ARAM - Something Is Actually Wrong

JustaNewb··4 votes·1,261 views

What are the odds of getting the exact same five champion team twice in a row?

It's high enough that a scientific calculator just produces E+297 after the rest of it's maxed out characters. The odds against getting the exact same five champions, in a random environment, is more astronomical than being struck by lightning 7 times on a clear day where you were abducted by aliens. (Actually looked this one up. It is still higher.) Being as this is the case, why do I invariably see a team comprised of 5 champions who can easily 1v5 in ARAM... repeatedly through out the day for the past month or so?

Jinx Corki Thresh Illaoi Zilean

This exact team, comprised of some of the (not absolutely the most broken, but...) most broken champions in ARAM right now due to their monsterous damage output, survivability, poke ability, and general 'rocking your ARAM world' ability. Somehow, in a 'random' environment, This precise team pops up 3-4 times a day. That is simply against -me-, let alone how many ARAM games are going each day.

This either means:

A: ARAM is getting less random with time. or B: People have discovered how to rig ARAM.

Both are major problems that needs RIOT attention.

25 Comments

Professor Ward2 votes

Over 4300 arams played, i can tell you save for free week, owned champs, and trades, aram is 100% random.

The chances* are actually the exact same as getting a different champion each roll

Ive explained this multiple times here, over 30 probably but it keeps being asked.

The champions you see in one game HAVE NO WEIGHT ON THE CHAMPIONS YOU SEE IN ANOTHER

IF YOU PULL A CHAMP OUT OF THE BAG, BEFORE YOU GO TO THE NEXT GAME YOU HAVE TO REPLACE THAT CHAMP TO THE BAG,

therefore making your chances the EXACT SAME AS THEY WERE BEFORE OF GETTING THE CHAMP

-------------------getting a different champ every single game for 500 games is the exact same chance as rolling gnar 500 times in a row.-------------------

Also of the champs you listed, none of them are the worst aram offenders. Sona, ziggs, teemo, and maokai are not listed, of which have the highest winrates (near 70%) and are basically auto losses at this point.

JustaNewb1 votes

I think the Professor doesn't understand how statistical probability is calculated... nor, apparently, understands the point here. I wasn't talking about champions that are OP in ARAM.

Maybe that is why you have to keep explaining it.

The algorithm is simple (and absolutely static in all Statistical math):

139^138^136^135^134 - Your odds of any one team, Ever. Ever. EVER. 139^138^136^135^134^2 - Odds of getting the same team twice. 139^138^136^135^134^3 - Odds of getting the same team thrice. etc etc etc

Ad Infinitum.

Professor Ward1 votes

{quoted}

I think the Professor doesn't understand how statistical probability is calculated... nor, apparently, understands the point here. I wasn't talking about champions that are OP in ARAM.

Maybe that is why you have to keep explaining it.

There is no statistical difference in the chances of rolling zilean or shen. They are identical. And they are identical game to game.

Your chances of rolling shen a second game, are identical to your chances of rolling shen the first game, and also identical to your chances of rolling zilean for a first, second, or third time.

and as for not talking about op champs in aram, wtf is this section about "broken champions" then -----

""This exact team, comprised of some of the (not absolutely the most broken, but...) most broken champions in ARAM right now due to their monsterous damage output, survivability, poke ability, and general 'rocking your ARAM world' ability."

JustaNewb1 votes

Yeah.. you just repeated my point for me. Except for the fact that, no, every time you get the same champion the odds grow against you getting the champion a second time. Thus statistics. I have an entire college textbook I can send you.

Professor Ward1 votes

{quoted}

Yeah.. you just repeated my point for me. Except for the fact that, no, every time you get the same champion the odds grow against you getting the champion a second time. Thus statistics. I have an entire college textbook I can send you.

i promise you that you are wrong, there is no magic thing that makes it harder to roll a champion twice in a row... How can you even believe that?

The odds are still 1 out of how many champs you own. There is no "weight" magically added to champs who you rolled previously. The system doesnt even know who you played last game when rolling? So how could it even effect it?

You are way off man.

Like after you roll gnar, does a point go into gnar, that makes him less likely to be picked next roll? No? So how in any way is he less likely to be rolled again? Hes not. The odds are the exact same every roll.

JustaNewb1 votes

Statistics =/= Champ Selection Statistics = Chances of, of 134 champions, getting 1 Champion Statistics = Greater Every time a randomized system produces the same result twice.

