The developers presumably intended for the machine's output to be random, so I started investigating the behaviour of simple pseudorandom number generators. We had no real leads to follow, so the next line of enquiry into best runes went right back to the drawing board. They confirmed that our current patterns would only last until the 11th February 2023 - could we find out what was going on and put it right by then? Analysing historical best rune data had proven that rougly every 3 years, the pattern shifted, making about 10% of the predictions incorrect from then on. However, we always knew that these patterns weren't perfect. Using these 40-day long patterns provided a very good fit (good enough, in fact, that they only started being wrong about 2 weeks ago, nearly 2 years later). Thankfully, our predictions were entirely correct. For clarity, I'll start calling the slot 2 possibilities 2a, 2b, and 2c. Here's a couple of examples.Īt the end of the last post, we were making predictions about the next day's slot 1 and 2 best runes, using 4 separate 40-day long repeating patterns - one for slot 1, and three for the three possibilities of slot 2. GID+1 refers to the value given to the rune that's directly after the best rune in the slot.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |