4 ms·

If one football team got beat 8-0, and another football team got beat 9-1, should we really be spending time arguing about which team is better when they're bot

If one football team got beat 8-0, and another football team got beat 9-1, should we really be spending time arguing about which team is better when they're both pretty obviously not very good?

SpicyLemonZest 6y ago
It's not about teams, it's about attempting to evaluate the counterfactual. If the difference between the US and Australia is primarily due to lockdowns, you would expect to see areas with more lockdown do better than areas with less. Much better, to a degree that drowns out any noise, since the overall difference is so vast. And that really doesn't seem to be what we observe.
- anoonmoose 6y ago
  You're acting like there were parts of the US in 100% lockdown vs 0% lockdown, when the difference is a lot more like 20% (max) vs 10%, and it's a lot harder to suss out any difference in result when the difference in practice is that small.
  - SpicyLemonZest 6y ago
    If lockdown explains most of the variance between Australia's 1k deaths and America's 500k deaths, then the difference between 10% and 20% lockdown should still be very visible.
    - mcguire 6y ago
      Few relationships in the real world are linear. That is just an assumption that makes the math easier.
      - SpicyLemonZest 6y ago
        Sure, but there's a point where the hypothesis just becomes unfalsifiable. Is there any evidence that would convince you lockdown policy isn't the primary factor here?
    - anoonmoose 6y ago
      That's a big "if" at the beginning of your sentence.
      - SpicyLemonZest 6y ago
        I'm not sure what you mean. That was the "if" that started this comment thread - an assertion that the US would be back to normal already if only it had done what Australia did.
    - zuminator 6y ago
      That doesn't follow. Imagine the effective spread ("Re") of a disease averages 1.5. Certain external factors, such as climate, population density, access to medicine, diet, etc. can cause a 0.2 plus or minus variance in Re. So, 1.3-1.7 range. Each 10% of lockdown will cause a 10% reduction in that amount. So if Florida does a 10% lockdown, its Re will be somewhere in the 1.3-1.7 range reduced by 10% or 1.17-1.53. If California does a 20% lockdown, its Re will be in the 1.04-1.36 range. California, if "unlucky" can still have a worse Re (1.36) than Florida if "lucky" (1.17). Meanwhile Australia at a say 70% lockdown would see numbers in the 0.39-0.51 range, well below sustained transmission levels even if "unlucky."
      - SpicyLemonZest 6y ago
        Makes sense. Thanks for doing out the math, that's pretty convincing to me.
mod 6y ago
No, we shouldn't argue about which one is better. If we're coaching the team that lost 0-8, we shouldn't try to run the playbook of the team that lost 1-9 in the next game. Especially when that playbook cripples our players in other ways--perhaps academically (to expand your analogy).

If one football team got beat 8-0, and another football team got beat 9-1, should we really be spending time arguing about which team is better when they're both pretty obviously not very good?

SpicyLemonZest 6y ago

It's not about teams, it's about attempting to evaluate the counterfactual. If the difference between the US and Australia is primarily due to lockdowns, you would expect to see areas with more lockdown do better than areas with less. Much better, to a degree that drowns out any noise, since the overall difference is so vast. And that really doesn't seem to be what we observe.

anoonmoose 6y ago
You're acting like there were parts of the US in 100% lockdown vs 0% lockdown, when the difference is a lot more like 20% (max) vs 10%, and it's a lot harder to suss out any difference in result when the difference in practice is that small.
- SpicyLemonZest 6y ago
  If lockdown explains most of the variance between Australia's 1k deaths and America's 500k deaths, then the difference between 10% and 20% lockdown should still be very visible.
  - mcguire 6y ago
    Few relationships in the real world are linear. That is just an assumption that makes the math easier.
    - SpicyLemonZest 6y ago
      Sure, but there's a point where the hypothesis just becomes unfalsifiable. Is there any evidence that would convince you lockdown policy isn't the primary factor here?
  - anoonmoose 6y ago
    That's a big "if" at the beginning of your sentence.
    - SpicyLemonZest 6y ago
      I'm not sure what you mean. That was the "if" that started this comment thread - an assertion that the US would be back to normal already if only it had done what Australia did.
  - zuminator 6y ago
    That doesn't follow. Imagine the effective spread ("Re") of a disease averages 1.5. Certain external factors, such as climate, population density, access to medicine, diet, etc. can cause a 0.2 plus or minus variance in Re. So, 1.3-1.7 range. Each 10% of lockdown will cause a 10% reduction in that amount. So if Florida does a 10% lockdown, its Re will be somewhere in the 1.3-1.7 range reduced by 10% or 1.17-1.53. If California does a 20% lockdown, its Re will be in the 1.04-1.36 range. California, if "unlucky" can still have a worse Re (1.36) than Florida if "lucky" (1.17). Meanwhile Australia at a say 70% lockdown would see numbers in the 0.39-0.51 range, well below sustained transmission levels even if "unlucky."
    - SpicyLemonZest 6y ago
      Makes sense. Thanks for doing out the math, that's pretty convincing to me.

mod 6y ago

No, we shouldn't argue about which one is better. If we're coaching the team that lost 0-8, we shouldn't try to run the playbook of the team that lost 1-9 in the next game. Especially when that playbook cripples our players in other ways--perhaps academically (to expand your analogy).