4 ms·

You should be a politician! Incredible work. Log scale is absolutely not the right measure, actually didn’t even see that nifty trick because I would never hav

You should be a politician! Incredible work. Log scale is absolutely not the right measure, actually didn’t even see that nifty trick because I would never have expected it used! No wonder your graph looked totally wrong. Log scale would work if they were staying constant so we could compare on long time scales, but that’s so far from the case. If you can’t admit that, we can’t have a discussion. You really should know that to be talking about this stuff. These are close values that are fluctuating too much. The swings over the last 10 years is relevant, in absolute values. I can write the numbers out for you, but the graphs are actually less refutable. You insinuate they aren’t, but they are they’re than numbers for this use: comparing relative performance. We want to see how they all are doing, in a straightforward comparison. No, my point is you’re choosing your window very selectively and distorting the numbers to fit your narrative, and when I take the window away your point doesn’t stand, as it doesn’t

mlyle 6y ago
When one wants to compare growth rates between things with different starting points, one uses a logarithmic scale. Things growing at the same rate are parallel then. I compared things on a 30 year, 20 year, and 10 year timescale. On each of them, Germany significantly outperformed Japan.
- nwienert 6y ago
  The premise and conclusion are wrong. You want to compare the whole picture not just growth form your arbitrary start at the peak of a mountain (your false framing). And you can just adjust them to the same Y axis, no need for log. We care about the marginal dollar. You’re doing a lot of work avoiding looking at the graph I posted in earnest. It’s the right graph, and again you can’t choose an outlier.
  - mlyle 6y ago
    I strongly disagree it's the right graph. I wonder if you're just trolling at this point. Comparing growth rates of countries by looking at absolute values of dollars from countries of different sizes is madness. I've given you 5 different framings at this point: * Real GDP PPP % change over 30, 20, and 10 years to the most recent fully reported year. * Share of high-income country GDP over last 20 years * Per capita PPP GDP over 30 years, where it slid from parity with Germany in 2005 to about 80% of Germany's. But they're all bad in favor of looking at a linear scale graph, where starting from a higher value in the green series obscures that the percentage of growth is low compared to most of the members of the set. If you're determined to believe a specific conclusion, there's not much point in discussion.
    - nwienert 6y ago
      The second graph I posted is GDP PPP, please tell me how the light blue line is an outlier.

mlyle 6y ago

When one wants to compare growth rates between things with different starting points, one uses a logarithmic scale. Things growing at the same rate are parallel then. I compared things on a 30 year, 20 year, and 10 year timescale. On each of them, Germany significantly outperformed Japan.

nwienert 6y ago
The premise and conclusion are wrong. You want to compare the whole picture not just growth form your arbitrary start at the peak of a mountain (your false framing). And you can just adjust them to the same Y axis, no need for log. We care about the marginal dollar. You’re doing a lot of work avoiding looking at the graph I posted in earnest. It’s the right graph, and again you can’t choose an outlier.
- mlyle 6y ago
  I strongly disagree it's the right graph. I wonder if you're just trolling at this point. Comparing growth rates of countries by looking at absolute values of dollars from countries of different sizes is madness. I've given you 5 different framings at this point: * Real GDP PPP % change over 30, 20, and 10 years to the most recent fully reported year. * Share of high-income country GDP over last 20 years * Per capita PPP GDP over 30 years, where it slid from parity with Germany in 2005 to about 80% of Germany's. But they're all bad in favor of looking at a linear scale graph, where starting from a higher value in the green series obscures that the percentage of growth is low compared to most of the members of the set. If you're determined to believe a specific conclusion, there's not much point in discussion.
  - nwienert 6y ago
    The second graph I posted is GDP PPP, please tell me how the light blue line is an outlier.