4 ms·

The median is like 20% in the chart. Not exactly a good look.

jvanderbot 3y ago
The median of the the top 25 most violent tribes is absolutely going to have a very high level of violence. You're right! But if there are ten million tribes not shown on this chart that have less than 0.01% chance of violent death, then the real median is much lower, right? This is why I find the chart misleading: It chooses the most violent non-state tribes as a basis. It therefore is not a valid way of concluding "How likely was violence" because it eschews states (which may have been more organized / less violent), and does not include anywhere near all tribes, and so a meaningful comparison with the rest of the population is impossible. Imagine we were discussing people. You show me the top 25 most murderous people, and say "The median murders by these people were 10!", this does not mean anything about the number of murders committed by anyone else.
- TulliusCicero 3y ago
  > top 25 most violent tribes Where'd you get this information? These are just the tribes they have data on, they didn't cherry pick it as far as I can see.
  - jvanderbot 3y ago
    The full data overview is pretty interesting: https://ourworldindata.org/ethnographic-and-archaeological-evidence-on-violent-deaths https://ourworldindata.org/ethnographic-and-archaeological-e... Reexamining the chart, you're right, it is in fact the full data analyzed. It's fair to reconsider at this point, but I think my initial reservations about the selection of the data still apply - author did not choose the 25 most deadly tribes, but he did choose only data for which violence was the subject. Ostensibly, there must be some under-representation among peoples that didn't leave mass graves, burned villages, etc. So we see a 100x increase over modern day USA murder rates among chiefdoms and anarchic tribes, which who knows how much of the population that represents. Again, I also take issue generalizing from that. For example, here's the same author estimating state deaths. https://ourworldindata.org/grapher/share-of-violent-deaths-state-societies https://ourworldindata.org/grapher/share-of-violent-deaths-s... So, how many lived in states vs tribes over time? And where are all the other states and tribes? I'd expect a power-law relationship, with the worst atrocities being highly visible, and a long tail of less-awful areas to live. The problem is did lots or a little people live in those places All I'm saying is "Was life basically nasty brutish and short for everyone?" I doubt it was all that bad, though I would never want to go back.

The median of the the top 25 most violent tribes is absolutely going to have a very high level of violence. You're right! But if there are ten million tribes not shown on this chart that have less than 0.01% chance of violent death, then the real median is much lower, right? This is why I find the chart misleading: It chooses the most violent non-state tribes as a basis. It therefore is not a valid way of concluding "How likely was violence" because it eschews states (which may have been more organized / less violent), and does not include anywhere near all tribes, and so a meaningful comparison with the rest of the population is impossible. Imagine we were discussing people. You show me the top 25 most murderous people, and say "The median murders by these people were 10!", this does not mean anything about the number of murders committed by anyone else.

TulliusCicero 3y ago
> top 25 most violent tribes Where'd you get this information? These are just the tribes they have data on, they didn't cherry pick it as far as I can see.
- jvanderbot 3y ago
  The full data overview is pretty interesting: https://ourworldindata.org/ethnographic-and-archaeological-evidence-on-violent-deaths https://ourworldindata.org/ethnographic-and-archaeological-e... Reexamining the chart, you're right, it is in fact the full data analyzed. It's fair to reconsider at this point, but I think my initial reservations about the selection of the data still apply - author did not choose the 25 most deadly tribes, but he did choose only data for which violence was the subject. Ostensibly, there must be some under-representation among peoples that didn't leave mass graves, burned villages, etc. So we see a 100x increase over modern day USA murder rates among chiefdoms and anarchic tribes, which who knows how much of the population that represents. Again, I also take issue generalizing from that. For example, here's the same author estimating state deaths. https://ourworldindata.org/grapher/share-of-violent-deaths-state-societies https://ourworldindata.org/grapher/share-of-violent-deaths-s... So, how many lived in states vs tribes over time? And where are all the other states and tribes? I'd expect a power-law relationship, with the worst atrocities being highly visible, and a long tail of less-awful areas to live. The problem is did lots or a little people live in those places All I'm saying is "Was life basically nasty brutish and short for everyone?" I doubt it was all that bad, though I would never want to go back.