Flaw in computing randomness

Talk about all things related to Conquer Club

Moderator: Community Team

Forum rules
Please read the community guidelines before posting.
zndow
Posts: 9
Joined: Thu Aug 21, 2008 8:57 pm

Flaw in computing randomness

Post by zndow »

Hi everyone!

Players are talking about streaks of bad dice rolls and I totally agree with them.
Players are talking about statistically correct randomness and I do not agree at 100%. Here why.

It is a fact that the dice are random without any doubt. But there is something missing in this randomness and it is the control over time of this actual randomness.

Let me clarify.

If you test CC randomness engine you will find out after a long series of throws that the randomness is absolutely correct on the long run (Statistically speaking).

Exactly on the long run, that is the problem, LONG RUN.

If there is no controls installed for short terms, then the dice are simply out of statics. You then see very often a player losing an almost winning position over a short term bad streak of dice rolls even with odds over 90% chance of winning a fight.

Example: losing 15 against 2, when you should often lose maybe 3-5 units. Do not get me wrong here but losing 5 units against 2 it is call bad streak in my book and it is what i expect when it is not going my way. Not losing 15 units. That is the kind of bad situation I think about when I prepare my tactics and strategies.

That is the short term problem of computing randomness.

Why do we see the highest rank players playing freestyle games only? Because they do not have to worry about the flaw in this dice randomness.
Why we do not see any very good strategists or tacticians on the top rank. Because there strategies or tactics often get kill by the flaw of the dice randomness.

Why those best tacticians always get a tremendous stack of units before doing a fight? Because they know if they want the best chance to win the fight they need to be able to do a lot of rolls in a single fight to truly hit there odds.

Right now I say the randomness is not good for the type of game CC offer. Until this flaw is fixed, only freestyle players will hit the top ranks because it is impossible to control your position correctly when the dice decide of your faith and not your better strategies or tactics.

Very good player understand they cannot win every fights easily and they prepare themselves for that. But when you cannot even control that then guess what!

"noooooooooo not again i had a winning position"
" Ah come on, I just lost 7 against 1, that's BS, now I am dead for sure when I worked so hard to get out of this bad position"
" blah blah blah blah....."

Have a nice day everyone,
User avatar
Johnny Rockets
Posts: 568
Joined: Sun May 13, 2007 9:58 pm
Gender: Male
Location: Winnipeg, Canada
Contact:

Re: Flaw in computing randomness

Post by Johnny Rockets »

I would tend to agree.


JRock
User avatar
Symmetry
Posts: 9255
Joined: Sat Feb 24, 2007 5:49 am

Re: Flaw in computing randomness

Post by Symmetry »

zndow wrote:Hi everyone!

Players are talking about streaks of bad dice rolls and I totally agree with them.
Players are talking about statistically correct randomness and I do not agree at 100%. Here why.

It is a fact that the dice are random without any doubt. But there is something missing in this randomness and it is the control over time of this actual randomness.

Let me clarify.

If you test CC randomness engine you will find out after a long series of throws that the randomness is absolutely correct on the long run (Statistically speaking).

Exactly on the long run, that is the problem, LONG RUN.

If there is no controls installed for short terms, then the dice are simply out of statics. You then see very often a player losing an almost winning position over a short term bad streak of dice rolls even with odds over 90% chance of winning a fight.

Example: losing 15 against 2, when you should often lose maybe 3-5 units. Do not get me wrong here but losing 5 units against 2 it is call bad streak in my book and it is what i expect when it is not going my way. Not losing 15 units. That is the kind of bad situation I think about when I prepare my tactics and strategies.

That is the short term problem of computing randomness.

Why do we see the highest rank players playing freestyle games only? Because they do not have to worry about the flaw in this dice randomness.
Why we do not see any very good strategists or tacticians on the top rank. Because there strategies or tactics often get kill by the flaw of the dice randomness.

Why those best tacticians always get a tremendous stack of units before doing a fight? Because they know if they want the best chance to win the fight they need to be able to do a lot of rolls in a single fight to truly hit there odds.

