This answer is very wrong. It says that the chance of finding a starter are A+B+C, which is fine, but then it completely miscalulates the values lmao.
C is the chance of one pokèmon being shiny and the other 2 not being shiny (they switched it up) which is:
1/8192 (the chance of a pokèmon being shiny) * 8191/8192 * 8191/8192 (chance of other two pokèmons being not shiny) * 100 (to make it a probability) = 0.012204051, which is their result, but this actually has to be multiplied by 3, because there's 3 ways in which this scenario could happen (only shiny chikorita, only shiny totodile, only shiny cyndaquil).
This error clearly propagates to the B case. Only the C case is correct, because there's a single way for the starters to all be shiny.
Also, this calculation is slow and can easily lead to errors (as you've just seen), so the smartest thing to do here would be to consider the complement: What's the chance for at least one pokèmon to be shiny? I don't know, but i know that:
(chance of at least one shiny pokèmon) + (chance of no shiny pokèmon) = 1, because these two events cover all the available possibilities (the sampling space).
so, (chance of at least one shiny pokèmon) = 1 - (chance of no shiny pokèmon) = 1 - (8191/8192)^3 = 1 - 0.99963383376 = 0.00036616623, which corresponds to 0.036616623%, which is approximately a 1/2731 chance