A quick proof of Bayes’ theorem | 3Blue1Brown

1. Audio
   Observation

   The spoken explanation at this interval discusses An event in the probability argument; visually associated with the yellow region/circle..

2. Formula
   Observation

   A appears in P(A), P(A and B), P(A|B), and is color-coded yellow.

3. Diagram
   Observation

   A is represented by a yellow circle and later a yellow vertical region within the sample-space square.

A

An event in the probability argument; visually associated with the yellow region/circle.

Event in a probability space

1. Audio
   Observation

   The spoken explanation at this interval discusses An event in the probability argument; visually associated with the blue region/circle..

2. Formula
   Observation

   B appears in P(B), P(A and B), P(B|A), and is color-coded blue.

3. Diagram
   Observation

   B is represented by a blue circle and later a blue horizontal region within the sample-space square.

B

An event in the probability argument; visually associated with the blue region/circle.

Event in a probability space

1. Formula
   Observation

   P(·) is used throughout for probabilities such as P(A), P(B), P(A and B), P(A|B), and P(B|A).

P

Probability operator.

Applies to events or conditional events

1. Audio
   Observation

   The spoken explanation at this interval discusses Logical conjunction of two events, corresponding to both events occurring..

2. Formula
   Observation

   The expression P(A and B) is shown explicitly.

and

Logical conjunction of two events, corresponding to both events occurring.

Used inside P(A and B)

1. Audio
   Observation

   The spoken explanation at this interval discusses Conditional-probability separator read as 'given'..

2. Formula
   Observation

   The notation P(B|A) and P(A|B) is shown explicitly.

|

Conditional-probability separator read as 'given'.

Used inside P(A|B) and P(B|A)

1. Audio
   Observation

   The spoken explanation at this interval discusses Probability of event A..

2. Formula
   Observation

   P(A) appears in the product P(A)P(B|A) and in the denominator of the rearranged formula.

3. Diagram
   Observation

   It is marked by a brace under the yellow vertical strip in the right-hand square.

P(A)

Probability of event A.

Real number in [0,1]

1. Audio
   Observation

   The spoken explanation at this interval discusses Probability of event B..

2. Formula
   Observation

   P(B) appears in the product P(B)P(A|B) and in the denominator of the other rearranged formula.

3. Diagram
   Observation

   It is marked by a brace beside the blue horizontal strip in the left-hand square.

P(B)

Probability of event B.

Real number in [0,1]

1. Audio
   Observation

   The spoken explanation at this interval discusses Joint probability that both A and B occur..

2. Formula
   Observation

   P(A and B) is displayed as the middle term in the chain of equalities.

3. Diagram
   Observation

   It corresponds to the overlap region in the Venn diagram and to the green intersection rectangle in the square diagrams.

P(A and B)

Joint probability that both A and B occur.

Real number in [0,1]

1. Audio
   Observation

   The spoken explanation at this interval discusses Conditional probability of B given A..

2. Formula
   Observation

   P(B|A) appears on the right side of P(A)P(B|A) and later as the isolated result after division by P(A).

3. Diagram
   Observation

   It is indicated by a brace along the lower portion of the yellow strip in the right-hand square.

P(B|A)

Conditional probability of B given A.

Real number in [0,1] when P(A)>0

1. Audio
   Observation

   The spoken explanation at this interval discusses Conditional probability of A given B..

2. Formula
   Observation

   P(A|B) appears on the right side of P(B)P(A|B) and later as the isolated result after division by P(B).

3. Diagram
   Observation

   It is indicated by a brace along the left portion of the blue strip in the left-hand square.

P(A|B)

Conditional probability of A given B.

Real number in [0,1] when P(B)>0

1. Diagram
   Observation

   The text 'Space of all possibilities' appears inside the gray square.

Space of all possibilities

Label for the full sample-space square used in the area model.

Visual label rather than algebraic symbol

1. Formula
   Observation

   The value (1/21) is placed above P(B) in the final numeric example.

(1/21)

Example numerical value assigned to P(B).

Rational number

1. Formula
   Observation

   The screen shows P(A|B) = P(A)P(B|A)/P(B).

2. Audio
   Observation

   The spoken explanation at this interval discusses Bayes' theorem formula.

The clip presents Bayes' theorem as a rearrangement of the joint probability identity, solving for one conditional probability in terms of the other conditional probability and the marginal probabilities.

1. A and B are events.

2. The displayed rearrangement requires P(B)\neq 0.

3. Both conditioning events have positive probability in the symmetric displayed identity, so both ordinary conditional probabilities and the two division forms are defined.

