Forum


the possibility you win
191078819 wrote
at 7:49 PM, Sunday October 19, 2008 EDT
hi all!

i calculated the possibility that attackers and defencer(?) will win. ill show below.

(if someone already calculated, sorry)

a = the number of dices that attacker owns.
d = the number of dices that defencer owns.
p(a,b) = the possibility that attacker will win.


p(8,8) = 0.470
p(8,7) = 0.673
p(8.6) = 0.843
p(8,5) = 0.947
p(8,4) = 0.989
p(8,3) = 0.999 (0.9990)
p(8.2) = 0.999 (0.99997)
p(8,1) = 1.000

p(7,8) = 0.274
p(7,7) = 0.468
p(7,6) = 0.685
p(7,5) = 0.862
p(7,4) = 0.961
p(7,3) = 0.994
p(7,2) = 0.999 (0.9998)
p(7,1) = 1.000

p(6,8) = 0.121
p(6,7) = 0.259
p(6,6) = 0.466
p(6,5) = 0.699
p(6,4) = 0.883
p(6,3) = 0.975
p(6,2) = 0.998
p(6,1) = 0.999 (0.99995)

p(5,8) = 0.036
p(5,7) = 0.103
p(5,6) = 0.242
p(5,5) = 0.463
p(5,4) = 0.717
p(5,3) = 0.909
p(5,2) = 0.987
p(5,1) = 0.999 (0.9998)

p(4,8) = 0.006
p(4,7) = 0.025
p(4,6) = 0.083
p(4,5) = 0.220
p(4,4) = 0.459
p(4,3) = 0.742
p(4,2) = 0.939
p(4,1) = 0.997

p(3,8) = 0.000 (0.0004)
p(3,7) = 0.003
p(3,6) = 0.014
p(3,5) = 0.060
p(3,4) = 0.191
p(3,3) = 0.453
p(3,2) = 0.778
p(3,1) = 0.973

p(2,8) = 0.000 (0.000005)
p(2,7) = 0.000 (0.00007)
p(2,6) = 0.000 (0.0007)
p(2,5) = 0.006
p(2,4) = 0.035
p(2,3) = 0.152
p(2,2) = 0.443
p(2,1) = 0.837

(needless to say, for example, p(7,8) + p(8,7) is not equal to 1)

what do you feel about these numbers??


« First ‹ Previous Replies 11 - 20 of 26 Next › Last »
ChristianSoldier wrote
at 10:23 AM, Monday October 20, 2008 EDT
The link is correct. I calculated the same values. Only it's not 36/10,000.. it is 36/10,000,000 as I said earlier (actually I was even more precise).

There is clearly a precision error in the OP's code.
Thraxle wrote
at 10:51 AM, Monday October 20, 2008 EDT
Well CS, you misquoted me a bit as I said it was per 100,000 rolls, but you are correct. It's every 10 million rolls. This means the 6v1 fails once every 277,778 rolls while the 8v2 fails once every 60,240 rolls.

I can't believe I actually agree with something you said!
ChristianSoldier wrote
at 11:15 AM, Monday October 20, 2008 EDT
Mixed emotions... Your agreeing with me increases the probability of error in my comments.
montecarlo wrote
at 11:58 AM, Monday October 20, 2008 EDT
can someone else (cus im too lazy) print out the least probable rolls in order with percentages, or wins out of 10,000,000? ofc, dont include the impossible rolls, i.e. 8v1 defense. so, basically, 2v8, 6v1, 8v2, 2v7, etc.... muchos gracias.
kam|k2 wrote
at 12:00 PM, Monday October 20, 2008 EDT
yeah, im interrested in the chance of winning 2v3, 3v4 etc, as well.
skrumgaer wrote
at 12:02 PM, Monday October 20, 2008 EDT
I have tried working it out on a spreadsheet of all 8-combinations versus all 2-combinations but I don't think I have the proper distribution of 8-combination. My series of frequencies for rolls of 8, 9, 10, etc. for 8 dice goes as

1 8 36 120 330 792 1716 3432 etc.

and I think that is not correct. Does someone have the correct series? For the series I have, an 8 v 2 loses 1,001 out of 191,814,480 rolls or once every 191,622 rolls.
Thraxle wrote
at 12:28 PM, Monday October 20, 2008 EDT
Possibility of failure

6v1 1/277,778
8v2 1/60,241
5v1 1/6,667
7v2 1/5,033
8v3 1/1,074
6v2 1/561
4v1 1/370
7v3 1/187
8v4 1/96
5v1 1/83
6v3 1/40
3v1 1/37
7v4 1/26
8v5 1/19
4v2 1/16
5v3 1/11

All of the above rolls have greater than 90% chance of winning.

Possibility of success

2v8 1/212,766
2v7 1/14,104
3v8 1/2,213
2v6 1/1,305
3v8 1/346
2v5 1/164
4v8 1/157
3v6 1/67
4v6 1/39
2v4 1/28
5v8 1/27
3v5 1/16
4v6 1/12

All of these rolls have less than 10% chance of winning.
Thraxle wrote
at 12:38 PM, Monday October 20, 2008 EDT
Feel free to debate whether my numbers are right. I realize that I have contradicted skrum's 8v2 statistic............which probably means I'm wrong. I used the percentages given on the wiki site and did simple division based on 5 decimal places for the percentage and 10,000,000 rolls.
skrumgaer wrote
at 12:41 PM, Monday October 20, 2008 EDT
Thraxle,

Your having contradicted skrum's 8 v 2 is not evidence that you are wrong.
Thraxle wrote
at 12:43 PM, Monday October 20, 2008 EDT
Just paying my due respect to the stat master Skrum.
KDice - Multiplayer Dice War
KDice is a multiplayer strategy online game played in monthly competitions. It's like Risk. The goal is to win every territory on the map.
CREATED BY RYAN © 2006 - 2026
GAMES
G GPokr
Texas Holdem Poker
K KDice
Online Strategy
X XSketch
Online Pictionary