Reign of Gaming

Quote from qucangel »

Ahri trashes fizz.

Gragas trashes fizz.

Zed trashes fizz.

Pantheon trashes fizz.

Never been beaten by the first three, I would go so far as to say zed is countered by fizz due to fizz's superior all-in and fizz's ability to completely nullify zed's ult with his e, whereas zed can only nullify fizz's ult with his own ult

Quote from C3rtitude »

Quote from StaleTaste »

 

Talon, Akali and Swain are good choices to fight fizz. Diana is a pretty even matchup as well. Also, if you are *very* good at cassiopeia, it can work

For Talon I haven't try him a lot vs FIzz but I had the feeling it could go either way, if Fizz manage to E Talon's ultimate, but I'll give it a try again. 

Concerning Swain, how were the good swain playing? Agressive early on ? What kind of summoners were they using (flash, barrier?). Same for items, that'd be cool if you could tell me a little more about the MU.

A very aggressive swain will beat a fizz 9/10 times. Swain both has a much stronger level one than fizz, and his e costs much less mana than fizz (which fizz has to dodge if he doesn't want to take tons of undue damage.) Because of this, you can easily burn out all his pots and mana and establish lane dominance, which you can transfer over into a lead post6.

I am a hardcore fizz main, so let me tell you what gives me trouble.

Ryze is not a good counter to fizz. Never pick him. Fizz will outtrade you, outroam you, and your team will be finished before 15 minutes.

NEVER PICK KARTHUS OR KASSADIN EVER NEVER

Lb can do nothing against fizz unless he's really really retarded

Talon, Akali and Swain are good choices to fight fizz. Diana is a pretty even matchup as well. Also, if you are *very* good at cassiopeia, it can work

Quote from Scildecido »

I enjoy Fizz top lane. I run tele/barrier but whatever summoners you take is preferable to you. 
He has some pretty good match-ups and he some extremely difficult match-ups.  And then there are match-ups where you have to win pre6 or suffer the rest of the lane. Akali comes to mind. 
When I first tried to play him, I did some wonky tanky build stuffs. 9/21/0 masteries, rush RoA into Rylais, then Lich. That worked out okay. I'm sure that sounds like a completely stupid set up to a lot of you guys. I've lately just been doing what I do if I played him mid. 30/0/0, rush an armguard or double Dorans rings, slide right into that sheen and lichbane. I don't get rod anymore simply because if it's not my first item, I won't play it. People generally finish their first big items around 15-21 mins. That means the rods full effect won't finish until 30 minutes in the game. So even trying to build it after a first item really doesn't guarantee that I'll get the full effect. And besides, people say not to get rod on him anyway =D
It's a lot harder to play the tankier fizz and to be able to snowball like he's known for. He's certainly harder to kill, no doubt. I still build a little beef on him though, I enjoy rylais on him. 

The nice thing about Fizz top is you aren't required to max E on him. It's still really good but fighting melee champs, maxing W first makes him trades better and that means you just keep your E for safety/chase. If you do max E first, just remember how mana hungry it is if you intend to harass with it. I personally max E mainly for the wave clear. People love to push on you if you don't. 

I hope that's a little insight for anyone interested. I'm just a silly silver player after all.

Don't get rylais, it doesn't proc on half his kit (only on the half that already has slows lol). Get roa if you want to build health on fizz, people are stupid saying it isn't good on him

As a hardcore near-plat fizz main, let me say that fizz has many positive matchups top and can run the lane extremely hard if he gets any sort of lead at all. However, you want to be roaming on fizz, because his roam is so exceptionally strong. If you put him top lane, he can't roam as well. So while he is a viable top lane, he is so much stronger mid lane with his lane dominance and roam that you should almost always try to go mid

According to your graph, the "average" league player should be in low silver, which is what we find.