Odds of getting any one champion in League of Legends, assuming all champions owned (actually common): 134 Champions + Number on Team 5

The algorithm is the exact same as determining your odds of winning a lottery:

x^x-1^x-2^x-3^x-4

The algorithm then is fed to itself once again if one wishes to replicate the results in an exact pattern.

x^x-1^x-2^x-3^x-4^2

This repeats forever.

Basic Statistical Math you learn on Day One. It isn't magic, it is probability.

Khristophoros1 votes

{quoted}

Statistics =/= Champ Selection Statistics = Chances of, of 139 champions, getting 1 Champion Statistics = Greater Every time a randomized system produces the same result twice.

Odds of getting any one champion in League of Legends, assuming all champions owned (actually common): 134 Champions + Number on Team 5

The algorithm is the exact same as determining your odds of winning a lottery:

x^x-1^x-2^x-3^x-4

The algorithm then is fed to itself once again if one wishes to replicate the results in an exact pattern.

x^x-1^x-2^x-3^x-4^2

This repeats forever.

Basic Statistical Math you learn on Day One. It isn't magic, it is probability.

COOL! Where's your actual data to apply that math to?

JustaNewb1 votes

Emailed to support, along with my match history for the past 3 years. I don't need your validation of my math, I am explaining to him how the math -works-.

Khristophoros1 votes

{quoted}

Emailed to support, along with my match history for the past 3 years. I don't need your validation of my math, I am explaining to him how the math -works-.

Why are you posting on the forum making these wild claims without any data?

Look, my guess is that you saw the same team comp 2 games in a row and you're freaking out.

Look at it this way. What are the odds that out of all the billions of ARAM games played in total that two of them would have the same team comp?

What are the odds that this would happen and also one of the 10 people from it would make a big deal about it on the forums?

Well probably not too bad, considering how many total games of ARAM are played.

The odds that it was you to experience this are the same odds as any other player.

JustaNewb1 votes

K. Enjoy your evening.

Khristophoros1 votes

{quoted}

K. Enjoy your evening.

Wait so when somebody is skeptical about your claims you just tell them to fuck off? what are you even trying to accomplish with this thread?

JustaNewb1 votes

A: Didn't say that. B: You defaulted to combative from your first reply to this thread. Counter productive, and absolutely no reason for me to continue with you.

Khristophoros1 votes

{quoted}

A: Didn't say that. B: You defaulted to combative from your first reply to this thread. Counter productive, and absolutely no reason for me to continue with you.

I told you that you have no data which is why I don't believe your claim.

Professor Ward1 votes

{quoted}

Statistics =/= Champ Selection Statistics = Chances of, of 139 champions, getting 1 Champion Statistics = Greater Every time a randomized system produces the same result twice.

Odds of getting any one champion in League of Legends, assuming all champions owned (actually common): 134 Champions + Number on Team 5

The algorithm is the exact same as determining your odds of winning a lottery:

x^x-1^x-2^x-3^x-4

The algorithm then is fed to itself once again if one wishes to replicate the results in an exact pattern.

x^x-1^x-2^x-3^x-4^2

This repeats forever.

Basic Statistical Math you learn on Day One. It isn't magic, it is probability.

So youre telling me, that if i go to a pile of marbles, pick one up, and place it back, Just by picking it up, i have decreased my chances of selecting it again next trip, and increased the chance of choosing another, without effecting any matter in the universe?

I just cant wrap my head around that. When i return to the pile, nothing has changed, and the chance that i will grab the same marble, should be the exact same as grabbing any other.

JustaNewb1 votes

Each marble is a variable... the number grows to the total marbles, thus providing 'x.' Each time you pick a new marble, if say blindfolded and they are in random order, is a new 'x' minus the one you took away, or two you took away, etc. Each selection therefore IS against the odds of 'x,' and each selection is x^x to retrieve the same marble... because you have the same odds of grabbing any -other- marble as well.

If you choose 5, replace them, and choose 5 random marbles again, your odds are the end results of the probability of the first pattern, squared as you are now performing it twice. Cubed if you perform it three times. Etc.

ComeGankMeBro2 votes

It needs to just be every single Champion as a possibility.

What possible reason is there for keeping it as owned and free-week Champs only?

There isn't one, Rioters are just a lazy excuse for game designers completely apathetic towards anything that isn't their "main" game mode.

