In Old Champ Select, You Got Your Desired Role ~64.8% of the Time

ElysMustache··60 votes·4,733 views

Over 155 ranked games, I got my desired role 70.97% of the time. What was my desired role? Mid. Supposedly that lane is the most in-demand. I never dodged based on lane that I got, and I never insta-locked.

Here's the math of the old system, assuming each role had 20% of players preferring it:

When I am first pick, the chance I get my role is 1. When I am second pick, the chance I get my role is 0.8. For 3rd pick, 0.64. For 4th, 0.48. and for last pick, I still have a 0.32 chance of getting my desired role.

Weighting each of these factors given that they each occur in one out of five games, we arrive at the following: The chance that I get first pick and also get my desired role is 0.2 The chance that I get 2nd and my role is 0.16. For 3rd pick, 0.128 Fourth, 0.096 And last, 0.064.

Summing all of these together results in a probability of 0.648 that I get my role, if the underlying assumption of equal role distribution is true, and that players will always take their first choice if a team mate also wants it who is after them in pick order.

I think that 155 games is a reasonable sample size. I got my role more often than my model here predicts. Generally the only thing I would say in champ select was "Prefer mid".

I am interested in seeing how often I get my primary role in the new system. I don't have enough info yet, having only played eleven games with the new system... but there is at least one thing I can be confident in... I'll get my worst role less. I'm pretty happy about that. Now I can set up all my rune pages for my two preferred roles. New players will also have an easier time acquiring champion pools for their role of choice. That's cool.

Thanks for reading, hope you found it somewhat interesting.

Edit: Slight math error as posted above. Thanks to Llanite for helping me realize error. Final chance should be 60%, not 64.8%.

87 Comments

Puddytat19 votes

{quoted}

Here's the math of the old system, assuming each role had 20% of players preferring it:

Emphasis mine. Your assumption is wrong, therefore your math can't be correct. As shown in Teambuilder, many more players want to play mid than support. If you got mid 70% of the time, I would credit that to "you must be a polite person, and have met many friendly teammates". It would be very interesting to see actual data on this, but I'm not sure they are available.

Penguin Eclipse5 votes

{quoted}

Emphasis mine. Your assumption is wrong, therefore your math can't be correct. As shown in Teambuilder, many more players want to play mid than support. If you got mid 70% of the time, I would credit that to "you must be a polite person, and have met many friendly teammates". It would be very interesting to see actual data on this, but I'm not sure they are available.

That's because support is ultimately the least requested roll. That doesn't mean you couldn't get the roll you wanted. People went to team builder specifically so they could get mid 100% of the time. Those of us who play support extensively had probably about a 90%+ chance of getting support no matter what queue we choose. We had no reason to choose team builder other than fast queues and to play with a typically much politer group of people.

The OP's math seems pretty accurate to me, but it will be higher depending on the roll. Whenever I wanted mid, I got it most of the time. Whenever I wanted support I got it almost all the time.

ElysMustache5 votes

I know the assumption is wrong.

I remember seeing an API sourced data analysis of role preference in the past, but I can't seem to find it right now. But from what I recall the role rates were reasonably close. Surveys that I have seen on the web range from 14-22% per role, with Top generally lowest, but I find surveys less trustworthy.

I think there is something to be learned even though the base assumption is wrong.

magic pole15 votes

except that maybe only 5% of players want support

Penguin Eclipse26 votes

[{quoted}](name=magic pole,realm=NA,application-id=3ErqAdtq,discussion-id=EzTt2kIi,comment-id=0003,timestamp=2016-01-19T20:28:16.765+0000)

except that maybe only 5% of players want support

There is a difference between wanting to play support and playing it anyway for the better of the team.

magic pole3 votes

yup

Jack º4 votes

I'm one of them 5% Sona ( ˘ ³˘) ~♥

TheDiff3 votes

I love playing support! But in my elo it is legit hell. You can't really carry a team when playing a support. You can do so much in your power to carry the lane, which may result in you and your adc into getting fed, and that may result into you guys carrying, but it is SO hard in Silver/gold.

Xicashina1 votes

Im in the same boat support is by far my best role but its often really hard to carry a game from it.

TheDiff1 votes

I preffer Support/mid if i ever join a ranked team, but In solo que i will always do Mid/top...something i can carry with. I mean there is only so much in your power to do in fights and in lane. You can't save everyone. You can't set up every single teamfight. I mean that's why it's hard. Whereas if you play mid/adc you can just destroy their carries, or if you play top/jungle can set up teamfights in favor of your team.

SH Azykan2 votes

Annie Blitzcrank Brand Kennen Lux Velkoz Zyra are all good "supports" to carry with. Most of them can completely dominate a lane on their own, or remove an enemy from a fight to give your team a numbers advantage.

