1. Brownian Motion (Introduction)

1. Audio
   Observation

   The speaker says he will call the random variable X sub T, with capital X as a random variable and subscript t meaning it is indexed at time t.

2. Formula
   Observation

   He writes $X_t$ on the whiteboard below the title and again in the distribution statement.

$X_t$

Random variable of the stochastic process observed at time t.

$t \ge 0$; $X_t$ is real-valued.

1. Audio
   Observation

   The speaker identifies the x-axis as time.

2. Diagram
   Observation

   The horizontal axis of the first coordinate sketch is labeled t.

t

Time index of the stochastic process, drawn on the horizontal axis.

Nonnegative time values are used in the examples.

1. Audio
   Observation

   The speaker says $X_t$ has a normal distribution mean zero variance t.

2. Formula
   Observation

   The whiteboard shows $X_t \sim N(0,t)$.

$N(0,t)$

Normal distribution notation with first parameter the mean and second parameter the variance.

The displayed variance parameter is t.

1. Audio
   Observation

   The speaker states that the mean is zero and that on average the value will be zero.

2. Formula
   Observation

   The first argument in $N(0,t)$ is 0.

0

Mean of the normal distribution for $X_t$.

Applies to the marginal distribution shown for $X_t$.

1. Audio
   Observation

   The speaker states that the variance is t and explains that as this t increases, the variance goes up one-to-one.

2. Formula
   Observation

   The second argument in $N(0,t)$ is t.

t

Variance parameter of the normal distribution for $X_t$.

Same time index as the subscript in $X_t$.

1. Formula
   Observation

   Written on the whiteboard as $X_t$.

2. Audio
   Observation

   The speaker refers to it as 'the value of X at time t'.

$X_t$

The value of the Brownian motion process at time t.

$t \ge 0$

1. Formula
   Observation

   Written on the whiteboard as $N(0, t)$ and $N(0, t-s)$.

2. Audio
   Observation

   The speaker says 'normal distribution'.

N

Notation for a normal distribution, with parameters for mean and variance.

$\mathbb{R}$

1. Formula
   Observation

   Written on the whiteboard as a parameter in $N(0, t)$ and as an axis label.

2. Audio
   Observation

   The speaker refers to it as 'time'.

t

Time variable.

$t \ge 0$

1. Formula
   Observation

   Written on the whiteboard as a parameter in $X_t - X_s$ and $N(0, t-s)$.

2. Audio
   Observation

   The speaker refers to it as 'time s'.

s

An earlier time variable, where $s < t$.

$0 \le s < t$

1. Formula
   Observation

   Written on the whiteboard as a double vertical line symbol between two expressions.

2. Audio
   Observation

   The speaker says 'This is the sign I'm going to use for independent'.

$\perp$\!\!$\perp$

Symbol denoting statistical independence between two random variables or events.

$N/A$

1. Formula
   Observation

   Whiteboard shows $X_t$ in the expressions $X_t \sim N(0,t)$, $X_t - X_s \sim N(0,t-s)$, and $X_t - X_s \perp X_w - X_v$.

2. Audio
   Observation

   The lecturer repeatedly refers to X sub t as the value of the process at time t.

$X_t$

Value of the Brownian motion process at time t.

t is a nonnegative time parameter; $X_t$ is real-valued.

1. Formula
   Observation

   Whiteboard shows $X_t \sim N(0,t)$.

2. Audio
   Observation

   The lecturer says X sub t is normal, the mean is zero, and the variance is time.

$N(0,t)$

Normal distribution with mean 0 and variance t.

Requires $t \ge 0$ for the variance to be nonnegative.

1. Audio
   Observation

   The speaker introduces a new type of stochastic process called Brownian motion and says it has applications in physics, finance, and statistics.

2. Diagram
   Observation

   The whiteboard title reads "Brownian Motion".

Brownian motion is presented as a popular stochastic process, meaning a random variable that evolves over time. The speaker frames this clip as an introduction and says later videos will show interesting results.

1. The topic is introduced informally in this clip.

2. Applications mentioned include physics, finance, and statistics.

1. Audio
   Observation

   The speaker says a stochastic process means a random variable that evolves throughout time, and that the subscript t means it is indexed at time t and has different values.

2. Diagram
   Observation

   A coordinate system is drawn with the horizontal axis labeled t.

