Exponential Smoothing Method in Forecasting | Forecasting Techniques - Exponential smoothing method

1. Diagram
   Observation

   Text box states 'Smoothing parameters (smoothing constants) — usually denoted by α'.

2. Audio
   Observation

   Speaker says 'This is denoted as alpha.'

$\alpha$

Smoothing constant or parameter that determines the weights for observations.

$0 \le \alpha \le 1$

1. Formula
   Observation

   Formulas use subscript t and $t+1$.

2. Audio
   Observation

   Speaker says 't plus one minus t denotes the time'.

t

Time period index.

Integer representing periods (e.g., years).

1. Diagram
   Observation

   Table column header is 'Forecast Demand ($F_t$)'.

2. Formula
   Observation

   Formulas use $F_t$ and $F_{t+1}$.

$F_t$

Forecast demand for period t.

Real numbers

1. Diagram
   Observation

   Table column header is 'Actual Demand ($A_t$)'.

2. Formula
   Observation

   Second formula uses $A_t$.

$A_t$

Actual demand for period t.

Real numbers

1. Formula
   Observation

   The slide shows the formula $F_{t+1} = \alpha A_t + (1-\alpha)F_t$.

2. Audio
   Observation

   The instructor refers to this as the formula and substitutes values into it for each period.

$F_{t+1}$

Forecast demand for the next period $t+1$.

Numerical forecast value; in this example, demand units per year.

1. Formula
   Observation

   The formula contains $A_t$.

2. Diagram
   Observation

   The table column heading is Actual Demand ($A_t$).

$A_t$

Actual demand in period t.

Given numerical demand data for years 1 through 6 in this example.

1. Formula
   Observation

   The formula contains $F_t$.

2. Diagram
   Observation

   The table column heading is Forecast Demand ($F_t$).

$F_t$

Forecast demand for period t, used as the previous forecast when computing $F_{t+1}$.

Numerical forecast value; initialized at 310 for period 1 in this example.

1. Formula
   Observation

   The formula uses $\alpha$ and ($1-\alpha$).

2. Audio
   Observation

   The instructor says, "we are given with alpha 0.4".

3. Diagram
   Observation

   Handwritten annotation shows $\alpha = 0.4$.

$\alpha$

Smoothing constant applied to the actual demand term.

In this example, fixed at 0.4.

1. Formula
   Observation

   The formula contains ($1-\alpha$).

2. Audio
   Observation

   The instructor says, "one minus alpha is one minus zero point four will give us zero point six".

3. Diagram
   Observation

   Handwritten annotation shows ($1-\alpha$)=$1-0.4=0.6$.

($1-\alpha$)

Complementary weight applied to the previous forecast term.

In this example, equals 0.6.

1. Formula
   Observation

   Subscripts t and $t+1$ appear in $F_{t+1}$, $A_t$, and $F_t$.

2. Diagram
   Observation

   The table lists Year 1 through Year 7.

t

Period index, here corresponding to year number.

Integer period labels shown from 1 to 7 in the table.

1. Formula
   Observation

   Formula shown as $F_{t+1} = \alpha A_t + (1-\alpha)F_t$; handwritten notes show $\alpha = 0.4$ and ($1-\alpha$) = $1 - 0.4 = 0.6$.

2. Audio
   Observation

   Speaker refers to 0.4 as the alpha constant and 1 minus alpha as 0.6.

$\alpha$

Smoothing constant used to weight the previous period's actual demand.

Fixed at 0.4 throughout the example.

1. Formula
   Observation

   ($1-\alpha$) appears in the formula and is evaluated as 0.6.

2. Audio
   Observation

   Speaker says one minus alpha is 0.6.

$1-\alpha$

Complementary weight applied to the previous forecast.

Fixed at 0.6 throughout the example.

1. Diagram
   Observation

   Text boxes define exponential smoothing as assigning exponentially decreasing weights from newest to oldest observations.

2. Audio
   Observation

   Speaker explains assigning more weights to actual data and less to forecasted data depending on forecast error.

Exponential smoothing is a time series forecasting technique that assigns exponentially decreasing weights to past observations. Newer data is given more priority (weight), while older data is given less weight.

1. Applicable to time series data.

1. Formula
   Observation

   Formula displayed: $F_{t+1} = F_t + \alpha(A_t - F_t)$.