Sad thing is most of the more fun supports rely on your adc being competent to do well with. Great for duo's, not so much for randoms. Alistar Bard Braum Morgana Thresh Shen being among those.

TheDiff1 votes

I actually prefer to paly Braum support because I can protect my adc....from everything. Lol :P

SH Azykan1 votes

yeah but relying on an adc to follow up on anything, or they go ham when your skills are on cd in a matchup like Jinx Morgana vs Vayne Braum is just annoying.

xxxDogmanxxx2 votes

Play Garen support. Nobody knows what to do against you bot lane. Does the enemy support engage? You'll melt them if they're near you. Does the enemy ADC kite you? Your ADC gets free hits on them while you tank the damage. Do they ignore you and focus on your ADC? You chunk huge amounts of health off the enemy ADC.

Alëk1 votes

O_o in my experience this is the same if not worse than Volibear support. If you have any idea how to play you just kite the fuck out of them and shove them out of lane. the only time they are able to do anything is if they flash Q you. otherwise it's obscenely easy to deal with them in my experience.

xxxDogmanxxx1 votes

Support Garen is better than support Volibear because Garen needs no mana and Garen's Passive is like a free Warmog's passive. Shove Garen out of lane? For all of ten seconds and then he's back at decent health.

city of refuge3 votes

that's why maining support is so OP, no one will ever contest your role and if you're first pick just pick Annie or Lux or Zyra so they think you're mid to fuck up their comp.

AMYS GRAVE2 votes

64 percent of all the world's statistics are made up right there on the spot 82.4 percent of people believe 'em whether they're accurate statistics or not

GenghisEmo1 votes

and im that 5% XD. But after i get a few bad ADCs, i'll rather carry as Mid or Jungler. Braum Janna Lulu Lux Nami Thresh Velkoz Volibear Leona Nautilus Morgana Sejuani (Yes Sej Supp, used it with Ashe as ADC, worked quite well quite a few times. Fun too XD.

VictoryDance965 votes

Maths.

Lazlow5 votes

Pick order > New system.

DrCyanide15 votes

Pick Order meant the following (over 182 games):

  • Main Role (mid for me): 64.3%
  • Secondary Roles (ADC/Support, no particular order): 13.2% + 9.3% = 22.5%
  • A Role I completely suck at: 13.2%

The new system guarantees that that 13.2% never happens. In my top 3 preferred roles I have a 58.2% win rate, compared to the 23.8% win rate of the others. Swapping that 13.2% of games out for games in a position I'm more comfortable with brings my overall win rate from 53.9% to 60.4% win rate.

ExoNautic1 votes

you like math dont you?

DrCyanide4 votes

As a subject? No. Applying it to things I care about? You bet!

I'm working on a League of Legends stat based website, but with school and life it's probably a few months away from being ready to show to the public. My own stats were pretty easy to pull thanks to the work I've done on that site.

Lao Fu the Tiger1 votes

What's the site? I'm looking to put together some stats from the API as well and I'd love to help out if you'd like.

Just add me in game if you'd like!

Loco Mot1ves4 votes

Pretty similar story here. So far, in the amount of games I've played in the new champ select, I've never not gotten one of my two roles selected, solo or duo.

One thing I wish was changed is that in a duo Q, both of us select bot and support, so that no matter what, we will always have bottom lane together and can trade for our first pick. That's 90% the reason we duo. If we want to have variety, pick whatever and see what you get.

Spoofghoul3 votes

In Old normal blindpick Champ Select, You Got Your Desired Role ~64.8% of the Time But. . .

Someone else got that role 64.8% of the time too

AsphodelNext6 votes

Over 155 ranked games

Ranked is not normal blindpick

kaltenashes3 votes

In Old Champ Select, You Got Your Desired Role ~64.8% of the Time

NOPS, I only got my desired role when i instalocked

I was never able to be one of the firsts to join in the lobby, i could never get my wanted role... if i could play jungle all the time i'd be happier.

i had to become a filler or rage every single time i joined late into the lobby and its already decided who's going where

New system > Pick order

The Bearded Bard3 votes

Not counting the games I've played on the new draft, I've had 55 games in mid (the position I wanted) vs 145 games in other positions this season.

With the new draft mode, I've had 13/13 games in positions I want to play.

Titanium703 votes

I guess the amount of times you get Mid if choosen as primary will go down a little. However the main advantage of the new Champselect will be the possiblity to NOT play three entire roles. You can Duo-Lane only, you can Solo-Lane only. Basicly you can be a TwoTrickPony 100% of the time. 'One' too if your champ is able to play two positions.