3. Formula
   Observation

   The symbol $X_t$ is written on the board.

The notation $X_t$ denotes a random variable indexed by time t. The horizontal axis in the sketch represents time, while the vertical direction represents the value of the random variable at that time.

1. t is the time index.

2. $X_t$ is a random variable whose value depends on t.

1. Audio
   Observation

   The speaker says Brownian motion can be defined as X sub T having a normal distribution mean zero variance t.

2. Formula
   Observation

   The whiteboard displays $X_t \sim N(0,t)$.

1. The video gives only the marginal distribution statement in this clip and does not state independent increments or continuity conditions here.

In this clip, Brownian motion is introduced through the displayed marginal rule $X_t \sim N(0,t)$: at each time t, the random variable $X_t$ is normally distributed with mean 0 and variance t. The speaker emphasizes that this single formula succinctly captures the distributional behavior being discussed.

1. t is the time index.

2. The first parameter of N is the mean.

3. The second parameter of N is the variance.

4. The variance equals the same t appearing in the subscript of $X_t$.

1. Stochastic process notation $X_t$

1. Audio
   Observation

   The speaker says that as this t increases, this t goes up one-to-one, and that the longer time we're looking forward in the future, the higher the variance will be.

2. Formula
   Observation

   The displayed formula $X_t \sim N(0,t)$ contains t as the variance parameter.

Because the variance parameter in $N(0,t)$ is exactly the time index t, larger times correspond to larger variances. The speaker interprets this as greater uncertainty farther into the future.

1. Uses the displayed convention that the second argument of N is variance.

2. Applies to the marginal distribution stated in this clip.

1. Marginal normal-distribution definition shown for Brownian motion

1. Formula
   Observation

   Title 'Brownian Motion' written on the whiteboard.

2. Audio
   Observation

   The speaker discusses properties of Brownian motion.

A stochastic process characterized by normally distributed values with a mean of zero and variance proportional to time, and independent increments.

1. $X_t \sim N(0, t)$

2. $X_t - X_s \sim N(0, t-s)$ for $s < t$

3. Increments are independent

1. Formula
   Observation

   Written as $X_t \sim N(0, t)$ and $X_t - X_s \sim N(0, t-s)$.

2. Audio
   Observation

   The speaker mentions 'normal distribution'.

A continuous probability distribution characterized by its mean and variance. In this context, it describes the distribution of the Brownian motion process at a given time or over a time interval.

1. $\mu$ is the mean

2. $\sigma^2$ is the variance

1. Formula
   Observation

   Written as $X_t - X_s$.

2. Audio
   Observation

   The speaker says 'increment basically means... the change in some small time interval'.

The difference in the value of the process between two times, s and t, representing the change over that interval.

1. $s < t$

1. Brownian Motion

1. Formula
   Observation

   Whiteboard displays $X_t \sim N(0,t)$.

2. Audio
   Observation

   The lecturer recaps that X sub t is normal, the mean is zero, and the variance is time.

The video states that for Brownian motion, the value at time t is normally distributed with mean 0 and variance t. The lecturer emphasizes that the mean being 0 means the most likely value is 0, not that every realized path stays at 0 on average in a sample-path sense.

1. $t \ge 0$

2. Mean is 0

3. Variance is t

1. Formula
   Observation

   Whiteboard displays $X_t - X_s \sim N(0,t-s)$ with $t > s$.

2. Audio
   Observation

   The lecturer says this is for t greater than s, otherwise we would have a negative variance.

For two times s and t with $t > s$, the increment $X_t - X_s$ is normally distributed with mean 0 and variance equal to the elapsed time t-s. The condition $t > s$ is necessary so that the variance is positive.

1. $t > s$

2. Increment mean is 0

3. Increment variance is t-s

1. Marginal distribution of Brownian motion at time t

1. Formula
   Observation

   Whiteboard displays $X_t - X_s \perp X_w - X_v$ with $t > s$ and $w > v$.

2. Audio
   Observation

   The lecturer says this is for non-overlapping intervals and explains that two non-overlapping intervals are independent.

3. Diagram
   Observation

   The lecturer draws two separate intervals on the upper graph to illustrate non-overlap.

1. The board also writes t-s, u-v under the independence statement, while the symbols in the formula use v and w; the intended second interval is clear from speech, but the written u is inconsistent.

