Genshin Impact Wishing Probability Calculator

Notes

• This calculator accounts for "single pulls", and not "multis (ten pulls)". With the mechanics, there is effectively no difference to the calculator in chance between going for, for example, C0 Mualani + C0 Kinich + C0 Xilonen (3 five stars) versus C2 Xilonen (3 five stars). In other words, while the target character copies selector is displayed as "go from constellation number to constellation number", it actually functions as "go from number of character copies to number of character copies", which means it would be solving for both scenarios of (a) going for multiple copies of just one character, and (b) any combination of multiple copies of different characters adding up to that number. However, for the latter scenario, there may be a slight difference in the actual chance as a real player is likely to do multis and not single pulls, where the calculator would be predicting for some ideal combination, while ten pulls could throw a few pulls off course.
• The weapon banner does not have the same "luxury" as the character banner with being able to also solve for getting from different banners. A significant part, specifically the epitomized path, of the weapon's pity mechanic can reset too variably.
• The calculator considers radiance to work as such: radiance pity goes up by one when 50/50 is lost, and it resets to zero when it's won. When a 50/50 is lost at 0/1/2/3 radiance pity, 0%/5%/50%/100% chance for the loss to turn into a win (or a total of 50%/52.5%/75%/100% chance to win the 50/50). Reference.
• At the time of writing this, there appears to be a newer model the calculator doesn't operate under. While its implementation is possible and may be, not sure, a rather simple modification to the calculator, that model is yet to completely provide the chances of winning from each count for radiance. For now, it's not such a high priority as it could just be a less than 0.1% difference in result for the calculator, because of how statistics average out over higher iterations. Its most important difference however is what the user should input as their pity for radiance. If you do not understand that model, but believe it more, and don't have a way to so easily find your pity for it, it is recommended to just input zero as your radiance pity, and just treat the radiance happening as a lucky happenstance, or input three if you lost three 50/50s in a row.
• For the current version, this calculator does not automatically account for "return wishes", wherein "starglitter" can be acquired from pulling, and be exchanged for more pulls to be done for the target(s).
• Per technical limitations, the calculator loses accuracy as a chance becomes less and less likely. Specifically, its result will be rounded to 12 decimal places (in decimal, non-percent form). This means to say that something with 1 in a trillion or less likely odds will probably just be rounded to zero.
• The calculator has some set bounds for setting inputs to the necessary minimum or maximum.

Github Repository

This calculator's files are posted on Github. You may see the files (and code) in the github repository. If issues are found, it should be possible to post in the issues section.