1. Joint probability as two equivalent products

1. Audio
   Observation

   The spoken explanation at this interval discusses Joint probability as two equivalent products.

2. Formula
   Observation

   The top line displays P(B)P(A|B)=P(A and B)=P(A)P(B|A).

3. Diagram
   Observation

   Two area decompositions of the same green overlap region are shown in the square diagrams.

The central fact of the clip is that the probability that both A and B occur can be decomposed in two symmetric ways: first take A and then the part of B inside A, or first take B and then the part of A inside B.

1. A and B are events in the same probability space.

2. Both conditioning events have positive probability in the symmetric displayed identity, so both ordinary conditional probabilities and the two division forms are defined.

1. Conditional probability as a restricted proportion

1. Audio
   Observation

   The spoken explanation at this interval discusses Conditional probability as a restricted proportion.

2. Diagram
   Observation

   The right square marks P(A) as a vertical strip and P(B|A) as the lower fraction of that strip; the left square marks P(B) as a horizontal strip and P(A|B) as the left fraction of that strip.

In this clip, P(B|A) is explained as the fraction of the A-region that also lies in B, while P(A|B) is explained as the fraction of the B-region that also lies in A. The visual model treats conditioning as restricting attention to one region and then measuring a subregion inside it.

1. For P(B|A), the narration implicitly restricts to cases where A is true.

2. For P(A|B), the narration implicitly restricts to cases where B is true.

3. Both conditioning events have positive probability in the symmetric displayed identity, so both ordinary conditional probabilities and the two division forms are defined.

1. Audio
   Observation

   The spoken explanation at this interval discusses Symmetry of the joint event 'A and B'.

2. Formula
   Observation

   The equality chain places P(A and B) in the middle between the two product expressions.

The proof method is to observe that the event 'A and B' is the same as 'B and A', so any valid decomposition of its probability must agree. This symmetry forces the equality of the two product formulas and yields Bayes' theorem after division.

1. The argument uses commutativity of logical conjunction for events.

1. Joint probability as two equivalent products

1. Diagram
   Observation

   A Venn diagram of overlapping circles is transformed into two square area models labeled by P(A), P(B), P(A|B), and P(B|A).

2. Audio
   Observation

   The spoken explanation at this interval discusses Area-model derivation of Bayes' theorem.

The clip proves the theorem geometrically by representing the whole sample space as a square, event regions as strips, and intersections as smaller rectangles. The same intersection area is computed in two different orders, giving the algebraic identity.

1. The visualization assumes probabilities can be represented by relative areas.

1. Joint probability as two equivalent products
2. Conditional probability as a restricted proportion

1. Formula
   Observation

   The source displays an A/B Bayes identity; H/E are editorial hypothesis/evidence labels.

Bayes' theorem describes the probability of an event, based on prior knowledge of conditions that might be related to the event.

1. P(H)>0 and P(E)>0 for the ordinary conditionals used here.

1. Formula
   Observation

   P(A and B) ???= P(A)P(B)

The joint probability of two independent events is the product of their individual probabilities.

1. Events A and B are independent.

1. Formula
   Observation

   Corrected formula P(A and B) = P(A)P(B|A) is displayed with a green checkmark.

2. Audio
   Observation

   The spoken explanation at this interval discusses General Multiplication Rule for Joint Probability.

For ordinary conditioning on an event A of positive probability, the joint probability is P(A)P(B|A), whether the events are independent or dependent. Conditional probabilities condition on events, without requiring a temporal order or causal effect.

1. The conditioning event A has positive probability for the ordinary conditional formula.

2. Independence is not required.

1. P(A and B)
2. P(A)
3. P(B|A)

1. Audio
   Observation

   The spoken explanation at this interval discusses Definition of Statistical Independence.

2. Diagram
   Observation

   Coin flip grid (HH, HT, TH, TT) and dice roll grid are shown as examples where this condition holds.

Events are independent when their joint probability equals the product of their marginal probabilities. When A has positive probability, this is equivalent to P(B|A)=P(B). Independence is not confined to coins or dice, and fairness alone does not establish independence.

1. The conditional equality requires A to have positive probability.

2. The product-of-marginals definition also covers zero-probability events.

3. Independent fair coin and die trials are examples, not the only possible independent events.

1. P(B)
2. P(B|A)

1. Formula
   Observation

   Step-by-step visual derivation builds the equation P(π|+) = [P(π)P(+|π)] / P(+).

2. Audio
   Observation

   The spoken explanation at this interval discusses Bayes' Theorem.

A fundamental theorem in probability theory that describes the probability of an event based on prior knowledge of conditions that might be related to the event. It allows for the calculation of a reverse conditional probability, such as finding the probability of being sick given a positive test result, by using the likelihood of the test result given sickness and the prior probability of sickness.

1. Both event probabilities are positive for the ordinary conditionals displayed in this symmetric formula.

2. The illustrated application involves testing; Bayes theorem also remains valid for independent events, when the posterior equals the prior.

1. P(π|+)
2. P(π)
3. P(+|π)
4. P(+)

1. Audio
   Observation

   The spoken explanation at this interval discusses Bayes' theorem follows from equality of the two joint-probability decompositions.

2. Formula
   Observation

   The screen rewrites the equality into P(A|B)=P(A)P(B|A)/P(B) and then into P(B)P(A|B)/P(A)=P(B|A).

From P(B)P(A|B)=P(A)P(B|A), one obtains P(A|B)=P(A)P(B|A)/P(B) and equivalently P(B|A)=P(B)P(A|B)/P(A).

1. A and B are events.

2. For the displayed division forms, the relevant denominator probability is nonzero: P(B)\neq 0 for the first rearrangement and P(A)\neq 0 for the second.

3. Both A and B have positive probability for the ordinary conditionals used here.

For events A and B of positive probability in the same probability space.

1. Formula
   Observation

   The top equation explicitly states P(B)P(A|B)=P(A and B)=P(A)P(B|A).

2. Audio
   Observation

   The spoken explanation at this interval discusses Equality of the two product decompositions of P(A and B).

P(B)P(A|B)=P(A and B)=P(A)P(B|A).

1. A and B are events in the same probability space.

2. Both A and B have positive probability for the ordinary conditionals used here.

For events A and B of positive probability in the same probability space.

1. Audio
   Observation

   The spoken explanation at this interval discusses Correlation affects joint probability.

If two events are correlated, the joint probability is not simply the product of their individual probabilities.

1. Events A and B are correlated.

For all correlated events A and B.

1. Audio
   Observation

   The spoken explanation at this interval discusses Gamified Examples Exhibit Genuine Independence.

2. Diagram
   Observation

   The display shows a100-flip tail-count probability histogram concentrated near50; under independent fair flips the exact finite-count distribution is binomial, not a continuous normal distribution.

The source uses standard independent coin and dice models to caution against applying independence automatically to every application. Fairness and independence are different assumptions.

1. These examples assume independent trials; fairness alone is insufficient.

Many introductory examples.

1. Audio
   Observation

   The spoken explanation at this interval discusses Dependence makes conditional updating informative.

2. Formula
   Observation

   Derivation of Bayes' theorem using medical test symbols (π and +).

The narrator emphasizes applications where evidence and hypotheses are dependent. This is pedagogical motivation, not a restriction on Bayes theorem: the theorem is valid for independent events too, with no posterior change.

1. The displayed reverse conditional probabilities are defined.

A motivation for useful conditional updates, not a theorem requiring dependence.

1. Audio
   Observation

   The spoken explanation at this interval discusses Proof of Bayes' theorem from symmetry of 'and'.

2. Formula
   Observation

   The screen shows the chain P(B)P(A|B)=P(A and B)=P(A)P(B|A), then rearranges to P(A|B)=P(A)P(B|A)/P(B), and then to P(B)P(A|B)/P(A)=P(B|A).

3. Diagram
   Observation

   The two square decompositions visually represent the same intersection area computed in two orders.

1. The video does not explicitly state the denominator-nonzero assumptions on screen or in speech.

2. The transition from the equality chain to the divided forms is shown visually but not verbally justified step by step.

3. Editorial condition: both conditioning events have positive probability for this symmetric ordinary-conditional proof. The actual source omits the explicit condition.

1. Expression

   $$P(A\text{ and }B)=P(A)P(B|A)$$
   Explanation

   Compute the probability that both events occur by first taking the overall proportion of cases where A is true, then multiplying by the proportion of those A-cases where B is also true.

   Justification

   Narration defines P(A) as the proportion of all possibilities where A is true and P(B|A) as the proportion of those events where B is also true.

   Shown in the video
2. Expression

   $$P(A\text{ and }B)=P(B)P(A|B)$$
   Explanation

   Compute the same joint probability by first taking the overall proportion of cases where B is true, then multiplying by the proportion of those B-cases where A is also true.

   Justification

   Narration explicitly mirrors the previous decomposition with the roles of A and B exchanged.

   Shown in the video
3. Expression

   $$P(B)P(A|B)=P(A)P(B|A)$$
   Explanation

   Since both products equal the same joint probability, they are equal to each other.

   Justification

   Transitivity of equality applied to the two expressions for P(A and B).

   Shown in the video
4. Expression

   $$P(A|B)=\frac{P(A)P(B|A)}{P(B)}$$
   Explanation

   Solve the equality for P(A|B) by dividing both sides by P(B).

   Justification

   Algebraic rearrangement of the previous equality; the clip displays this form directly.

   Shown in the video
5. Expression

   $$P(B|A)=\frac{P(B)P(A|B)}{P(A)}$$
   Explanation

   Alternatively, solve the same equality for P(B|A) by dividing both sides by P(A).

   Justification

   Algebraic rearrangement of the same equality; the clip displays this second form directly.

   Shown in the video

Bayes' theorem is obtained as the rearrangement of the symmetric identity for the joint probability P(A and B).

1. Formula
   Observation

   The final screen annotates P(B) with (1/21), P(A|B) with (4/10), and P(A) with (24/210) in the equation P(B)P(A|B)/P(A)=P(B|A).

2. Audio
   Observation

   The spoken explanation at this interval discusses Numerical substitution into the rearranged Bayes identity.

1. The simplified result is not spoken or shown explicitly in the clip.

2. The source of the specific numbers is not explained within this fragment.

1. Expression

   $$\frac{(1/21)(4/10)}{24/210}=P(B|A)$$
   Explanation

   Substitute the displayed example values into the rearranged formula for P(B|A).

   Justification

   Direct replacement of symbols by the annotated rational numbers on screen.

   Shown in the video
2. Expression

   $$\frac{4/210}{24/210}=P(B|A)$$
   Explanation

   Multiply the numerator fractions.

   Justification

   Standard arithmetic on rational numbers.

   Derived from the video
3. Expression

   $$P(B|A)=\frac{4}{24}=\frac{1}{6}\approx 0.1667$$
   Explanation

   Cancel the common denominator 210 and simplify the remaining fraction.

   Justification

   Arithmetic simplification not explicitly written in the video.

   Derived from the video

With the displayed example values, the resulting conditional probability is P(B|A)=1/6≈0.1667.

1. Formula
   Observation

   P(\mathrm{Book}) P(\mathrm{Magnifier}|\mathrm{Book}) / P(\mathrm{Magnifier}) = P(\mathrm{Book}|\mathrm{Magnifier})

1. Expression

   $$P(\mathrm{Book}) P(\mathrm{Magnifier}|\mathrm{Book}) / P(\mathrm{Magnifier}) = P(\mathrm{Book}|\mathrm{Magnifier})$$
   Explanation

   The video shows a visual proof of Bayes' theorem using icons for a book and a magnifying glass.

   Justification

   Definition of conditional probability.

   Shown in the video

Bayes' theorem is proven visually.

1. Animation
   Observation

   The target, prior, likelihood and denominator are assembled into the final testing formula.

1. The strict algebraic proof linking the general multiplication rule to Bayes' theorem is skipped in the visual animation; it is presented as a direct formula construction.

1. Expression

   $$P(\pi|+)$$
   Explanation

   Start with the target conditional probability: the probability of being sick given a positive test result.

   Justification

   Problem setup defined by the visual arrows and text labels.

   Shown in the video
2. Expression

   $$P(\pi)$$
   Explanation

   Add the prior as one numerator component; this piece alone is not an equality to the target.

   Justification

   Numerator component of Bayes' theorem.

   Supplementary explanation
3. Expression

   $$\times P(+|\pi)$$
   Explanation

   Multiply by the likelihood: the probability of getting a positive test result given that the person is actually sick.

   Justification

   Numerator component representing the true positive rate.

   Shown in the video
4. Expression

   $$/ P(+)$$
   Explanation

   Divide the entire product by the marginal probability of getting a positive test result.

   Justification

   Denominator normalizes the probability over all possible ways to get a positive result.

   Shown in the video

The final constructed equation is P(\pi|+) = \frac{P(\pi)P(+|\pi)}{P(+)}, which is the specific instance of Bayes' theorem for the medical testing scenario.

1. Formula
   Observation

   The equation P(B)P(A|B)/P(A)=P(B|A) is annotated with (1/21), (4/10), and (24/210).

2. Audio
   Observation

   The spoken explanation at this interval discusses Worked numeric use of the rearranged Bayes formula.

1. The video does not explain where the numbers come from.

2. The final simplified value is not shown on screen.

Use the displayed values to compute P(B|A) from the rearranged identity.

1. P(B)=1/21

2. P(A|B)=4/10

3. P(A)=24/210

4. Formula shown: P(B)P(A|B)/P(A)=P(B|A)

Find the numerical value of P(B|A).

1. Expression

   $$P(B|A)=\frac{(1/21)(4/10)}{24/210}$$
   Explanation

   Insert the given numbers into the displayed formula.

   Justification

   Direct substitution into the rearranged Bayes identity shown on screen.

   Shown in the video
2. Expression

   $$P(B|A)=\frac{4/210}{24/210}$$
   Explanation

   Multiply the numerator fractions to obtain a single fraction.

   Justification

   Arithmetic multiplication of rational numbers.

   Derived from the video
3. Expression

   $$P(B|A)=\frac{4}{24}=\frac{1}{6}\approx 0.1667$$
   Explanation

   Cancel the common denominator and reduce the fraction.

   Justification

   Standard simplification of rational numbers.

   Derived from the video

P(B|A)=1/6≈0.1667

Substituting the answer back gives (24/210)(1/6)=4/210=(1/21)(4/10), matching the displayed equality.

1. Formula
   Observation

   P(\mathrm{Heart}\mathrm{Heart}) = 1/4 * 1/4 = 1/16

In an illustrative grid assigning marginal probability 1/4 to each event, what joint probability would an independence assumption give?

1. Teaching-model marginal P(heart disease) = 1/4.

2. Independence is a provisional assumption for this grid, later challenged by the sibling example.

Calculate P(both die of heart disease).

1. Expression

   $$P(\mathrm{Heart}\mathrm{Heart}) = 1/4 * 1/4 = 1/16$$
   Explanation

   Assuming independence, multiply the probabilities.

   Justification

   Multiplication rule for independent events.

   Shown in the video

Under this illustrative independence assumption: 1/16. No determined joint risk is supplied for dependent siblings.

The multiplication is correct within the assumed independent teaching grid. The later dependence discussion removes that assumption; the source supplies no replacement numerical joint risk.

1. Formula
   Observation

   P(TT) = 1/2 * 1/2 = 1/4

What is the probability of getting tails on two successive coin flips?

1. P(tails) = 1/2 under a fair-coin model.

2. The two trials are assumed independent.

Calculate P(two tails).

1. Expression

   $$P(TT) = 1/2 * 1/2 = 1/4$$
   Explanation

   Multiply the probabilities of each independent flip.

   Justification

   Multiplication rule for independent events.

   Shown in the video

1/4

The result holds under both fairness and independence assumptions; fairness alone does not establish independence.

1. Formula
   Observation

   P(\mathrm{One}\mathrm{One}) = 1/6 * 1/6 = 1/36

What is the probability of rolling two ones on a pair of dice?

1. P(one) = 1/6 under a fair-die model.

2. The two trials are assumed independent.

Calculate P(two ones).

1. Expression

   $$P(\mathrm{One}\mathrm{One}) = 1/6 * 1/6 = 1/36$$
   Explanation

   Multiply the probabilities of each independent die roll.

   Justification

   Multiplication rule for independent events.

   Shown in the video

1/36

The result holds under both fairness and independence assumptions; fairness alone does not establish independence.

1. Diagram
   Observation

   Text labels 'You are sick' and 'Positive test result' connected by arrows to the symbols π and +.

2. Formula
   Observation

   Bayes' theorem is explicitly written out using these specific symbols.

1. No numerical values for the probabilities (e.g., base rate of disease, test accuracy) are provided in this clip.

Determine the probability that a person is actually sick given that they have received a positive medical test result.

1. Event π: You are sick.

2. Event +: Positive test result.

3. Goal is to find P(π|+).

Express the probability of sickness given a positive test in terms of the displayed prior, likelihood and positive-result probability; numerical values are not supplied.

1. Expression

   $$P(\pi|+) = \frac{P(\pi)P(+|\pi)}{P(+)}$$
   Explanation

   Apply Bayes' theorem to express the reverse conditional probability using the prior probability, the likelihood of the test, and the marginal probability of a positive test.

   Justification

   Direct application of the derived formula shown in the video.

   Shown in the video

P(\pi|+) = \frac{P(\pi)P(+|\pi)}{P(+)}

The formula correctly maps the visual definitions: P(π) is the prior chance of sickness, P(+|π) is the test's ability to catch the sickness, and P(+) is the overall chance of a positive result.

1. Diagram
   Observation

   The title 'Bayes' theorem' appears at upper left with a small inset probability diagram labeled P(E|H), P(H), and P(E|-H).

2. Animation
   Observation

   Four pi-shaped characters occupy the lower half of the frame while the main formula appears to their right.

