Sir, am I correct in understanding that the trend line calculated using moving averages is not a linear trend and that is why a line of best fit must be estimated?
Given that they are all variations about the average then automatically they must add up to zero (some are below average and some are above average). They cannot possibly all be positive!!!
In the multiplicative model, I understood where you take the total of average figures for each quarter. But where did 400.00 come from?
In additive model, to make the total of seasonal variations “zero” we deducted the same figure,isn’t it? And then took the average so as to distribute among the 4 quarters.
Each season is a bit more or a bit less than 100% – if there was no seasonality then they would all equal 100%. So since there are 4 of them, they should add up to 4 x 100 = 400%
It was quarter 4, not quarter 3. I divided 0.69 by 4, which strictly means subtracting 0.1725 from each quarter. However to stick to 2 decimal places I subtracted 0.17 from 3 of them and 0.18 from the other.
Sir i really do not understand the averaging part, i am trying so hard to see what point am i missing but i do not, why do not we find the average of each year individually instead for example we are taking the second quarter of the first year and the first quarter of second year and finding their average and so on… Also the moving average part.
why didnt we find the next 4 moving average of the moving averages to find the Trend but rather find the average of 2 moving averages. If the trend find is to be plotted on a graph, the line is not perfectly straight, so can one continue to average the and if yes will it be average of 2 of the trend figures or what?. I realised that if the sales is averaged on every 4 sales basis throughout until two point are gotten, the line will be straight. Can it be done practically and academically?
The 4 quarter moving average was to get the average sales per quarter. The only reason for centering it is because the middle quarter did not coincide with any one actual quarter. After centering it, we can then compare the actual sales in a quarter with the ‘average’ sales in the quarter.
There would be no point in keep averaging. The only point is to average out the seasonality. In practice, the trend may be a smooth pattern (either linear or curved) and would be easier then to forecast than the actual sales which keep going up and down because of the seasonality.
You only need to centre the average if it is an even number of periods. If it is an odd number of periods then the moving average already corresponds to one of the periods.
haddock says
Sir, am I correct in understanding that the trend line calculated using moving averages is not a linear trend and that is why a line of best fit must be estimated?
John Moffat says
Correct 🙂
haddock says
Sir, why must the average seasonal variations cancel each other out? (add up to 0) What if they were all positive variations?
John Moffat says
Given that they are all variations about the average then automatically they must add up to zero (some are below average and some are above average). They cannot possibly all be positive!!!
haddock says
Thank you. I find your lectures very clear and informative
John Moffat says
You are welcome 🙂
msk29 says
In the multiplicative model, I understood where you take the total of average figures for each quarter. But where did 400.00 come from?
In additive model, to make the total of seasonal variations “zero” we deducted the same figure,isn’t it? And then took the average so as to distribute among the 4 quarters.
Kindly explain me about the 400 part.
Thank you.
John Moffat says
Each season is a bit more or a bit less than 100% – if there was no seasonality then they would all equal 100%. So since there are 4 of them, they should add up to 4 x 100 = 400%
mango1991 says
Sir will they be telling us in the exam to calculate the trend based on how many periods/ months/ seasons moving average ?
For example – will they be mentioning like this – ‘calculate the trend line based on a three month moving average’
If not, how will we know how many periods to take ?
John Moffat says
It will be made clear in the exam 🙂
etiamm says
Hi,
Just wondering why is the average of 84.75 right in the middle of Q2 and Q3 ? could it not be between Q1 and Q2 also ?
Thanks
John Moffat says
No. We averaged Q’s 1 to 4, and the ‘middle’ of those 4 is effectively the ‘middle’ of Q’s 2 & 3.
etiamm says
Perfect thanks
John Moffat says
You are welcome 🙂
malikabdullah0666066 says
hello,
in quater 3 (time 24:31) why you took 0.18 and how?
John Moffat says
It was quarter 4, not quarter 3. I divided 0.69 by 4, which strictly means subtracting 0.1725 from each quarter. However to stick to 2 decimal places I subtracted 0.17 from 3 of them and 0.18 from the other.
Joseph says
Multiplicative Q3 Table .. it was 95.35% , in the table u wrote 91.35 and solve it for 91.35%
im really benefiting from ur lectures but this one is a bit too long , would be better if u hinted out at the question areas
Thanks alot
akhalid93 says
Mr. John, the lecture was delivered very well.
I want to clear my doubt why you have shown the final result zero in the additive model?
What is the advantage of using Multiplicative model over the additive model and how we would come to know which methods is feasible to use?
Thank you
keshav11 says
very thankful to you sir.. now i am able able to understand it well
John Moffat says
Thank you for the comment 🙂
keshav11 says
welcome 🙂
mubezi1 says
great thanks sir i really understand now
John Moffat says
I am pleased 🙂
winters says
Sir i really do not understand the averaging part, i am trying so hard to see what point am i missing but i do not, why do not we find the average of each year individually instead for example we are taking the second quarter of the first year and the first quarter of second year and finding their average and so on…
Also the moving average part.
lordlawrence79 says
now I am getting there thank you Lecturer
tijani says
why didnt we find the next 4 moving average of the moving averages to find the Trend but rather find the average of 2 moving averages.
If the trend find is to be plotted on a graph, the line is not perfectly straight, so can one continue to average the and if yes will it be average of 2 of the trend figures or what?. I realised that if the sales is averaged on every 4 sales basis throughout until two point are gotten, the line will be straight. Can it be done practically and academically?
John Moffat says
The 4 quarter moving average was to get the average sales per quarter.
The only reason for centering it is because the middle quarter did not coincide with any one actual quarter. After centering it, we can then compare the actual sales in a quarter with the ‘average’ sales in the quarter.
There would be no point in keep averaging. The only point is to average out the seasonality. In practice, the trend may be a smooth pattern (either linear or curved) and would be easier then to forecast than the actual sales which keep going up and down because of the seasonality.
kittygr says
This Q&A also solved my problem, thanks a lot.
John Moffat says
You are welcome 🙂
issack says
thank you sir,i realy loved your lecture.very clear and easy to understand,
Mohammad Ibrahim says
Also why are the first two left out?
John Moffat says
Because there is no average to compare them to.
(They are not left out in calculating the moving averages)
Mohammad Ibrahim says
How do know that the centered average is for a certain month in case of even or odd number of total periods?
John Moffat says
You only need to centre the average if it is an even number of periods. If it is an odd number of periods then the moving average already corresponds to one of the periods.
M. Osman Kamran says
I am really thanksfull for Opentition.com site for real free accounting resources.
mvgino says
very helpful!
hassita says
Great!
zaidh11 says
Impressive!