The video presents the independent-increments property of Brownian motion: if two time intervals do not overlap, then the corresponding increments are independent random variables. The lecturer contrasts this with overlapping intervals, where observing one increment gives information about the other.

1. $t > s$

2. $w > v$

3. The intervals determined by (s,t) and (v,w) are non-overlapping

1. Distribution of a Brownian increment

1. Audio
   Observation

   The lecturer says the variance grows with time, the variance is time, and the farther time goes on the wider the distribution gets.

2. Diagram
   Observation

   The lower sketch shows a narrow bell curve near early time and a wider bell curve later in time.

The lecture interprets the formula $X_t \sim N(0,t)$ visually: as time increases, the spread of the distribution increases. Early in time the process is more concentrated near 0; later in time large values become more likely because there has been more time to drift.

1. Applies to the marginal law $X_t \sim N(0,t)$

1. Marginal distribution of Brownian motion at time t

1. Diagram
   Observation

   Whiteboard title "Brownian Motion" is visible above a hand-drawn path.

2. Formula
   Observation

   Whiteboard states $X_t \sim N(0,t)$.

1. The clip does not explicitly restate the initial condition $X_0=0$, although it is consistent with the displayed formula at $t=0$.

For Brownian motion, the value at time $t$ is normally distributed with mean $0$ and variance $t$. This is the marginal distribution of the process at a fixed time.

1. $t\ge 0$ is implicit from the time-indexed setting shown on the board.

2. The clip presents this as part of the introductory description of Brownian motion.

1. Audio
   Observation

   The speaker says Brownian motion has applications in all different fields: physics, finance, obviously statistics.

Brownian motion has applications in physics, finance, and statistics.

1. The term under discussion is Brownian motion.

No formal quantifier is stated; this is an informal list of application areas.

1. Audio
   Observation

   The speaker says as this t increases, this t goes up one-to-one, and the longer time we're looking forward in the future, the higher the variance will be.

2. Formula
   Observation

   The board shows $X_t \sim N(0,t)$.

For the displayed definition $X_t \sim N(0,t)$, increasing t increases the variance of $X_t$ one-to-one.

1. The second parameter of N denotes variance.

2. t is the same time index in $X_t$ and in the variance parameter.

For the times considered in the lecture, larger t gives larger variance.

1. Formula
   Observation

   Written as $X_t \sim N(0, t)$.

2. Audio
   Observation

   The speaker explains that at time 10, the variance is wider than at time 5, but the mean is still zero.

For a Brownian motion process $X_t$, the value at any time t follows a normal distribution with mean 0 and variance t.

1. $X_t$ is a standard Brownian motion process

For all $t \ge 0$.

1. Formula
   Observation

   Written as $X_t - X_s \sim N(0, t-s)$.

2. Audio
   Observation

   The speaker states that the increment is a normal distribution with mean zero and variance t-s.

For a Brownian motion process $X_t$, the increment $X_t - X_s$ for $s < t$ follows a normal distribution with mean 0 and variance t-s.

1. $X_t$ is a standard Brownian motion process

2. $s < t$

For all $0 \le s < t$.

1. Formula
   Observation

   Written as $X_t - X_s \perp$\!\!$\perp X_v - X_w$.

2. Audio
   Observation

   The speaker states that the increment $X_t - X_s$ is independent of $X_v - X_w$.

1. The specific ordering of the time points s, t, v, w is not explicitly stated, though non-overlapping intervals are implied by the concept of independent increments.

For a Brownian motion process $X_t$, non-overlapping increments are statistically independent.

1. $X_t$ is a standard Brownian motion process

For appropriate time points defining non-overlapping intervals.

1. Audio
   Observation

   The lecturer says, 'this is also for t greater than s otherwise we'd have a negative variance.'

2. Formula
   Observation

   Whiteboard shows $X_t - X_s \sim N(0,t-s)$ with $t > s$.

For the increment formula $X_t - X_s \sim N(0,t-s)$ to make sense as a normal variance, one must have $t > s$; otherwise t-s would be negative.

1. The increment variance is t-s

2. A normal distribution variance must be nonnegative

For the displayed increment law with times s and t.

1. Audio
   Observation

   The lecturer says it makes sense that if intervals are overlapping then they're dependent, because observing part of one interval gives information about the other.