And they don't even take care of that one well~

Sir Fuzzi2 votes

What they need is a separate rotation for ARAM, one wherein free picks comprise low playrates (or winrates for padding)... Every champion will backfire because now everyone can trade everything to anyone. I think by unlocking the full roster you just enable strategic rerolling.

Realistically, the problem is high performers being picked constantly into teams that don't have them, or other high performers. I think parity in the other direction-- giving both teams a flat likelihood of 'low impact' picks-- makes it feel a little less like Lux vs some clowns without ranged cc. Or Sona vs a team with bad aoe.

But we had best take care to remember it's not really meant to be balanced...

Done251 votes

Pics or it didn't happen.

Khristophoros1 votes

So do you have actual data on this? I have two important questions for you. If you can't answer these questions definitively, you don't have anything statistically valid to talk about.

  1. How many ARAM games in a row have you seen this team comp?

  2. How many times have you seen this team comp in your past 100 games?

Mysticman891 votes

I suspect you're not calculating odds correctly, and/or making some incorrect assumptions there. (For example, it seems likely you're implicitly assuming that all 10 people in the match own all champs, which is almost certainly false.)

Accounting for rerolls also makes determining odds tricksy.

Rioters have occasionally chimed in on ARAM randomness (maybe it was in an ask riot or a /dev blog or something, or maybe posts I don't remember), and they consistently say they've looked at the code and it is random, it's just theres a lot of confounding factors that make the champs you ultimately see see less random (not to mention the human tendency to find patterns even in random data).

For example, some champions are just way more popular than others (e.g. jinx/lux) independent of their power in ARAM, so a lot more people are going to own those champions than own someone like udyr. Those champs also happen to be strong in aram, so if someone gets them (somewhat likely given most players will own them), they're likely to keep them. By contrast if someone gets a bottom tier aram champion, theres a pretty good chance someone will reroll them, so you're likely to see strong aram champs more so than not so strong ones.

On top of that, dodges are pretty common, so if a team comp doesn't seem strong, it's not uncommon for someone to dodge, so the champs that actually see play on the abyss tends to be biased towards champs that are strong there, despite those champs being weighted exactly the same as far as the code randomly picking champs is concerned.

Free week champs are also naturally going to show up more, since those are the only champs everyone is guaranteed to own. Both zilean and jinx are free week currently, so it's no surprise you see them often enough. Thresh is one of the most popular supports, so also prone to showing up. Corki is somewhat popular too, and it's only really illaoi that seems kindve like an oddball, but with enough games played and the other 4 being somewhat common, it's suddenly far more probable than you'd face the team comp you describe than the multiple lightning strike scenario you're toting.

If they want to make the aram experience really feel random, they'd have to unlocked the entire champ pool for everyone there, and remove champ rerolls, and prevent dodges from rerolling your champ. And even that would probably not necessarily feel random to everyone, since people are bad at recognising randomness.

tl;dr: It's uniformly random across a players available champions (not all champions), so it's "random", but what champions a given player has available is not random, which naturally biases peoples initial comps towards popular champs (many of which happen to be strong aram champs), and then champion rerolls bias that already biased result towards comps that are strong in ARAM.

JustaNewb

Unless we assume that all 10 players ONLY own these champions, and no others, the odds actually -grow- if I could calculate the chances of them OWNING only these champions, and it still turning up this exact team... right?

JustaNewb1 votes

Unless we assume that all 10 players ONLY own these champions, and no others, the odds actually -grow- if I could calculate the chances of them OWNING only these champions, and it still turning up this exact team... right?

Mysticman892 votes

We can prove the assumption that they only have those champs and no others wrong immediately because of free week champions.

Despite that, if all 10 players only had the same 10 champs between them, getting a repeat of a comp featuring a particular subset of 5 of them would be relatively likely, with those odds diminishing as the number of champs owned increases.

Point is, if you did any calculations where you're using the number 139 (i.e. the number of league champs currently existing), theres a strong chance that even if the calculation is correct, it does not model the situation you experienced.

JustaNewb1 votes

Unfortunately it is easier to assume that players post 30 have most champions (especially when they have skins for every champion tossed at them). With the Loot System making farming of champions vastly easier, I assumed for PoR that they have near all champions. Otherwise we can never actually calculate the statistics with a set baseline, without RIOT explaining the collection of all players in the games.

However, the likelihood of players only having free week champions, and those 5 champions, is very low.

Sõna1 votes

Just fyi, there's some professional ARAM hunter who use ARAM account to dominate your game.

More from Gameplay

More by JustaNewb