The next period's forecast equals the current period's forecast plus the smoothing constant multiplied by the forecast error (actual minus forecast).

1. Requires knowing current forecast $F_t$, actual demand $A_t$, and smoothing constant $\alpha$.

1. Definition of Exponential Smoothing

1. Formula
   Observation

   Formula displayed: $F_{t+1} = \alpha A_t + (1 - \alpha)F_t$.

The next period's forecast is a weighted average of the current actual demand and the current forecast. The speaker prefers this form for calculation.

1. Requires knowing current actual demand $A_t$, current forecast $F_t$, and smoothing constant $\alpha$.

1. Definition of Exponential Smoothing

1. Formula
   Observation

   Visible formula on the slide: $F_{t+1} = \alpha A_t + (1-\alpha)F_t$.

2. Audio
   Observation

   The instructor repeatedly substitutes into this formula for periods 2, 3, 4, and begins period 5.

The video presents a one-step-ahead exponential smoothing rule in which the next forecast is a weighted combination of the current actual demand and the current forecast. The actual demand receives weight $\alpha$, and the previous forecast receives weight $(1-\alpha)$.

1. A forecast for period $t+1$ is computed from period t data.

2. The same smoothing constant is used throughout the worked example.

3. The initial forecast $F_1$ must be specified before recursion can begin.

1. Audio
   Observation

   The instructor says, "Whatever is your actual demand, straight away you can post that as a forecasted demand."

2. Diagram
   Observation

   The first row shows Actual Demand 310 and the instructor writes 310 in the Forecast Demand column for Year 1.

For the first period, the video uses the simplest initialization rule: take the observed actual demand in period 1 and place it directly into the forecast column as $F_1$. This creates the seed needed for the recursive calculation starting in period 2.

1. Used when no prior forecast is available.

2. Applied specifically to period 1 in this example.

3. After initialization, later forecasts are generated recursively.

1. Exponential smoothing forecast formula

1. Audio
   Observation

   The instructor says, "Starting from period two onwards, you can take this."

Once $F_1$ has been initialized, the exponential smoothing formula is applied beginning with the calculation of the forecast for period 2. Each new forecast depends on the immediately preceding actual demand and preceding forecast.

1. Requires an already initialized $F_1$.

2. Each step uses only the previous period's actual and forecast values.

1. Initialization by setting the first forecast equal to the first actual demand
2. Exponential smoothing forecast formula

1. Audio
   Observation

   The instructor states that the problem gives alpha 0.4 and computes 1 minus alpha as 0.6.

2. Diagram
   Observation

   Handwritten annotations show $\alpha = 0.4$ and ($1-\alpha$)=$1-0.4=0.6$.

This example fixes the smoothing constant at $\alpha=0.4$, so the complementary weight on the previous forecast is $1-\alpha=0.6$. These two weights remain unchanged across all subsequent calculations in the clip.

1. Specific to the given problem.

2. Weights sum to 1.

3. The same pair of weights is reused for every period in this example.

1. Exponential smoothing forecast formula

1. Formula
   Observation

   Printed formula at top center: $F_{t+1} = \alpha A_t + (1-\alpha)F_t$.

2. Audio
   Observation

   Speaker repeatedly substitutes values into this structure.

The next-period forecast equals a weighted sum of the current period's actual demand and the current period's forecast. The weights are $\alpha$ and $1-\alpha$, which must sum to 1.

1. $\alpha$ is fixed across periods in this example.

2. $A_t$ and $F_t$ refer to the same period $t$.

3. The result is the forecast for period $t+1$.

1. Formula
   Observation

   Handwritten note: $\alpha = 0.4$ and $(1-\alpha) = 1 - 0.4 = 0.6$.

2. Audio
   Observation

   Speaker states '0.4 is the alpha constant' and 'one minus alpha 0.6'.

For this problem, the smoothing constant is set to $\alpha=0.4$, making the complementary weight $1-\alpha=0.6$. These values remain unchanged for all calculations.

1. Applies specifically to the worked example shown.

1. Exponential Smoothing Formula

1. Audio
   Observation

   Speaker states: 'When you are looking for stability, assign more weights to the actual demand. When you are looking for responsiveness, assign more weights to the forecasted demand.'

2. Diagram
   Observation

   Handwritten text appears: 'Stability -> Actual demand', 'Responsiveness -> Forecast demand'.

1. Standard forecasting theory often associates higher alpha with responsiveness to recent changes and lower alpha with stability (smoothing out noise). The video's statement appears to contradict standard terminology, but it is recorded as stated in the source.

To achieve stability, assign more weights to actual demand; to achieve responsiveness, assign more weights to forecasted demand.

1. Using exponential smoothing method.

Universal within the context of the method's application.

1. Formula
   Observation

   Both formulas are presented side-by-side with 'or' between them.

1. Expression

   $$F_{t+1} = F_t + \alpha(A_t - F_t)$$
   Explanation

   Start with Formula 1.

   Justification

   Given in video.

   Shown in the video
2. Expression

   $$F_{t+1} = F_t + \alpha A_t - \alpha F_t$$
   Explanation

   Distribute alpha.

   Justification

   Algebraic distribution.

   Supplementary explanation
3. Expression

   $$F_{t+1} = \alpha A_t + (1 - \alpha)F_t$$
   Explanation

   Factor out $F_t$ from the terms involving $F_t$.

   Justification

   Algebraic factoring.

   Supplementary explanation

The two formulas presented in the video are algebraically equivalent.

1. Audio
   Observation

   The instructor substitutes 0.4, 310, 0.6, and 310 into the formula and says the answer is 310.

2. Formula
   Observation

   Handwritten expression reads $0.4(310)+0.6(310)$.

3. Diagram
   Observation

   A calculator window appears and shows the computation $0.4 \times 310 + 0.6 \times 310 = 310$.

4. Diagram
   Observation

   The result 310 is written in the Forecast Demand column for Year 2.

1. Expression

   $$F_2=0.4(310)+0.6(310)$$
   Explanation

   Substitute $\alpha=0.4$, $A_1=310$, $1-\alpha=0.6$, and $F_1=310$ into the exponential smoothing formula.

   Justification

   Direct substitution into $F_{t+1}=\alpha A_t+(1-\alpha)F_t$ with $t=1$.

   Shown in the video
2. Expression

   $$F_2=310$$
   Explanation

   The weighted sum evaluates to 310, and this value is entered as the forecast for Year 2.

   Justification

   Arithmetic evaluation shown on the calculator and written into the table.

   Shown in the video

The forecast for period 2 is $F_2=310$.

1. Audio
   Observation

   The instructor says the next calculation uses previous demand 365 and previous forecast 310, giving 332.

2. Formula
   Observation

   Handwritten expression reads $0.4(365)+0.6(310)$.

3. Diagram
   Observation

   A calculator window shows $0.4 \times 365 + 0.6 \times 310 = 332$.

4. Diagram
   Observation

   The value 332 is written in the Forecast Demand column for Year 3.

1. Expression

   $$F_3=0.4(365)+0.6(310)$$
   Explanation

   Use the actual demand from Year 2 and the forecast just obtained for Year 2 to compute the forecast for Year 3.

   Justification

   Apply the same recurrence with $t=2$, where $A_2=365$ and $F_2=310$.

   Shown in the video
2. Expression

   $$F_3=332$$
   Explanation

   The arithmetic result is 332, which is placed in the Year 3 forecast cell.

   Justification

   Calculator display and handwritten table entry both show 332.

   Shown in the video

The forecast for period 3 is $F_3=332$.

1. Audio
   Observation

   The instructor says the next period uses actual demand 395 and forecasted demand 332, then states the result is 357.2 and rounds it to 357.

2. Formula
   Observation

   Handwritten expression reads $0.4(395)+0.6(332)$.

3. Diagram
   Observation

   A calculator window shows $0.4 \times 395 + 0.6 \times 332 = 357.2$.

4. Diagram
   Observation

   The rounded value 357 is written in the Forecast Demand column for Year 4.

1. The exact rounding convention is not formally stated beyond the spoken phrase "round it off".

1. Expression

   $$F_4=0.4(395)+0.6(332)$$
   Explanation

   Substitute the Year 3 actual demand and the Year 3 forecast into the recurrence to obtain the Year 4 forecast.

   Justification

   Apply the formula with $t=3$, where $A_3=395$ and $F_3=332$.

