Personal score as a replacement to ELO

Comment below rating threshold, click here to show it.

TheCowness

Senior Member

11-20-2011

This is silly, and overcomplicated. Without even discussing in detail how the original post assigns Janna players roughly zero points, you can't really keep track of how well a person did by looking at the small things. There's too much to consider, and a computer can't measure you with that much accuracy. For example, if I'm Morg or Lux and throw my shield on an ally, but no enemy hits them, and there are indeed enemies near us, do I deserve a "point" in my personal score? Were they even in danger? Or is the shield the reason my enemy didn't attack them? A human referee could probably make the call, but a computer couldn't. It seems crazy to try to measure everything a player does in-game, and that'll just lead to things like "Wait don't kill dragon yet; lemme get over there so I get some points." And then the enemy team has time to stop you from getting it altogether. I've lost the chance to kill Baron before because we waited for someone to respawn so he could have the buff, too. I don't want that happening with dragon.


It's gotta stay simple, so that it isn't abused. I think a better alternative would be to just use a formula based on your final k/d/a and maybe a few other things (creep score?) compared to everyone else on your team. You also might have to factor in the champion's role -- supports and tanks get more "points" from assists than carries, maybe, and it doesn't matter if they don't kill creeps. But you'd still have to keep it simple, because some champs can take more than one role. The most important thing, I think, is that you compare to everyone else on your team when judging how you did, because in a defeat you'll have more deaths than a loss, but so will everyone else. You can't compare your own score in one game against your own score in another, because they were completely different games, with different teammates and enemies.


Comment below rating threshold, click here to show it.

mdigibou

Senior Member

11-20-2011

1.) your true elo will take significantly longer to show, and is reliant VERY much upon your team. The insane fluctuations are what is known as ELO hell.
2.) I mentioned getting points for casting allied buffs didnt i? Does janna have a buff? Does janna have a shield? Does janna have a heal? Does janna have cc? Oh wait, she does. In addition the system IDEALLY would assign a point for the shield no matter where you used it. Can this be abused? Yeah of course. Thats where your tweaking ideas come to play. For instance, make it so the point is awarded only if enemies are near. Thats easy to track as well.

The dragon situation ive considered both ways to be honest. By one person killing it, awarding the whole team points seems unfair, but the other situation also seems unfair. Im not sure, but I do think that this is likely the most complex situation that comes up for this system.

The CV situation isnt necesarily bad. the CV bonus isnt intended to make or break your points, only a little flavor so to speak. I think that spamming CV to see the enemy support in the bushes is acceptable. The goal is enhanced map awareness.


Comment below rating threshold, click here to show it.

Coreille

Member

11-20-2011

Quote:
Originally Posted by mdigibou View Post
I think that spamming CV to see the enemy support in the bushes is acceptable. The goal is enhanced map awareness.
Sure doing this helps you, but it is MUCH more usefull to find the enemy jungler or to save cv for a gank to reveal a brush where the enemy might walk into.
And the saving lifes point is still open... This should be worth just as much as a kill (maybe even more if the person you saved is on killing spree) but is impossible to measure.
I really think finding a good personal score is unrealistic. I'd be happy if you could prove me wrong there, but I think your current system is not enough and that any fair system would be far too complicated.


Comment below rating threshold, click here to show it.

MrSpaz

Senior Member

11-21-2011

Quote:
Originally Posted by mdigibou View Post
1.) your true elo will take significantly longer to show, and is reliant VERY much upon your team. The insane fluctuations are what is known as ELO hell.
This is inaccurate on two counts:

1. It would take just as long. The factor here is number of games played, not the metric by which you're ranking players. The end goal is the same in both styles -- you seek an elo at which you're consistently being paired up with players of the same skill level. Matchmaking is gonna take just as long to provide a result here than it will in the simpler victory-based elo.

2. Elo hell has nothing to do with fluctuations. Elo hell is players thinking that they're better than they are, and being impatient with the fact that they have to work their way up over a long period of games.

also ..
Quote:
the goal is enhanced map awareness
it wouldn't enhance your map awareness. You already know the support is in the bush, that's the situation that Coreille set up. If you know Sona's sitting in that bush, CVing her does literally nothing for you.


