Tired of being demoted because of bad players and trolls. Rank system needs a complete overhaul

Silver Strings··6 votes·2,817 views

LP needs to stop being solely based on whether you win or lose. That's the main point of what I'm about to say. At one point I was s2. I have been demoted all the way down to the bottom of s5 because of the teammates that I have been getting. Yes, I have my off games and I will claim full responsibility for that, but when nearly 90% of my lost games have been because of other players, and I'm the one getting punished for it, there's a big problem. I would propose that the rank system is based on your in game performance. So that bad players can't coast into higher ranks for free and better players aren't punished so severely when a loss isn't their fault. Before the argument comes up that, "what if people go afk or leave so they start dying less?" leaving or going afk, feeding, or just stopping helping the team would result in the maximum penalty and possibly a short ban from ranked. I think this would encourage some players to be less reckless and others to be more determined. The base idea would look a bit like this: 100 cs = 1 LP 5 kills = 1 LP 10 = assists 1 LP 5 deaths = -1 LP firstblood = + or - 1 LP depending on which side of it you're on dragon = 1 LP baron = 2 LP loss = -10 LP win = +5 LP tower kill/assist = 1 LP Win streaks grant +5 LP for each win starting at 3 wins Loss streaks cost -5 LP for each loss starting at 4 losses. LP would cap at + or - 30 Being in the lowest rank and having a loss streak in excess of -20 LP each game would result in a temporary ranked ban.

I did the math with a few of my games to test how this would turn out.

10/11/10 with 134 cs 1 baron 2 dragons and I assisted with 6/9 towers. +1 for cs +2 for kills +1 for assists -2 for deaths +2 for dragons + 2 for baron +5 for winning +6 for towers = 16 LP for a fairly average and balanced game which is about what the average is per game anyway.

6/5/8 with 170 cs and 3/4 towers +1 for cs +1 for kills +0 for assists -1 for deaths +0 for dragons + 2 for baron -10 for losing +2 for towers = -5 LP for a fairly average game. It balances out decently well compared to the flat 15+ loss on average with the current system.

25/5/2 with 205 cs and 6/9 towers +2 for cs +5 for kills +0 for assists -1 for deaths +1 for first blood +2 for dragons + 4 for baron +5 for winning +5 for win streak +6 for towers = 29 LP for an exceptional game.

Edit: I would like to point out that I only got this score because this particular team worked well together. It wasn't just me getting fed. While I got the most kills, the rest of the team was not starved either

Unfortunately I could only find one loss of my own that balanced out with less than -10 at an even -15 due to a bad loss streak and my ADC leaving the game which in my opinion serves as proof of this systems effectiveness. People who work with their teams and fight for objectives and play the game as it's meant to be played aren't punished as severely as those who purposefully ruin games.

To reiterate what was said earlier, leaving a game, going afk, or being reported for legitimate reasons will result in an automatic -30 LP for that game as well as a temporary ranked ban for repeat offenders.

Obviously this would need some improvements and changes from riot to make it more reliable but as a base idea this has a lot of potential.

EDIT: I do not think that this system is at all necessary in team ranked matches.

23 Comments

Deep Terror Nami3 votes

Having good scores does not equate to helping win the game. The only point that matters is the Nexus kill, so unless what you did accomplished that, it's all irrelevant.

Rewarding players based on their own score instead of the win promotes greedy play that, as I already said, is not the same as winning the game; as a matter of fact, that kind of play results in losing. This is a team game.

Silver Strings1 votes

the problem is some people will not work with their teams regardless. you're right. being a selfish player will not win you a game. i won my most recent game because of a selfish jungler. in that case, he isn't rewarded. they still lost and he lost LP. you cannot progress while still losing. which is the whole point. you're not rewarded for losing in any way. it does however lessen the blow. people who play with their teams will win more often and be rewarded for it. what you seem to be ignoring is the fact that all objectives and player kills will contribute to taking the nexus. playing selfishly or going for kills and dying, which i see often, will balance itself out.

this idea isn't perfect. i recognize that. but its an idea. and its more than riot is doing

Silver Strings1 votes

in fact i did the math with a lot of my lost games. none of them rose above 0. even the ones where my personal score was rather high.

TrulyBland3 votes

[{quoted}](name=Silver Strings,realm=NA,application-id=cIfEodbz,discussion-id=4VLnrsBJ,comment-id=00000000,timestamp=2015-10-08T14:29:57.548+0000)

the problem is some people will not work with their teams regardless. you're right. being a selfish player will not win you a game. i won my most recent game because of a selfish jungler. in that case, he isn't rewarded. they still lost and he lost LP. you cannot progress while still losing. which is the whole point. you're not rewarded for losing in any way. it does however lessen the blow. people who play with their teams will win more often and be rewarded for it. what you seem to be ignoring is the fact that all objectives and player kills will contribute to taking the nexus. playing selfishly or going for kills and dying, which i see often, will balance itself out.

this idea isn't perfect. i recognize that. but its an idea. and its more than riot is doing

The problem is, that one single action will hardly ever determine the outcome of the game, so even if winning is the main contribution towards points, it can still happen that the logically best course of action (from a selfish point of view) is detrimental to the team. Let's say the enemy nexus is already pretty damaged, we can definitely end the game after the teamfight is over, and there is only one enemy left alive that you are 100% certain you can kill... then simply because the ranked system is a relative system, the best idea is to get your teammates to die, before you finish the game.

That's why Riot is not doing anything in this regard. The goal above all other things has to be to win the game, not to optimize your points.

Silver Strings1 votes

I don't disagree. winning is always the ultimate goal. although im not entirely sure what you mean by this: then simply because the ranked system is a relative system, the best idea is to get your teammates to die, before you finish the game. I don't see exactly where getting your teammates killed is ever a good thing

TrulyBland2 votes

[{quoted}](name=Silver Strings,realm=NA,application-id=cIfEodbz,discussion-id=4VLnrsBJ,comment-id=0000000000010000,timestamp=2015-10-08T14:55:29.876+0000)

I don't disagree. winning is always the ultimate goal. although im not entirely sure what you mean by this: then simply because the ranked system is a relative system, the best idea is to get your teammates to die, before you finish the game. I don't see exactly where getting your teammates killed is ever a good thing

I made that point a bit hasty; considering somebody had to get those kills as well that point was in fact flawed.

Let me try again, this time not in champion select :) Let's say I can either save my carry by sacrificing myself (-0.2 points difference), and let's say that by letting my teammate die, I lower our chance of winning by a percent (statistically: -0.15 points difference) then the obviously better choice for me is to let my teammate die.

Of course, you won't do maths like that; and you can't really estimate it this precisely either. But generally, in situations that are not decisive to the game, I'll be more inclined to act selfish in order to optimize my points. Let's say the game is practically lost and I don't see any chance of winning it. The perfect way of dealing with a situation like this under your system is to wait until an enemy is almost dead, then killsteal him; while never engaging in teamfights so I don't die.

But wait, there's more: What if I play a tank, and each fight ends with me dying, because the enemy team for some odd reason decides to use all their available ultimates on me... then they proceed to lose the teamfight (naturally). The enemy team does a mistake, but I have to pay for it. I will, even if I did everything correctly, receive less points than my teammates. But hey, you could argue, it's your enemies' job to make you lose... okay, let's swap that. I make the most amazing initiate. I manage to cc their entire team, soak up a few skillshots and... die. Because my teammates didn't follow up. My teammates get away, I get -0.2 points. And there are more scenarios like this, where you are punished for other peoples' mistakes. At least, under the current system, if my teammates fuck up I only get punished by the same amount as them; instead of being punished extra.

But that's not all... If we use the system you are suggesting, you are obviously encouraging kill steals; and thus champions that are easy lasthitters or commonly accepted to get kills. Don't even get me started on how unfair to supports it is that you are rewarding cs. Your system also favours certain types of champions across all classes: Some simply have easier ways of gaining assists. Large AoEs or AoE-buffs are pretty awesome in that regard.

And lastly: You are at all times favoring late game champions. Why? Because lategame champions will make the game last longer and can, often, risk intentionally not yet killing the nexus, in order to rack in some more kills.

I could on for days like this. This is not the first time this has been suggested, and I'm 100% certain Riot has seen these suggestions before. The reason why they're not implementing them is always the same: They are impossible to balance perfectly and thus will always have weak spots where people are treated unfairly or worse even, where the system can be exploited intentionally to get an unfair advantage. The best way to create a fair system is by creating a simple system. One that takes only what matters into account (how high are your chances of winning with an equal team against a team with X MMR?), one where the only variables are randomized and thus equalized over time. Say what you will, but the current system is statistically fair. And personally, I value long term accuracy and fairness over short term sort-of-accuracy.

