I was thinking about yesterday’s handicap history chart and it was bothering me. I should have cleaned up the x-axis and shaded the years. So I did.
I made two new columns:
Q7: =(MOD(YEAR(P7),2))
R7: =YEAR(P7)
I added a new data series using those columns and:
- Move the new data series to a secondary axis
- Changed the chart type for that series to a column chart
- Change the fill to pale yellow and the border to 25% gray
- Change the gap width to zero
- Set the secondary y-axis scale to 0-1 and removed the labels
- Removed the tick marks and data labels for the primary x-axis
- Removed the tick marks for the secondary x-axis
- Move the data labels from the secondary x-axis to Low
May I make another suggestion, remove the trend line. Looking at the data provided, there is no reason to suspect the trend. Most you can say is there is the mean changes in 2007 and 2008. There is no statistically supportable trend, especially a complex trend such as a quadratic.
I’d change the transparacy of the bars so the gridlines are visible in the even years.
What does “no statistically supportable trend” mean? How can I know when a trendline is appropriate and when it’s not?
Thufir –
I believe you are correct. Dick shows a parabolic trend, and of course his handicap must asymptotically approach zero. :roll:
… mrt
Who says? there is no reason to believe he will always get better? Don’t we all get older, or perhaps if you overlaid info on hours billed or such, you might find a trend that says when Dick is busy he can’t practice his golf and goes off form, potentially wildly of in the long term as he gets older.
Having indirectly criticised his golf I would say that there is nothing wrong with his trending. If it goes more than 5% off you might even be tempted to call it statistically significant (albeit the sample set of 1 is a bit too low in this case)
Jan – There was no seriousness there. My tongue was in my cheek when I said that. If there had been an emoticon for that, I’d have used it, instead of rolling my eyes.
… mrt