2. Diagram
   Observation

   He contrasts non-overlapping intervals with overlapping examples such as 8 to 2 and 11 to 4.

1. This is presented as intuition rather than a formal proof.

When two time intervals overlap, the corresponding increments are generally dependent because they share common time evolution.

1. The two intervals overlap in time

For overlapping time intervals in the Brownian motion setting.

1. Formula
   Observation

   Whiteboard shows $X_t - X_s \perp X_w - X_v$ with non-overlapping written underneath.

2. Audio
   Observation

   The lecturer says two non-overlapping intervals are going to be independent.

For Brownian motion, increments over non-overlapping time intervals are independent.

1. $t > s$

2. $w > v$

3. The intervals determined by (s,t) and (v,w) do not overlap

For any two non-overlapping time intervals in the process.

1. Audio
   Observation

   Speaker says the variance is "just the time that elapses from time S to time T, which is T minus S."

2. Formula
   Observation

   Whiteboard shows $X_t - X_s \sim N(0,t-s)$.

For $t>s$, the increment $X_t-X_s$ has variance $t-s$.

1. $t>s$.

2. $X$ is Brownian motion as introduced on the board.

For all times $s,t$ with $t>s$.

1. Audio
   Observation

   Speaker says increments are independent "as long as they're not overlapping" and explains that overlapping increments would give information about one another.

2. Formula
   Observation

   Whiteboard shows $X_t - X_s \perp X_w - X_v$ with "non-overlapping".

Increments of Brownian motion over non-overlapping time intervals are independent.

1. The two intervals are each ordered in time.

2. The intervals do not overlap.

For pairs of non-overlapping increments of the process.

1. Audio
   Observation

   The speaker starts from the displayed formula, points out that the mean is zero and the variance is t, then links the two occurrences of t and concludes that farther future times have higher variance.

2. Formula
   Observation

   The board shows $X_t \sim N(0,t)$.

1. Expression

   $$X_t \sim N(0,t)$$
   Explanation

   Begin with the displayed marginal distribution for Brownian motion in this clip.

   Justification

   Directly written on the whiteboard and spoken by the lecturer.

   Shown in the video
2. Expression

   $$\mathbb{E}[X_t]=0$$
   Explanation

   The first parameter of the normal distribution is the mean, so the expected value is zero.

   Justification

   Standard interpretation of N(mean, variance) plus the speaker's explicit statement that the mean is zero.

   Derived from the video
3. Expression

   $$\operatorname{Var}(X_t)=t$$
   Explanation

   The second parameter is the variance, so the variance equals the time index t.

   Justification

   Standard interpretation of N(mean, variance) plus the speaker's explicit statement that the variance is t.

   Derived from the video
4. Expression

   $$t_1<t_2 \Rightarrow \operatorname{Var}(X_{t_1})<\operatorname{Var}(X_{t_2})$$
   Explanation

   Since variance is exactly t, larger time gives larger variance.

   Justification

   Monotonicity of the identity function $t \mapsto t$ for positive times.

   Derived from the video

The displayed definition implies that the process remains centered at zero while its spread increases linearly with time.

1. Audio
   Observation

   The lecturer says, 'doesn't mean on average it's zero, it's just mean that the mean is zero, the most likely value is zero.'

1. This is a verbal clarification rather than a formal derivation.

1. Expression

   $$$$
   Explanation

   Start from the displayed law $X_t \sim N(0,t)$.

   Justification

   Whiteboard formula and spoken recap.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Interpret the first parameter 0 as the distributional mean, i.e., the center of the normal law.

   Justification

   Standard meaning of $N(\mu,\sigma^2)$.

   Supplementary explanation
3. Expression

   $$$$
   Explanation

   State that this does not assert each realized path has average zero in a sample-path sense; it says the most likely value / center of the distribution is 0.

   Justification

   Lecturer's explicit verbal clarification.

   Shown in the video

The statement 'mean is zero' refers to the center of the normal distribution of $X_t$, not to a pathwise average interpretation.

1. Audio
   Observation

   Speaker explains that the increment remains normally distributed regardless of whether $X_s$ starts in positive or negative territory.

2. Formula
   Observation

   Whiteboard displays $X_t - X_s \sim N(0,t-s)$ with $t>s$.

