Statistics

[MartijnF]

Hello,

While playing a game of RISK at Conquer Club, players (as usual) were stacking up units in a single area, making it harder for the other players to take this area from them. I started to wonder how many units I would need for a succesful attack on such a stack. How many units should I stockpile to stand a good chance at winning from a stack? And how many units would I have left?

To solve this problem, I wrote a little simulator which performs attacks using different numbers of units. To make sure the results are reliable, each attack was repeated 100,000 times, using reasonably random dice rolls. I've put the the data that is the result of this simulation here for everyone to look at: http://mspaint.kicks-ass.org/martijn/risk/

I've got no idea if this is of any use or help for anyone, but I thought it was pretty interesting.

Cheers,
MartijnF

[azezzo]

yah well 10 against 2, and i lose 8 supposedly i had a 98.96% chance of winning, or conversely a 1.04% chance of losing

thats why i do not gamble anymore, i can be a "mush"

[Qwert]

this is pointles.
Sometime with 10 army you can kill 9 territories,and sometime with 10 you can not kill one territory who defend 1 army.
Any statistic who tell you that you have 98% of chances to kill some territory,in CC not relevant.

[Fruitcake]

this is pointles.
Sometime with 10 army you can kill 9 territories,and sometime with 10 you can not kill one territory who defend 1 army.
Any statistic who tell you that you have 98% of chances to kill some territory,in CC not relevant.

the legend that is qwert has spoken...nuff said

[Wolffystyle]

Qwert, this is the fundamental concept of PROBABILITY. 98 times out of a 100 you will be successful. The remaining 2 times will result in failure. No, nothing is ever CERTAIN nor GUARANTEED. However, those who roll their dice understanding the governing odds are playing providently.

[Timminz]

Hello,

While playing a game of RISK at Conquer Club, players (as usual) were stacking up units in a single area, making it harder for the other players to take this area from them. I started to wonder how many units I would need for a succesful attack on such a stack. How many units should I stockpile to stand a good chance at winning from a stack? And how many units would I have left?

To solve this problem, I wrote a little simulator which performs attacks using different numbers of units. To make sure the results are reliable, each attack was repeated 100,000 times, using reasonably random dice rolls. I've put the the data that is the result of this simulation here for everyone to look at: http://mspaint.kicks-ass.org/martijn/risk/

I've got no idea if this is of any use or help for anyone, but I thought it was pretty interesting.

Cheers,
MartijnF

That's impressive, and looks like it took a fair bit of work. Myself, I usually use http://www.gamesbyemail.com/Games/Gambit/BattleOdds, or the Assault Odds plugin, you can get in the plugins & add-ons sub-forum of General Discussion.

[Georgerx7di]

this is pointles.
Sometime with 10 army you can kill 9 territories,and sometime with 10 you can not kill one territory who defend 1 army.
Any statistic who tell you that you have 98% of chances to kill some territory,in CC not relevant.

This post has no value. Tim's odds above will help you find out odds of success. People who know nothing about math should not post about such subjects.

[MartijnF]

That's impressive, and looks like it took a fair bit of work. Myself, I usually use http://www.gamesbyemail.com/Games/Gambit/BattleOdds, or the Assault Odds plugin, you can get in the plugins & add-ons sub-forum of General Discussion.

Thanks! Most work was done by my CPU though, it took about 4 days of processing to simulate all of the dice rolls and calculate the results afterwards

[Qwert]

by Wolffystyle » Thu Nov 26, 2009 1:19 am

Qwert, this is the fundamental concept of PROBABILITY. 98 times out of a 100 you will be successful. The remaining 2 times will result in failure. No, nothing is ever CERTAIN nor GUARANTEED. However, those who roll their dice understanding the governing odds are playing providently.

Ofcourse if i attack 50 against 1,then 100% im sure that i will kill these 1.
I only know that theory its one thing,and in reality its diferent.

[Georgerx7di]

by Wolffystyle » Thu Nov 26, 2009 1:19 am

Qwert, this is the fundamental concept of PROBABILITY. 98 times out of a 100 you will be successful. The remaining 2 times will result in failure. No, nothing is ever CERTAIN nor GUARANTEED. However, those who roll their dice understanding the governing odds are playing providently.

Ofcourse if i attack 50 against 1,then 100% im sure that i will kill these 1.
I only know that theory its one thing,and in reality its diferent.

no, they are not. and 50v1 is not 100%

[Qwert]

no, they are not. and 50v1 is not 100%

how do you mean its not?
IF you play 100 games,and if you have situation 50 against 1,then probably you will in all 100 games kill this 1,and this is 100% succes.

[Qwert]

Actualy not games-i meant to say-100 times

[AndyDufresne]

That's impressive, and looks like it took a fair bit of work. Myself, I usually use http://www.gamesbyemail.com/Games/Gambit/BattleOdds, or the Assault Odds plugin, you can get in the plugins & add-ons sub-forum of General Discussion.

Thanks! Most work was done by my CPU though, it took about 4 days of processing to simulate all of the dice rolls and calculate the results afterwards

Nice work! Always interesting to see such statistics. Though some might have preferred you used Random.org for rolls---since every now and again there are a few users who believe that it is Random.org that isn't random!

But good work.


--Andy

[targetman377]

good job dam that is intresting you put alot of work into it.

[johnkelly00]

no, they are not. and 50v1 is not 100%

how do you mean its not?
IF you play 100 games,and if you have situation 50 against 1,then probably you will in all 100 games kill this 1,and this is 100% succes.

Seriously? You don't understand probabilities very well qwert. You state, "probably" you will [win] in all 100 games. "probably" is not the same thing as "will definitely" and so therefore, the original poster is correct.

It may be semantics but it may also save you from blowing a blood vessel if you ever lose 50 v 1.

[e_i_pi]

I think the crucial point people are missing here is that nothing is relevant

[MartijnF]

no, they are not. and 50v1 is not 100%

how do you mean its not?

When using dice, nothing is 100%. There's always a chance the defender will keep rolling 6's.

Altough when you look at my data, it actually does say there's a 100% chance of winning with 11 against 1. This is (of course) not really true, but happened because I rounded the numbers to two digits after the dot.

I just looked it up in the database and there were actually 5 out of the 100,000 recorded situations in which the attacker (with 11 units) didn't managed to overtake the defended area (with 1 unit). That means the chance of winning is (99,995 / 100,000) which is 0.99995 or 99.995%. This rounds up to 100% in the chart on my web page, but that's a slight distortion of the truth.

[karelpietertje]

this is pointles.
Sometime with 10 army you can kill 9 territories,and sometime with 10 you can not kill one territory who defend 1 army.
Any statistic who tell you that you have 98% of chances to kill some territory,in CC not relevant.

You would think that somebody with such low expectations of the dice wouldn't have to complain about them in every single game, and in every single thread.
Apparently not.

[nippersean]

Oh for goodness sake just download the battle odds plug in.....sacrifice a virgin, and pray to the dice Gods...