d20 statistics in FGU, need your opinion on this!

[Meric Wolf]

Hello everyone!

I’m just posting to have your opinion on this. You see, I’ve started a new campaign, my first campaign in FGU, using 4e, and although my players seem to be rolling really well, I, the DM, cannot roll above a 10 to save my life.

So I was wondering if I was dreaming this, because I cannot believe I have this amount of bad luck to last me several weeks in a row so… I made stats. I rolled 500 times to see what would come out.

Here is where I need the devs or anybody to tell me what am I looking at. Because I do not know if those stats are normal since I have nothing to compare them to. I hope I can find an answer here!

Here is the screenshot of my results:

Capture d'écran 2024-04-21 115722.png

[damned]

It looks pretty well spread to me.
If you are expecting to see 25 results in each column then your data set is too small.

MoreCore has some roll scripts that allow you to roll 10,000/100,000/1,000,000 dice (in lots of 50) and give you all the stats.
They still dont come out at 4% each but it isnt a long way off.

In your example you rolled 11 or above more than 50% of the time.

[HywelPhillips]

With each bin having an expected value of 25, and assuming the statistical error comes from a normal distribution (which they almost certainly are, because of something called the central limit theorem), the error on that is the square root of the expected value ie +/- 5.

That means we'd expect to get about 68% of the results between 20 and 30, and 95% between 15 and 35. With 20 measurements, we might expect one bin maybe to be outside that. That looks pretty close to what you got.

So those numbers look consistent with the hypothesis that the dice are fair and are distributed as you'd expect to me.

If you want to really test it, as Damned says, you need to run with with much higher statistics.

Cheers, Hywel