3 ms·

For that 2nd graph, he chose buckets that are uniquely poorly suited to evaluating his statement about whether returns tend toward 10%. It looks to me like, if

For that 2nd graph, he chose buckets that are uniquely poorly suited to evaluating his statement about whether returns tend toward 10%. It looks to me like, if he had instead made them (5%)-5%, 5%-15%, 15%-25%, etc., then the mode would indeed have been the 5-15% bracket.

ricardobeat 6y ago
The mode is already in the 10-20% bucket in the graph, changing the buckets would not affect the argument that follows, which does mention the 8-12% and 10-15% ranges: > If we look at the calendar year returns plus or minus 2% from the 10% average (so 8% to 12%) this has happened in just 5 calendar years > Just 18% of returns have been between 5% to 15% in any given year. The main point being that the odds of you seeing any returns near the 10% number are quite low (18/100). It's just a lot more spiky than most people realize and not a safe way to get 10% returns unless you're looking at 30+ year periods.
- mumblemumble 6y ago
  Yes, that's all true, but none of it changes the point that it's a poorly constructed visualization that is ill-suited to how the author is trying to frame their story. Edward Tufte would weep.
  - ricardobeat 6y ago
    The first graph establishes that there is no clear trend or clustering in the data points. The second one lets you visualize just how many years are around the 10% average or not, and then the argument is expanded from there. They seem to serve their purpose pretty well. There's probably a nicer way to display that, and maybe combine them as one of Tufte's principles would suggest, but I don't see anything that would make him weep here :) What do you have in mind?
ellisv 6y ago
Agreed. I’ve seen that figure before (or some variation) and it is a bad way to try to argue his statement.
TrueDuality 6y ago
That range is covered in the article: > Just 18% of returns have been between 5% to 15% in any given year.

ricardobeat 6y ago

The mode is already in the 10-20% bucket in the graph, changing the buckets would not affect the argument that follows, which does mention the 8-12% and 10-15% ranges: > If we look at the calendar year returns plus or minus 2% from the 10% average (so 8% to 12%) this has happened in just 5 calendar years > Just 18% of returns have been between 5% to 15% in any given year. The main point being that the odds of you seeing any returns near the 10% number are quite low (18/100). It's just a lot more spiky than most people realize and not a safe way to get 10% returns unless you're looking at 30+ year periods.

mumblemumble 6y ago
Yes, that's all true, but none of it changes the point that it's a poorly constructed visualization that is ill-suited to how the author is trying to frame their story. Edward Tufte would weep.
- ricardobeat 6y ago
  The first graph establishes that there is no clear trend or clustering in the data points. The second one lets you visualize just how many years are around the 10% average or not, and then the argument is expanded from there. They seem to serve their purpose pretty well. There's probably a nicer way to display that, and maybe combine them as one of Tufte's principles would suggest, but I don't see anything that would make him weep here :) What do you have in mind?

ellisv 6y ago

Agreed. I’ve seen that figure before (or some variation) and it is a bad way to try to argue his statement.

TrueDuality 6y ago

That range is covered in the article: > Just 18% of returns have been between 5% to 15% in any given year.