3. Diagram
   Observation

   Speaker points to the time axis and the increment formula while talking.

1. Expression

   $$X_t - X_s \sim N(0,t-s)$$
   Explanation

   The displayed increment law says the change from time $s$ to time $t$ is Gaussian with mean $0$ and variance $t-s$.

   Justification

   Read directly from the whiteboard formula.

   Shown in the video
2. Expression

   $$\text{mean}=0,\quad \text{variance}=t-s$$
   Explanation

   The speaker identifies the two parameters of the normal law: mean zero and variance equal to the elapsed time.

   Justification

   Audio explanation accompanying the formula.

   Shown in the video
3. Expression

   $$\text{sign of }X_s\text{ does not matter}$$
   Explanation

   Whether the path is currently in positive or negative territory does not change the distribution of the future increment.

   Justification

   Speaker explicitly contrasts positive and negative starting territory.

   Shown in the video

The increment distribution is determined by the length of the time interval, not by the current position of the path.

1. Audio
   Observation

   Speaker says, "if I said there was an increment like that mostly overlapped, obviously knowing about one of the increments would give information about the other one."

2. Diagram
   Observation

   Speaker gestures over the time-axis sketch while contrasting separated and overlapping intervals.

1. Expression

   $$\text{non-overlapping intervals} \Rightarrow \text{independent increments}$$
   Explanation

   The board and speech present independence under the non-overlap condition.

   Justification

   Directly stated in the lecture segment.

   Shown in the video
2. Expression

   $$\text{mostly overlapped intervals} \Rightarrow \text{information shared}$$
   Explanation

   If two increments share part of their time interval, observing one reveals something about the common portion and hence about the other.

   Justification

   Speaker's verbal explanation of dependence through overlap.

   Shown in the video

Independence is tied to disjointness of the time intervals; overlap creates statistical dependence.

1. Audio
   Observation

   The speaker says he is drawing one sample path of Brownian motion and describes it as a random thing bumping along.

2. Diagram
   Observation

   A coordinate system with horizontal axis t is drawn, and a jagged continuous-looking curve is sketched above it.

1. The hand-drawn curve is schematic rather than a numerically exact realization.

Illustrate what one sample path of Brownian motion looks like.

1. Horizontal axis represents time t.

2. The process is denoted $X_t$.

3. Brownian motion is introduced as a stochastic process.

Provide a visual sketch of a typical trajectory.

1. Expression

   $$$$
   Explanation

   Draw axes with time on the horizontal axis.

   Justification

   The speaker explicitly identifies the x-axis as time.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Write the process label $X_t$ near the sketch.

   Justification

   The speaker introduces the notation for the random variable at time t.

   Shown in the video
3. Expression

   $$$$
   Explanation

   Sketch an irregular, wandering curve as a possible realized path.

   Justification

   The speaker describes the path as a random thing bumping along.

   Shown in the video

A jagged, irregular curve over the time axis serves as the visual example of one sample path.

The sketch matches the spoken description of a random path evolving over time, though it is qualitative rather than computed.

1. Audio
   Observation

   The speaker compares time t equals 3 to time t equals 10 and says that at time 3 the value is a normal distribution centered at zero and does not have too much variance.

2. Diagram
   Observation

   A second coordinate sketch is drawn with tick marks labeled 3 and 10 on the horizontal axis.

3. Formula
   Observation

   The earlier displayed rule $X_t \sim N(0,t)$ remains on the board.

1. The clip ends before the speaker completes the comparison at $t=10$.

2. No explicit numerical density is written for the $t=3$ case.

Use the displayed definition to compare the behavior of $X_t$ at two different times.

1. $X_t \sim N(0,t)$.

2. Two times are marked on the axis: 3 and 10.

Explain how the distribution changes between these times.

1. Expression

   $$X_3 \sim N(0,3)$$
   Explanation

   Substitute $t=3$ into the displayed marginal rule.

   Justification

   Direct substitution into $X_t \sim N(0,t)$.

   Derived from the video
2. Expression

   $$X_{10} \sim N(0,10)$$
   Explanation

   Substitute $t=10$ into the displayed marginal rule.

   Justification

   Direct substitution into $X_t \sim N(0,t)$.

   Derived from the video
3. Expression

   $$$$
   Explanation

   The speaker begins describing the $t=3$ case as centered at zero with not too much variance.

   Justification

   This follows from mean 0 and variance 3 in the displayed formula.

   Shown in the video

At $t=3$ the distribution is narrower than at $t=10$ because the variance is 3 instead of 10; the clip stops before the full verbal comparison is completed.

The conclusion follows immediately from the displayed formula and the marked times on the second sketch.

1. Audio
   Observation

   The lecturer gives numerical examples: $t=5$, $s=3$, $w=10$, $v=7$ are non-overlapping; 8 to 2 and 11 to 4 overlap from 4 to 8.

2. Diagram
   Observation

   He points to the board and draws separate intervals to illustrate the idea.

1. The spoken ordering '8 to 2' and '11 to 4' is informal; the intended intervals are [2,8] and [4,11].

Use concrete numbers to distinguish non-overlapping time intervals from overlapping ones in the independent-increment property.

1. First interval endpoints: $t=5$, $s=3$

2. Second interval endpoints: $w=10$, $v=7$

3. Contrasting example: intervals [2,8] and [4,11]

Determine which pairs of intervals are non-overlapping and therefore independent in increment, and which overlap and are dependent.

1. Expression

   $$[3,5]$$
   Explanation

   Rewrite the first pair as the time interval from $s=3$ to $t=5$.

   Justification

   Directly from the lecturer's spoken example.

   Shown in the video
2. Expression

   $$[7,10]$$
   Explanation

   Rewrite the second pair as the time interval from $v=7$ to $w=10$.

   Justification

   Directly from the lecturer's spoken example.

   Shown in the video
3. Expression

   $$[3,5] \cap [7,10] = \varnothing$$
   Explanation

   These two intervals do not share any time points, so they are non-overlapping.

   Justification

   Set-intersection check of the two intervals.

   Derived from the video
4. Expression

   $$X_5 - X_3 \perp X_{10} - X_7$$
   Explanation

   Therefore the corresponding increments are independent by the Brownian independent-increments property.

   Justification

   Application of the stated theorem for non-overlapping intervals.

   Shown in the video
5. Expression

   $$[2,8] \cap [4,11] = [4,8]$$
   Explanation

   In the contrasting example, the intervals overlap on [4,8].

   Justification

   Direct interval intersection calculation from the spoken numbers.

   Derived from the video
6. Expression

   $$X_8 - X_2 \not\perp X_{11} - X_4$$
   Explanation

   Because the intervals overlap, the increments are dependent rather than independent.

   Justification

   Lecturer's intuitive explanation that shared time creates dependence.

   Shown in the video

Non-overlapping example: [3,5] and [7,10] give independent increments. Overlapping example: [2,8] and [4,11] share [4,8], so their increments are dependent.

Checked by comparing interval intersections: empty intersection implies non-overlap, nonempty intersection implies overlap.

1. Diagram
   Observation

   The whiteboard initially shows the title "Brownian Motion" at the top.

2. Diagram
   Observation

   The presenter draws a vertical axis and a horizontal axis beneath the title.

1. Whiteboard title "Brownian Motion"

2. Vertical axis

3. Horizontal axis

1. A blank titled board becomes a coordinate sketch.

2. The horizontal axis is later identified as time.

1. The topic remains Brownian motion.

2. The sketch is placed below the title.

The visual setup introduces a graph whose horizontal direction will represent the time parameter of the stochastic process.

1. Diagram
   Observation

   A jagged curve is drawn across the first coordinate system.

2. Audio
   Observation

   The speaker says a Brownian motion might look something like this and describes it as a random thing bumping along.

1. The exact shape is hand-drawn and not quantitatively specified.

1. Jagged curve

2. Time axis

3. Vertical value axis

1. An empty coordinate plane gains an irregular path.

2. The path moves up and down as time increases.

1. The curve remains a single continuous-looking trajectory.

2. The horizontal axis still represents time.

The animation-like live drawing conveys that an individual realization of Brownian motion is irregular and fluctuating rather than smooth or deterministic.

1. Diagram
   Observation

   A second coordinate system is drawn below the formula.

2. Diagram
   Observation

   Tick marks labeled 3 and 10 appear on the horizontal axis.

3. Audio
   Observation

   The speaker compares time t equals 3 to time t equals 10.

