I used a Works spreadsheet and entered in my scoring out of 100 for each episode of Stargate SG-1. I then depicted this as a bar graph. Obviously, there are loads of thin little bars all over the place.

In order to get a more general picture of how the series was doing, I made a column of data that takes the average of five consecutive episodes centred on the specific datapoint. The resulting line graph had many fluctuations but the extremes were smoothened out. I the did a similar thing for 9 episodes and 19 episodes. Each time, the extremes are dulled down but the lines is still eratic.

I then took a different approach. For a 9 episode trend, instead of using the episode scores directly, I took an average of the surrounding 5 averages of 5 episodes. I did a similar thing to this to get a 19 episode trend. The weird thing is that the line is certainly a smoother curve, but has a greater tendency to bounce more between extremes.

What's going on here and what is the best way to represent the series trend?

__________________
Freedom For Fission A breath of fresh Iodine-131