4 ms·

This isn't like trying to beat the stock market. There's no intrinsic reason why their predictions should be any good, and in fact they've been comically bad, c

This isn't like trying to beat the stock market. There's no intrinsic reason why their predictions should be any good, and in fact they've been comically bad, coming in too low every year for 21 consecutive years: https://www.pv-magazine.com/2018/10/09/iea-low-balls-solar-growth-again/ https://www.pv-magazine.com/2018/10/09/iea-low-balls-solar-g... (this is from 2018, but the same trend has continued up through 2023)

JumpCrisscross 3y ago
> no intrinsic reason why their predictions should be any good This is where the "generaly credible" bit comes in. They're good forecasters in other domains. Part of being a good forecaster is adjusting as you make errors. So we'd need a hypothesis for why they wouldn't be making those adjustments. As a trivial example, Hoekstra's chart shows the WEO serially predicting false plateaus between 2006 and 2018. Those plateaus are no longer the case for the IEA's solar predictions [1]. Again, not rejecting that this may be undershooting reality. But if your only evidence is past performance, that's likely oversimplistic. [1] https://www.iea.org/reports/renewable-energy-market-update-june-2023/executive-summary https://www.iea.org/reports/renewable-energy-market-update-j...
- scythe 3y ago
  The link you're referring to only goes to next year, whereas the infamous pessimistic graphs tried to extrapolate for around a decade. I suppose capitulation would count as an adjustment.
  - JumpCrisscross 3y ago
    > link you're referring to only goes to next year, whereas the infamous pessimistic graphs tried to extrapolate for around a decade This CleanTechnica article references the short-term energy outlook. The context is near-term forecasts. > capitulation would count as an adjustment What else would it be?
- epistasis 3y ago
  If they are specifically bad in an area, with a very strong reputation, why should we assume that specific reputation would change to match the general reputation? Also, I'm not familiar with any projections that are generally good. Collecting existing data? Sure they are good at that. But projection is an entirely different skill set and I have not ever heard somebody cite great projections from EIA.

JumpCrisscross 3y ago

> no intrinsic reason why their predictions should be any good This is where the "generaly credible" bit comes in. They're good forecasters in other domains. Part of being a good forecaster is adjusting as you make errors. So we'd need a hypothesis for why they wouldn't be making those adjustments. As a trivial example, Hoekstra's chart shows the WEO serially predicting false plateaus between 2006 and 2018. Those plateaus are no longer the case for the IEA's solar predictions [1]. Again, not rejecting that this may be undershooting reality. But if your only evidence is past performance, that's likely oversimplistic. [1] https://www.iea.org/reports/renewable-energy-market-update-june-2023/executive-summary https://www.iea.org/reports/renewable-energy-market-update-j...

scythe 3y ago
The link you're referring to only goes to next year, whereas the infamous pessimistic graphs tried to extrapolate for around a decade. I suppose capitulation would count as an adjustment.
- JumpCrisscross 3y ago
  > link you're referring to only goes to next year, whereas the infamous pessimistic graphs tried to extrapolate for around a decade This CleanTechnica article references the short-term energy outlook. The context is near-term forecasts. > capitulation would count as an adjustment What else would it be?
epistasis 3y ago
If they are specifically bad in an area, with a very strong reputation, why should we assume that specific reputation would change to match the general reputation? Also, I'm not familiar with any projections that are generally good. Collecting existing data? Sure they are good at that. But projection is an entirely different skill set and I have not ever heard somebody cite great projections from EIA.