1. The clip ends before a full distributional sketch at both times is completed.

1. Second coordinate system

2. Tick mark 3

3. Tick mark 10

4. Horizontal time axis

1. A new axis pair is added below the formula.

2. Specific time values 3 and 10 are marked for comparison.

1. The same definition $X_t \sim N(0,t)$ governs both times.

2. The horizontal axis continues to represent time.

The second diagram shifts from a single sample path to a comparison of marginal distributions at two different times.

1. Diagram
   Observation

   A graph with a horizontal axis labeled 't' and a vertical axis. Two bell curves are drawn, one centered at $t=5$ and a wider one at $t=10$, both centered vertically at 0.

1. horizontal axis (time t)

2. vertical axis

3. bell curve at $t=5$

4. bell curve at $t=10$

1. The speaker draws a second, wider bell curve at $t=10$ to illustrate increasing variance.

1. Both curves are centered at a vertical value of 0, representing the mean.

The graph visually demonstrates that as time t increases, the variance of the Brownian motion $X_t$ increases, while the mean remains zero.

1. Diagram
   Observation

   A jagged line graph representing a sample path of Brownian motion, with points marked at times s and t.

1. jagged line

2. point at time s

3. point at time t

1. The speaker points to the values at times s and t to explain the increment.

The graph illustrates the concept of an increment $X_t - X_s$ as the vertical difference between the process values at two different times.

1. Diagram
   Observation

   The whiteboard title reads 'Brownian Motion'. Below it are a jagged path sketch, the formula $X_t \sim N(0,t)$, a time-axis sketch with marked points, the formula $X_t - X_s \sim N(0,t-s)$ with $t > s$, and the formula $X_t - X_s \perp X_w - X_v$ with $t > s$, $w > v$, non-overlapping, and t-s, u-v written nearby.

1. Some handwritten labels on the time axis are small and not fully legible beyond the main formulas.

1. Title 'Brownian Motion'

2. Jagged path sketch

3. Formula $X_t \sim N(0,t)$

4. Time-axis sketch

5. Formula $X_t - X_s \sim N(0,t-s)$

6. Formula $X_t - X_s \perp X_w - X_v$

7. Handwritten conditions $t > s$, $w > v$, non-overlapping, t-s, u-v

1. The lecturer adds the conditions $t > s$ and $w > v$ beside the increment and independence formulas.

2. He writes 'non-overlapping' and 't-s, u-v' beneath the independence statement.

3. He later draws extra interval marks on the upper graph to illustrate non-overlap.

1. The central topic remains Brownian motion throughout the clip.

2. The marginal law $X_t \sim N(0,t)$ stays visible on the board.

The board organizes the lesson into three layers: the marginal distribution, the increment distribution, and the independence of increments over disjoint time intervals.

1. Diagram
   Observation

   A lower sketch shows a narrow bell-shaped curve near early time and a wider bell-shaped curve later along the time axis.

2. Audio
   Observation

   The lecturer says the variance grows with time and the farther time goes on the wider the distribution gets.

1. The exact horizontal positions of the drawn curves are schematic rather than numerically calibrated.

1. Narrow bell curve near early time

2. Wider bell curve at later time

3. Time axis

1. As one moves to later times on the sketch, the drawn normal curve becomes wider.

1. Both curves are centered at 0.

2. The qualitative message is that variance increases with time.

The drawing illustrates that $X_t$ remains centered at 0 but becomes more spread out as t increases, matching $X_t \sim N(0,t)$.

1. Diagram
   Observation

   The lecturer draws two separate intervals on the upper path sketch while explaining non-overlapping increments.

2. Audio
   Observation

   He says, 'I can draw two non-overlapping intervals, which I can draw here, are going to be independent.'

1. The exact endpoint labels on the newly drawn intervals are not all clearly legible.

1. Upper jagged Brownian path

2. Two separately marked intervals

1. Two distinct intervals are marked on the path to represent disjoint time windows.

1. The intervals are drawn without overlap.

2. The purpose is to visualize independence of increments.

The visual separates the time line into disjoint pieces to show that the corresponding increments evolve independently.

1. Diagram
   Observation

   A whiteboard shows the title "Brownian Motion", a jagged sample path, a horizontal time axis marked with $s$ and $t$, and three formulas: $X_t \sim N(0,t)$, $X_t - X_s \sim N(0,t-s)$ with $t>s$, and $X_t - X_s \perp X_w - X_v$ with side conditions including "non-overlapping".