On a sidenote, I'm about to be in the >98% of players :D :D :D

Quote from Yunooh »

i like how this turned into a more constructive discussion

 

i would really like to thank for the link and comments as i feel it nicely complements the discussion and also verifies some of the things i have mentioned (high variance of skill interpretation)

 

there are still a few things which I feel have been overexaggerated

 

it was mentioned that a very good player should overcome the odds the teammates throw at him and various other people have suggested that "its their own fault"

 

this, as I mentioned in my original post, is highly subjective and current data + riots actions do not indicate this (please check the link above, it clearly mentions that under the current system there is a 6.5%chance for a 1100elo player to reach gold)

 

simply said you can only improve yourself and not the system. extreme case scenario even a pro cannot win a 5v1 in LoL (4ppl afk). therefore all a player can do is contribute a certain % to win, we have to acknowledge this fact. for all of you crying now i repeat: skill does matter to influence your win%

 

but

 

- there are many ways your win % can be reduced without you being able to control it

 

- the system can not exactly pinpoint your elo and often mismatches you (30games ttl player & 600games ttl player same team)

 

- the player distribution is not representative of what people have in their heads.

example:

5players, 1beginner (1game played) 1 pro (2000games played), the remaining 3 are split between 1average and the rest between beginner average and the pro.

 

bronze 1

silver 1 

gold 1

plat 1

diamond 1

 

this would mean that 20% of the population is split evenly. now as we all know the pros should be rare so a game like sc2 fixes it as 2% of population (master division) and tries to set the average player in the middle (remember the gauss) which is the gold division in sc2 (roughly 32%)

 

now the problem in LoL

bronze 〜30%

silver 〜60%

gold 〜8%

plat 〜1.5%

diamond 〜0.5%

 

this distribution does not apper to be like the gauss (if you dont believe feel free to recheck with sc2, where the gaussian "bump" is in the middle, namely gold) nor is it I think what ppl except. if its not like the gauss which is also called the "normal distribution" it indicates that something is not right

 

it shows that players are concentrated in the lower tiers.

now before you jump to conclusions like "they all bad" please remember that it is a distribution, meaning that if that were the case even within the baddies you would have better baddies and badder baddies.

 

this all together is why there is a "elo hell" mindset

 

it is not so much that once you are inside you cant get out. it is just as if something is trying to pull you down (statistically distort) which leaves you and the system puzzled (unable to pinpoint) about your true skill level and you feel that within 5 ranks you should be comparably higher than you were last year, IF there are 5 ranks...(when in actuality your true skill level has never been captured without distortion)

 

the perception "i am better than my team (justified or unjustified) and the above all together creates the illusion of elo hell

 

but i would never statistically agree to its existence

 

here i find the last comment from enthalpire very fitting. as almost all players are in silver league it really doesnt matter from a personal perspective, which division you are in (or how often you drop and rise), as it just feels too random (which is statistically the most distorted place to be according to above numbers)

Your percentages are skewed, I think. You're making the assumption that the distribution is not normal because bronze and silver have so many players. However, let's look at the elo ranges for each division.

Bronze: 0-1250

Silver:1250-1500

Gold:1500-1750

Plat:1750-2000

Diamond:2000+

As can be clearly seen, looking at this graph, one can see that the distribution is indeed *very* normal; however, bronze simply encompasses much more of a range than any other area. I would posit that there would be comparable percentages of people playing in low bronze, lower bronze, and very low bronze, as gold, plat and diamond respectively.

One must be wary of how they interpret statistics.

Quote from Yunooh »

i acompletely agree with your statement

 

i would also agree that sample size does matter and that very exceptional players will move faster

 

the problem I see is in sheer randomness of the matchmaking, the inablity to effectively influence the outcome every game. there just simply are way too many random factors which will distort your performance and eventually how the system is rating you. this is most true for the majority of players.

as indicated in my original post, it just will take a lot of time for the system to recognize your true skill level, especially for the not so exceptional player base

 

therefore please note that my argument is not that no skill is involved or that a pro can not influence the outcome

 

what I am trying to say is that for the majority of players the actual experience will feel pretty much random and difficult to influence which leads to the very real feeling of "elo hell"

 

to illustrate what I mean imagine a player whose skill is exactly that of 60% the player base, lets say silver

 

let us assume that we have all the knowledge necessary to judge this and we judge the player to be silver level. the problem is that for that player (who does not have all the knowlege) due to inherently flawed system and randomness, it will not feel that he is in control

 

I,e. even at his skill level ppl will go afk or start insulting which will lead to short term decline in elo/mmr

 

would this player like to advance? yes.