   Shown in the video
2. Expression

   $$F_4=357.2$$
   Explanation

   The calculator returns 357.2 before rounding.

   Justification

   Visible calculator output in the video.

   Shown in the video
3. Expression

   $$F_4\approx 357$$
   Explanation

   The instructor rounds the decimal result and writes 357 in the table.

   Justification

   Spoken instruction to round it off, followed by the handwritten table entry 357.

   Shown in the video

The unrounded forecast for period 4 is $357.2$, and the value recorded in the table is $357$.

1. Audio
   Observation

   The instructor says, "Let's calculate the forecasted demand for the fifth year," then identifies actual demand 415 and forecasted demand 357.

2. Formula
   Observation

   Handwriting begins with $0.4(415)$+ and the next line is cut off before completion.

3. Diagram
   Observation

   The Year 5 row is being prepared using the Year 4 values already written in the table.

1. The final numeric result for Year 5 is not shown within this clip.

1. Expression

   $$F_5=0.4(415)+0.6(357)$$
   Explanation

   The instructor starts the next recursive step by substituting the Year 4 actual demand and the rounded Year 4 forecast into the same formula.

   Justification

   This follows directly from $F_{t+1}=\alpha A_t+(1-\alpha)F_t$ with $t=4$, using the values visible in the table.

   Derived from the video

The clip establishes the setup for $F_5$ but does not reach a completed numerical answer before ending.

1. Audio
   Observation

   Speaker calculates year 5 forecast using 415 and 357.

2. Formula
   Observation

   Handwritten line: $\rightarrow 0.4(415) + 0.6(357)$.

3. Animation
   Observation

   Calculator shows input `$0.4 * 415 + 0.6 * 357$` resulting in `380.2`; table entry written as `380`.

1. Rounding from 380.2 to 380 is done by truncation or nearest integer without explicit rule stated.

1. Expression

   $$F_5 = 0.4(A_4) + 0.6(F_4)$$
   Explanation

   Substitute period 4 values into the exponential smoothing formula.

   Justification

   Definition of the method with $\alpha=0.4$.

   Shown in the video
2. Expression

   $$F_5 = 0.4(415) + 0.6(357)$$
   Explanation

   Insert actual demand $A_4=415$ and forecast $F_4=357$.

   Justification

   Values read from the table.

   Shown in the video
3. Expression

   $$F_5 = 166 + 214.2 = 380.2$$
   Explanation

   Compute products and sum.

   Justification

   Arithmetic evaluation.

   Derived from the video
4. Expression

   $$F_5 \approx 380$$
   Explanation

   Record rounded value in the table.

   Justification

   Presenter writes 380 in the Forecast column for Year 5.

   Shown in the video

The forecast for Year 5 is recorded as 380.

1. Audio
   Observation

   Speaker says 'Let's move on to the calculation for the sixth year...' and uses 450 and 380.

2. Formula
   Observation

   Handwritten line: $\rightarrow 0.4(450) + 0.6(380)$.

3. Animation
   Observation

   Calculator shows `$0.4 * 450 + 0.6 * 380$` resulting in `408`; table entry written as `408`.

1. Expression

   $$F_6 = 0.4(A_5) + 0.6(F_5)$$
   Explanation

   Apply formula for period 6 using period 5 data.

   Justification

   Recursive nature of exponential smoothing.

   Shown in the video
2. Expression

   $$F_6 = 0.4(450) + 0.6(380)$$
   Explanation

   Substitute $A_5=450$ and the newly calculated $F_5=380$.

   Justification

   Table values and previous step result.

   Shown in the video
3. Expression

   $$F_6 = 180 + 228 = 408$$
   Explanation

   Calculate the weighted sum.

   Justification

   Arithmetic.

   Derived from the video

The forecast for Year 6 is 408.

1. Audio
   Observation

   Speaker says 'Let's calculate the demand for the last year...' using 465 and 408.

2. Formula
   Observation

   Handwritten line: $\rightarrow 0.4(465) + 0.6(408)$.

3. Animation
   Observation

   Calculator shows `$0.4 * 465 + 0.6 * 408$` resulting in `430.8`; table entry written as `431`.

1. Rounding from 430.8 to 431 implies rounding to the nearest integer.

1. Expression

   $$F_7 = 0.4(A_6) + 0.6(F_6)$$
   Explanation

   Apply formula for the final period.