1. Title text 'Bayes' theorem'

2. Inset square diagram with labels P(E|H), P(H), P(E|-H)

3. Four pi-shaped characters

4. Main formula P(A|B)=P(A)P(B|A)/P(B)

1. The main Bayes formula appears on the right side of the frame.

2. The inset diagram remains visible while the formula is introduced.

1. The opening establishes that the clip concerns Bayes' theorem before the proof begins.

The opening visually frames the topic as Bayes' theorem and previews the later area-based interpretation through the inset diagram.

1. Animation
   Observation

   A yellow circle labeled A and a blue circle labeled B appear overlapping, with a white arrow pointing to the intersection.

2. Diagram
   Observation

   The formula P(B)P(A|B)=P(A and B)=P(A)P(B|A) is already present above the diagram.

1. Yellow circle A

2. Blue circle B

3. Intersection region

4. White downward arrow

5. Top formula chain

1. The two circles are drawn and overlapped.

2. The arrow highlights the intersection corresponding to 'A and B'.

1. The color coding links A to yellow and B to blue throughout the clip.

The Venn diagram identifies the target quantity P(A and B) as the overlap of the two event regions.

1. Animation
   Observation

   The Venn diagram morphs into a square sample space with a yellow vertical strip labeled P(A) and a lower portion of that strip labeled P(B|A).

2. Diagram
   Observation

   The text 'Space of all possibilities' appears inside the gray square.

1. Gray square sample space

2. Yellow vertical strip for A

3. Lower subregion within the yellow strip

4. Brace labels P(A) and P(B|A)

1. The circle picture is replaced by a rectangular area model.

2. The yellow strip is subdivided to show the fraction of A that also lies in B.

1. The total square still represents the whole probability space.

2. The green overlap area continues to represent P(A and B).

This visualization encodes P(A and B) as the area of the yellow strip multiplied by the fractional height P(B|A).

1. Animation
   Observation

   A second square appears on the left with a blue horizontal strip labeled P(B) and a left subregion labeled P(A|B).

2. Diagram
   Observation

   Both squares are shown side by side with the same top equality chain.

1. Second gray square sample space

2. Blue horizontal strip for B

3. Left subregion within the blue strip

4. Brace labels P(B) and P(A|B)

1. A mirrored decomposition is added on the left.

2. The same intersection area is now represented as a horizontal strip times a fractional width.

1. Both squares depict the same sample space and the same intersection area.

2. The top equality chain remains unchanged.

The left model computes the same joint probability by conditioning on B instead of A.

1. Animation
   Observation

   The two squares shrink to the bottom corners while the algebra is rearranged on screen.

2. Formula
   Observation

   The display changes first to P(A|B)=P(A)P(B|A)/P(B) and then to P(B)P(A|B)/P(A)=P(B|A).

1. Shrunken left and right area diagrams

2. Central algebraic expressions

1. The equality chain is collapsed into a single solved form for P(A|B).

2. The solved form is then rewritten as a solved form for P(B|A).

1. The underlying equality P(B)P(A|B)=P(A)P(B|A) is preserved through the rearrangements.

The animation emphasizes that Bayes' theorem is just an algebraic rearrangement of the symmetric joint-probability identity.

1. Formula
   Observation

   The actual display is book given magnifier = book probability times magnifier given book, divided by magnifier probability.

1. Book event icon

2. Magnifier event icon

3. Book-given-magnifier formula

1. Event names are replaced with icons while preserving the reverse-conditional identity.

1. The algebraic structure of the formula remains the same after substitution.

The identity survives event relabeling; earlier letter assignments are not imposed globally.

1. Animation
   Observation

   Four stylized pi characters are shown at the bottom of the screen, reacting to the formulas above.

1. Four pi characters

1. They look around and react to the formulas.

1. Their positions at the bottom of the screen.

They serve as visual mascots to engage the viewer.

1. Diagram
   Observation

   A Venn diagram showing two overlapping circles labeled A and B.

1. Two overlapping circles

1. The intersection is highlighted.

1. The labels A and B.

Illustrates the concept of joint probability P(A and B).

1. Diagram
   Observation

   Grids showing combinations of heart disease, coin flips, and dice rolls.

1. Grids with icons

1. Cells are highlighted to show specific outcomes.

1. The structure of the grids.

Visually demonstrates the multiplication of probabilities for independent events.

1. Animation
   Observation

   Transition from a complex 4x4 human figure grid to simple 2x2 coin and 6x6 dice grids.

2. Formula
   Observation

   Red cross appears over P(A and B) = P(A)P(B); green check appears next to P(A and B) = P(A)P(B|A).