Llanite3 votes

Faulty math here.

Chance of you having first pick = 0.2 Chance of you having second pick AND first pick isn't mid = 0.2 x0.8 = 0.16 Chance of you having third pick AND first pick isn't mid AND 2nd pick isn't mid = 0.2 x0.8 x0.6 Then 0.2 x0.8 x0.6 x0.4 Then 0.2 x0.8 x0.6 x0.4 x0.2 Total = 50.28%

The assumption being, every role has equal 20% weight which we know isn't the case, the distribution is somewhere 35% mid, 30% top, 15% adc, 15% jungle and 5% support. The final result will be ~25% using above distribution.

The new queue gets you primary 66% of the time, from my experience queuing mid/top and adc/jungle. Anything that has equal demand should follow 2:1 split rule.

Do yourself a favor, do NOT queue secondary support if you don't mean it, they have very high demand and low supply, resulting anyone queuing support being called into it.

qetzel1 votes

Isn't The chance of third pick not being mid is 0.8, not 0.6 ? I think the OP is right given the assumptions they made. Silly assumptions though :P

ElysMustache1 votes

You're not factoring in that each player has fewer choices of which role to take. In other words, you are assuming that Player2 can take my lane when Player1 already took my lane.

Llanite2 votes

You're right. 1 a.m. math didnt work. The number of choices go down everytime.

Chance of you having first pick = 0.2 Chance of you having second pick AND first pick isn't mid = 0.2 x0.8 = 0.16 Chance of you having third pick AND first pick isn't mid AND 2nd pick isn't mid = 0.2 x4/5 x3/4 Then 0.2 x4/5 x3/4x 2/3 Then 0.2 x4/5 x3/4 x2/3 x1/2 Total = 60%

ElysMustache1 votes

0.2 x 1 = 0.2 0.2 x 1 x 0.8 = 0.16 0.2 x 1 x 0.8 x 0.75 = 0.12 0.2 x 1 x 0.8 x 0.75 x 0.666... = 0.08 0.2 x 1 x 0.8 x 0.75 x 0.666.... x 0.5 = 0.04

sum: 0.2 + 0.16 + 0.12 + 0.08 + 0.04 = 0.6

I agree. I had an error in my OP. I put a 0.8 in a spot I should have had a 0.75. Thanks for making me look at it again.

Steel Blossom2 votes

Man I love these types of posts. Actual math, coherent logic, and no biases. The board needs more people like you.

DeynaTaggerung2 votes

Legit math and admits multiple sides to an argument...nice post

DorkunedAuras2 votes

Dude, good math. I'm happy intelligent people still exist to prove points.

Jonxx2 votes

Actually I get my preferred role 100% of the time now, because I can play multiple roles [slayer-jinx-catface]

Rock MD1 votes

Roughly 50% of my games are on one champion, somewhere between 85-95% of my games were mid.

In my last 5 games, 20% of my games were mid (1 game).

In 10 games, 40% were mid.

Hotlin3 Fling2 votes

10 games is still too few to make an accurate judgement IMO

Rock MD1 votes

{quoted}

10 games is still too few to make an accurate judgement IMO

5 games today. Mid for first game, then all the rest were ADC.

I think it's proven quite well what happens to you if you get to high elo as a mid main.

Hotlin3 Fling1 votes

Not that I know what high elo is like, but it's still possible to get shafted for a few games in a row. After more games you should get mid more consistently

Rock MD1 votes

{quoted}

Not that I know what high elo is like, but it's still possible to get shafted for a few games in a row. After more games you should get mid more consistently

If you don't know what high elo is like why are you trying to tell me what it's like?

Hotlin3 Fling2 votes

TBH it's because it seems like you can't grasp the concept of a reasonable sample size. i'm trying to rectify that

DracosKasu1 votes

But in the old system people who get slower computer can't get the role they wanted because they actually enter after fast pc who rush and CALL MID!!!

I don't have this problem because I upgrade my pc when I need, but you can't hide the fact this is true. Now everyone have the same amount of chance to get the role.

[slayer-jinx-catface]

Dan Abnormal2 votes

draft pick is pick order you walnut. You're talking about blind pick

DracosKasu1 votes

You know there still people who call position in queue, just because you pick before someone you are allow to steal it because you pick before him. Now you can play mid even if you are last pick also you can ask the other if he want to switch. I generally play in draft mode also so don't BM.

Dan Abnormal2 votes

"because they actually enter after fast pc who rush and CALL MID!!!" calling order doesnt matter in draft pick. People called their role preference yes, but calling it first or last didnt matter to the outcome. What this means is that connecting to champ select earlier in draft mode didnt increase your chance of getting your role.

More from Gameplay

More by ElysMustache