I made a small script that gathered results from that result text box. I couldn't find the full data, but I found the results based on a smaller sample I had emailed some guy about a year ago. It's not statistically significant (has only some 1800 results) but already shows some trends.

Notice how even in this small sample the 8v8 is the most common attack type and its success rate is abyssmal (as expected).


Overall statistics by attack type:

attacking dices / number of attacks / % of all attacks / overall success % of attacks

2vX 275 14.64% 62.9%
3vX 366 19.48% 77.3%
4vX 351 18.68% 81.1%
5vX 153 8.14% 90.2%
6vX 123 6.60% 86.3%
7vX 113 6.01% 87.6%
8vX 497 26.45% 61.8%

So attacks by 8 dices is the most common attack type and it's also the least succesful one. Attacks by 5,6 and 7 dices are the most succesfull ones but also the least common ones. This describes how people play - not probabilities of success.


Here's the raw data:

attack wins/attacks success %

2v1 131/162 80.86419753086419%
2v2 42/96 43.75%
2v3 1/10 10%
2v4 0/3 0%
2v6 0/1 0%
2v7 0/3 0%
3v1 129/131 98.47328244274809%
3v2 135/182 74.17582417582418%
3v3 19/45 42.22222222222222%
3v4 0/5 0%
3v5 0/3 0%
4v1 86/86 100%
4v2 109/118 92.37288135593221%
4v3 82/122 67.21311475409836%
4v4 7/22 31.818181818181817%
4v5 0/2 0%
4v6 1/1 100%
5v1 29/29 100%
5v2 48/49 97.95918367346938%
5v3 33/37 89.1891891891892%
5v4 28/35 80%
5v5 0/3 0%
6v1 5/5 100%
6v2 32/32 100%
6v3 32/34 94.11764705882352%
6v4 25/29 86.20689655172414%
6v5 9/16 56.25%
6v6 3/6 50%
6v7 1/1 100%
6v8 0/1 0%
7v1 12/12 100%
7v2 18/18 100%
7v3 15/15 100%
7v4 22/25 88%
7v5 14/15 93.33333333333334%
7v6 7/8 87.5%
7v7 6/6 100%
7v8 5/14 35.71428571428571%
8v1 20/20 100%
8v2 23/23 100%
8v3 22/22 100%
8v4 23/23 100%
8v5 33/33 100%
8v6 20/27 74.07407407407408%
8v7 18/27 66.66666666666666%
8v8 148/322 45.962732919254655%
Sample size 1879
Success 1393:486 74.13517828632251%

lumentum,

Does your data support the idea that attacks by 8s is the most common attack type with also the lowest success rate and attacks by 5,6,7 are the least common but also with the highest success rate?

hippo, the 8vX will probably have the lowest success rate because it has the highest ratio of even rolling. people 8v8 TONS of times. but its rare to see a 7v7 or 6v6, etc... the next most common roll that is even is prolly 2v2.

anyways, the success rate of 8vX makes sense that its the lowest. the 8v8 success seems a bit low, but is that outside of expectation? i dunno.

Try to have a resolution to the game before it reaches 8v8 suckfest. It's the only way to be sure.

Yeah, 8v8 causes the small success rate. I am commenting on play style, not the attack results themselves. Results are ok, randomness is not broken.


Of course there's logic behind the 8v8 style. Even if only considering attacking, it's a way of maximizing your attack potential while minimizing the risk - ironically by making an attack that most likely fails. ^^


To explain that intuitive common sense with math:

Example 1: For example with six fully stacked lands your 8v8 attack loss is only 17% of the full stack (lost 8 of 48).

Example 2: If your same six lands are stacked on average to threes (18 dice total) and you lose a 6v5 attack, you lose 33% of the total stack (6 of 18). The loss of total stack is 94% higher even while you actually lost 2 dice less. O_o Loss per single dice is up from 2% to 5.5%, or in other words you suffer a 158% increase in stack loss per dice. X_x Not optimal.


That will do for examples showing the logic behind intuitive common sense. What now follows is the hippo example. It shows the mind freed from the suffering caused by past experiences of fails of the type of example 2 - those caused by the early game unexpected plus attack fails that made you lose a great proportion of your total stack. It shows the mind freed from the suffering caused by the false vision of example 1 - the one with false sense of security and minimized loss.

Thus witness the mystical revelation! O holy o holy! Lubricate yourselves!

Example 3: Now, consider the same six lands stacked on average to sixes (36 dice) and you lose a 6v5 attack. Loss relative to the total stack is 17%. OMG it's just like it was in the 8v8 example! BUT your chance for success with 6v5 is 50% higher. ^^

Thus the sermon of the great wise sexy hippo: as long as the number of your attacking dice makes the same portion of the total current stack as an 8v8 with full stacks would make later, it's favorable to make a +1 (or better) attack. To wait for the eight is a mistake because it makes your attack more likely to fail (by, GOSH OMG OMG, 50% more likely!) AND you lose more when it fails.


Of course there's much more to consider in a real game, but all else being equal, you should do like the wise hippo does and pity the fools that wait for 8v8.

The most negative trade off you get with the example 3 style attack, even when all other things are equal, is that you lose a bit in the stacking up race and there's a chance that the enemy counters your non-8 attack with a non-8 attack of his - though he is most likely a fool and a coward who chooses instead to try to stack up to 8 first (which is favorable to you).