Dear Mr, I'd like to send the greatest thank to your dedication on these useful explanations about this forecasting method. I wonder where you were for these previous years, why didnt you continue, Mr?
Was going to answer Alpay Dincer's question from 3 years ago, but maybe it's useful for others. Not offering a declarative answer here, just sharing my experience. Paraphrased: "What does it mean if solver gives values of 0 for the smoothing constants?" Solver can certainly give you a values of the smoothing constants at or near zero. You can check this by increasing the number of decimal places excel displays for those cells. However, if you allow unconstrained variables to be negative (by unchecking "Make Unconstrained Variables Non-Negative)", you may get a clearer picture. In my experience this can happen in one of four scenarios: (a) The overall trend is negative or almost non-existent, and needs to be smoothed that way; (b) There aren't enough data points to meaningfully plot, which is confusing solver (do a regression on the forecast and check the f test and R^2); (c) there is little or no seasonality in your data (Same check as in point (b)); or (d) some combination of the previous three. Point (c) can happen if a business isn't properly marketing or promoting their product to create a steadily increasing level of demand; they will be (in effect) subject to exogenous factors (whims of consumers, weather, retail partner sales, etc.). Anyone else have any insights or experience they can share?
When F17 was calculated why did you use S13 and not S17 (after you calculate S17) ? For future forecasts values the formula says F(t+1) = )Lt + l*Tt) * (St+l). and what does S(in subscript (t+l)) even stand for?
Hei. Thank you for the video. Please correct me if I'm wrong, the highlighted section in column 3 starts two cells after and ends two cells before the cells in column two because "p" is equal to 4. What if "p" is 7? How many cells after and before should be the highlighted area? Thank!
In the solver, you have mentioned only one criteria,i.e. less than equal to 1(=0). If it is less than one, it could also pick up values less than 1 i.e a negative number. Although, the criteria for Alpha, Beta and Gamma is that the values should be between 0 and 1
Great explanation but your formulas notations seem incorrect. It seems like you are forecasting the future based on future data. Is it that your t+1 refers to cell positioning as opposed to the actual period coding?
Thanks for this thorough walk through! If I wanted to forecast for week 21+, do I just drag and pull from formula I22 and adjust the changed value in the formula accordingly or is it a different process? Thanks!