2. Animation
   Observation

   The lecturer points first to the increment formula, then to the time axis, then to the independence line, and finally gestures broadly while concluding.

1. The exact handwritten side condition under the independence line includes $u>v$ even though the displayed second increment uses $w$ and $v$; this appears to be a notation inconsistency on the board.

1. Title "Brownian Motion"

2. Jagged sample path

3. Horizontal time axis with marks $s$ and $t$

4. Formula $X_t \sim N(0,t)$

5. Formula $X_t - X_s \sim N(0,t-s)$

6. Formula $X_t - X_s \perp X_w - X_v$

7. Handwritten conditions $t>s$, $u>v$, and "non-overlapping"

1. The lecturer shifts attention from the marginal law to the increment law, then to the independence condition.

2. Pointing gestures connect the spoken explanation to specific formulas and to the time-axis sketch.

1. The board content itself remains fixed throughout the clip.

2. The central theme remains the defining probabilistic structure of Brownian motion.

The visual layout organizes Brownian motion into three layers: marginal normality at a fixed time, Gaussian increments over an interval, and independence between increments over disjoint intervals.

1. Audio
   Observation

   After drawing a path that looks like a random thing bumping along, the speaker says it is really cool that probability and statistics tools can succinctly define Brownian motion.

One might think Brownian motion is just an irregular random-looking curve with no precise structure.

The speaker contrasts the rough sketch with a concise probabilistic definition, namely the normal-distribution rule written on the board.

1. Audio
   Observation

   The speaker explicitly notes that as t increases, the variance goes up one-to-one.

2. Formula
   Observation

   The board shows $X_t \sim N(0,t)$.

One might assume the spread of $X_t$ stays the same at all times.

In the displayed definition, the variance equals t, so the distribution becomes more spread out as time increases.

1. Audio
   Observation

   The lecturer explicitly corrects himself: 'doesn't mean on average it's zero, it's just mean that the mean is zero, the most likely value is zero.'

Thinking that saying the mean is zero means each realized trajectory has average zero over time.

The video clarifies that mean zero refers to the center of the normal distribution of $X_t$, i.e., the most likely value is 0, not a sample-path average statement.

1. Audio
   Observation

   The lecturer says that if t is not greater than s, 'otherwise we'd have a negative variance.'

Using $X_t - X_s \sim N(0,t-s)$ without requiring $t > s$.

The video stresses that t must exceed s so that t-s is a valid positive variance.

1. Audio
   Observation

   The lecturer contrasts non-overlapping intervals with overlapping ones and says overlapping intervals are dependent.

Believing increments are independent even when their time intervals overlap.

Only increments over non-overlapping intervals are independent; overlapping intervals share time evolution and therefore create dependence.

1. Audio
   Observation

   Speaker says that even if the increment starts in negative territory or positive territory, the change between $s$ and $t$ is still normally distributed with mean zero and variance $t-s$.

One might think that if $X_s$ is already positive or negative, then the future increment should have a different distribution.

The increment $X_t-X_s$ remains $N(0,t-s)$ regardless of whether the current position $X_s$ is positive or negative.

1. Audio
   Observation

   Speaker explains that if increments mostly overlap, knowing one gives information about the other.

One might assume all increments of Brownian motion are independent without checking the intervals.

Independence is asserted only for non-overlapping increments; overlapping increments are dependent because they share common time.

1. Audio
   Observation

   The speaker defines a stochastic process as a random variable that evolves throughout time and then introduces $X_t$ as the indexed random variable.

The general idea of a stochastic process is made concrete through the time-indexed notation $X_t$.

1. Audio
   Observation

   The speaker moves from the notation $X_t$ to saying that Brownian motion can be defined by a normal distribution with mean zero and variance t.

2. Formula
   Observation

   The board shows $X_t \sim N(0,t)$ after the notation $X_t$ has been introduced.

The time-indexed notation is used to state the marginal distribution law for Brownian motion.

1. Audio
   Observation

   The speaker reads consequences directly from the formula, noting that the same t appears in the subscript and as the variance.

2. Formula
   Observation

   The displayed formula is $X_t \sim N(0,t)$.