Is the show worth carrying on with
Your data
The rating can be on any scale you like, as long as it is the same one all the way down. Anything before the last two numbers is treated as the name of the season.
Results
Hours still ahead of you if you carry on
—
| Episodes still ahead of you | — |
| The rating you would be watching for those hours | — |
| Hours you have already spent on it | — |
| The rating those hours were worth | — |
| Average taken season by season | — |
| Average taken episode by episode | — |
What each season costs you in hours
| Season | Rating | Episodes | Hours |
|---|
The two averages are different numbers and people quote whichever one suits them. Taking it season by season gives every season one vote, however long it was; taking it episode by episode gives every episode one vote. They only agree when all the seasons are the same length, and they drift apart most when the weak seasons are the long ones, which is the common case.
And the question is not really whether the rating fell. It is whether the hours still ahead of you are worth the rating still ahead of you, which is why this page puts the two side by side. A weaker season that is six episodes long asks much less of you than a slightly better one that runs to twenty two.
Why are there two different averages?
Because there are two different questions. Averaging season by season gives every season one vote no matter how long it ran; averaging episode by episode gives every episode one vote. With the seasons on this page the first comes to 7.84 and the second to 7.62.
They agree only when every season is the same length, and they drift furthest apart when the weak seasons are the long ones, which is the usual shape of a show that overstayed. People quote whichever of the two suits the argument they are already making.
How does the page decide whether it is really declining?
By fitting a straight line through the season ratings and comparing the change that line predicts across the whole run against how far the seasons scatter around it. A slide only counts as a direction when it is bigger than the bouncing around it.
The obvious test, which is to take the peak, subtract the latest season and compare that with the spread, gets it backwards. On a show that merely wobbles between 7.7 and 8.2 that test shouts decline, and on a show climbing steadily from 6.8 to 8.4 it says nothing is happening. The line through all the seasons gets both right.
Is that a proper statistical test?
No, and the page does not pretend otherwise. With five or six seasons there is not enough to infer anything in the formal sense, and any page that hands you a confidence level from that many points is selling certainty it does not have.
It is a comparison between the trend and the noise around the trend, which is still a great deal more than looking at the most recent season and feeling disappointed.
So should I keep watching or not?
That is not arithmetic, and the page will not pretend it is. What it can do is put the price next to the purchase: the hours you would still be spending, and the rating you would be spending them on.
That framing changes answers more often than the rating alone does. A weaker season six episodes long asks very little of you, while a slightly better one that runs to twenty two asks for a working week.
Reactions
0
0 Comments
Be the first to comment