News for hackers. Stuff that matters.
Re:Fewer than 3% of cars sold in the U.S. have manual transm... by crispyambulance on Thursday November 17, 2016 @08:11AM (#12977108)
The key figure, I think, is number of fatalities per billion vehicle Kilometers. For the USA that is 7.1. Unfortunately, Americans drive A LOT MORE than most other countries, so raw annual fatality count is very high (~34K). By comparison UK has 3.6 fatalities/billion KM and fatality count of 1.8K. Population ratio is ~6:1 US:UK. (https://en.wikipedia.org/wiki/List_of_countries_by_traffic-r...)

There are other countries where fatalities are much higher.

Re:Fewer than 3% of cars sold in the U.S. have manual transm... by torrent-of-ions on Thursday November 17, 2016 @08:38AM (#12977273)
It seems that the fatalities per billion vehicle kilometres will be skewed simply by the average trip length being longer in the US. Fatalities per number of trips is probably a better metric since longer trips tend to use safer roads like motorways. It's just not possible to rack up that many "safe" miles somewhere like the UK, while the number of more high risk miles, ie. towns and country roads, is probably more similar. As suspected, that metric is the one the US does the best on in the article linked.

The best metric is probably fatalities per capita. It's not fair to take into account only the number vehicles because the whole population is at risk when it comes to cars. And it's not fair to take into account number of miles travelled due to the difference in trip lengths.

Re:Fewer than 3% of cars sold in the U.S. have manual transm... by hx87 on Thursday November 17, 2016 @11:38AM (#12978914)
> It seems that the fatalities per billion vehicle kilometres will be skewed simply by the average trip length being longer in the US.

Trip length figures into the denominator (deaths/billion kilometers), so everything else being the same, longer trip lengths won't skew the statistic.

Re:Fewer than 3% of cars sold in the U.S. have manual transm... by Dylan16807 on Thursday November 17, 2016 @05:23PM (#12982290)
Not that the statistic will be skewed with respect to itself. But that it will be skewed against what really matters. It might be the wrong decision to factor out distance.