Explore GameReplays...

Battle for Middle Earth

Ladder error

Closed Topic Start new topic
# 1HungryTroll Jan 26 2018, 20:51 PM
Today i played vs RPO and ladder generated 3 times in a row RPO hosts:
Spoiler
Attached Image



I decided to check how it works:
Spoiler
Attached Image
Attached Image

He generated 6 host games and only 2 offhost (look at the screenshot). Why he not generated host/offhost > host/offhost games for two players?
This generating is very bad, playing 6 games in offhost.

It would be good to correct this issue

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 2Earendil Jan 26 2018, 21:05 PM
host is random.


Posts: 5,704

Clan: Sons of Elendil

Game: Battle for Middle Earth


+
# 3HungryTroll Jan 26 2018, 21:16 PM
QUOTE(Earendil @ Jan 27 2018, 02:05 AM) *

host is random.

ohmy.gif Is this rightly honestly of 2 players?

But if it comes 10 times offhost in a row?

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 4Earendil Jan 26 2018, 22:10 PM
then you got bad luck tongue.gif


Posts: 5,704

Clan: Sons of Elendil

Game: Battle for Middle Earth


+
# 5HungryTroll Jan 26 2018, 23:10 PM
It's like something silly invented. We do not play roulette.
Should be fair to both players

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 6Beterwel Jan 27 2018, 00:35 AM
QUOTE(HungryTroll @ Jan 27 2018, 01:10 AM) *

It's like something silly invented. We do not play roulette.
Should be fair to both players

Random host is probably the most fair way on condition that the empiric chance (chance established after a lot of repetition) is near 50%. I assume this is the way host is determined.

On EA server some names were 'uber-hosters' that would host like 95% of the games in QM. This system doesn't seem so bad imo.

Posts: 391

Game: Battle for Middle Earth


+
# 7HungryTroll Jan 27 2018, 01:19 AM
QUOTE(Beterwel @ Jan 27 2018, 05:35 AM) *

Random host is probably the most fair way on condition that the empiric chance (chance established after a lot of repetition) is near 50%. I assume this is the way host is determined.

On EA server some names were 'uber-hosters' that would host like 95% of the games in QM. This system doesn't seem so bad imo.

Than this system better host/offhost/host/offhost?
From point of view justice and mathematical statistics, this system is bad.
I know, if the number of games → ∞, then on the average 50/50.

But, at small game numbers, deviations may be greater (6< games in offhost in at once).
It's not very pleasant to play a few games in a row on the offhost wacko.gif


Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 8suturbs Jan 27 2018, 02:00 AM
Hungry is right to a point. It should be random but not mere random. I.E: In 50 count, it should converge to 25 to 25. Meanwhile possible to have offhost/onhost 10+ times in a row.

In real life it will never be infinite amount of games. Hence will never be "justice". Luck will rule.. Need such programming codes into the generator to make it a more meaningful QM. Programmer needed. May pay programmers instead of tourney winners.

This post has been edited by suturbs: Jan 27 2018, 02:07 AM

Posts: 861

Game: Battle for Middle Earth


+
# 9Amateur`Hour Jan 27 2018, 02:11 AM
There's gonna be streaks like that, you hosted 3 out of the 4 vs me today. In a smaller sample size you can get results like that, but after hundreds of games most players should be around 50% host in 1v1

Posts: 2,994

Game: League of Legends


+
# 10suturbs Jan 27 2018, 05:38 AM
It will practically never converge %50-50, coz infinity is a theorical issue, as the number incereases has nothing to do with infitinity. In fact, most likely there will never be a constant %50-%50 picture. (I guess it will go around between %30-%70 most of the times)

Moreover, not only the host is random, the army is also randomly generated. Which multiplies the uncertainity.

Conclusionally saying, designing and improving meaningful tournmanets (in terms of army and host) is easier than to hire a software developer for Ladders.

Posts: 861

Game: Battle for Middle Earth


+
# 11RP0 Jan 27 2018, 09:06 AM
QUOTE(HungryTroll @ Jan 26 2018, 20:51 PM) *

Today i played vs RPO and ladder generated 3 times in a row RPO hosts:
Spoiler
Attached Image

I decided to check how it works:
Spoiler
Attached Image
Attached Image

He generated 6 host games and only 2 offhost (look at the screenshot). Why he not generated host/offhost > host/offhost games for two players?
This generating is very bad, playing 6 games in offhost.

It would be good to correct this issue


Lol we didnt even play the 3rd game I got on my host. Instead, we drew again. You got gon mirror host on fangorn forest which you also lost badly. So I dont see the point of your complaining...

To he able to judge the randomness of the qm system you have to at least have a sample of at least 100 games, not like 3.

Posts: 352


+
# 12HungryTroll Jan 27 2018, 09:35 AM
Immediately visible people familiar with mathematics (Suturbs)

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 13HungryTroll Jan 27 2018, 09:36 AM
QUOTE(RP0 @ Jan 27 2018, 14:06 PM) *

Lol we didnt even play the 3rd game I got on my host. Instead, we drew again. You got gon mirror host on fangorn forest which you also lost badly. So I dont see the point of your complaining...