   Justification

   Method definition.

   Shown in the video
2. Expression

   $$F_7 = 0.4(465) + 0.6(408)$$
   Explanation

   Use $A_6=465$ and $F_6=408$.

   Justification

   Table and previous calculation.

   Shown in the video
3. Expression

   $$F_7 = 186 + 244.8 = 430.8$$
   Explanation

   Perform multiplication and addition.

   Justification

   Arithmetic.

   Derived from the video
4. Expression

   $$F_7 \approx 431$$
   Explanation

   Round the result for the table entry.

   Justification

   Presenter writes 431.

   Shown in the video

The forecast for Year 7 is recorded as 431.

1. Diagram
   Observation

   Table shows Years 1-6 with Actual Demands 310, 365, 395, 415, 450, 465. Year 7 is blank.

2. Diagram
   Observation

   Text box states smoothing constant (alpha) equal to .4.

3. Audio
   Observation

   Speaker explains setting up the calculation for year 7 using the second formula.

1. The video cuts off before the final numerical answer for Year 7 is calculated. It only sets up the equation $F_7 = 0.4(465) + (1-0.4)F_6$.

Calculate the forecast demand for Year 7 given 6 years of actual demand data and a smoothing constant alpha = 0.4.

1. Actual demands for Years 1-6: 310, 365, 395, 415, 450, 465.

2. Smoothing constant $\alpha = 0.4$.

3. Initial condition: No forecast for Year 1 is provided, requiring an assumption (not fully detailed in clip).

Find $F_7$ (Forecast Demand for Year 7).

1. Expression

   $$F_{t+1} = \alpha A_t + (1 - \alpha)F_t$$
   Explanation

   Select the preferred weighted average formula.

   Justification

   Speaker states preference for this simpler method.

   Shown in the video
2. Expression

   $$F_7 = 0.4(A_6) + (1 - 0.4)F_6$$
   Explanation

   Substitute $t=6$ into the formula.

   Justification

   Algebraic substitution.

   Shown in the video
3. Expression

   $$F_7 = 0.4(465) + 0.6(F_6)$$
   Explanation

   Substitute known values $A_6 = 465$ and alpha = 0.4.

   Justification

   Data from table and problem statement.

   Shown in the video

Incomplete in video. The setup $F_7 = 0.4(465) + 0.6(F_6)$ is shown, but $F_6$ is not yet calculated.

Cannot verify final result as the video ends before calculation completion.

1. Diagram
   Observation

   A table titled Exponential Smoothing Method lists Year, Actual Demand ($A_t$), and Forecast Demand ($F_t$).

2. Diagram
   Observation

   Visible actual demand entries are 310, 365, 395, 415, 450, and 465 for Years 1 through 6.

3. Audio
   Observation

   The instructor works through forecasts for successive years using these table values.

1. Year 7 has no actual demand shown in the visible table.

2. The clip ends before the Year 5 forecast is completed.

Given annual actual demand values and $\alpha=0.4$, compute successive forecast demands using exponential smoothing, starting by setting the first forecast equal to the first actual demand.

1. Actual demand values shown in the table: Year $1 = 310$, Year $2 = 365$, Year $3 = 395$, Year $4 = 415$, Year $5 = 450$, Year $6 = 465$.

2. Smoothing constant $\alpha=0.4$.

3. Complementary weight $1-\alpha=0.6$.

4. Initialization rule: $F_1=A_1=310$.

Fill the Forecast Demand column recursively for the next periods using the exponential smoothing formula.

1. Expression

   $$F_1=310$$
   Explanation

   Place the first actual demand directly into the forecast column as the starting forecast.

   Justification

   Stated initialization rule in the audio and shown by writing 310 in the Year 1 forecast cell.

   Shown in the video
2. Expression

   $$F_2=0.4(310)+0.6(310)=310$$
   Explanation

   Compute the Year 2 forecast from Year 1 actual demand and Year 1 forecast.

   Justification

   Direct substitution into the recurrence, confirmed by calculator display and table entry.

   Shown in the video
3. Expression

   $$F_3=0.4(365)+0.6(310)=332$$
   Explanation

   Compute the Year 3 forecast from Year 2 actual demand and Year 2 forecast.

