Tested the randomness of the RNG.

[narutobarrage17]

The results were a little surprising for me.

I played as the warrior, and went through 20 turns, using it's 3X reroll each turn. So 60 rerolls.

In general it was pretty random. Only 8 times did a reroll give me the same number back, and most of them were pretty spread out. One turn I rolled all 5's, and all 3 times it gave me a 5 back, so that was the biggest anomaly there. Maybe it bugged out a bit. Either way, that's 13%.

The second odd thing was number switching. That is, rerolling it once, then rerolling it again to get the same number you started with back. This happened on 7 out of 20 turns. That's 35%. So a little more than a third of the time rerolling the same dice twice got you the original dice back.

I saw a post someone made about frequent number switching, and someone told him it was just confirmation bias. Someone else suggested crunching the numbers to see what happens, and so I did. It seems the OP wasn't just making it up after all if a third of the pie has the same results. But 60 is hardly enough for a conclusive answer.

I've already 100% completed the game, and I only did this today to satisfy a curiosity I had since I viewed that post, so I don't feel like going crazy and doing 300 rerolls just to test things more thoroughly. I wonder if anyone else wants to try this experiment too and see if their percentages vastly differ from mine.

If anyone is interested here is the exact rolls per turn so they can see the data for themselves.
Turn 1 --- -- Turn 2 -- --- Turn 3 --- -- Turn 4
1. 1 to 2 -- 4. 4 to 5 -- 7. 2 to 5 -- 10. 3 to 3
2. 1 to 3 -- 5. 4 to 6 -- 8. 5 to 6 -- 11. 3 to 4
3. 2 to 2 -- 6. 5 to 5 -- 9. 6 to 5 -- 12. 4 to 6
----------------------------------------------------------------------
Turn 5 ------ Turn 6 ------ Turn 7 ----- Turn 8
13. 3 to 1 -- 16. 5 to 5 -- 19. 4 to 1 -- 22. 4 to 3
14. 1 to 4 -- 17. 5 to 5 -- 20. 1 to 4 -- 23. 3 to 3
15. 4 to 6 -- 18. 5 to 5 -- 21. 4 to 2 -- 24. 3 to 2
----------------------------------------------------------------------
Turn 9 ------ Turn 10 ------ Turn 11 ------ Turn 12
25. 2 to 6 -- 28. 1 to 6 -- 31. 2 to 3 -- 34. 2 to 4
26. 6 to 4 -- 29. 6 to 3 -- 32. 3 to 2 -- 35. 4 to 6
27. 4 to 6 -- 30. 3 to 2 -- 33. 2 to 1 -- 36. 6 to 1
----------------------------------------------------------------------
Turn 13 --- -- Turn 14 ------ Turn 15 ----- Turn 16
37. 4 to 1 -- 40. 3 to 1 -- 43. 4 to 1 -- 46. 4 to 2
38. 1 to 5 -- 41. 1 to 3 -- 44. 1 to 1 -- 47. 2 to 5
39. 5 to 6 -- 42. 3 to 1 -- 45. 1 to 3 -- 48. 5 to 1
---------------------------------------------------------------------
Turn 17 ------ Turn 18 -- -- Turn 19 ---- Turn 20
49. 3 to 6 -- 52. 2 to 1 -- 55. 3 to 2 -- 58. 1 to 4
50. 6 to 3 -- 53. 1 to 2 -- 56. 2 to 3 -- 59. 4 to 1
51. 3 to 2 -- 54. 2 to 2 -- 57. 3 to 2 -- 60. 1 to 1

[terry]

I get that it's tempting to dig into this, but there's nothing to be found: the game is using a very standard random number generator. You can even read the source code for it here!

https://twitter.com/terrycavanagh/status/1183808073342898177

[terry]

And to add: in general, when somebody comes along and thinks they've seen some kind of a pattern in the dice rolls (oh I got five 2s in a row, that's so unlikely), I usually just suggest that they continue to track rolls and see what happens. I guarantee any patterns you think you've spotted will vanish.