Comment below rating threshold, click here to show it.

mdigibou

Senior Member

11-21-2011

Too much opposition at this time for me to further this. More eloquent speakers than I are necesarry to further this cause. I'm going to give up for now because the thread is inhabited primarily by people who are STILL incessant upon discussing the merit of whether this is a good system or not and not the ACTUAL topic, which is to discuss how to implement or keep this concept going. Consider it a thought excersize if you don't necesarilly agree with the topic. It's honestly not that hard of a concept to grasp.

Im out, have fun continually trolling, mrspaz. you have derailed a potentially exceptional thought excersize a few pages back. Oh and for the record, see that gold icon under my name? "being impatient with the fact that they have to work their way up". Yeah.... ok. You got it champ. Gold icon... no icon. Hrm, curios.


Comment below rating threshold, click here to show it.

MrSpaz

Senior Member

11-21-2011

Quote:
Originally Posted by mdigibou View Post
raaaaaaaaaage
Man, chill. I backed out of the discussion because you didn't want to discuss anymore, which was fair. It's your deal.

I just wanted to point out two technically incorrect things, because I don't really like people getting away with posting misinformation. It spreads likes a virus on forum environments like this.


Comment below rating threshold, click here to show it.

TrojanHorse

This user has referred a friend to League of Legends, click for more information

Senior Member

11-21-2011

The better team does not always win. Just look at baseball -- it's a game of individuals, but you don't get to go to the playoffs just because you're the best person on your team, you go because you work well together as a team. The more in doubt any particular game is, the more games you need to play to get a representative idea of how good your team is in relation to others.

Now, take that to LoL -- you now join a different team every game. The best players will tend to win more games than the worst players. Why? They are a positive factor towards their teams -compared to the average opponent-. Since your team is completely random, it will take more time to determine how good each player is, but if you play enough games, you will approach your "true elo."

Sports, games, etc. can't be boiled down to a series of stats. If they could be, you could predict the winner of every competitive game. Since it's so complex, the most basic stat of all -- whether you won or you lost over the course of a long period of time -- will determine whether you're a champion or a pretender.


Comment below rating threshold, click here to show it.

IS1879fbf8ece39275db4d9

Senior Member

11-22-2011

I have played over 800 games and I still get people on my team who are not even close to my own skill level. Thank God Matchmaking works.


Comment below rating threshold, click here to show it.

IS1f48700f529fbcca97996

Senior Member

12-09-2011

Yab.


Comment below rating threshold, click here to show it.

Ziramond

Senior Member

04-04-2012

I'm extremely sad that this thread has died, this is almost exactly the same system that some friends of mine and I came up with, though our system was a bit more simplistic (i.e. we didn't think about CV ward placing or neutral buff jazz). The idea behind taking tower damage into effect would also be a bit too complex i think. By simply destroying the tower, either by a champion or a minion, the entire team benefits, so it would make sense to reward the entire team. Same goes with dragon and baron. There are many times when a team chooses to get dragon because both solo tops are in their lanes and the enemy jungler is ganking top lane, this generates a 3v4 fight at dragon, or a free dragon. Why should the top laner not receive the benefit simply because he didn't hit it. All structures should add everyone's total score because by destroying a structure, the entire team benefits.

To the gentleman who thinks that W/L is the better system, i propose this question. Lets assume there was a ladder reset so everyone is 1200. One team has 3 people on it that are, in reality, 800 ELO players, 1 1100 ELO player, and 1 1800 ELO player. Note that these are their proper ELO's and not their actual in game ELO. On the other team there are 3 people that are 1400 ELO players, 1 1000 ELO player, and 1 1800 ELO player. Again, i note that these ELO's are not what is tracked in game but rather what their ELO "should" be given enough time, as so you think. Now, obviously the second team would win, giving them additional ELO and removing ELO from the other team. lets say 12. Now there is a player that should be at 1800 ELO that is now 1188. Like it or not this now makes it more likely for him to be paired with people that of "actual" lower ELO as compared to him, but again whether he wins or loses is completely up to the other 4 people on his team being "better" than the 4 worst people on the opposing team (because obviously the 1800 player will most likely be "better" than the "best" player on the other team). Now, what are the odds him actually winning those matches? Higher? or lower? Tell you what, i'll give you the answer. Neither, it's completely random, and that fact coupled with him now being farther below his "proper" ELO actually removes his contribution to the team entirely, proving that a W/L style system simply doesn't work to properly rate someone's actual skill. Also this is a perfect example of the phenomenon known as ELO hell. It is were you are being randomly paired with people that are worse than the opposing team in reality, but their ELO shows otherwise, and when that happens to that person he gets punished for every loss.

How is that fair?