   Justification

   Direct substitution into the recurrence, confirmed by calculator display and table entry.

   Shown in the video
4. Expression

   $$F_4=0.4(395)+0.6(332)=357.2\approx 357$$
   Explanation

   Compute the Year 4 forecast from Year 3 actual demand and Year 3 forecast, then round the result for the table.

   Justification

   Calculator shows 357.2; the instructor says to round it off and writes 357.

   Shown in the video
5. Expression

   $$F_5=0.4(415)+0.6(357)$$
   Explanation

   Begin the Year 5 forecast using Year 4 actual demand and the rounded Year 4 forecast.

   Justification

   The instructor explicitly starts this next calculation, but the clip ends before the result is completed.

   Shown in the video

Within this clip, the completed forecast entries are $F_1=310$, $F_2=310$, $F_3=332$, and $F_4=357$ after rounding; the Year 5 calculation is only set up, not finished.

Each completed value is checked against the visible calculator output and the handwritten entries in the Forecast Demand column.

1. Diagram
   Observation

   Complete table with Years 1-7, Actual Demand, and Forecast Demand columns filled out progressively.

2. Audio
   Observation

   Narrator walks through each row calculation sequentially.

Given a time series of actual demands for years 1-6 and an initial forecast, compute forecasts for years 2-7 using $\alpha=0.4$.

1. $\alpha = 0.4$

2. $1-\alpha = 0.6$

3. Initial Forecast $F_1 = 310$

4. Actuals: $A_1=310, A_2=365, A_3=395, A_4=415, A_5=450, A_6=465$

Fill in the Forecast Demand column for all available periods.

1. Expression

   $$F_2 = 0.4(310) + 0.6(310) = 310$$
   Explanation

   Calculate Year 2 forecast using Year 1 data.

   Justification

   Formula application.

   Shown in the video
2. Expression

   $$F_3 = 0.4(365) + 0.6(310) = 332$$
   Explanation

   Calculate Year 3 forecast using Year 2 data.

   Justification

   Formula application.

   Shown in the video
3. Expression

   $$F_4 = 0.4(395) + 0.6(332) = 357.2 \rightarrow 357$$
   Explanation

   Calculate Year 4 forecast, rounding down/truncating.

   Justification

   Formula application and observed table entry.

   Shown in the video
4. Expression

   $$F_5 = 0.4(415) + 0.6(357) = 380.2 \rightarrow 380$$
   Explanation

   Calculate Year 5 forecast.

   Justification

   Formula application.

   Shown in the video
5. Expression

   $$F_6 = 0.4(450) + 0.6(380) = 408$$
   Explanation

   Calculate Year 6 forecast.

   Justification

   Formula application.

   Shown in the video
6. Expression

   $$F_7 = 0.4(465) + 0.6(408) = 430.8 \rightarrow 431$$
   Explanation

   Calculate Year 7 forecast.

   Justification

   Formula application.

   Shown in the video

Forecasts: $F2=310$, $F3=332$, $F4=357$, $F5=380$, $F6=408$, $F7=431$.

Results match the handwritten entries in the table and calculator outputs shown on screen.

1. Diagram
   Observation

   A table with columns 'Year', 'Actual Demand ($A_t$)', and 'Forecast Demand ($F_t$)' is displayed.

1. Table with 7 rows and 3 columns.

2. Red handwritten annotations.

1. Red bracket drawn around Years 1-6.

2. Question marks written in Year 7 Forecast cell.

3. Handwritten notes on stability/responsiveness appear and disappear.

4. Formulas appear on the right side.

5. Red handwriting calculates $F_7$ setup.

1. The core data values in the Actual Demand column remain unchanged.

The visual aids structure the problem data and highlight the target variable (Year 7 forecast) and the methodological choices (formulas).

1. Diagram
   Observation

   The slide title reads Exponential Smoothing Method.

2. Diagram
   Observation

   A three-column table is on the left and the formula is on the right.

3. Diagram
   Observation

   Red handwritten marks, arrows, circles, and numbers are added progressively over the static slide.