Silver Strings1 votes

you're the first person to actually have a logical argument. and I respect that. I of course do see the flaws. I knew of them when I was suggesting this. what I think is my biggest personal problem with the current simple system is that it is inconsistent. I could win 2 games for a total of 30 LP and then lose one for a total of 27. and that has happened. the inconsistency is quite possibly what ruins it the most.

Echo Kazem3 votes

CS, Kills, Objectives, and any other components of the game are only important insofar as they allow you to destroy the enemy nexus before yours is destroyed.

Any system that rewarded/removed LP based on anything other than W/L would be defeat the entire purpose of the game.

Silver Strings1 votes

how would it? how would rewarding players for doing things that contribute to winning the game defeat the purpose of the game.

Echo Kazem1 votes

[{quoted}](name=Silver Strings,realm=NA,application-id=cIfEodbz,discussion-id=4VLnrsBJ,comment-id=00030000,timestamp=2015-10-08T15:27:43.085+0000)

how would it? how would rewarding players for doing things that contribute to winning the game defeat the purpose of the game.

your assumption that these things contribute to winning all of the time is incorrect. cs, kills, and objectives can all be obtained in ways which hurt your chances of winning. additionally, there are ways to win without necessarily doing well in any of these categories.

consider proxy singed (think DruidDroid). a typical gamescore for singed would be something like 0/15/0. conversely, a top laner going against proxy singed may end with a score like 10/2/10 and a high creep score. however, proxy singed's win rate in the right hands was well above 60%. (in fact they instituted a number of changes to the game to deal with him)

split pushing comps are similar in this regard. if a Nasus is 2/10/2 but he takes a turret on his own even when the other team sends 2+ members to deal with him, opening the other side of the map for his 4 team members, is he really having a poor game? likewise, is the 9/4/9 top laner who went against him but was unable to stop him from taking objectives actually having a good game?

to give you one more example, think about Renekton vs Wukong. Renekton is a huge lane bully and massive counter to wukong, so you would expect him to out cs the wukong (and probably end up with a good number more kills/assists in laning phase) but Renekton falls off later in the game while Wukong scales very well. Often times in lower elos the Renekton player will win lane hard but be unable to transition this into other objectives around the map/gold for his teammates and as such he will get outscaled an eventually lose the game.

even consider a 21/7/10 Leblanc. If 5 of her deaths were because she got caught by the enemy team, causing her team to lose many objectives and eventually lose (which is especially likely if she got caught this often because she probably has most of her teams gold), did she actually play well?

it would also be wise to consider that any system based on more than w/l will add some criteria to judge people on (in your case cs/kills/objectives). these aren't the only criteria that decide who wins the game though. by creating a system which relies on these criteria, you're effectively stating that these criteria are more important than others which is simply not the case.


the rank system measures how well all of your decisions/skills affect your team's ability to win the game. if the sum of all these does not lead to you winning more than you lose, it means that your overall gameplay is weaker than that of the peers at your elo. it's entirely possible for you to win your lane more often than not and have higher KDAs than average but to still be a well below average player (relative to your elo). if you find this is the case, I would focus on improving your gameplay and asking yourself why you lose often even when you have high KDAs/cs scores as opposed to asking for a new ranked system.

Silver Strings1 votes

you have literally ignored half of what i have said. objectives are the most important things to get. and if you're getting a lot of cs, 7/10 times thats because you're pushing for a tower. why do you think i said that every 5 kills vs ever 1 tower dragon or baron. and in order to get these you have to work as a team. also lane pushers like tristana and nasus, proxy singed like you said, wouldn't be punished by this because they would be pushing, getting cs, and taking towers. objectives are the most important criteria of the game with the nexus being the main objective, which you cannot get without taking the others first.

and no honestly the ranked system only loosely measures how your decisions effect a game. a support can make every play in the world but that will not pull every team together. one man cannot carry 4 no matter how hard they try. it requires all five to make a winning game. that includes ganks and lane rotation. not every game requires a high kda. but every game requires a minimum objective score. but then think again. how hard is it actually to take a tower when you don't have any farm or any kills or assists, meaning you have no items and have to push the lane slowly and the tower takes forever. and as well are prone to ganks? how important is the KDA then? you are trying to completely discount the value of doing well personally. or especially in the case of bot lane, working with your lane partner.

