PY-L2-18 · Random Module — Dice, Cards & Coins

Learning Goals

3 min

By the end of this lesson you can:

Match each random tool to the question it answers: dice, coin, single pick, weighted pick, shuffle, multi-pick, random fraction.
Use random.seed(n) to make "random" runs repeatable while you're debugging.
Avoid the duplicate trap by choosing random.sample over random.choice in a loop.

Warm-Up

5 min

What does each of these print? Try to answer without running the code first.

import random

print(random.randint(1, 6))           # ?
print(random.choice(["heads", "tails"]))   # ?
print(random.sample([1, 2, 3, 4, 5], 3))   # ?

cards = ["A", "K", "Q", "J"]
random.shuffle(cards)
print(cards)                           # ?

What you should see (yours will vary)

4
heads
[2, 5, 3]
['Q', 'A', 'J', 'K']

Each call answered a different question: roll a die, flip a coin, pick three different cards, shuffle a deck. Same module, four jobs.

Today's big idea

Don't default to random.randint. Each random tool has a question it's perfect for — and a few it's wrong for. Picking the right one usually halves the code.

New Concept · Seven Random Tools

14 min

1 · `randint(a, b)` — pick a whole number

Inclusive at both ends. random.randint(1, 6) can give you a 1 or a 6 — unlike range which stops before its end.

import random
die  = random.randint(1, 6)
coin = random.randint(0, 1)        # 0 or 1

2 · `choice(seq)` — pick one item

Pass any list, tuple or string. Returns one item, uniformly at random.

drinks = ["cocoa", "coffee", "hot chocolate", "horlicks"]
print(random.choice(drinks))

# A coin flip from a list of two:
print(random.choice(["heads", "tails"]))

3 · `choices(seq, k=n)` — pick many, with replacement

Notice the plural choices. You get a list of n picks; duplicates are allowed.

# Roll three dice
print(random.choices([1,2,3,4,5,6], k=3))     # e.g. [3, 5, 3]

# With weighted picks (loaded die!)
weights = [1, 1, 1, 1, 1, 5]    # 6 is five times more likely
print(random.choices([1,2,3,4,5,6], weights=weights, k=10))

4 · `sample(seq, k=n)` — pick many, no duplicates

Same shape as choices, but never picks the same item twice. Perfect for dealing cards or drawing raffle tickets.

deck = ["A", "K", "Q", "J", "10", "9", "8", "7"]
hand = random.sample(deck, 3)
print(hand)        # e.g. ['Q', '7', 'A']  -- all different

5 · `shuffle(lst)` — scramble in place

Rearranges the list itself. Returns None — never write lst = random.shuffle(lst), that wipes it out.

deck = ["A", "K", "Q", "J"]
random.shuffle(deck)
print(deck)        # e.g. ['J', 'Q', 'A', 'K']

6 · `random()` — a fraction between 0 and 1

Returns a float in [0.0, 1.0). Great for percentage chances.

if random.random() < 0.3:
    print("Rare drop! 30% chance.")
else:
    print("Common.")

7 · `seed(n)` — make randomness repeatable

Random numbers are actually deterministic — every run starts from a hidden "seed". random.seed(42) fixes the seed so the same calls always give the same answers. Brilliant for debugging.

random.seed(42)
print(random.randint(1, 100))     # always 82 on Python 3.x
print(random.randint(1, 100))     # always 15
print(random.randint(1, 100))     # always 4

random.seed(42)                    # reset!
print(random.randint(1, 100))     # 82 again

Use seed while you're hunting a bug — once it's fixed, comment the line out so the program is "random" again.

The decision tree

Question                                     Tool

Roll a die / pick a whole number in range    randint(a, b)
Pick one item from a list                    choice(seq)
Pick n items, duplicates OK (with weights?)  choices(seq, k=n, weights=...)
Pick n different items                       sample(seq, k=n)
Scramble a list in place                     shuffle(lst)
A fraction between 0 and 1 (for chance)      random()
Make a run reproducible while debugging      seed(n)

The duplicate trap

The classic bug: dealing five cards with a loop of random.choice(deck). You can get the same card twice. Use random.sample(deck, 5) instead. The plural-vs-singular and the choices-vs-sample distinctions are exactly what trips most beginners up.

