AI summary by TwinMind
Live 2020-04-20!!! Expected Values
StatQuest with Josh Starmer
33m 0s
Introduction and Setup
- Josh Starmer hosting StatQuest live stream
- Global audience: Egypt, India, Poland, Colombia, Brazil, Uruguay
- Technical issues: power outage, transformer explosion nearby
- Topic: probability and expected values as fundamental building blocks for statistics
Basic Probability Concepts Using StatLand Example
- StatLand Population Data:
- Total population: 213 people
- 37 people heard of movie "Troll 2" (1990)
- 176 people never heard of "Troll 2"
- Percentage Calculations:
- 17% heard of Troll 2 (37/213 × 100)
- 83% never heard of Troll 2 (176/213 × 100)
- 100% total coverage (17% + 83%)
- Probability Values:
- P(heard of Troll 2) = 0.17
- P(never heard of Troll 2) = 0.83
- P(either category) = 1.0
First Betting Scenario
- Bet Terms: Friend bets $1 that next person has heard of Troll 2
- Outcomes:
- If person heard of Troll 2: lose $1 (outcome = -1)
- If person never heard of Troll 2: win $1 (outcome = +1)
- Probability Analysis:
- P(lose $1) = 0.17
- P(win $1) = 0.83
- 100 Bets Scenario:
- Expected losses: 0.17 × 100 = 17 times = $17 lost
- Expected wins: 0.83 × 100 = 83 times = $83 won
- Net expected gain: $83 - $17 = $66
- Average per bet: $66/100 = $0.66
Expected Value Formula Introduction
- Definition: E(X) = 0.66 (expected value notation)
- Mathematical Formula: E(X) = Σ(x × P(x))
- Sum of each outcome × probability of that outcome
- Calculation:
- E(X) = (-1)(0.17) + (1)(0.83) = 0.66
- Interpretation: Average expected gain per bet = 66 cents
Second Betting Scenario
- Modified Bet Terms:
- Win $10 if person heard of Troll 2
- Pay $1 if person never heard of Troll 2
- Expected Value Calculation:
- E(X) = (10)(0.17) + (-1)(0.83)
- E(X) = 1.7 - 0.83 = 0.87
- Result: Expected gain of 87 cents per bet
Six-Sided Die Example
- Standard Die Properties:
- Each side probability = 1/6
- Outcomes: 1, 2, 3, 4, 5, 6
- Expected Value Calculation:
- E(X) = (1)(1/6) + (2)(1/6) + (3)(1/6) + (4)(1/6) + (5)(1/6) + (6)(1/6)
- E(X) = (1+2+3+4+5+6)/6 = 21/6 = 3.5
- Interpretation: Never roll 3.5, but average of many rolls approaches 3.5
Final Betting Scenario with Die
- Bet Terms:
- Win $100 if roll a 6
- Pay $10 for any other roll (1,2,3,4,5)
- Probability Analysis:
- P(roll 6) = 1/6
- P(roll anything else) = 5/6
- Expected Value:
- E(X) = (100)(1/6) + (-10)(5/6)
- E(X) = 16.67 - 8.33 = 8.33
- Result: Expected gain of $8.33 per bet
Channel Updates and Future Content
- Membership Progress: Approaching 200 channel members
- Upcoming Content:
- Jupyter notebook tutorials for machine learning in Python
- Zoom classes for XG boost implementation
- PDF study guides (downloadable)
- Next live stream: First Monday in May at noon
- Future Topics:
- Conditional probability
- Joint and marginal distributions
- Bayesian statistics
- Variance equation explanation (why n-1 instead of n)