Right now I say the randomness is not good for the type of game CC offer. Until this flaw is fixed, only freestyle players will hit the top ranks because it is impossible to control your position correctly when the dice decide of your faith and not your better strategies or tactics.

Very good player understand they cannot win every fights easily and they prepare themselves for that. But when you cannot even control that then guess what!

"noooooooooo not again i had a winning position"
" Ah come on, I just lost 7 against 1, that's BS, now I am dead for sure when I worked so hard to get out of this bad position"
" blah blah blah blah....."

Have a nice day everyone,


I'm kind of unsure about what you're suggesting here. Maybe that certain types of attack should not have a random element? Or that the dice aren't random at the moment?

Maybe you feel that there should be a limit on the random element? That in your 15 vs 2 scenario, it should be random whether you lose 2 or 5 troops, but you should win nevertheless.

Still, I'd recommend chess.
zndow
Posts: 9
Joined: Thu Aug 21, 2008 8:57 pm

Re: Flaw in computing randomness

Post by zndow »

I played chess for several years actually.

What i am saying is if odds tell me that with 5 units against 1 units i should win the fight 95% of the time then i expect exactly that. And right now it is not the case because of this flaw in the randomness. I need like 40 units against 8 (x8) to start hitting those odds.
User avatar
Robinette
Posts: 2944
Joined: Sat Apr 29, 2006 1:32 pm
Location: Northern California

Re: Flaw in computing randomness

Post by Robinette »

If you have 6 fronts, each with a 90% chance of winning...
then you really only have a 53% statistical chance of succeeding...
(.90 x.90 x.90 x.90 x.90 x.90 = .53)

In other words... you will LOSE almost HALF the time, even though the odds appeared to be 90%



and 5 vs 1 is 97%
Last edited by Robinette on Sat Sep 26, 2009 11:23 am, edited 1 time in total.
Image
User avatar
Symmetry
Posts: 9255
Joined: Sat Feb 24, 2007 5:49 am

Re: Flaw in computing randomness

Post by Symmetry »

zndow wrote:I played chess for several years actually.

What i am saying is if odds tell me that with 5 units against 1 units i should win the fight 95% of the time then i expect exactly that. And right now it is not the case because of this flaw in the randomness. I need like 40 units against 8 (x8) to start hitting those odds.


I'm still not sure how you propose fixing this. Are you saying that 5 rolls out of every hundred in the theoretical 5 vs 1 scenario above should be set as losses? Doesn't seem right to me. What if you won the previous 95 encounters? You'd know that the next 5 would be losses.

I think you might have run in to a common problem when people think about randomness. Previous rolls don't affect the outcome of future rolls.

I'd suggest taking a deep breath, a glass of decent Scotch (or if not a drinker, a mug of hot cocoa) and stepping back a bit.
zndow
Posts: 9
Joined: Thu Aug 21, 2008 8:57 pm

Re: Flaw in computing randomness

Post by zndow »

no robinette i belive you are mistaking.

If i have 6 fronts with 90% of winning them each then there is 53% chance that 1 of them will be lost and not losing half of the time.
zndow
Posts: 9
Joined: Thu Aug 21, 2008 8:57 pm

Re: Flaw in computing randomness

Post by zndow »

Actually it is the other way. 47% chance for the defender to stay alive on 1 of the 6 fronts and 53% chance to win them all for the attacker.
User avatar
Robinette
Posts: 2944
Joined: Sat Apr 29, 2006 1:32 pm
Location: Northern California

Re: Flaw in computing randomness

Post by Robinette »

zndow wrote:no robinette i belive you are mistaking.

If i have 6 fronts with 90% of winning them each then there is 53% chance that 1 of them will be lost and not losing half of the time.



Let me clarify...

If you are going for the kill
, then you will lose nearly Half the time, despite the 90% odds...
I can see no other reason to attack someone so hard, but then again, i've never played 1v1
Image
User avatar
obliterationX
Posts: 953
Joined: Tue Dec 23, 2008 1:52 pm
Gender: Female
Location: Yeah

Re: Flaw in computing randomness

Post by obliterationX »

100% correct.
User avatar
hwhrhett
Posts: 3120
Joined: Fri Jun 02, 2006 8:55 pm
Gender: Male
Location: TEXAS --- The Imperial Dragoons

Re: Flaw in computing randomness

Post by hwhrhett »

Robinette wrote:
zndow wrote:no robinette i belive you are mistaking.

If i have 6 fronts with 90% of winning them each then there is 53% chance that 1 of them will be lost and not losing half of the time.



Let me clarify...

If you are going for the kill
, then you will lose nearly Half the time, despite the 90% odds...
I can see no other reason to attack someone so hard, but then again, i've never played 1v1



never?
Image
User avatar
Georgerx7di
Posts: 2277
Joined: Mon Dec 11, 2006 7:11 pm
Gender: Male

Re: Flaw in computing randomness

Post by Georgerx7di »

zndow wrote:I played chess for several years actually.

What i am saying is if odds tell me that with 5 units against 1 units i should win the fight 95% of the time then i expect exactly that. And right now it is not the case because of this flaw in the randomness. I need like 40 units against 8 (x8) to start hitting those odds.



Btw, just so everyone know, 5v1 is way less than 95%. 6v1 is only about 90%.
User avatar
the.killing.44
Posts: 4724
Joined: Thu Oct 23, 2008 7:43 pm
Gender: Male
Location: now tell me what got two gums and knows how to spit rhymes
Contact:

Re: Flaw in computing randomness

Post by the.killing.44 »

hwhrhett wrote:
Robinette wrote:
zndow wrote:no robinette i belive you are mistaking.

If i have 6 fronts with 90% of winning them each then there is 53% chance that 1 of them will be lost and not losing half of the time.



Let me clarify...

If you are going for the kill
, then you will lose nearly Half the time, despite the 90% odds...
I can see no other reason to attack someone so hard, but then again, i've never played 1v1



never?

http://www.conquerclub.com/player.php?m ... &so=D&np=2
User avatar
owenshooter
Posts: 13311
Joined: Wed Mar 07, 2007 6:01 pm
Gender: Male
Location: Deep in the Heart of Tx

Re: Flaw in computing randomness

Post by owenshooter »

this would be a great discussion to have here if CC were responsible for the dice, but they aren't. shoot your concerns to random.org, since they supply the dice... good luck...-0

p.s.-they also have a nifty little area where they explain the randomness of their generated dice...
Image
Thorthoth,"Cloaking one's C&A fetish with moral authority and righteous indignation
makes it ever so much more erotically thrilling"
User avatar
Robinette
Posts: 2944
Joined: Sat Apr 29, 2006 1:32 pm
Location: Northern California

Re: Flaw in computing randomness

Post by Robinette »

Georgerx7di wrote:
zndow wrote:I played chess for several years actually.

What i am saying is if odds tell me that with 5 units against 1 units i should win the fight 95% of the time then i expect exactly that. And right now it is not the case because of this flaw in the randomness. I need like 40 units against 8 (x8) to start hitting those odds.



Btw, just so everyone know, 5v1 is way less than 95%. 6v1 is only about 90%.



Pfft... don't even need a calculator to see you're wrong george...

5v1... just do it in your head...
1st roll has a 1/3rd chance of losing,
2nd roll also has a 1/3rd chance of losing (so 1 in 9 is about 11% at this point)
3rd roll has about a 50/50 chance (so our 11% is now about 5 1/2%
4th roll has a 2/3rds chance of losing (so remove a third of that 5 1/2 and you've got about 3 1/2%)

100 - 3.5 = 96.5% Approx chance of success....... lol... and without a calculator!


so lets do the simple math in our head for 6 v 1, shall we...

same as above, except 3rd roll has 1/3rd chance of losing (so now it's about 1 in 27, which is LESS than 4%)
4th roll is 50/50 (so now we are LESS than 2%)
and the 5th roll, with a 2/3rds chance of losing, brings us to what... 4/3rds... so LESS than 1.33%

100 - 1.33% = 98.7% approx chance of success............
go ahead, grab a calculator, check my estimated math...
it won't be exact... but i expect it will be close...
Image
User avatar
Timminz
Posts: 5579
Joined: Tue Feb 27, 2007 1:05 pm
Gender: Male
Location: At the store

