4 ms·

Having all lines normalized so they start from a common point at x = 0 would be helpful. That would allow reading the total returns as percentage of the initial

Having all lines normalized so they start from a common point at x = 0 would be helpful. That would allow reading the total returns as percentage of the initial (1987) investment and compare them. Similarly in the drawdown chart, if it starts at 1987, it makes no sense to start the dollar at -81%. If I am looking at a chart from 1987 to 2022, I want to see how the dollar did compared to other assets in that time period.

nonethewiser 4y ago
Is that even possible? Wouldn't that suggest something that is not true?
- epgui 4y ago
  How is it not true? All these numbers are relative, and you can make the choice of comparing them to whatever reference point you want.
  - rm999 4y ago
    The y-axis has actual meaning: it tells you the purchasing power of a public stock, which will never be 0. What you could do is normalize it by the value of the stock at the start of the chart, which would make the charts start at 1. On a log plot this is the equivalent of dividing all the values by the starting value, which moves the lines up/down but does not change their shape. This could make it easier to compare the lines, but in doing so, you throw out information (the real value of the y axis).
    - pedrosorio 4y ago
      > The y-axis has actual meaning: it tells you the purchasing power of a public stock, which will never be 0. If we are trying to visualize the total return of a given asset since 1987, how is the price of a single stock (an arbitrary unit) in 1987 or any time since, relevant data? The ROI expressed as a percentage on the y axis (with 0% at the beginning of the period) would be a much better visualization of the relative returns among asset classes throughout the period. Currently I get USD < VBMXFX < VFINX at the beginning of the chart, for reasons that have nothing to do with the total return since 1987. The drawdown chart is even more confusing with the USD starting at -81%. If we were plotting a chart of drawdown of multiple currencies, the older ones would start lower (since they've had more time to be affected by inflation) which only makes it hard to visualize the answer to the question "how did they do since 1987".
    - epgui 4y ago
      As another commenter points out, it completely depends on what feature of the data you're trying to highlight, or what question you're trying to answer. I don't think anyone looking at this chart really cares about the inflation-adjusted value of one share in a specific year, I think the main point of this chart is the real returns of stocks, bonds and cash (or an approximation of such represented by the selected indices), over time.
ricardobayes 4y ago
I had the same thought. Having each assets start from the same value (eg. 100k USD) would really help.
kurthr 4y ago
You don't really need that on a logarithmic plot since the ratio of stock prices is always a constant vertical separation. It is very necessary on linear pricing plots.
- pedrosorio 4y ago
  Can you tell at a glance which asset had a better return since the start date until March of 2009? A plot of total ROI (with all assets starting at y = 0) would make that obvious (for any given end date). As far as I can tell, the different intercepts just add noise (price of a single stock) that is not helpful to visualize what the plot is supposed to be showing (total returns).
  - kurthr 4y ago
    To within a fraction of a percent? Probably not, but to within a reasonable fraction of total performance over 13 years... yes. Look at the initial separation and the final separation and see if the are larger or smaller or look the same. If you can't see it, use a larger monitor. I'd see the point of scaling them when looking at VFISX vs GOOG since there would be a huge range on the scale, but even there you can see the current draw downs are very similar (38% vs 28%). So you can see a 10% (<1% per year) difference in performance between things with a 250x range. Look at the link for VFISX,VFITX,VUSTX where they start with a ~100% separation (4,5.5,8) and end with very low separation (<+/-3%). In June of 2020 there's a blip where VUSTX goes up 20-25% over VFISX (about a quarter of the original separation) and VFITX is about a third of that or 7-8% over VFISX. So you can see ~3% difference in performance over 12 years or (0.25% annualized).
    - pedrosorio 4y ago
      > To within a fraction of a percent? Probably not, but to within a reasonable fraction of total performance over 13 years... yes. You didn't answer my question. Could you tell which one had larger returns from the beginning of the period until March of 2009? I couldn't easily tell, and yet VMBFX has a total return of 127% and VFINX 89%

nonethewiser 4y ago

Is that even possible? Wouldn't that suggest something that is not true?

epgui 4y ago
How is it not true? All these numbers are relative, and you can make the choice of comparing them to whatever reference point you want.
- rm999 4y ago
  The y-axis has actual meaning: it tells you the purchasing power of a public stock, which will never be 0. What you could do is normalize it by the value of the stock at the start of the chart, which would make the charts start at 1. On a log plot this is the equivalent of dividing all the values by the starting value, which moves the lines up/down but does not change their shape. This could make it easier to compare the lines, but in doing so, you throw out information (the real value of the y axis).
  - pedrosorio 4y ago
    > The y-axis has actual meaning: it tells you the purchasing power of a public stock, which will never be 0. If we are trying to visualize the total return of a given asset since 1987, how is the price of a single stock (an arbitrary unit) in 1987 or any time since, relevant data? The ROI expressed as a percentage on the y axis (with 0% at the beginning of the period) would be a much better visualization of the relative returns among asset classes throughout the period. Currently I get USD < VBMXFX < VFINX at the beginning of the chart, for reasons that have nothing to do with the total return since 1987. The drawdown chart is even more confusing with the USD starting at -81%. If we were plotting a chart of drawdown of multiple currencies, the older ones would start lower (since they've had more time to be affected by inflation) which only makes it hard to visualize the answer to the question "how did they do since 1987".
  - epgui 4y ago
    As another commenter points out, it completely depends on what feature of the data you're trying to highlight, or what question you're trying to answer. I don't think anyone looking at this chart really cares about the inflation-adjusted value of one share in a specific year, I think the main point of this chart is the real returns of stocks, bonds and cash (or an approximation of such represented by the selected indices), over time.

ricardobayes 4y ago

I had the same thought. Having each assets start from the same value (eg. 100k USD) would really help.

kurthr 4y ago

You don't really need that on a logarithmic plot since the ratio of stock prices is always a constant vertical separation. It is very necessary on linear pricing plots.

pedrosorio 4y ago
Can you tell at a glance which asset had a better return since the start date until March of 2009? A plot of total ROI (with all assets starting at y = 0) would make that obvious (for any given end date). As far as I can tell, the different intercepts just add noise (price of a single stock) that is not helpful to visualize what the plot is supposed to be showing (total returns).
- kurthr 4y ago
  To within a fraction of a percent? Probably not, but to within a reasonable fraction of total performance over 13 years... yes. Look at the initial separation and the final separation and see if the are larger or smaller or look the same. If you can't see it, use a larger monitor. I'd see the point of scaling them when looking at VFISX vs GOOG since there would be a huge range on the scale, but even there you can see the current draw downs are very similar (38% vs 28%). So you can see a 10% (<1% per year) difference in performance between things with a 250x range. Look at the link for VFISX,VFITX,VUSTX where they start with a ~100% separation (4,5.5,8) and end with very low separation (<+/-3%). In June of 2020 there's a blip where VUSTX goes up 20-25% over VFISX (about a quarter of the original separation) and VFITX is about a third of that or 7-8% over VFISX. So you can see ~3% difference in performance over 12 years or (0.25% annualized).
  - pedrosorio 4y ago
    > To within a fraction of a percent? Probably not, but to within a reasonable fraction of total performance over 13 years... yes. You didn't answer my question. Could you tell which one had larger returns from the beginning of the period until March of 2009? I couldn't easily tell, and yet VMBFX has a total return of 127% and VFINX 89%