Discuss this with me, please. (3 game streak, just like this)

123
Comment below rating threshold, click here to show it.

Sydrian

Member

02-03-2013

Quote:
Originally Posted by Seminole Sun View Post
This is a really good point. You are the only common thread in your games. Yes, that's cold comfort in any individual game. But over the long run you will be in the place you should be.

The problem with basing on performance is that it would require a different scoring algorithm based on role. Heck, even within a role, it would be different for different champions. I frequently go 2/6/20 with Malphite because I build tanky enough to survive awhile but enough AP that I can't be ignored. So I'm happy to die in a team fight in which we got 1-3 or 1-4 because I made a clutch initiate.

And once you score differently based on role, you lock in a specific meta. Can't play two ADCs. What about a double jungle? Nope. All out the window.

The other problem is that your k/d/a is a pretty bad indicator of how good you are at League. I suppose it's a good indicator (along with CS) of one type of ability but there's so much it doesn't measure. On top of that, it would encourage more bad habits. Solo queue already has enough people that focus exclusively on themselves. Adding a further incentive for that behavior is not a positive change.

I'd like to see IP be rewarded partially based on your performance, however. I think it would take some of the sting out of a rough loss when you found yourself with 300 IP instead of 150.
That comes back to that the sample here tells us... that 75% of the information contributing to Elo is corrupted due to serious mismatching. Obviously... that's super high and probably too small a sample to give us definite numbers. In my entire ranked career of league, I think the mismatching is actually closer to 65-70%. That's wins, and losses that i think were not properly matched. (Meaning one team crushed the other, rather than the struggle game that it's supposed to be.)

A large volume, in the Elo system CANNOT be accurate, if the data corrupted by poor match-makings, exceeds 50%. It's not a difficult concept to grasp. At higher ratings, obviously... the numbers are more accurate, because matches are much closer, more often. But the bottom of that mountain, is a eroded horrifyingly broken mess.

It would be REALLY simple for Riot to prove it works, or anyone to prove it works as intended; and that would be to create a scientific sample for everyone to see. Testing each of the tiers of ranked play, with say 10 or 100 people, providing the reports with the important information and replays. A scientifically acceptable sample size would be a minimum of 100 ranked games. Easiest to do these in increments of tens. Given the size the community, it would actually be better to test 1000 people or even 10000 people... but such a group gets a little ridiculous.

If I were Riot, I would've done this at least 2 years ago, and thrown up huge banners over my forums and website leading to this information to dissuade all naysayers to the system and settle that it works. Why doesn't such a report exist? Certainly enough time has passed for it to be created. I think someone actually did employ a scientific method some months ago or even last year to actually test the system. As I recall... Their testing accounts were banned and their actual account was banned. I do not recall the results of their test.

[EDIT] In point of fact... I'm sure they store a VERY large library of the game histories(Far beyond your last 10 games). It would not surprise me if they have information on nearly every game back to the start of LoL, given that their corporation is still intact.

I think I'm done here... Gonna get back to doing whatever it is I do when I'm not wasting my time on video games.

Have fun with your future games.


Comment below rating threshold, click here to show it.

NotSoSuperMan

Member

02-04-2013

To prove how wrong you are I've knocked up an Elo simulator.

Elo Simulator


Comment below rating threshold, click here to show it.

Hippodotamus

Junior Member

02-04-2013

If ure standings are supposed to be based on k/d/a then it doesn't really make sense. A game needs an ultimate goal. The enemy nexus is your goal. Not the no. of kills on you. It's like running a raid on BT, going halfway, wiping and getting a Legendary weapon just because you healed well. It doesn't really work that way. I know it's hard when there are trolls and people on your team who suck. But don't tell me that you've never gone through a game when you've sucked the most. Remember, you win some, you lose some. Be happy that you were the best in the game. And if you are indeed the best, then you will not stay in a bad division for long. Peace.


Comment below rating threshold, click here to show it.

Jesustim

Senior Member

02-04-2013

you shouldve tried a lane swap for your bot lane, kat and talon doing instant gap closes is a phenomenol early kill and zoning lane. you could have sent garen 2v1 against them or swapped garen and nidalee to give them a bigger threat when they jump in.


Comment below rating threshold, click here to show it.

Grundle69

Member

02-04-2013

If your so butt hurt about the uncontrollable actions of your four teammates, then play a single player game.

you get bad teammates/team comps sometimes, just get over it. There's no hidden factor or algorithm that can explain why you feel you get crappy teammates all the time, its just the luck of the draw sometimes


Comment below rating threshold, click here to show it.

Sydrian

Member

02-04-2013

@Hippodatomus Unfortunately... I will be stuck here, because that very game, cost me a division. Sure I'm happy I performed well under bad conditions, but I'm unhappy, because in a rated/ranked, competitive game; I lost points on account of a minimum of 2 people who didn't function at all, and a third who functioned VERY poorly. The only one in the list I can't attack is Garen, because he had give and take going on. Ultimately, I still feel he did poorly too though, because each death is worth a LOT of minions, and he gave up as many as any other member of the team.

I am a player who doesn't measure performance by kills, I measure them by deaths. In a 24 minute game, 26 team deaths... is terrible. Equates to more than 1 death per minute. Let me continue by saying we had 8 champion kills. Which means... The 1 kill per minute advantage the enemy had, was negated once every 3 minutes. If I apply these numbers together, the enemy team, in addition to creep farm, was receiving 0.7466666 champion kills, every minute; If you want to know how much of that they received from me, it was 0.083333 every minute. Which means... 0.66333 Champion kills per minute, was being handed to them, by my team; Two of the others on my team, gave up... 0.20833 Champion kills per minute/per player, and the final two.. Gave up 0.291666 champion kills per minute.

It is FACT, that your deaths. are your own responsibility. Regardless of other players, your map awareness, your efforts to defend your champion are the last and ultimate line of defense, for your death count. A responsible and capable player knows this, and will employ this knowledge. Situations vary in difficulty, but a capable player can adequately protect his own champion's life either through pre-emptive retreat, ward coverage, other precautions most reasonable players take to defend themselves. The defense isn't always perfect, but a player working effectively and efficiently to keep their champion alive, will do just that. It is courtesy to warn other players that an attack on them is imminent, and there is no reason not to obviously, since every death on your team constitutes a considerable advantage gained by the enemy.

It is common knowledge that champion kills make up a LARGE chunk of player advantage in games, they are the easiest to obtain because most of the time, they result from player error. I.e. my 2 deaths were attempts to rescue other players, ended up costing both of us deaths. The other player's error was in failing to retreat when ward coverage revealed the enemy's presence and clear pursuit; My error was in thinking I could successfully save him and myself, or even just save him at the cost of myself or in my execution of said rescue. Regardless... There were errors, and those are only examples of errors, taken from the game sample I provided before. There are any number of errors, offensive play errors and defensive play errors.

So... to recap this...

Unit: Champion Kills, per minute.

Enemy team advantage: 1.08 ~ Kills per minute.
26/24

My contribution to that advantage, 0.08333 Per minute.
2/24
=========
The enemy team advantage I was fighting to resist = 1~ Kill per minute.
24/24
=========
2 Players on my team, 0.20833 per player.
Total 0.416666 for those two.
5/24 *2.

2 final players: 0.291666 per player.
Total 0.5833333 For these two.
7/24 *2.
--------------------------------------------
@Notsosuperman. The simulator is something I can appreciate, course...I will look at this tool, and compare it to my experiences, from that I will determine how accurate I find it to be.

@Grundle
So... What you are saying Grundle, is that the luck of the draw then dictates my progress? The simulator I can actually appreciate, because it's constructive information for me to work with. I can't help but feel like... you're ignoring the issue completely, in order to attack me here; Get past targeting me, and actually concentrate on what the discussion is about.

Sometimes is not 75%. It's not 50%. I would accept sometimes, if the number was closer to 10-20%. That might qualify, as sometimes.

@Jesustim Those are arguably viable solutions, that could've changed things. I did suggest I move to bot and send kat mid, didn't take. I believe my words were along the line of, "This isn't working... We have to change the way we are doing this, or we'll lose." With more typing errors, because I won't take time to correct it during game... But, here we are.

The bottom line is, the system isn't working for me. It works properly for one person I know; and my list of contacts that play league is upwards of 80. Skill levels are plain to see in games. I can't judge whether or not the people against me in that game belong with me in group, but I know that at very least, 3 of the players matched me, do NOT belong matched with me, or the people we were against. And because of that skill difference, I didn't stand a reasonable, or even fair chance to protect my division standing.

How do you know when a system like matchmaker functions properly? Fundamentally, no victory should be easy in this system, because you are supposed to be matched with people adjacent or near to your rating(And by extension, your skill level.) That isn't the case, there are landslide victories/defeats left and right. Landslides are indications of the system not functioning, as it was designed. I suppose however I should define landslide as an adjective to make it clear.

When describing wins losses here; Landslide is said to mean that one team, stood small or minimal chances of victory/defeat. Example: See page 1. That's an example of one such victory/defeat.

And it's easy to speculate what happened or could have happened to defend the argument against mine for this particular instance, but stay to the subject. Work with what is here, if you insist on continuing. If you are here, it's not to attack me. It's to discuss Elo. Just the same, I'm only here to hear what you have to say, apply it if useful, listen to it if constructive/respectful, or ignore it if belligerent and destructive to the conversation.


Comment below rating threshold, click here to show it.

NotSoSuperMan

Member

02-04-2013

If you're interested here's a Elo Simulator with a match making system that tries to keep all players on both teams within a 100 Elo range (Expanding in 100 Elo steps).

It shows that a decent matchmaker improves the speed at which players find their true ranking.

Elo Simulator


Comment below rating threshold, click here to show it.

SammyChestBrah

Junior Member

02-06-2013

mate I'm having a similar problem that seems to get brushed aside for a lot of players now.. I'm not here to brag about skill because I know I'm definitely not remotely good enough to be in gold class etc.

HOWEVER, my main mechanics of this game like tactics, lane control, reactions and map awareness are not bad. I try to stay positive but it pains me to play this game sometimes when you get people who don't speak, don't ward and feed then say "whatever I don't care"
I'm not a hardcore gamer, but I find I'm losing interest in this game because ranked ISN'T FUN ANYMORE. I'm stuck in bronze with nowhere to go and as hard as I try I don't think I'm getting out.

Would love to see something done but as always Riot will please the masses over us guys. Really hope the next teemo skin is worth it ^.^


Comment below rating threshold, click here to show it.

f3zz1k

Senior Member

02-06-2013

I agree there is something wrong here... Some games I can take a 1v5 fight.... and win.... How the heck am I paired against people so incompetent? Sometimes I'm the only one on my team who doesn't have at least 10 more deaths than kills... same here, how does that happen?
So tired of people who are playing their 3rd ranked game, ever, coming into a game complaining they are gonna troll for whatever reason... then just get raped by the enemy team... or having to watch it happen to the other team. Just sucks imo.


123