The line "X Lanterns buys Y Bronze Chests (not counting Spring Lantern prize)" is the initial purchase, and is included to show the difference with the next line "Average Chests actually opened" which does take into account lantern reward "chains". It would defeat the entire purpose of the simulator to ignore it (this is also obvious from the source code).
What are the effects of running large numbers of simulations and gaining averages?
Does this increase the likelihood in the chance of attaining a reward (like the Daily Special)?
As a result, would finding the mean like the simulator's results are, be a true representation of central tendency?
I would predict no, because by taking into account all results gained through Spring Lantern chains you would also be naturally including the outliers.
Outliers can ONLY skew the result further and further upwards because results in the negative are not possible. For this reason, the average would be skewed much higher than the real / actual result.
If this is the case what you would find, is that with increasing simulation number, overall results of the simulation would produce results that oscillate either side of a rising logarithmic curve.
My prediction would be, the mode of these results would be a better measure of central tendency instead of average.
However, what I say here would be useless if you have taken into account outliers / results that would say be above 6 daily specials being returned.
Are you able to do a simulation where the mode is produced?????