Jubbinaut2 votes

[{quoted}](name=Silver Strings,realm=NA,application-id=cIfEodbz,discussion-id=4VLnrsBJ,comment-id=0003000000000000,timestamp=2015-10-08T17:43:38.025+0000)

proxy singed like you said, wouldn't be punished by this because they would be pushing, getting cs, and taking towers.

Example win:

+5 (win) +5 (towers) +4 (cs) -3 (deaths ~15) -1 (fb) = 10 LP.

Example loss:

-10 (loss) +3 (towers) +3 (cs) -3 (deaths ~15) -1 (fb) = -8 LP

And this is the problem: yes, it's a gain, but it's less of a gain than other strategies. Which means the system itself would act to remove perfectly valid and effective strategies, because - despite high win percentages - they undervalue statistics that the system prizes.

That's meta-locking. And Riot doesn't like it.

EDIT: I just realized, both of my calculations there assume that the Singed got takedown credit for his towers. Which actually isn't all that common - a large portion of what made it work was simply allowing the minions to shove towers by themselves, by preventing the enemy minions from ever reaching them. So the gain on a win could be as little as +5, while the penalty on a loss could be as much as -12.

Silver Strings1 votes

riot doesn't like it but lets all be honest with ourselves. it happens anyway. thats why people get so angry when you try to play anybody "out of place"

Jubbinaut2 votes

There's a difference, though. And it's an important one.

Currently, the meta exists as a function of champion strength and play styles - it consists of what's most effective in a particular match, based more-or-less exclusively on those two things.

With a system such as you suggest, the meta would be a systemically enforced thing. It's true that different champions may rise and fall based on overall strength, but the available strategies would be significantly limited based on what offers the best LP gain on a win and least LP penalty on a loss.

Specifically, given the way you've weighted things, we'd be looking at a focus on strong duelists, assassins, and burst mages. And pretty much nothing else.

Silver Strings1 votes

I did say that mostly as a joke but you bring up a point, but thinking about it, aren't we already a part of that meta? fiora, nasus, riven, and darius dominate top lane. mid you always see yasuo, zed, veigar, diana and ahri. The only lane that seems to have any variance is bot. At least in my opinion.

I sometimes feel like I'm the only one that chooses different champions specifically just to see something new

Jubbinaut2 votes

[{quoted}](name=Silver Strings,realm=NA,application-id=cIfEodbz,discussion-id=4VLnrsBJ,comment-id=00030000000000000000000000000000,timestamp=2015-10-08T21:02:25.287+0000)

I did say that mostly as a joke but you bring up a point, but thinking about it, aren't we already a part of that meta? fiora, nasus, riven, and darius dominate top lane. mid you always see yasuo, zed, veigar, diana and ahri. The only lane that seems to have any variance is bot. At least in my opinion.

What I tend to see top-lane (last two being my own picks) based on my last 4 SR matches:

Gangplank Kayle Lulu Nasus Olaf Shen Chogath

What I tend to see mid-lane (last one being my own pick) based on same:

Fizz Syndra Malzahar Ahri Varus Veigar Riven Annie

You're right in that those champions are popular right now, but I wouldn't say they're all we see.

The difference, though, is simple: are they played because of themselves? Or are they played because an external system rewards their playstyle more than others?

Because one of those is indicative of imbalance, and the other of a profound flaw in the system.

Silver Strings1 votes

Same meta. different reason. and you're right. they are played because of themselves. I for one particularly enjoy ahri's skill set. as well as varus and riven. personally im tired of seeing nasus altogether but thats an unrelated personal grudge. one which i feel most players share. I've enjoyed this conversation; I apologize if I came of as being rude or sarcastic in anyway. Just my arguing style. Mainly since you're one of only two that brought up valid points, based on everything that I brought up. I suppose one thing I didn't account for would be difference in play styles. This system would work phenomenally for me because of the kind of player that I am, whereas it would hurt some others and obviously riot has to cater to everybody

XxobscenexX2 votes

[{quoted}](name=Silver Strings,realm=NA,application-id=cIfEodbz,discussion-id=4VLnrsBJ,comment-id=,timestamp=2015-10-08T14:16:51.220+0000)

