Examples¶

Short recipes for things people use PyProbs for. Each one assumes this import:

>>> from pyprobs import Probability as pr

Loot drops in a game¶

Check the rarest reward first. Because the checks run in order, the ring below has a 30% chance only when the sword did not drop:

>>> def open_chest():
...     if pr.prob("5%"):
...         return "legendary sword"
...     if pr.prob("30%"):
...         return "silver ring"
...     return "gold coins"
>>> for _ in range(5):
...     print(open_chest())
gold coins
silver ring
gold coins
gold coins
gold coins

Checking a probability by simulation¶

Draw many outcomes and look at the share that came out True. sum() counts the True values, so the share gets closer to the probability as the number of draws grows:

>>> outcomes = pr.prob("1/3", num=100_000)
>>> sum(outcomes) / len(outcomes)
0.33453

Sampling a fraction of items¶

Keep roughly one item in ten, for example to log the details of a sample of requests:

>>> sampled = []
>>> for request_id in range(1, 31):
...     if pr.prob("10%"):
...         sampled.append(request_id)
>>> sampled
[19, 20]

Free throws with a fixed success rate¶

An instance with a constant remembers the rate, and its history keeps the score. Ten throws are few, so the count can stray quite far from 72%:

>>> player = pr()
>>> player.set_constant("72%")
>>> player.iprob(num=10)
[False, True, True, False, True, True, False, True, False, False]
>>> player.count_values()
{True: 5, False: 5}
>>> player.iprob(num=10)
[True, True, False, False, False, True, True, True, False, True]
>>> player.count_values("all")
{True: 11, False: 9}

Stable outcomes in tests¶

Seed the random module before each test, and every run draws the same outcomes. With pytest, an autouse fixture does it for all tests in a module:

import random

import pytest


@pytest.fixture(autouse=True)
def seed_random():
    random.seed(1234)