Season One Draft Order Mistake

First Riot Post
123
Comment below rating threshold, click here to show it.

MattMarcou

eSports Manager

Follow RiotMarcou on Twitter

06-24-2010
2 of 4 Riot Posts

Quote:
Originally Posted by sturgeon View Post
Why have you chosen to go with ABBA bans? Most other MOBAs use an ABAB system of bans.
I don't think the order has a huge impact in this case, but we wanted to continue with the snake system. Most competitive tournaments (Counter-strike, Warcraft 3) use the ABBA system when selecting maps and continue that system in the transition from ban to pick.

One of my thoughts is that the team who bans first might feel that their ban is wasted in certain scenarios. There are always common heroes that both teams may want banned and the meta game can become a game of chicken (i.e. who's going to flinch and ban this hero). Since A is going first, it's good to offer them the last ban, so they are not in a situation where they feel 2 bans were thrown away.

Overall, I feel that the desire to be team A or team B is mostly psychological and similar to winning the coin flip in an NFL game. The odds are as even as they can be.


Comment below rating threshold, click here to show it.

MattMarcou

eSports Manager

Follow RiotMarcou on Twitter

06-24-2010
3 of 4 Riot Posts

Dreampod, Makbeth, Fuya, and anyone else that has commented or comments

I appreciate all of your points as well as the thought and passion that went into them. I think we can all agree that this would be a very complex mathematical equation that none of us can address in any reasonable time frame.

I believe that the system is as balanced as it's going to get. If there was even a 3-5% variance with either side, then I would be happy with the balance. Recommended reading for you all would be debates on going 1st or 2nd in chess. See this wikipedia article as a nice introduction: http://en.wikipedia.org/wiki/First-m...ntage_in_chess

Makbeth, I am going to address your points directly since I am in support of the status quo and you offered a viewpoint to change the status quo. Please do not take my comments as being a personal attack, because I appreciate the discussion.

1. "This is a strong advantage, as team two will always win the game 100% of the time with this model. A game with 50 champions is more complicated, but it would take a year of explaining game theory models to prove a 10% advantage."

--Absolutisms are dangerous. Any model that spells out a 100% victory over an appropriate sample size has to be flawed or incorrectly applied. The advantage of going first in chess has yet to be proven or disproven, so I don't expect us to resolve this anytime soon.

2. "You
must, in any game theory model, assume each pick's quality is equal. That is to say, each champion is exactly balanced."

--I understand this thought in game theory, but people are not perfect and I don't think we can make this assumption on their picks or on the balance of the heroes themselves. All choices are made by people and inherently not equal.

3. "As such, the first pick
is not an advantage. "

--It can be if their truly is an "S class" hero who stands out as the best pick regardless of the composition of the remaining 4 heroes. Also, if two teams with similar individual strengths meet, then they may have the desire to play a similar lineup that has one or two heroes as their cornerstone.

4. "Any player selecting a champion randomly will be as equally successful as a person who really really really really wants to play a jungling warwick (something I hate Riot for, by the way!)."

--I think this is a clear case where theory is diverging with reality. This is clearly not the case at the lower levels of play and although this become more and more true at the upper levels of play, if we look back to dota there was not a single professional player who could claim mastery of even 20 heroes.

5. "The last pick, therefore, has the absolute most value."

--See the above points. It could be argued that the first has a good amount of value OR we could just say that each pick has a significant and nearly equal value.

I consider this matter to be closed, but I will be monitoring the game play data. If I notice anyway to further even the odds, then I will bring it up to you all.


Comment below rating threshold, click here to show it.

Makbeth

Member

06-30-2010

What you have done here is cement my point. You stated that the first pick in chess is debated to this day, thousands of years later. That was my entire point; the first pick is not of greater value. The points you do make are from either experience or opinion, something that has never quite held a candle to math.

Please publish your data as games are logged (and put me on team 2 every time).


Comment below rating threshold, click here to show it.

