think adding more information to the system (length of time played) would make it more or less arbitrary?
It would make it more arbitrary. Win rate obviously correlates to winning chances, time played, as discussed, does not correlate enough.
+0 / -0
4 years ago
quote:
Gonna have to explain this phrase better.

Example (not sure is the best solution, but to explain the principle): you get x*x of the gain you would get if you would stayed the whole game, where x is the proportion of time played (so if you stayed 0.1 of time you get 0.01, if you stayed 0.5 of the game you get 0.25, and so on). Various other formulas can be devised.

quote:
time played, as discussed, does not correlate enough.

That is a very generic statement to which everybody can have an opinion but until it is tested we don't know for sure. I am not sure it is worth the effort in implementing for the difference in ranking, but I do think there will be cases for which it will make a difference. Example: if you resign in the first minute, and fight takes more than 10 minutes, your skill does not influence the outcome. I have seen this cases, for example people that get in the game by mistake and then leave because they can't play, but the exit vote does not succeed... Or people that stay afk for the whole game.
+0 / -0
Page of 4 (66 records)