Another variation of simulation, in Python, and bit more defined strategy-abstraction (though written for improving own understanding anyway):

from __future__ import division
import random
def run(strt): # compute (possibly stochastic) outcome of strategy execution
docont = strt() # ... no arguments.
if not docont: # at intersection X
return 0
docont = strt()
if not docont:
return 4
return 1
def evst(strt, fun=run, n=100): # evaluate strategy by running `n` times
r = 0.0
for i in range(n):
r += fun(strt)
return r/n
""" # sample use:
> ra = [(v/30, evst(lambda: random.random() < v/30, n=30000)) for v in range(30)]
> max(ra, key=lambda v: v[1]) # maximal stochastic value and its `p`
(0.6666666666666666, 1.3515666666666666)
"""