Management Cheat Sheet

Types of Forecasts: Judgmental Time Series Associative Models
Judgmental Forecasts: Executive opinions Sales force Composite Consumer surveys Outside opinion Opinions of managers/staff Delphi technique
Time Series Forecasts Level-Long-term “base” of the data Trend- long-term upward or downward movement in data Seasonability- short-term regular variations in data at constant time intervals Cyclicity- long term variations due to economic cycle Random variations- Caused by chance. Unpredictable- not subject to modeling.
T = Index for any time period
At = Demand in period t
Ft = Forecast for period t (made at the end of period t-1)
Ft+k = Forecast for period t+k made at the end of period t
St = Estimate of the smoothed value of forecast for period t
Tt = Estimate of the trend value of forecast for period t
TAFt = Trend Adjusted Forecast for Period t made at the end of t-1)
TAFt+k=Trend Adjusted Forecast for period t+k made at the end of period t
Exponential Smoothing Ft+1 = aAt + (1 - a) Ft Forecasting next period= Forecast for the current period+ a fraction of the error for the current period
Trend Adjusted Exponential Smoothing St = Smoothed Forecast at the end of period t Tt = Trend Estimate at the end of period t St = a1 At + (1 - a1) TAFt Tt = a2 (TAFt -TAFt -1) + (1 - a2) Tt -1 TAFt+1 = St + Tt
Error - difference between actual value and predicted value
Mean forecast error: Average error
Mean absolute deviation (MAD) Average absolute error
Mean squared error (MSE) Average of squared error
Mean Absolute Percent error (MAPE) Tracking signal
Ratio of cumulative error and MAD
Time Series Forecasting Naïve (Just move the At value over 1 and down 1 to the Ft column) Moving Average Weighted Moving Average Exponential Smoothing Trend Adjusted Forecasting Moving Average N=3 (493+498+492)/3=494.33

Weighted Moving Average .2, .3,.5
(.2*493)+(.3*498)+(.5*492)=494

Exponential Smoothing