Board Thread:Questions and Answers/@comment-25345226-20140825000544/@comment-24577221-20140828044706

KickMeElmo wrote: Occurs to me, I may end up abusing save backup for something in the future, when I'm less focused on just playing the game. Being able to redo a breeding attempt once every minute or so with no drawbacks means I could use it to get us some statistics on breeding rates and torch effectiveness. It's not something I'll do any time soon, but I have a feeling it'd be useful data to have, especially for people considering attempting to eventually permanently light ten torches per island (like me). It would be very nice to have good stats on the effectiveness of the breeding torches. We know for sure that they do have a significant effect, so that won't be as speculative an effort as for most of the other variables, but it's something that would be expensive for most people to play with. Just be warned that it'll take a lot of tries to get decent statistics. It looks like the "default" probability of success for breeding a Ghazt is close to 1.0% (Entbrat + T-Rox, both Level 4 and 0% happy, sales tax not included). If you use 10 torches and each torch adds 2%, and want the relative uncertainty down to 10%, you'd need about 375 trials. If you use 10 torches and each adds only 1% (and that's my intuitive guess, based on my own breeding attempts), and want the relative uncertainty below 10%, you'd need about 810 tries.

If you'll settle for less precision, things get somewhat easier. If you're okay with 20% relative uncertainty, decrease the above by a factor of 4. The relationship is an N2 thing.

[N = (1 - Pm) / (  Pm * (&sigma;P m /Pm)2  )  ]