Everything below is what the code does. The constants marked "unsourced" are this tool's own assumptions.
pace0 = last-30-day followers / 30 (if entered; bursts inside that window are removed)
= (followers - past bursts) / days since page open (otherwise)
tau = half-life / ln 2
pace(t) = pace0 x (floor + (1 - floor) x e^(-t / tau))
organic(T) = pace0 x (floor x T + (1 - floor) x tau x (1 - e^(-T / tau)))
burst(T) = sum of capped boost x min(1, (T - burst day) / 5) (5-day linear ramp)
followers(T) = followers today + pace scale x organic(T) + burst scale x burst(T)
wishlists = followers(T) x multiplier
low = pace x low %, bursts x low %, low multiplier
ratio = actual wishlists / followers today (if entered; floor and ceiling = ratio -/+ spread %)
base = pace x 100%, bursts x 100%, base multiplier
high = pace x high %, bursts x 100%, high multiplier
gap = base wishlists - target
followers needed = target / multiplier; extra per week = (needed - projected base) / (days to launch / 7)
Low and high stack every pessimistic or optimistic assumption together. They are a spread of scenarios, not a probability interval.