I guess someone put this to good use
http://forums.ddo.com/showthread.php?t=389632[QUOTE=xxx13;4655148]== The Obvious ==
When you attempt to strike your opponent, an attack roll is made. A natural 1 on an attack roll is always a miss.
When you roll a fair 20-sided die (D20), each value is equally likely. Probability of rolling a 1 is 5%.
== The Unknown ==
When you attack with a falchion, the game uses a different die than when you attack with a dagger.
The "falchion die" is heavily loaded.
Probability of rolling a 1 is > 8%.
(Most likely this is a more general case: one handed vs two handed weapons.)
== The proof ==
Let's attack something many times. Many many times. Like: 10_000 times.
First with a dagger:
Then with a falchion:
== A bit of statistics ==
The chart below might be helpful when deciding how many rolls we need when testing dice.
After each 100 rolls I computed a test of goodness of fit (using Pearson's chi-squared test).
On this chart we have 3 hypotheses:
* yellow - "dagger D20" is fair
* blue - "falchion D20" is fair
* red - "falchion D20" is significantly unfair: 1 rolls 8% times, any other number 4.84%
Values under 0.05 are statistically significant "proof" that given hypothesis is not true.
As you can see the only hypothesis we can reject is the blue one. For example:
* after 5000 rolls p = 0.000000002
* after 6000 rolls p = 0.0000000000047 [/quote]
Pages: 
You people need to go read this (Read 19439 times)