Worked Example · The Casino Demo

12 min

Run all seven tools in one file. Save as casino.py:

Code

# casino.py — every random tool in one demo

import random

# Use a seed for the demo so the output is the same every run
random.seed(42)

# 1 — dice
print("Dice roll  :", random.randint(1, 6))

# 2 — coin
print("Coin flip  :", random.choice(["heads", "tails"]))

# 3 — three dice rolls, duplicates allowed
print("3 dice     :", random.choices([1,2,3,4,5,6], k=3))

# 4 — loaded die: 6 is five times more likely
print("Loaded 1d6 :", random.choices([1,2,3,4,5,6],
                                     weights=[1,1,1,1,1,5], k=10))

# 5 — deal three different cards
deck = ["A", "K", "Q", "J", "10", "9", "8", "7"]
print("Hand       :", random.sample(deck, 3))

# 6 — shuffle the whole deck
random.shuffle(deck)
print("Shuffled   :", deck)

# 7 — random chance
if random.random() < 0.3:
    print("Bonus drop!")
else:
    print("No drop.")

Output (same every run because of the seed)

Dice roll  : 1
Coin flip  : heads
3 dice     : [4, 1, 1]
Loaded 1d6 : [6, 5, 6, 6, 5, 6, 6, 6, 6, 4]
Hand       : ['7', 'K', '10']
Shuffled   : ['8', 'A', '7', 'Q', 'K', '9', 'J', '10']
Bonus drop!

Read the diff

Notice the loaded-die output — five of the ten rolls landed on a 6, because the weights tilted toward it. That's how loot tables and rare-drop chances work in real games. Notice also that the seed fixes the entire run — comment out random.seed(42) and every line will change next time.

Basic

5 min

01 🟢 Roll five dice

Roll five dice in one call and print the list. Then print the total and the maximum.

Hint

import random
rolls = random.choices([1,2,3,4,5,6], k=5)
print("Rolls :", rolls)
print("Total :", sum(rolls))
print("Best  :", max(rolls))

choices with k=5 handles all five rolls in one line.

Challenge 1

4 min

02 🟡 Deal a poker-ish hand

Build a 52-card deck (4 suits × 13 ranks) and deal a 5-card hand using random.sample. Confirm there are no duplicates.

Hint

import random

suits = ["♠", "♥", "♦", "♣"]
ranks = ["A", "2", "3", "4", "5", "6", "7", "8", "9", "10", "J", "Q", "K"]

deck = [r + s for s in suits for r in ranks]
print("Deck size :", len(deck))    # → 52

hand = random.sample(deck, 5)
print("Hand      :", hand)
print("Unique?   :", len(set(hand)) == len(hand))

The double list comprehension [r + s for s in suits for r in ranks] builds all 52 cards in one line. Get used to it; it'll feel weird at first.

Challenge 2

4 min

03 🔴 Loot drop simulator (stretch)

Build a loot table where common items appear often and rares appear rarely. Open 50 chests and print how many of each item appeared.

items   = ["potion", "sword", "shield", "gold ring", "dragon scale"]
weights = [50,        20,      15,        10,           5]

Hint

import random

items   = ["potion", "sword", "shield", "gold ring", "dragon scale"]
weights = [50,        20,      15,        10,           5]

opens = random.choices(items, weights=weights, k=50)

counts = {}
for it in opens:
    counts[it] = counts.get(it, 0) + 1

for it in items:
    print(f"{it:<14}: {counts.get(it, 0)}")

The output won't exactly match the weights — randomness is wobbly — but over 50 opens the distribution should be roughly in proportion. weights + choices is the standard pattern for loot, gacha pulls and ad placements.

Challenge 3 · The Loaded-Coin Lab

8 min

Build coin_lab.py. Flip a coin a lot of times and confirm that the proportion of heads matches the weight you set. This is the core of all probability simulation.

Your file must:

Define a function flip(n_flips, weight_heads) that returns the proportion of heads in n_flips flips. Use random.choices(["H", "T"], weights=[weight_heads, 1 - weight_heads], k=n_flips).
Call it with (100, 0.5), (100, 0.7), (10_000, 0.7). Print each.
Observe that the proportion is closer to the target with more flips.

Stretch goal. Use random.seed(0) so the run is repeatable. Build a small table comparing n_flips = 10, 100, 1000, 10000 against target 0.7.