1. Complex 4x4 grid of human figures with varying colors and icons.

2. Simple 2x2 grid of coin flips (HH, HT, TH, TT).

3. Simple 6x6 grid of dice rolls.

4. Crossed-out incorrect formula.

5. Checked correct formula.

1. The complex human grid is replaced by the simpler coin and dice grids.

2. The incorrect multiplication formula is visually negated with a red cross.

3. The correct conditional multiplication formula is validated with a green checkmark.

1. The core concept of calculating joint probability P(A and B) remains the focus throughout the transition.

The visual shift demonstrates that while the simple P(A)P(B) rule works for highly structured, independent games like coins and dice, it fails for complex, dependent real-world scenarios represented by the human grid. The correct universal rule must account for dependence via P(B|A).

1. Animation
   Observation

   Stacks of red and blue coins drop and accumulate, while a yellow vertical line tracks the current count on a bell-curve histogram.

1. The diagram illustrates an independent fair-coin model; its appearance alone does not prove independence.

1. Stacks of red and blue coins.

2. Bell-curve histogram with axes '# of tails' and probability values.

3. Moving yellow vertical tracking line.

1. Coins continuously drop and stack up.

2. The yellow line moves left and right along the x-axis, updating its position based on the cumulative number of tails.

1. Total flips remain fixed at 100.

2. The displayed finite tail-count distribution stays fixed.

Under independent fair trials, the tail count has a binomial distribution centered at expectation 50. The bell-like display is an illustration of that assumed model, not a proof of independence.

1. Animation
   Observation

   Symbols π and + appear with descriptive text, followed by the sequential building of the fraction.

1. Green pi symbol (π).

2. Blue plus symbol (+).

3. Text labels 'You are sick' and 'Positive test result'.

4. Fraction bar and probability terms.

1. Arrows link the abstract symbols to their real-world meanings.

2. The equation is built piece by piece: starting with the LHS, adding the prior, multiplying by the likelihood, and finally dividing by the marginal probability.

1. The semantic mapping of π to sickness and + to a positive test remains constant.

The gradual assembly demystifies Bayes' theorem, showing it not as an arbitrary formula, but as a logical combination of prior beliefs, test reliability, and overall base rates to update our understanding of a dependent event.

1. Audio
   Observation

   The spoken explanation at this interval discusses Mistaking the asymmetric-looking product formula for a genuinely asymmetric joint probability.

One might think P(A)P(B|A) and P(B)P(A|B) are different quantities because they are written in opposite orders.

They are equal because both compute the same event probability P(A and B); the apparent asymmetry is only in the chosen conditioning order.

1. Formula
   Observation

   The clip divides by P(B) and later by P(A) without stating the nonzero condition aloud.

1. This caution is not explicitly mentioned in the video.

The rearranged formulas can be applied without checking whether the denominator probability is zero.

To divide by P(B) or P(A), one needs P(B)\neq 0 or P(A)\neq 0 respectively; the clip displays the divisions but does not state this condition.

1. Audio
   Observation

   The spoken explanation at this interval discusses Misconception about Joint Probability.

Assuming P(A and B) = P(A)P(B) for all events.

This formula only holds if events A and B are independent. Correlated events require different calculations.

1. Formula
   Observation

   The formula P(A and B) = P(A)P(B) is prominently displayed and then aggressively crossed out with a red X.

2. Audio
   Observation

   The spoken explanation at this interval discusses Misconception: Joint Probability is Always the Product of Marginals.

Believing that the probability of two events happening together can always be found by simply multiplying their individual probabilities, i.e., P(A and B) = P(A)P(B).

This rule is only valid when events A and B are strictly independent. For dependent events, you must use the general multiplication rule: P(A and B) = P(A)P(B|A), which accounts for how the occurrence of A changes the likelihood of B.

1. Audio
   Observation

   The spoken explanation at this interval discusses Misconception: Introductory Examples Represent Real-World Complexity.

2. Diagram
   Observation

   Transition from simple coin/dice grids to the complex human grid and medical testing formula.

Assuming that because introductory probability relies heavily on independent events like coin flips and dice rolls, real-world probabilistic reasoning operates under the same simple independence rules.

Do not infer independence merely because introductory models use independent coin or die trials. Dependence must be assessed from the actual model; Bayes theorem applies to either case, and may leave probabilities unchanged under independence.

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

2. Formula
   Observation

   The screen moves from P(B)P(A|B)=P(A and B)=P(A)P(B|A) to the solved Bayes forms.