1. Title text: Exponential Smoothing Method.

2. Table with columns Year, Actual Demand ($A_t$), Forecast Demand ($F_t$).

3. Printed formula $F_{t+1}=\alpha A_t+(1-\alpha)F_t$.

4. Handwritten parameter notes for $\alpha$ and $1-\alpha$.

5. Red arrows and circles linking table values to the formula.

6. Calculator pop-up windows during arithmetic checks.

1. The Year 1 forecast cell is filled with 310.

2. Handwritten notes add $\alpha=0.4$ and ($1-\alpha$)=0.6.

3. Successive substitution lines are written beside the table for Years 2, 3, and 4.

4. Calculator windows appear during the arithmetic for Years 2, 3, and 4.

5. Forecast values 310, 332, and 357 are written into the table rows for Years 2, 3, and 4.

6. The setup for Year 5 begins but remains incomplete by the end.

1. The printed formula stays visible throughout.

2. The actual demand column values remain unchanged.

3. The smoothing weights remain 0.4 and 0.6 throughout the worked steps.

The visual structure shows a recursive forecasting procedure: each new forecast is built from the previous row’s actual demand and previous forecast, with annotations tracing exactly which table entries feed each substitution.

1. Diagram
   Observation

   A Windows scientific calculator appears during the Year 2 calculation and shows $0.4 \times 310 + 0.6 \times 310 = 310$.

2. Diagram
   Observation

   The calculator reappears during the Year 3 calculation and shows $0.4 \times 365 + 0.6 \times 310 = 332$.

3. Diagram
   Observation

   The calculator reappears during the Year 4 calculation and shows $0.4 \times 395 + 0.6 \times 332 = 357.2$.

1. Scientific calculator window.

2. Typed arithmetic expressions matching the handwritten substitutions.

3. Displayed results 310, 332, and 357.2.

1. The calculator opens for each major numeric step.

2. Each displayed expression corresponds to the current period’s substitution.

3. Results are then transferred into the forecast column.

1. The calculator always evaluates the same weighted-sum structure.

2. The coefficients remain 0.4 and 0.6 in every displayed check.

These pop-ups function as explicit numerical verification of the recursive formula rather than introducing a new method.

1. Animation
   Observation

   A Windows Calculator window appears on the right side of the screen, displaying arithmetic operations corresponding to the spoken numbers.

1. Calculator Window

2. Numeric Keypad

3. Display Screen

1. Window appears when calculation starts.

2. Numbers are typed sequentially matching the formula terms.

3. Result is displayed before being written to the table.

4. Window disappears between major steps.

1. The underlying slide content remains visible behind the calculator.

2. The formula $\alpha=0.4$ stays constant.

Visual aid demonstrating the exact arithmetic being performed for each recursive step.

1. Animation
   Observation

   Yellow highlighter marks specific numbers in the handwritten equations (e.g., 0.4, 0.6, 310, 365) as they are discussed.

1. Handwritten Equations

2. Yellow Highlighter Stroke

1. Specific coefficients and variables are highlighted in sync with narration.

2. Highlights persist after the term is mentioned.

1. The structure of the equation does not change, only emphasis shifts.

Directs viewer attention to the specific components of the formula being substituted in the current step.

1. Audio
   Observation

   Speaker claims stability comes from weighting actual demand more, and responsiveness from weighting forecasted demand more.

1. This contradicts standard statistical definitions where high alpha (weight on actual) increases responsiveness to recent changes, and low alpha (weight on forecast/history) increases stability/smoothing.

Assigning more weight to actual demand creates stability, while assigning more weight to forecasted demand creates responsiveness.

In standard exponential smoothing, a higher alpha (more weight on actual demand $A_t$) makes the forecast more responsive to recent changes. A lower alpha (more weight on previous forecast $F_t$) makes the forecast more stable by smoothing out short-term fluctuations.

1. Audio
   Observation

   The instructor distinguishes the first period from later ones by saying the actual demand can be posted straight away as the forecast, and only starting from period two can the formula be taken.

One might think the formula itself determines the first forecast from scratch.

The video instead initializes $F_1$ directly from $A_1$, and only then applies the recurrence from period 2 onward.

1. Audio
   Observation

   The instructor says the result is 357.2 and then says, "so we can round it off 357".

2. Diagram
   Observation

   The calculator shows 357.2 while the table entry written afterward is 357.

1. The clip does not state a general rounding policy for all future steps beyond this example.

One might assume the number written in the table is always the exact unrounded result.

For Year 4, the exact computed value shown is 357.2, but the instructor records the rounded value 357 in the table.

