I read that book in the 1970s. Now I feel old.

It's a fantastic book, and everyone should read it. However, none of the rules are 100%.

The downside to starting every graph at zero is that it compresses the behavior of the data so if you're interested in the change in the data, then it's hard to see.

For example, if we looked at atmospheric CO2 levels on a graph that started at 0 ppm, then CO2 would look almost flat. However, we're interested in changes in CO2 level.

But what about the graph in the article? In this case, your criticism is pretty bang on, IMHO!

Starting the graph at 20 instead of 0 doesn't really help you read it better, and it makes the changes "more dramatic".

Thanks for writing, and have an excellent new year!