 

does he feel that he can influence it?

no.

 

is he aware of his true skill level?

no.

 

naturally he will feel stuck and not at his skill level. this is for me also the main reason riot implemented the new ladder system away from elo, so players have a "feeling" of more progress

 

I think this is true to some extent. I will posit that, given a person's true elo is X, at approximately X-100, they will feel stuck as their skill level will be unable to cope with the random variables of the game. However, I think at X-200 elo, or anything lower than that, the individual can carry very easily.

For an extreme example, let's take 0 elo. I got my account reset to 0 elo at the end of season 2, like a few other people. At zero elo, if I played 1000 games, I would win 1000 games. There is no contest there (and likely isn't for anyone in this forum, it is hilarious). At each 100 elo increment, if I played 1000 games, I would lose progressively more games, until I got to my true elo, where I would theoretically win 50% and reach a carrying capacity.

While it is understandable to feel powerless in these situations close to your elo where you feel that other people's actions influence the game more than your own, you can easily feel less powerless if you work to get better at the game, thereby advancing your true elo to a higher point and making it easier to carry games where you are.

Quote from Yunooh »

i think i can agree with the fact that exceptional players can influence the outcome of the match more if say almost everyone is way beyond their skill level. also you mentioned an interesting point which i didnt further specify, an exceptionally good player will also know which champ to choose, which imo is one of the main differentiators. this should make it easier, to move to the tier where the player base is concentrated.(lets not talk about champ op-ness, but let us just acknowledge the fact its there)

but, again. just because you achieved it multiple times proves nothing. if you roll a dice 10 times and you get 7times a 6, you will tend to overerstimate your skill when by definition it was 90% luck (which same as with the dice you simply can not know in LoL).the same applies for players who feel that they are in elo hell. they roll 7times 1. naturally ther experience will differ from yours. and you both wont be able to understand each other (sure, skill can inlfluence, but it is not the main factor)

for me the question is not whether there are skilled players,(I find myself to be very skilled for instance :) the question is whether in a team game with so many distortional and random factors (and such biased player base distrbution) is a good indicator for skill.

if all this statistical jibberish is too much, think of it this way. isnt the mere fact this debate is ongoing and you always get a different answer depending on who you ask already showing that it isnt a good system? I do not see this kind of debte in sc2 or tennis and in any other established team sport it is common sense that only one player can not influence the outcome alone

I think you're overgeneralizing a bit.

Two things impact a game at any elo: individual skill level and random variables thrown in by your team.

If your skill level is high enough, you can override the random variables thrown in by your team. Simply put, you are so much better than everyone that their randomness will not stop you from winning the game.

However, there is also the scenario where people's skill levels may not be high enough to overcome the random variables. These are the people that complain about elo hell; they percieve an inability to rise, simply because they do not "carry hard enough" to win a large portion of their games.

And I find a game where you can control your individual actions and the consequences of them quite different than rolling the dice a  couple of times

Quote from Yunooh »

<Very long, do not read if not interested in objective discussion>

All this talk and no statistics makes me doubt the fact that LoL is played by many educated people

I would like to start the discussion in a more objective way. I want you all for a moment to forget all emotional crap and it doesn't matter if some black dude makes an emotionally loaded (and funny) video about it, it is basically just an empty show by a kid for kids. The video has absolutely no objective value because it can not prove anything. Please keep in mind that only because an opinion is expressed the loudest does not necessarily make it true. I have yet to see a clear mathematical explanation showing that your elo (or now division + rank) are a good indicator for your skill level. This being said, for me the mere fact that Riot changed from elo to the new ladder system indicates that something is fundamentally wrong and they are trying to fix it in a way.