To he able to judge the randomness of the qm system you have to at least have a sample of at least 100 games, not like 3.

Why a hundred? 10 is enough to see.
Check out my second spoiler

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 14suturbs Jan 27 2018, 13:22 PM
If there could be some regulation, in every 20 games, 10 could be hosted by each. As extreme condition, after 10host in a row, next 10 games could be offhost and the balance in every 20 games to be met.

(20 coz considering 100 games barely played by someone.)

Posts: 861

Game: Battle for Middle Earth


+
# 15Beterwel Jan 27 2018, 13:30 PM
QUOTE(HungryTroll @ Jan 27 2018, 11:36 AM) *

Why a hundred? 10 is enough to see.
Check out my second spoiler

lol. I can tell you are not familiar with statistics tho. 10 is not a large enough sample size to draw statistical conclusions from, to assume normality at least 30 are needed and even this might not be representative.

Posts: 391

Game: Battle for Middle Earth


+
# 16suturbs Jan 27 2018, 14:02 PM
when calculating average values from randomly generated data, especially natural stuff, enough sample needed. depends of the dimension of the sample. 10 may not be enough. like climate determination using rain-temperature regimes, I guess they take at least 30 years.

but this is a litle different. coz its not average issue, not climate average determination over rain-temperature regime's average values. Also considering average players game count in a year, i would say 20 is ok. guys let someone able to have 10 onhost in a row max, it doesnt hurt. It urges people to play at least 15 20 games which fits with the sense of activity promotion..

This post has been edited by suturbs: Jan 27 2018, 14:07 PM

Posts: 861

Game: Battle for Middle Earth


+
# 17HungryTroll Jan 27 2018, 14:07 PM
QUOTE(Beterwel @ Jan 27 2018, 18:30 PM) *

lol. I can tell you are not familiar with statistics tho. 10 is not a large enough sample size to draw statistical conclusions from, to assume normality at least 30 are needed and even this might not be representative.

Im rly sorry, but u talking wrong opinion.
Do u know what is statistic?

It is applicable for a large number of repetitions, 100+ experimentation. Statistics is right when number of tests goes to infinity. Or if u want in mathematic symbols:
lim random (host game; offhost game) = 50/50
games → ∞
When we talking about game, we can not rely on statistics, coz the number of games can not be large.
I cant play for 2 hours 100 games. I can play only 5-6 games, and as can be seen in the screenshot, i can get all games in offhost.


Hold 1 more screenshot:
Spoiler: 4games ooffhost 1 host
Attached Image

Tell me we playing roulette or in serious sports competition? At tournaments it is accepted host/offhost then neutral host right? Why in QM no?

This post has been edited by HungryTroll: Jan 27 2018, 14:40 PM

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 18Fishy Jan 27 2018, 16:06 PM
How can people sounds so "smart" (suturbs and HungryTroll) and yet they be so wrong??

There is a simple random generating program that work as it should. Implementing more functions like remembering your onhost/offhost situation would need to implement much more completely unnecessary work.


The more games you play the closer you will be to 50-50 host situation. That is how this random fairness works.

Play 10 games and you probably can end up 7-3..that is harsh...but for ex. 52-48 is ok yet its the same difference.

Guys dont make fool of yourself and just suck up these exceptions. Next time it might be you hosting 4 games in a row. But then you wont be coming here crying about unfairness I bet.

Posts: 1,918

Clan: Sons of Elendil

Game: Battle for Middle Earth


+
# 19HungryTroll Jan 27 2018, 16:09 PM
QUOTE(Fishy @ Jan 27 2018, 21:06 PM) *

How can people sounds so "smart" (suturbs and HungryTroll) and yet they be so wrong??

There is a simple random generating program that work as it should. Implementing more functions like remembering your onhost/offhost situation would need to implement much more completely unnecessary work.
The more games you play the closer you will be to 50-50 host situation. That is how this random fairness works.

Play 10 games and you probably can end up 7-3..that is harsh...but for ex. 52-48 is ok yet its the same difference.

Guys dont make fool of yourself and just suck up these exceptions. Next time it might be you hosting 4 games in a row. But then you wont be coming here crying about unfairness I bet.

Okay, I just offered my opinion to make a ladder better.
I doubt that remembering the sequence 1 0 1 0 will create a serious load on the server...

Posts: 1,000

Clan: The Last Knights of Middle-Earth

Game: Battle for Middle Earth


+
# 20Xehanort Jan 27 2018, 16:40 PM
I assume you made this topic, not because you were offhost 3/4, but because you also lost these games.

If you draw 10 games the probability of getting 5 on host and 5 offhost is like 25%. It's hardly enough data to come to a conclusion. Over time these things eventually even out...

Posts: 1,374

Game: Battle for Middle Earth


+

1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)