Hacker News new | ask | show | jobs
by lmilcin 1879 days ago
> More than 3,000 boxes dropped into the sea last year, and more than 1,000 have fallen overboard so far in 2021.

Umm... we have end of April (the article is dated April 29th) or 1/3rd of the year. The last time I checked a third of 3000 is around 1000.

Anyway, nowhere in the article I see any numbers to support that there is a "sudden rise in accidents". Even the graphs themselves seem to show entirely different picture -- that this might just be a fluke.
2 comments
jlund-molfese 1879 days ago
But the graph only shows YTD losses, right? If you make the 2021 graph 3 times larger to compensate for the roughly 1/3rd of the year we've seen so far[1], it looks like more than a fluke than just 2020 alone. Both 2020 and 2021 should have more cargo loss than any year since 2013.

[1] https://i.imgur.com/k5hWs8u.jpg

link
boomboomsubban 1879 days ago
There's no reason to suspect that we'll see two more 250+ crate loss accidents this year leading to results like your graph predicts.
link
nieve 1879 days ago
Nor is there any reason to supposed that the increased accidents are pure flukes that can't possibly happen again this year. If there's increased _systemic_ risk due to risky sailing, bad cargo management, etc. it'd be understandable for more outliers to occur. I don't think we can handwave away that somehow we're done with major accidents for the year only a third of the way in.
link
boomboomsubban 1879 days ago
>I don't think we can handwave away that somehow we're done with major accidents for the year only a third of the way in.

These kinds of events are rare, even if something is causing more systemic risk there's no way to predict how many more will happen this year. I'm not handwaving anything away, I'm saying projections aren't useful.

link
boomboomsubban 1879 days ago
> Even the graphs themselves seem to show entirely different picture

The graph for 2020 is roughly double what you'd see in every year except 2013, and 2021 has already matched most other years.

I think they're trying to make too much of a trend out of years dominated by one major accident, but the data definitely shows two worse than usual years in a row.

link