1. Formula
   Observation

   Calculator shows 380.2 but table has 380. Calculator shows 430.8 but table has 431.

2. Audio
   Observation

   Speaker says 'This will give us a value of 380.2... 380' and later 'result of 430.8... 431'.

1. The video does not explicitly state a rounding rule (e.g., 'always round to nearest'), though the examples imply standard rounding or truncation depending on the specific number.

Viewers might assume exact decimal answers are required or that there is a complex rounding rule.

The presenter rounds to the nearest whole number for the table entries (380.2->380, 430.8->431), treating demand as discrete units.

1. Formula
   Observation

   Two formulas presented with 'or'.

The error correction form and the weighted average form are algebraically identical.

1. Audio
   Observation

   Speaker introduces the concept then shows formulas to calculate it.

The formulas are the mathematical implementation of the exponential smoothing definition.

1. Audio
   Observation

   The instructor first explains the initialization and then says the formula is used starting from period two onwards.

The recursive forecasting process depends on having an initial forecast value before later periods can be computed.

1. Formula
   Observation

   All worked substitutions follow the printed formula $F_{t+1}=\alpha A_t+(1-\alpha)F_t$.

2. Audio
   Observation

   Each spoken calculation restates the same weighted combination with different period values.

The worked example is a direct application of the exponential smoothing formula to a demand table.

1. Diagram
   Observation

   The handwritten values $\alpha=0.4$ and $(1-\alpha)=0.6$ are placed next to the printed formula.

2. Audio
   Observation

   The instructor computes these values before substituting them into the formula.

The example instantiates the general formula with the specific smoothing constant $\alpha=0.4$ and complementary weight $0.6$.

1. Audio
   Observation

   Speaker uses the result of the previous calculation (e.g., 380) as the input for the next ($F_5$ becomes part of calculating $F_6$).

2. Formula
   Observation

   $F_{t+1}$ depends on $F_t$.

The output of one period's calculation serves as the input ($F_t$) for the subsequent period's calculation, creating a recursive chain.

1. Audio
   Observation

   Speaker walks through choosing a formula and substituting values.

1. Exponential Smoothing Formula 1 (Error Correction Form)
2. Exponential Smoothing Formula 2 (Weighted Average Form)
3. Setting up Exponential Smoothing Calculation

1. Diagram
   Observation

   Text box defining alpha.

1. Definition of Exponential Smoothing
2. $\alpha$

1. Audio
   Observation

   Speaker discusses trade-offs.

1. Trade-off between Stability and Responsiveness
2. Misunderstanding of Stability vs. Responsiveness

1. Formula
   Observation

   The printed formula is visible throughout the clip.

1. Exponential smoothing forecast formula

1. Audio
   Observation

   The instructor explains that the first actual demand is posted directly as the forecast and that the formula is used from period two onward.

1. Initialization by setting the first forecast equal to the first actual demand
2. Recursive use begins from period 2 onward

1. Formula
   Observation

   The handwritten substitution is $0.4(310)+0.6(310)$.

2. Diagram
   Observation

   The calculator displays 310 as the result.

1. Exponential smoothing forecast formula
2. Derivation of the Year 2 forecast

1. Formula
   Observation

   The handwritten substitution is $0.4(365)+0.6(310)$.

2. Diagram
   Observation

   The calculator shows 332.

1. Derivation of the Year 3 forecast

1. Audio
   Observation

   The instructor says the result is 357.2 and then rounds it to 357.

2. Diagram
   Observation

   The calculator shows 357.2 while the table entry is 357.

1. Derivation of the Year 4 forecast
2. Rounded table entries may differ from the raw computed value

1. Audio
   Observation

   The instructor identifies alpha as 0.4 and computes 1 minus alpha as 0.6.

2. Diagram
   Observation

   Handwritten notes show these values beside the formula.

1. Parameter values used in the worked example
2. Exponential smoothing forecast formula

1. Audio
   Observation

   Full walkthrough of calculations.

1. Exponential Smoothing Formula
2. Year 5 Forecast Calculation
3. Year 6 Forecast Calculation
4. Year 7 Forecast Calculation

1. Formula
   Observation

   Handwritten notes defining alpha.

1. Smoothing Constant Assignment
2. Exponential Smoothing Formula