LP needs to stop being solely based on whether you win or lose. That's the main point of what I'm about to say. At one point I was s2. I have been demoted all the way down to the bottom of s5 because of the teammates that I have been getting. Yes, I have my off games and I will claim full responsibility for that, but when nearly 90% of my lost games have been because of other players, and I'm the one getting punished for it, there's a big problem. I would propose that the rank system is based on your in game performance. So that bad players can't coast into higher ranks for free and better players aren't punished so severely when a loss isn't their fault. Before the argument comes up that, "what if people go afk or leave so they start dying less?" leaving or going afk, feeding, or just stopping helping the team would result in the maximum penalty and possibly a short ban from ranked. I think this would encourage some players to be less reckless and others to be more determined. The base idea would look a bit like this: 100 cs = 1 LP 5 kills = 1 LP 10 = assists 1 LP 5 deaths = -1 LP firstblood = + or - 1 LP depending on which side of it you're on dragon = 1 LP baron = 2 LP loss = -10 LP win = +5 LP tower kill/assist = 1 LP Win streaks grant +5 LP for each win starting at 3 wins Loss streaks cost -5 LP for each loss starting at 4 losses. LP would cap at + or - 30 Being in the lowest rank and having a loss streak in excess of -20 LP each game would result in a temporary ranked ban.

I did the math with a few of my games to test how this would turn out.

10/11/10 with 134 cs 1 baron 2 dragons and I assisted with 6/9 towers. +1 for cs +2 for kills +1 for assists -2 for deaths +2 for dragons + 2 for baron +5 for winning +6 for towers = 16 LP for a fairly average and balanced game which is about what the average is per game anyway.

6/5/8 with 170 cs and 3/4 towers +1 for cs +1 for kills +0 for assists -1 for deaths +0 for dragons + 2 for baron -10 for losing +2 for towers = -5 LP for a fairly average game. It balances out decently well compared to the flat 15+ loss on average with the current system.

25/5/2 with 205 cs and 6/9 towers +2 for cs +5 for kills +0 for assists -1 for deaths +1 for first blood +2 for dragons + 4 for baron +5 for winning +5 for win streak +6 for towers = 29 LP for an exceptional game.

Edit: I would like to point out that I only got this score because this particular team worked well together. It wasn't just me getting fed. While I got the most kills, the rest of the team was not starved either

Unfortunately I could only find one loss of my own that balanced out with less than -10 at an even -15 due to a bad loss streak and my ADC leaving the game which in my opinion serves as proof of this systems effectiveness. People who work with their teams and fight for objectives and play the game as it's meant to be played aren't punished as severely as those who purposefully ruin games.

To reiterate what was said earlier, leaving a game, going afk, or being reported for legitimate reasons will result in an automatic -30 LP for that game as well as a temporary ranked ban for repeat offenders.

Obviously this would need some improvements and changes from riot to make it more reliable but as a base idea this has a lot of potential.

EDIT: I do not think that this system is at all necessary in team ranked matches.

Amen

Billco1 votes

I don't think the points themselves need reworking, so much as the awarding of those points. Everyone has had the experience where you play really well and get good teams, then hit promos and find yourself with duo mid feeders, one AFK and the last idiot is lagging on McDonalds' Wi-Fi. The result is your promos are almost entirely about luck and you never get the chance to rise out of Silver.

It is not confined to Silver either. I often spectate Plat games, and they suffer from the same problems. At first, I thought these were people who bought Plat accounts, but after looking at their stats and history, they simply have no clue what they are doing but still won against AFKs, ragers & feeders.

Riot needs to take a long hard look at the ranking system, because right now it's punishing a lot of good players while not doing enough to screen out the bad ones. I really should not need to meticulously check people's stats in champ select and second-guess their rank.

Silver Strings1 votes

yeah. my idea isn't perfect. i know that. but it would be a step forward. or at least the bottom line of a much better idea that could spring forth from it. its true. a lot of good players are punished severely against their will

DárkFlameMáster 1 votes

100% agreed just made a post similar. Demotion isn't even fun. This game rewards what other people can do to you. I've told myself I've been bad at league for 2 years now and that hasn't gotten me anywhere. There is no progress, its like playing at recess and never getting punished for knocking kids off the play place.

Silver Strings1 votes

well the real problem is theres no way to balance a bad game for yourself. and theres hardly any incentive for teams to try to salvage a loss.

the other problem is that your W/L does actually tend to be nearly 50/50 but that doesn't promote progress in any way either

Adeptus Terra1 votes

And what would happen to the higher mmr players? In some of my ranked games there are a total of 100 kills (between both teams) so if you are in higher tiers and there are less you would drop mmr. That hardly sounds fair. Im in Silver so i am by no means credible or have a just reason to say this but i would assume you are a low mmr player due to your opinion on how points should be given.

More from [ARCHIVED] Miscellaneous