Therefore please let us try to stop all the noob and trash talk for a moment and just try to elaborate on the facts of the current system.

The premise: the better a player, the more games played, the more accurately can the system determine the skill level (elo or mmr)

First we need to gain an understanding what skill means. In order to be able to show skill, there are some basic requirements. First the bad news, a game like LoL is a game which is very bad at this inherently. Imagine for a moment there is only one champ to choose from, let us say Master Yi. If every game were just Master Yi vs. Master Yi without any changes made to the game (i.e. only blue team Master Yi gets nerfed) then players, over a long enough time spam could develop skill. Now , in the game's current state, a player can develop skill when it comes to overall game mechanics, timings, positioning etc. but the inherent flaw starts with a) champ selection (i.e. Lux has 4 skill shots, the requirement of playing a Lux is way more difficult than a Master Yi, because Master Yi does not have the "skill shot" component. Other aspects like timing, positioning ceteris paribus). This is one of the major and basic reasons for distortion in displaying a players "true skill level". To make this very important aspect even clearer lets take Tennis as an example. The only differentiation in skill is due to body physique, skillfulness in hitting the ball (99% of skill) and some other components like terrain preference and maybe gear. (1%~5%) (which basically means it fully depends on the players skill). This premise is not given in LoL as, let us be honest, it does not take that much skill to play the game (this is also the reason why it is only a silly game, it is designed to be accessible to everyone)

Unfortunately, this is one of the major issues in determining the skill level of players in digital games and this is also the reason why some pros in SC2 might lose to a 6 pool rush (i.e. you can not do this to a Federer, this is why a lot of people pay so much money to see Tennis pro, it just is a spectacular event)

I think overall a good example would be SC2 as it is an established system, with 3 races which are statistically(!) balanced and a clear 1v1 situation in most of the games. I think it is currently one of the most, if not the only balanced competitive game out there.

Now LoL

1) as many of you know, also from the pros, the general consensus is that champ select is the biggest variable when estimating win probabilty (comparing lolking win% per champ gives u an idea of the irregularities. The average win% per champ varies, depending on the patch fix/nerf/buff/op item, this can be proven in the past (just think of talon win % before and after black cleaver nerf)

2) different skill levels in the team are a regular case (not talking about elo or mmr). Currently a player with 600games played and a player with 30games played can get matched in one team by the random matchmaking which only seems to apply the current mmr as basis. Double this variable as enemy team is also prone to this randomness (ergo u don not necessarily play against people with similar skill level). The fact that you can observe this absurdity at entry level bronze already indicates that something is wrong (i.e. in sc2 even the worst players after 500games will usually be gold lvl, due to statistical distribution of player base.) This - I assume - is also the reason why Riot changed it to the current system. Per definition, when you are playing at your skill level you would win 50% of the games played. Riot saw that this was not indicating any progress for the players, this is why now, if you have a win streak you might get it to the next division and can not get back. Riot just made it in 5 steps (divisions), in order to ensure that you can not just skip it with pure luck, as you can fall between the division, but not in the league (once silver you stay silver).

3) As already broadly discussed, if each role is equally important and able to carry, each player should be able to contribute evenly to 20% of the probability to win. Now, in a balanced game with similar skill level this would mean that the initial standard win chance is 50% (in simple words if everyone in the team plays 100% of their skill, the win % is still only 50%). Unfortunately a lot of variables enter the game (as oppose to playing alone). As already thoroughly discussed, team mentality, feeding, insults, afk, champ select disputes and the list goes on and on, heavily influence performance of each player. As Riot already admitted, the player base is extremely immature and because of the internet there are no repercussion in insulting and going afk. People can basically do whatever they want without having to face any social pressure. Due to this fact (i have no statistical numbers in my mind), there are so many ways in which the team performance can be influenced, that - as every active player knows - it is more surprising to see a team which has no dispute and just play without much talk, than to see yet another disaster. 