MattMarcou

eSports Manager

Follow RiotMarcou on Twitter

06-30-2010
4 of 4 Riot Posts

Quote:
Originally Posted by Makbeth View Post
Please reconsider your draft mode order, it is balanced improperly.


Team 1 must pick a hero first, which (if everything is considered balanced) is a disadvantage because Team 2 will have the opportunity to counter.

To equalize this, Team 1 would conceivably be allowed to pick last, but this is not the case currently. As it is, Team 2 will pick last and thus have the opportunity to counter Team 1's choices every step of the way.

You can fix it now, or you can wait for the logs to prove the statistical conclusion that I have made a very nice living being on the proper side of
Makbeth, your argument was "[our draft mode] is balanced improperly." You then offered a partial solution "to equalize this ...," and the you offered to prove the math behind it.

You did not provide either a full solution or the math showing a flaw. No point was cemented.

I acknowledged that there could be a slight statistical advantage one way or the other based on arguments in chess. This does not cement your point either.

As the kids would say: Math or GTFO


Comment below rating threshold, click here to show it.

Dreampod

Senior Member

06-30-2010

Quote:
Originally Posted by Makbeth View Post
What you have done here is cement my point. You stated that the first pick in chess is debated to this day, thousands of years later. That was my entire point; the first pick is not of greater value. The points you do make are from either experience or opinion, something that has never quite held a candle to math.

Please publish your data as games are logged (and put me on team 2 every time).
Did you actually stop to read and then think about what he said? I have a lot of trouble believing that you both make a living using game theory AND actually have a vague understanding of this game.


Comment below rating threshold, click here to show it.

Dysber

Senior Member

07-01-2010

Just let me give my 2 or 3 cents to this.

Doing the amth for it is pretty useless becasue there are no numbers to start with. All you can do is assume one number or another.
If you start by saying "everything is fair" than chosing a champion and banning him is equal to "choosing to win with 50% probability".

Now the statement says "counterpicking increases you chance to win". So if you not "choose to win" but instead "choose to counterpick" your chances will be 50%+x.
X is unknown and as far as i am aware there is no statistics on counterpicking. So we only can assume it. So let X be 1%.

Now the math should be doable by everyone. If one pick gets 51% of winning, the other should be 49%.
Ok, here is where the complicated stuff begins:

Let there be 20 heroes. Let every hero have at least one counterpick.(no overpowered champion). Furthermore to make it easy, let the counterpick always be the hero+1.
Lets now see the order of the picks:
[--Bans--Picks---------]
ABBA ABBAABBAAB

Worst case scenario
A picks Hero 1
B picks Hero 2
B picks Hero 13
A picks Hero 5
A picks Hero 7
B picks Hero 6
B picks Hero 8
A picks Hero 10
A picks Hero 12
B picks Hero 11

This order would give B 5 heroes with each the ability "counterpick". while A has 0 counterpicks. In forms of Math: Devastating for Team A since B now can arrange their laning with the apropiate counterpick!

Best Case scenario
A picks Hero 1
B picks Hero 2
B picks Hero 3(to make sure A doesnt get counter)
A picks Hero 4
A picks Hero 5(to make sure B doesnt get counter)
B picks Hero 6
B picks Hero 7(to make sure A doesnt get counter)
A picks Hero 8
A picks Hero 9(to make sure B doesnt get counter)
B picks Hero 10

Team A now has 2 counterpicks
Team B now has 3 counterpicks

Since it is a odd number of heroes, there will always be a chance one has more counterpicks than others!
But as it is now, team B has ALWAYS the CHANCE to be this one group!


So, lets get in there a little bit further, as draft mode is like this:
[--Bans--Picks---------]
ABBA ABBAABBAAB

Team A starts off with the bans.
Question is: waht do you want to accomplish with your banns?
Do you want to bann heroes you dont want to play against?
Do you want to bann heroes you might think are OP?
Do you want to bann heroes you know are a counterpick to your team?

if you jsut bann heroes you dont WANT to play against you are nothing more than flipping a coin.(its not a strategy: "ahhhh i dont want to play against ashe, becasue i hate big boobed women")
As team A, you definitively do NOT want to bann an OP champion!. At least not the most OP one. You are first to pick a champion, so you can pick the OP champion! As team A you want to bann OP number 2 and 3.
On the other side, as Team B you want to get rid of the OP champion because team A is going to pick first!