Bayes' theorem is derived directly from the equality of the two product decompositions of the joint probability.

1. Diagram
   Observation

   The square diagrams label subregions with P(A), P(B|A), P(B), and P(A|B).

2. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

The area model applies the verbal notion of conditional probability as a restricted proportion to a geometric representation.

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

2. Formula
   Observation

   The equality chain centers on P(A and B).

The symmetry of the event 'A and B' is the conceptual reason the two product formulas for the joint probability must coincide.

1. Formula
   Observation

   The numeric annotations are placed onto the rearranged formula for P(B|A).

2. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

The worked numeric substitution demonstrates how to use the rearranged Bayes formula in practice.

1. Formula
   Observation

   The actual source displays an A/B Bayes identity; H/E are editorial labels for hypothesis and evidence, preserving the same identity.

Bayes' theorem uses conditional probabilities, which are derived from joint probabilities.

1. Formula
   Observation

   Visual contrast between the crossed-out P(A)P(B) and the checked P(A)P(B|A).

2. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

The rule for independent events (P(A and B) = P(A)P(B)) is merely a special case of the general multiplication rule (P(A and B) = P(A)P(B|A)) that applies only when the conditional probability equals the marginal probability.

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

2. Diagram
   Observation

   Shift from the coin flip histogram (independent) to the medical testing formula (dependent).

Independence is a special relation that leaves the conditional probability unchanged. The source contrasts simple independent examples with dependent applications, where Bayes updating can be informative; Bayes theorem is not limited to dependent events.

1. Formula
   Observation

   Both the general multiplication rule and Bayes' theorem rely fundamentally on the concept of conditional probability P(B|A).

1. The video does not explicitly show the algebraic steps deriving Bayes' theorem from the multiplication rule, though it is standard mathematical knowledge.

Bayes' theorem is mathematically derived by applying the general multiplication rule to both P(A and B) and P(B and A), recognizing that the joint probability is commutative, and then solving for the reverse conditional probability.

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

2. Formula
   Observation

   The full derivation from the joint equality to the Bayes forms is displayed.

1. Bayes' theorem formula
2. Joint probability as two equivalent products
3. Proof of Bayes' theorem from symmetry of 'and'

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

2. Formula
   Observation

   The equality chain uses P(A and B) as the common middle term.

1. Symmetry of the joint event 'A and B'
2. Joint probability as two equivalent products
3. Proof of Bayes' theorem from symmetry of 'and'

1. Diagram
   Observation

   The square diagrams label P(A), P(B|A), P(B), and P(A|B) as nested proportions.

1. Conditional probability as a restricted proportion
2. Area-model derivation of Bayes' theorem
3. Right-hand area model for P(A)P(B|A)
4. Left-hand area model for P(B)P(A|B)

1. Formula
   Observation

   The screen shows both P(A|B)=P(A)P(B|A)/P(B) and P(B)P(A|B)/P(A)=P(B|A).

1. Bayes' theorem formula
2. Proof of Bayes' theorem from symmetry of 'and'

1. Formula
   Observation

   The final equation is annotated with (1/21), (4/10), and (24/210).

1. Worked numeric use of the rearranged Bayes formula
2. Numerical substitution into the rearranged Bayes identity

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

2. Diagram
   Observation

   A becomes a book icon and B becomes a magnifying glass icon.

1. book icon
2. magnifying glass icon
3. Icon-labeled reverse conditional

1. Formula
   Observation

   The display gives book-given-magnifier as the book marginal times magnifier-given-book, divided by the magnifier marginal.

1. Bayes' Theorem

1. Formula
   Observation

   P(A and B) ???= P(A)P(B)

1. Joint Probability of Independent Events

1. Formula
   Observation

   Red cross over P(A and B) = P(A)P(B).

2. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

1. General Multiplication Rule for Joint Probability
2. Definition of Statistical Independence
3. Misconception: Joint Probability is Always the Product of Marginals

1. Diagram
   Observation

   Coin histogram animation.

2. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

1. Definition of Statistical Independence
2. Misconception: Introductory Examples Represent Real-World Complexity
3. s152-cr-independence-to-bayes

1. Animation
   Observation

   Step-by-step formula construction.

2. Diagram
   Observation

   Arrows defining π and +.

1. Bayes' Theorem
2. Medical Testing Scenario for Bayes' Theorem
3. Step-by-Step Assembly of Bayes' Theorem Formula

1. Audio
   Observation

   The spoken explanation at this interval discusses this probability step.

1. Bayes' Theorem
2. Dependence makes conditional updating informative
3. s152-cr-independence-to-bayes