Show one possible solution

# coin_lab.py — confirm that simulation approaches the true weight

import random

def flip(n_flips, weight_heads):
    out = random.choices(
        ["H", "T"],
        weights=[weight_heads, 1 - weight_heads],
        k=n_flips,
    )
    heads = out.count("H")
    return heads / n_flips

random.seed(0)

print(f"100 flips, fair    : {flip(100,    0.5):.3f}")
print(f"100 flips, 70% H   : {flip(100,    0.7):.3f}")
print(f"10000 flips, 70% H : {flip(10_000, 0.7):.4f}")

# Stretch — convergence table
print()
print(f"{'n':>7}{'observed':>12}{'target':>10}")
for n in (10, 100, 1000, 10_000):
    print(f"{n:>7}{flip(n, 0.7):>12.4f}{0.7:>10.4f}")

Non-negotiables: a function that flips with a weight, the out.count("H") / n_flips trick to measure heads, and a comparison that shows higher n_flips = closer-to-target. The Law of Large Numbers in five lines.

Recap

3 min

Seven random tools cover almost every randomness question. randint(a, b) for whole numbers in a range (inclusive both ends!). choice for one item, choices for many with duplicates and optional weights, sample for many without duplicates. shuffle reorders in place. random() gives a fraction for percentage chances. seed(n) locks the run so you can debug. Match the tool to the question and the code halves.

Vocabulary Card

random.randint(a, b): A whole number between a and b, both inclusive.
random.choice(seq): One item, uniformly at random.
random.choices(seq, weights=..., k=n): n picks with duplicates; weights tilts the odds.
random.sample(seq, k=n): n different picks — never the same item twice.
random.shuffle(lst): Rearrange the list in place. Returns None.
random.seed(n): Make the random sequence reproducible. Comment out for real randomness.

Extra Mission

4 min

Build quiz_picker.py. You have ten quiz questions. Pick three of them without repeats for tonight's class, then shuffle their order.

questions = [
    "Capital of the world?",
    "Capital of Indonesia?",
    "Square root of 81?",
    "Largest planet?",
    "Smallest prime?",
    "Name a noble gas.",
    "Year of independence?",
    "What does HTML stand for?",
    "Highest mountain where you live?",
    "Programming language we're using?",
]

Your file must:

Use random.sample to draw 3 questions.
Use random.shuffle on the drawn list — yes, even after the sample.
Print them numbered 1. ... 2. ... 3. ... using f-strings and the padding from PY-L2-16.

Stretch. At the top of the file, random.seed(int(input("Seed (press enter for random): ") or 0)) — if the user enters a number, the same questions will appear next time. Why might a teacher want that?

Sample · quiz_picker.py

# quiz_picker.py — pick + shuffle a small quiz

import random

questions = [
    "Capital of the world?",
    "Capital of Indonesia?",
    "Square root of 81?",
    "Largest planet?",
    "Smallest prime?",
    "Name a noble gas.",
    "Year of independence?",
    "What does HTML stand for?",
    "Highest mountain where you live?",
    "Programming language we're using?",
]

# Stretch — optional seed
raw = input("Seed (enter to randomise): ")
if raw.strip():
    random.seed(int(raw))

picked = random.sample(questions, 3)
random.shuffle(picked)

print()
for i, q in enumerate(picked, start=1):
    print(f"{i:02}. {q}")

Non-negotiables: random.sample(..., 3), random.shuffle, and an enumerate(..., start=1) printed loop. Why might a teacher want a seed? So that two classes get the same quiz on Monday and Tuesday — fairness.

import random print(random.randint(1, 6)) # ? print(random.choice(["heads", "tails"])) # ? print(random.sample([1, 2, 3, 4, 5], 3)) # ? cards = ["A", "K", "Q", "J"] random.shuffle(cards) print(cards) # ?

# Roll three dice print(random.choices([1,2,3,4,5,6], k=3)) # e.g. [3, 5, 3] # With weighted picks (loaded die!) weights = [1, 1, 1, 1, 1, 5] # 6 is five times more likely print(random.choices([1,2,3,4,5,6], weights=weights, k=10))

random.seed(42) print(random.randint(1, 100)) # always 82 on Python 3.x print(random.randint(1, 100)) # always 15 print(random.randint(1, 100)) # always 4 random.seed(42) # reset! print(random.randint(1, 100)) # 82 again

Question Tool Roll a die / pick a whole number in range randint(a, b) Pick one item from a list choice(seq) Pick n items, duplicates OK (with weights?) choices(seq, k=n, weights=...) Pick n different items sample(seq, k=n) Scramble a list in place shuffle(lst) A fraction between 0 and 1 (for chance) random() Make a run reproducible while debugging seed(n)

