I’ve run into something I don’t understand. This code:

import random


class MyRandom(random.Random):
    pass


r = random.Random()
r.seed(10)
print(r.randint(0, 20))
print(r.randint(0, 20))
print(r.randint(0, 20))

print("------------")

r2 = MyRandom()
r2.seed(10)
print(r2.randint(0, 20))
print(r2.randint(0, 20))
print(r2.randint(0, 20))

…prints:

18
1
13
------------
18
1
13

This is as expected, since it’s the same seed. But if I innocuously override the random() method like this:

import random



class MyRandom(random.Random):

    def random(self):
        return super().random()



r = random.Random()
r.seed(10)
print(r.randint(0, 20))
print(r.randint(0, 20))
print(r.randint(0, 20))

print("------------")
r2 = MyRandom()
r2.seed(10)
print(r2.randint(0, 20))
print(r2.randint(0, 20))
print(r2.randint(0, 20))

Suddenly it prints:

18
1
13
------------
20
9
6

Why on earth would that change anything? How is the call to super somehow upsetting the internal state?

Even weirder, there isn’t a discrepancy if I’m calling r.random() instead of r.randint(). Perhaps that clue will be helpful.