4) Coming back to the skill level: If one player can realistically only contribute to 20% of a 50% win chance (balanced game premise), it is clear that no single player can effectively carry a game. Sure, in some instances this might happen to a certain (percentage) extent, but in the majority of the games this will simply not be possible, because no matter how fed, in a extreme case 1v5 scenario, the "5" will always win. Of course, a very good player might influence the game. An exceptionally good player (i.e. pro) might even add 5% more win chance to the total game. The fact remains that there is only so much a player can do against all the above mentioned distortions. The only way you can effectively get higher elo or mmr under the current system is play a lot ranked games. Let say it will take the average player in SC2 to get from silver to gold about 200 played games. In LoL this number would probably be 4 to 5 times higher due to the above distortions. Mr. Pro might do this in only 500 games to get to platinum, because the 5% of super skill (plus of course the connections which good players have between each other - they can team up, they just naturally play a lot) will make the whole advance easier. But this does not mean that 80 to 90% of the player base will have the same experience. They might just feel "stuck".

In Summary:

Imagine LoL as a fully random slot machine where you only have 10% chance to contribute to you winning the game. You will never be able to tell whether it was your 10% that contributed to winning or not, all you can do is try to improve your odds against the house. There is no quick way of winning, there is no way of escaping all the above mentioned factors which will inevitably distort your true skill level. All you can do is run it over and over again.

This leads to the conclusion that statistically, based on the total playing population, only a very few will get promoted to higher mmr as they simply can not influence all the outcomes (some very lucky guys very fast, some very unlucky guys almost never). This is what statisticians call the Gaussian distribution (https://en.wikipedia.org/wiki/Normal_distribution), meaning that roughly 80% will be stuck in the middle of 50% win chance, while some exceptionally few will get promoted and also stay there (as you can not really drop). 

This means, that elo hell does exist. The only problem is what this word defines. Hell is a very emotional word, meaning that you do not want to be there (and you maybe do not deserve being there). All LoL players will just have to come to terms with the fact that most of us are simply stuck. It doesnt matter if it is bronze, silver or platinum. It will just take a lot of effort and time invested to somehow turn the odds to your favor with that little that you can influence.

Finally one thing we need to be perfectly clear is that it is just a flawed game. And let us be honest. Isnt this the reason why most of us are playing it? We expect to get fed, we want to have that feeling that we can actually influence the outcome although we know that so many lucky factors are involved. So it is only natural that we get disappointed if it not happens. Well tough luck, but that is the cruel and unfortunately very real world of numbers in a game which is simply not fit to reward 100% to skill.

I'm sorry, but I think this is just incorrect. According to what you're saying, because of random outcomes, an individual player starting 10 different accounts would fluctuate in the number of games it took him to reach his actual elo. When in practice, this is simply not the case. I will *never* get stuck in the bronze elo and I will rarely if ever go less than very positive in silver elo. I've done this with three accounts and I've gotten out of bronze an d silver in no less than 60 games in each.

Am I a good player? No, I suck. That's why I'm in gold. A diamond player would have no trouble getting out of gold (I know, I've played with diamond smurfs.) The fact of the matter is, you can influence a game MUCH more than 20%. You're assuming that the pros will basically make it 5% more likely to win, making each game a 55%/45% split. However, in practice, you will see that this value is much higher from basically any observer you take who makes a smurf (e.g. someone who *actually* doesn't belong in the elo they are currently in). If I play fizz in bronze elo, I will not lose a game. Almost ever. like, 95% of the time I win the game. It is like uncertainty. If your uncertainty is too large in comparison to your value, it will imply that your data is not good data. In the same way, if your individual skill level is not high enough to influence the fluctuations in the elo you're currently in, your skill level is unreliable and your true elo is difficult to read.