# casino.py — every random tool in one demo import random # Use a seed for the demo so the output is the same every run random.seed(42) # 1 — dice print("Dice roll :", random.randint(1, 6)) # 2 — coin print("Coin flip :", random.choice(["heads", "tails"])) # 3 — three dice rolls, duplicates allowed print("3 dice :", random.choices([1,2,3,4,5,6], k=3)) # 4 — loaded die: 6 is five times more likely print("Loaded 1d6 :", random.choices([1,2,3,4,5,6], weights=[1,1,1,1,1,5], k=10)) # 5 — deal three different cards deck = ["A", "K", "Q", "J", "10", "9", "8", "7"] print("Hand :", random.sample(deck, 3)) # 6 — shuffle the whole deck random.shuffle(deck) print("Shuffled :", deck) # 7 — random chance if random.random() < 0.3: print("Bonus drop!") else: print("No drop.")

import random suits = ["♠", "♥", "♦", "♣"] ranks = ["A", "2", "3", "4", "5", "6", "7", "8", "9", "10", "J", "Q", "K"] deck = [r + s for s in suits for r in ranks] print("Deck size :", len(deck)) # → 52 hand = random.sample(deck, 5) print("Hand :", hand) print("Unique? :", len(set(hand)) == len(hand))

import random items = ["potion", "sword", "shield", "gold ring", "dragon scale"] weights = [50, 20, 15, 10, 5] opens = random.choices(items, weights=weights, k=50) counts = {} for it in opens: counts[it] = counts.get(it, 0) + 1 for it in items: print(f"{it:<14}: {counts.get(it, 0)}")

# coin_lab.py — confirm that simulation approaches the true weight import random def flip(n_flips, weight_heads): out = random.choices( ["H", "T"], weights=[weight_heads, 1 - weight_heads], k=n_flips, ) heads = out.count("H") return heads / n_flips random.seed(0) print(f"100 flips, fair : {flip(100, 0.5):.3f}") print(f"100 flips, 70% H : {flip(100, 0.7):.3f}") print(f"10000 flips, 70% H : {flip(10_000, 0.7):.4f}") # Stretch — convergence table print() print(f"{'n':>7}{'observed':>12}{'target':>10}") for n in (10, 100, 1000, 10_000): print(f"{n:>7}{flip(n, 0.7):>12.4f}{0.7:>10.4f}")

questions = [ "Capital of the world?", "Capital of Indonesia?", "Square root of 81?", "Largest planet?", "Smallest prime?", "Name a noble gas.", "Year of independence?", "What does HTML stand for?", "Highest mountain where you live?", "Programming language we're using?", ]

Learning Goals

Warm-Up

New Concept · Seven Random Tools

1 · randint(a, b) — pick a whole number

2 · choice(seq) — pick one item

3 · choices(seq, k=n) — pick many, with replacement

4 · sample(seq, k=n) — pick many, no duplicates

5 · shuffle(lst) — scramble in place

6 · random() — a fraction between 0 and 1

7 · seed(n) — make randomness repeatable

The decision tree

Worked Example · The Casino Demo

Read the diff

Basic

Challenge 1

Challenge 2

Challenge 3 · The Loaded-Coin Lab

Recap

Vocabulary Card

Extra Mission

Sample · quiz_picker.py

Learning Goals

Warm-Up

New Concept · Seven Random Tools

1 · randint(a, b) — pick a whole number

2 · choice(seq) — pick one item

3 · choices(seq, k=n) — pick many, with replacement

4 · sample(seq, k=n) — pick many, no duplicates

5 · shuffle(lst) — scramble in place

6 · random() — a fraction between 0 and 1

7 · seed(n) — make randomness repeatable

The decision tree

Worked Example · The Casino Demo

Read the diff

Basic

Challenge 1

Challenge 2

Challenge 3 · The Loaded-Coin Lab

Recap

Vocabulary Card

Extra Mission

Sample · quiz_picker.py

1 · `randint(a, b)` — pick a whole number

2 · `choice(seq)` — pick one item

3 · `choices(seq, k=n)` — pick many, with replacement

4 · `sample(seq, k=n)` — pick many, no duplicates

5 · `shuffle(lst)` — scramble in place

6 · `random()` — a fraction between 0 and 1

7 · `seed(n)` — make randomness repeatable

1 · `randint(a, b)` — pick a whole number

2 · `choice(seq)` — pick one item

3 · `choices(seq, k=n)` — pick many, with replacement

4 · `sample(seq, k=n)` — pick many, no duplicates

5 · `shuffle(lst)` — scramble in place

6 · `random()` — a fraction between 0 and 1

7 · `seed(n)` — make randomness repeatable