AI summary by TwinMind

Pokémon and Geometric Distributions - Numberphile

Numberphile

23m 21s

Introduction and Problem Setup

  • Analyzing mathematics behind catching Pokemon in video games
  • Goal: determine expected number of encounters needed to catch all Pokemon
  • Featured Pokemon examples: Eevee, Fampir, Pancham
  • Core motto: "Gotta catch them all"
  • Original Pokemon Red game: 150 different Pokemon to catch
  • Current total across all generations: 1,025 Pokemon

Mathematical Assumptions and Simplifications

  • All Pokemon equally likely to appear (unrealistic but necessary for modeling)
  • Once Pokemon appears, automatically caught (no failures or escapes)
  • Unlimited supply of each Pokemon species in wild
  • Start simple, can refine model later
  • Common Pokemon like Pidgey and Rattata normally appear frequently

Three Pokemon Example Analysis

  • Starting with n=3 Pokemon for demonstration
  • Probability of any specific Pokemon appearing: 1/3
  • First encounter: guaranteed new Pokemon (T1 = 1)
  • Second new Pokemon (T2): probability 2/3 of getting new one
  • Third new Pokemon (T3): probability 1/3 of getting new one

Geometric Distribution Framework

  • T_n = number of encounters until seeing new Pokemon having already seen n-1
  • Follows geometric distribution pattern
  • Probability of taking exactly K encounters: (1-p)^(K-1) × p
  • For T2: p = 2/3, so probability = (1/3)^(K-1) × (2/3)
  • For T3: p = 1/3, so probability = (2/3)^(K-1) × (1/3)

Expected Value Calculation

  • Expected value of geometric distribution = 1/p
  • Derived through infinite series summation
  • Uses differentiation of geometric series formula
  • Key formula: E[X] = Σ(k × P(X=k)) for all k
  • Geometric series: Σ(a × r^k) = a/(1-r) when |r| < 1

Three Pokemon Results

  • Total expected encounters = 1 + E[T2] + E[T3]
  • E[T2] = 1/(2/3) = 3/2 = 1.5
  • E[T3] = 1/(1/3) = 3
  • Total = 1 + 1.5 + 3 = 5.5 encounters expected
  • Live demonstration resulted in 4 encounters (beat the odds)

General Formula for n Pokemon

  • First encounter always yields new Pokemon
  • Subsequent encounters follow geometric distributions
  • T2: probability (n-1)/n, expectation n/(n-1)
  • T3: probability (n-2)/n, expectation n/(n-2)
  • Continue until Tn: probability 1/n, expectation n/1 = n

Harmonic Number Connection

  • Total expected encounters = n × (1 + 1/2 + 1/3 + ... + 1/n)
  • Sum in parentheses = nth harmonic number (H_n)
  • H_n grows approximately like natural log(n)
  • H_n > ln(n) (proven through integral comparison)
  • Expected encounters ≈ n × ln(n)

Practical Results for Pokemon Games

  • n=3: approximately 5.5 encounters (matches calculation)
  • n=150 (original): approximately 752 encounters
  • n=1,025 (current total): approximately 7,716 encounters
  • 150 Pokemon requires almost 5× more encounters than Pokemon count
  • 1,025 Pokemon requires minimum 7,000+ encounters realistically
  • Each encounter takes approximately 1 minute of gameplay

Real-World Gaming Implications

  • Numbers represent lower bounds (equal probability assumption)
  • Actual gameplay involves rare vs common Pokemon frequency differences
  • Catching mechanics can fail in real games
  • Presenter completed original 150 Pokedex multiple times
  • Modern games too extensive for casual completion
  • Demonstrates significant time investment required for completion

TwinMind

Get summaries like this for your own meetings

TwinMind takes notes in the background during meetings, lectures, and calls, then writes the summary for you.

Free plan to get started

No bot joins your calls

Phone, desktop & Chrome apps

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome