Calculate the first passage time of a sequence consisting of prices

I want to calculate the hitting time of a sequence of prices. I define a log price corridor by for forming a double barrier, up_threshold and down_threshold.
When one of the barriers is touched, I note the time and reset the barriers around the last price

My script looks as follows:

import pandas as pd

def speed(prices: pd.DataFrame, threshold: float=0.05):
   times = []
   up_threshold, down_threshold = threshold, -threshold
   returns = (prices['Price'] / prices['Price'].shift(1)).apply(np.log)
   cum_returns = returns.cumsum(axis=0)
   for cum_return in cum_returns:
      if -threshold < cum_return < threshold:
         # this possibility will probably occur the most frequently
         continue
      elif cum_return > threshold:
         times.append(cum_return)
         up_threshold += threshold
         down_threshold += threshold
      else:
         times.append(cum_return)
         up_threshold -= threshold
         down_threshold -= threshold


if __name__ == '__main__':
    prices = np.random.random_integers(10, 25, 100)
    prices = pd.DataFrame(prices, columns=['Price'])
    speed(prices=prices, threshold=0.05)

How can I improve this script? Is it wise to also use an apply function for the cum_returns-part? The speed of the script is important. Thanks in advance

  if -threshold < cum_return < threshold:
     # this possibility will probably occur the most frequently
     continue
  elif cum_return > threshold:
     times.append(cum_return)
     up_threshold += threshold
     down_threshold += threshold
  else:
     times.append(cum_return)
     up_threshold -= threshold
     down_threshold -= threshold

  if -threshold < cum_return < threshold:
     # this possibility will probably occur the most frequently
     continue

  times.append(cum_return)

  if cum_return > threshold:
     up_threshold += threshold
     down_threshold += threshold
  else:
     up_threshold -= threshold
     down_threshold -= threshold