Re: Flaw in computing randomness

Post by Timminz »

6v1 = 99.03171%
5v1 = 97.15441%
User avatar
the.killing.44
Posts: 4724
Joined: Thu Oct 23, 2008 7:43 pm
Gender: Male
Location: now tell me what got two gums and knows how to spit rhymes
Contact:

Re: Flaw in computing randomness

Post by the.killing.44 »

Timminz wrote:6v1 = 99.03171%
5v1 = 97.15441%

Which in (CC) reality is total bullshit; no way I've won 5v1 97% of the time.
User avatar
Timminz
Posts: 5579
Joined: Tue Feb 27, 2007 1:05 pm
Gender: Male
Location: At the store

Re: Flaw in computing randomness

Post by Timminz »

the.killing.44 wrote:
Timminz wrote:6v1 = 99.03171%
5v1 = 97.15441%

Which in (CC) reality is total bullshit; no way I've won 5v1 97% of the time.

Proof? Or are you just making things up?
User avatar
the.killing.44
Posts: 4724
Joined: Thu Oct 23, 2008 7:43 pm
Gender: Male
Location: now tell me what got two gums and knows how to spit rhymes
Contact:

Re: Flaw in computing randomness

Post by the.killing.44 »

Timminz wrote:
the.killing.44 wrote:
Timminz wrote:6v1 = 99.03171%
5v1 = 97.15441%

Which in (CC) reality is total bullshit; no way I've won 5v1 97% of the time.

Proof? Or are you just making things up?

I would love to find proof for that.
slowreactor
Posts: 1356
Joined: Tue Jan 20, 2009 3:34 pm
Gender: Male
Location: Ithaca, NY

Re: Flaw in computing randomness

Post by slowreactor »

the.killing.44 wrote:
Timminz wrote:
the.killing.44 wrote:
Timminz wrote:6v1 = 99.03171%
5v1 = 97.15441%

Which in (CC) reality is total bullshit; no way I've won 5v1 97% of the time.

Proof? Or are you just making things up?

I would love to find proof for that.


Actually, those calculations assume that you keep attacking all the way down to 1. Seeing as just about everybody stops at 3, your percentage will be somewhat lower.
User avatar
xelabale
Posts: 452
Joined: Sun Sep 21, 2008 8:12 am

Re: Flaw in computing randomness

Post by xelabale »

So the OP is saying that randomness is only true over a large sample, but inaccurate for a small sample? He's saying that's why good players adopt strategies to deal with that? He's saying that sometimes the dice aren't dead on their predicted outcome, but that they even out?

Well I never, I'm gonna have to rethink this whole concept of randomness...
User avatar
Robinette
Posts: 2944
Joined: Sat Apr 29, 2006 1:32 pm
Location: Northern California

Re: Flaw in computing randomness

Post by Robinette »

xelabale wrote:So the OP is saying that randomness is only true over a large sample, but inaccurate for a small sample? He's saying that's why good players adopt strategies to deal with that? He's saying that sometimes the dice aren't dead on their predicted outcome, but that they even out?

Well I never, I'm gonna have to rethink this whole concept of randomness...


wait---- What?......
sooooo, does this mean we should actually adjust our strategy to anticipate that we will LOSE men when rolling...

i think i get it now...

With 3 rolls, in the LONG RUN the stats say i'll win... win 2, win 1, win 0

But if we look at this the other way --- the rolls actually look like this.... lose 2, lose 1, and lose 0

So with 6 v 2 you lose the 1st roll (lose 2), now it's 4 v2...
2nd roll at 4 v 2 and i'll anticipate that i lose 1, now it's 3 v 1
uh oh... 2 dice vs 1... i'm screwed... MAYBE if i go down to 1 i'll likely get it, but ouch...

So the above 6 v 2 odds are supposed to be 89%, and if i take the pain of going down to 1 it will likely happen just like that...

I really think what we are missing here, is that with an 89% chance of victory, we expect to win handily...
but look again at that 3v1 position after the 2nd roll... which is precisely within the odds... yet very disheartening!


It's like yesterday...
i caught every RED light...
some days it's all GREEN...
but usually it's 6 of one, half dozen of the other... lol...


so try this little exercise...
it only takes 2 rolls to completely win, or lose... and yet................

