Did all of you use the same ninjas and talents?
Some ninjas are more likely to combo than others, some abilities increase crit rate. I'm not entirely sure if any ninja is inherently more likely to crit than others, but it can be possible on certain ninjas that depends on more crit.
So you have to ensure you are comparing apples to apples and not apples to oranges.
Also, if you were comparing apples to apples(or recollect data to make it so) what is the size of your data collection and therefore the standard deviation? What is the alpha value of your observation?
On a side note, I'm not entirely sure if there is a "level suppression" sort of effect in game, so also ensure that your apples are all the same level (both main and ninja).
Edit: While NPC targets are a good start, as they are likely to have the same stat, there might be a difference between elite(named) units in a TI and regular(generic) units. So perhaps seperate those results as well. In case of brohters, for example, sasuke, itachi and deidara needs to be tracked seperately from the rest of the "stuff"(whom I assume have the same secondary stat, should be a safe enough assumption)
This post was last edited by PraiseLuka at 2017-5-8 12:55