Because you're not a random factor to yourself.
So there are 4 other random factors (aka players) on your team and 5 on the other. 4/9 = 44.44% < 50%.

Of course, if you're a bad player, the chance of your team having a bad player (100%) is higher than the other team's (<100%).

This, too funny, you can't possibly be serious? Come on, Please, here, google TwoplusTwo Forums, so you can see how much they complain about luck and randomness in poker, if you REALLY believe poker is purely determined by skill you are so lost.

The variance in poker due to random events is incomparably higher to the variance of league of legends elo due to random events(ie your teammates) the two are not even in the same realm.

It is based off your performance in the game, if you win, you go up.

Stop complaining about elo hell, I just won a 4v5 on my smurf as support sona, oh and it was actually 3v5 for about 10 minutes. Does this mean i'm amazing? No, it just means that I try to find a way to win no matter what, and that even as a support character, with top lane disconnecting after level 5, that i'm still able to be a big enough influence to turn a loss into a win.

Like ive said before, it takes a very big person to accept the concept that they might actually not be that good at this game, but it is necessary in order to get better.

alright, the way i see it, this idea is very nice
but, in my personal opinion, the way elo is right now, around 1200 would be the average skill level of a team split up by solo/duoqueues
going below this would make you below the average, and going above it would make you above the average
so 1200 is pretty much the zero that you're looking for, but there's a chance to get into "below 0", or "negative" elo
if the 0 elo you proposed was to be put into effect, there would be too big of a cesspool in skill level at 0 elo to create proper skill level matchmaking, unless of course, you COULD go below 0, but then nothing would change

i think a big misconception is that people would be stuck at zero

if u win one game against 1200 hidden elo players, and ur actual elo is 0, ur actual elo shoots up 100 or 150 or something, because of the difference in ACTUAL to HIDDEN elo for the winner is so stretched out

if u lose one game against 1200 hidden elo players and ur actual elo is 150, ur actual elo goes down by 1 or 2 points, becuase ur ACTUAL elo is so far below ur HIDDEN elo (which starts at 1200)

get it? ur actual elo and ur hidden elo catch up to each other whether u win or lose. when u lose u dont lose much actual elo (if any at all) until ur actual elo hits ur hidden elo (which starts at 1200)

if u start at 0 actual, and 1200 hidden, and u go on a winning streak, ur hidden elo starts to skyrocket, and ur actual elo skyrockets even harder to compensate

EDIT: for the people that win some lose some, they end up at 1200 pretty quickly, and for the people that just lose, they end up at 600 pretty quickly too

the benifits of this kind of system is that the new player doesnt have this idea that baddie players are dragging them down.

They should really teach more statistics in schools. Elo systems use a normal distribution to assign their rank values. Setting the low point to zero would just force you to include negative elos.

Elo works in team games because it is possible to measure, over time, the net impact that a player has towards his team's victory at a given level quite accurately. Your win % at a given level reflects your mobility- If you are not able to win more than 6/10 of your games, chances are your contribution to your team winning is insignificant at that level, and so you will stay there.

In S1, I was around 1800~ most of the time, and had around a 65% winrate with my AP mids of choice. The problem was, I didn't always get to mid, so I couldn't contribute as well to my team as I would like. At the moment, I am at a meager 1320 due to having thought I was not playing League again in the future and dicking around, but my anivia mid win% is around 75%~. If I played nothing but anivia mid, statistically speaking I would be likely to rise quite swiftly until I reached the point where I become better matched, and it would begin to decrease.

not if u use a hidden elo ranking system like World of Warcraft did.

Yeah I myself have been thinking of making such a thread. Glad to see someone else agrees with me for once. Throwing noobs into the middle of the ladder makes no sense.

No, you would not have negative elos. It would just be that ******* who consistently cost their team the game would just stay at zero and pile up there.


Why the **** would you use a normal distribution to assign RANK values? Thats like using a normal distribution to assign numbers to football players on a team, it makes no ****ing sense.

Whilst talking about teaching statistics, I think you seriously need to take a beginner course yourself cowboy. Normal distributions work for random variables. LoL is not a game of chance, its a game of skill. If this were roulette, assuming elo is normal would be more or less valid. This is most certainly not roulette.

So what you said just makes no sense, it just makes no goddamn sense.

Wiki Elo system.

Thanks, that only confirms my diagnosis.

People's performance in games is most certainly not random. If we all randomly right and left clicked after randoming champ select, the above person's assertions would be true. Thats definitely not the case. In fact, you're inherently not capable of engaging in 'random' behavior. (Which you probably should be aware of but I'm putting it out there just in case.)

So yeah.