Image
Image
User avatar
Crazyirishman
Posts: 1564
Joined: Sun Jan 04, 2009 8:05 pm
Gender: Male
Location: Dongbei China

Re: Flaw in computing randomness

Post by Crazyirishman »

Robinette wrote:
xelabale wrote:So the OP is saying that randomness is only true over a large sample, but inaccurate for a small sample? He's saying that's why good players adopt strategies to deal with that? He's saying that sometimes the dice aren't dead on their predicted outcome, but that they even out?

Well I never, I'm gonna have to rethink this whole concept of randomness...


wait---- What?......
sooooo, does this mean we should actually adjust our strategy to anticipate that we will LOSE men when rolling...

i think i get it now...

With 3 rolls, in the LONG RUN the stats say i'll win... win 2, win 1, win 0

But if we look at this the other way --- the rolls actually look like this.... lose 2, lose 1, and lose 0

So with 6 v 2 you lose the 1st roll (lose 2), now it's 4 v2...
2nd roll at 4 v 2 and i'll anticipate that i lose 1, now it's 3 v 1
uh oh... 2 dice vs 1... i'm screwed... MAYBE if i go down to 1 i'll likely get it, but ouch...

So the above 6 v 2 odds are supposed to be 89%, and if i take the pain of going down to 1 it will likely happen just like that...

I really think what we are missing here, is that with an 89% chance of victory, we expect to win handily...
but look again at that 3v1 position after the 2nd roll... which is precisely within the odds... yet very disheartening!


It's like yesterday...
i caught every RED light...
some days it's all GREEN...
but usually it's 6 of one, half dozen of the other... lol...


so try this little exercise...
it only takes 2 rolls to completely win, or lose... and yet................

Image


Wouldn't it take 4? Since you can only go one space at a time.
AAFitz
Posts: 7270
Joined: Sun Sep 17, 2006 9:47 am
Gender: Male
Location: On top of the World 2.1

Re: Flaw in computing randomness

Post by AAFitz »

zndow wrote:I played chess for several years actually.

What i am saying is if odds tell me that with 5 units against 1 units i should win the fight 95% of the time then i expect exactly that. And right now it is not the case because of this flaw in the randomness. I need like 40 units against 8 (x8) to start hitting those odds.


BS... ive played enough games to know 5 vs 1 does win 95%... I almost never ever lose 5 vs 1.. .it of course does happen occasionaly, but probably only 5% of the time.

If you want to prove your dice are not random...write down all of your rolls, then present them. But saying you lost 5 to 1 is just not information...or no more information than me saying they are random in any case.

Statistically over time the dice are random. Over the short period, you have absolutely no idea what rolls you may get....that is random. The fact that you cant predict them, or that they dont follow the stats only goes to prove they are random.

If they followed the exact expected statistical forecast, they would not be random.

Me and mine just played 10 games of duel solitaire. There are very little decisions to be made, so the game is mostly luck based, and I am certainly no better than she was.... yet... I won 9 games straight before she won. 9 games. On every game she had a 50/50 chance of winning, but I won each time for 9 straight.

The cards were random, they were shuffled, and I did not cheat. Statistically she should have won half, and conceptually thinking she should have maybe won 40% of them...but nope... she won 0%.

She of course did not take any time to write to bicycle and tell them that their cards were broken and not random.

She just denied me of sex for a week like any sane woman would do.
I'm Spanking Monkey now....err...I mean I'm a Spanking Monkey now...that shoots milk
Too much. I know.
User avatar
Robinette
Posts: 2944
Joined: Sat Apr 29, 2006 1:32 pm
Location: Northern California

Re: Flaw in computing randomness

Post by Robinette »

Crazyirishman wrote:
Wouldn't it take 4? Since you can only go one space at a time.


opps... looks like you are right... it can be over in 4 rolls....
But in which direction? Left or Right? You just never know...
That's the problem with Random... lol
Last edited by Robinette on Sun Sep 27, 2009 2:59 pm, edited 1 time in total.
Image
Post Reply

Return to “Conquer Club Discussion”