Comments
Congratulations New Zealand… We only killed one person!
We’ve had mega road tolls the last few long weekends (12 at Easter). Needless deaths caused by a lethal combo of alcohol, speeding and lack of seatbelts.
This Queens Birthday Weekend, the police had had enough of scraping us all off the roads and vowed to have a zero tolerance approach to speeding. This meant a combo of placing Police everywhere and having an absolute max of 4kms over the speed limit before fines were handed out.
So, despite some grumblings from those who think of themselves as race car drivers, on a weekend with terrible weather and a LOT of people on the roads, one solitary (self caused) death is not only a major achievement, it’s the lowest road toll we’ve had since records began.
I think we’re pretty proud of ourselves. We did good!
However, one needless death is still one too many… So while we should be patting ourselves on the back, we should also set a whole new target: Zero.
Hello, my name is Natalie, I have a business called 
June 8th, 2010 at 8:50 pm
Nat, I love you and your heart is in the right place, but you’re dead wrong on this and so are the police. There is no statistical evidence at all that anything the police do makes any difference.
You can see the historical data here:
http://www.transport.govt.nz/ourwork/Land/landsafety/HolidayRoadToll/
If you look at, say, the twenty years before 2010 you can see that the overall average number of deaths at Queen’s Birthday weekend is 5.6. This is a really small number compared to the number of people who are driving so it’s reasonable to model it with a binomial distribution.
Based on nothing but an average of 5.6 out of a lot of people at risk (at least 10,000 but 4 million gives the same result), the chances of having a given number of people die, from pure random variations, is:
0: 0.4%
1: 2.1%
2: 5.8%
3: 10.8%
4: 15.2%
5: 17.0%
6: 15.8%
7: 12.7%
8: 8.9%
9: 5.5%
10: 3.1%
11: 1.6%
So you’d expect to see the deaths as low as 1 about in year out of 48, and that’s pretty much exactly what has happened (it’s not the first time, contrary to what you said).
Things do actually seem to be getting better in recent times. I’d put that down to newer cars making crashes more survivable, but in any case the Queen’s Birthday average for the ten years before 2010 is 4.7.
That turns out to increase the chances of only 1 death in a particular long weekend to 4.3% (more than twice as likely than when the average is 5.6), but the other side of it is that it makes the chances of 10 deaths only 1.3%. Once in 77 years. And yet it happened last year.
If it’s really police activity affecting it then what the heck were they doing last year?
Of course they weasel out and say that they did their best but it was impatient and careless drivers.
Can’t have it both ways, guys.
If some factor was to truly reduce the risk enough that the average value was, say, 3, then you’d see numbers like this:
0: 5.0%
1: 14.9%
2: 22.4%
3: 22.4%
4: 16.8%
5: 10.1%
6: 5.0%
7: 2.2%
8: 0.8%
9: 0.3%
10: 0.1%
Single death long weekends would become common — at least one every decade — and 0 would happen from time to time. 8, 9, and 10 would be basically never. 7 would be as rare as 1 is now (in fact 7 happened in 2000, 1997, and 1994)
We’re not seeing that. Everything points to a slow but steady decline as cars (and roads) improve, not some new magic enforcement formula from the police.
A prediction to pull out in five months from now:
- the police will try the same thing in Labour weekend
- the number of deaths will be much higher
- the police will blame impatient drivers, not themselves.
June 9th, 2010 at 12:10 am
Oops seems you were right that it’s the first time since records began. I’d read a report which said the first time in 54 years or something like that and interpreted it to mean we had more records than that and it had happened before.
Regardless, it is unsurprising that it should happen once in 50-odd years and just chance that it happened to be this time.
Note that there were two more deaths shortly after the official end of the holiday weekend. They could so easily have been a few hours earlier.
June 9th, 2010 at 10:48 am
I’m more than happy to do another bet on this subject (my PT has 100 pushups coming this way after pretty much saying what you did about QB tolls).
In my mind, the reason the deaths happened AFTER the official weekend is that the police left the roads and the speed limit increased again.
I understand the statistical approach, but that also means you are saying we have very little control over road deaths, and we KNOW we do – I think Police presence and the resulting adherence to the speed limits is a key part of what we CAN do.
I also think they did something else that people hardly noticed. One day in, they were all over the media congratulating us all and making us all feel like we OWNED the problem and we all had our part to play. I certainly noticed a difference on the roads, normally everyone is speeding, this year 105kms was pretty much the max and the police were EVERYWHERE. Probably unsustainable, but maybe it will kick us into a new routine?
June 9th, 2010 at 10:49 am
Oh, and one death HAS happened before – the first year they recorded the number – but this is the equal lowest :)