So, lets check the math again. For easy numbers, lets say an OP champion has the winning chance of 60%. And lets assume there is no counterpick to an OP chmpion(he wouldnt be OP than, would he). Furthermore, let 4 of the 20 heroes be OP(hero 17-20, 20 beeing the most OP)
[--Bans--Picks---------]
ABBA ABBAABBAAB

Lets not check the worst case scenario, as this is something like "5 friends "just palying to win" vs "5vompetative players", its not really interesting to see who wins in this scenario. so lets assume team A and Team B are trying to make the best choices to win!

Bans:
A Bans 19
B Bans 20
B Bans 18
A Bans 5

A picks 17 (OP, no counter possible)
B picks 1
B picks 2 (to make sure A doesnt get counter)
A picks 3
A picks 4
B picks 6
B picks 7 (to make sure A doesnt get counter)
A picks 8
A picks 9
B picks 10

Lets see how it went:
A got: 1 OP; 2 counterpicks
B got: 1 counterpick!

Suddenly the chances of B winning have decreased to a minimum!
Team A winning strategy is to NOT TO BAN THE OP CHAMION AS SECOND BANN, INSTEAD BAN A COUNTERPICK OF THEIR OWN TEAM!

If there is only one OP champion, team A will not bann him! instead they will ban 2 counterpicks of their team!
Team B on the other side has to ban the OP champion.



So, what if there are no OP champions?
Well Team A has the first bann, randomly banning a champion is throwing a coin -> useless. SO Team a starts with banning counterpicks, now Team B has the chance to bann either counterpicks of themself, or the original hero so the counterpick ban of team a was useless, lets see what happens

Bans:
A Bans 2(counterpick to 1)
B Bans 1(ban of the original)
B Bans 4(counterpick of 3)
A Bans 3(ban of the original)

=> Its the same scenario as the beginning: B will end off better!


Bans:
A Bans 2(counterpick to 1)
B Bans 5(counterpick to 4)
B Bans 11(counterpick to 10)
A Bans 8

A picks 1
B picks 3
B picks 4 (to make sure A doesnt get counter)
A picks 6
A picks 7 (to make sure B doesnt get counter)
B picks 9
B picks 10 (to make sure A doesnt get counter)
A picks 12
A picks 13 (to make sure B doesnt get counter)
B picks 14

Team A:
0 counterpicks!

Team B:
1 counterpick!



So, now we can conclude:
- Two clever choosing teams will use the right ban strategy
- Having the "wrong number of OP champions" will result in an devastating disadvantage for Team B!
- Having no OP champions will result in Team B always beeing better of by 1 counterpick!
- Letting Team A have one ban more than team B COULD make the field even.
- Having ONE OP champion would have the same result as having one ban more for team A
- Team A should not Bann the most OP champion in the game!
- Team B ALWAYS should ban the most OP champion in the game!

Edit:
If team B gets the one counterpick, it should always try to counterpick Team As mid lane, that would be the biggest disadvantage of Team A


Comment below rating threshold, click here to show it.

Kayzar

Senior Member

07-02-2010

Quote:
Originally Posted by Makbeth View Post
Tell you what. Please put in a special exception just for me. I'd like to always be placed on Team 2. No more arguments from this end.
And I'd like the exception to always be on FuYA's team whenever (if ever) I play against Makbeth.


Comment below rating threshold, click here to show it.

Makbeth

Member

07-08-2010

I appreciate that it's hard to believe without a full out report, but the only thing that will convince the riot community is likely to be the statistical proof for both arranged and random teams... Unless you find a spare hundred grand laying around to pay for an analysis. Random teams will be most accurate I am sure, so please take the time to allow the logging and public posting of game results. Names aren't necessary, but team makeup and results would be appreciated. This will benefit both the community and Riot.


123