Safe Speed Forums

I read yesterday about a report by Napier University for "Virtual Fleet Risk manager":

http://www.vfrm.com

... where it is claimed that:

1% of drivers cause 10% of accidents
2% of drivers cause 20% of accidents
5% of drivers cause 45% of accidents

I have long suspected this sort of pattern, but this is the first time that I have heard about supporting research. I've been looking at curves of crash risk against driver quality frequency and the any curve that fits these datapoints is vicious. It seems extremely likely now that 20% of drivers cause 80% of accidents (which is an old theory of mine).

This new research (and I have requested a full copy, but not received it yet, let alone read it), if correct, should change our entire approach to road safety. For example:

* It's now reasonable for 85% of drivers to claim they are above average - because in terms of crash likelihood they are.

* Speed cameas have absolutely no chance of identifying the 5% or even the 20% danger groups. This is one more reason why it's a policy doomed to fail.

* We now know that we need to concentrate our road safety efforts on the low end of the scale of driver quality.

I can't wait to read the full report.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

I can't wait to find out where these "experts" did their maths degree's.

10%+20%+45%=85%

1%+2%+5%=8%.

Therefore, 92% of drivers cause the remaining 15% of accidents.

OK, i accept I may well be being pedantic, but I would be VERY worried about the 92% of drivers causing 15% of accidents

_________________
Semper in excreta, nur quantitat variat.

Ummm ... you can't say that. For example, what if the remaining 15% of accidents were caused by pedestrians, cyclists, and stray dogs? You'd have accounted for 100% of accidents but still only have 8% of drivers who cause accidents.

Even that doesn't quite fit because it doesn't allow for accidents that more than one driver cause, which would cause double counting because the same accident would be counted for more than one driver.

_________________
Will

I read it that the 5% causing 45% of accidents included the 2% / 20% which in turn included the 1% / 10% - in other words it plots three points on a steep exponential curve.

That's the way I understand it too - not least because the three figures are provided separately on different pages of the web site.

I'm still waiting for the full report...

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

Some questions to consider.
What if someone who is close to you is in that 85%?
Let's say they are crashed into by one of the 5%. If they were driving at a safe speed they could have avoided the crash even though they may legally have been 'in the right'.
Can we help the 85% better avoid the 5%?
Can we help the 85% learn behaviours that can make them safer than the remaining 85% (ie safe speed)?
Isn't there a difference between 'safer than average' and safer than the average of your particular group?

Careful now Alex, have you any idea how much time we can spend discussing this stuff?

One point is we're coming from different perspectives - your objective is to improve the individual. My objective is to improve the road safety system.

This is old unpublished work of mine that might see the light of day because there's now some science that may support it. Consider this graph:



We're showing a frequency plot (in green) of drivers ranked by quality - the low quality drivers are so rated because of their crash risk (in red).

Improving road safety absolutely depends on improving the mean driver's ability to avoid a crash. The three regions I have defined need differing sorts of treatment.

The low quality drivers need remedial training and the attentions of the law. If they can't make a reasonable standard of safety, let's get them off the road. This small group (apparently) is responsible for the majority of accidents.

The large middle group need to learn from their mistakes and to be encouraged to improve.

The top group need recognition, encouragement and privilege, partly to reward them for the standards achieved, but also to send signals to the medium group about the prospects for improvement.

Undoubtedly improvements in the middle and upper groups will make them far more able to avoid the errors of the bottom group.

The most wonderful thing about this is that our average driver has a very acceptable crash risk despite having masses of room for improvement. Hell, making the roads ten times safer is within the bounds of possibility and that's before we add in the benefits of vehicle engineering improvements.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

So the challenge is to find some indicator which can be used with a reasonable degree of confidence to identify this 5%, who can then be referred for assessment and retraining, and if all else fails taken off the road.

Regards,

Peter

_________________
"Show me someone who says that they have never exceeded a speed limit, and I'll show you a liar, or a menace." (Austin Williams - Director, Transport Research Group)

Any views expressed in this post are personal opinions and may not represent the views of Safe Speed

Just to be equally pedantic, and main discussion aside... in my day... 10%+20%+45% = 75% ...sorry Dratsabasti, couldn't resist


Kaz

Couldn't it also be that the worst 10% or 20%, for instance, of drivers cause 100% (or 99.9%) of accidents and that the remaining 80-90% of drivers never or very rarely cause accidents?

And that's even if we're only talking about accidents caused by people when they are driving, or if we don't take willcove's point into account:


It doesn't follow that if 5% of drivers cause 45% of accidents the remaining 55% of accidents are caused by randomly selected people in the remaining 95% of drivers; which is implied by a statement like '92% of drivers cause the remaining 15% of accidents'. I'm sorry if I misunderstood what you were saying.

I'm not sure it's much of a challenge - every second accident includes one of them. I also suspect that our excellent traffic officers were quite talented at spotting them, although they probably lacked the correct tools for dealing with the problem.

It's suddenly obvious exactly why we talk to so many drivers who claim: "I've been driving for 10/20/30/40/50 years and I've never had an accident." In order to balance the high risk folk at the bottom of the scale, there must be a great many very low risk folk in the top half.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

But we're not picking them up at present, so maybe we need something more to bring the information together - perhaps a combined database of all insurance claims and all accidents attended by the police.

We did have a debate in the ABD forum recently about whether it was reasonable to send drivers involved in accidents for retraining, irrespective of fault. My view was (and remains) no, but if drivers are frequently involved in accidents that aren't their fault it does suggest something amiss with their driving.

Once collected, that information needs to be turned into action by sending those with poor records for driver assessments.

Regards,

Peter

_________________
"Show me someone who says that they have never exceeded a speed limit, and I'll show you a liar, or a menace." (Austin Williams - Director, Transport Research Group)

Any views expressed in this post are personal opinions and may not represent the views of Safe Speed

Who says we're not picking them up? We might well be - what we're failing to do is take appropriate action. Of course restoring traffic police numbers and talents (with apologies to the small number of talented traffic police remaining) could well be the key to picking them up properly.

I'm 50/50 on the issue of sending not at fault accident involved drivers for training. If we dressed it up properly, we'd get an excellent opportunity to help them avoid being accident involved in the future. I believe that the typical accident involving two participants consists of one who causes and one who fails to avoid. Properly presented, anyone accident involved should be delighted to be given a greater opportunity to avoid in the future. The trick may be to present it as opportunity, not punishment. Perhaps the causer pays for both courses?

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

Perhaps it would be better to move completely away from the punishment culture and let the Government pay for both courses (using some of the huge pile of collected motoring taxation).

In a more positive culture I would see this is broadly equivalent to offering physiotherapy to someone suffering a sporting injury. The benefits of reduced casualties would surely make it self-financing in a short while anyway, from the Govt's point of view.

I couldn't agree more - but with 3 million accidents each year and courses at a rock bottom minimum of (say) £200 - I can't see politicians agreeing to pick up a £1.2 billion tab.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

They could start with "just" the fatals. Assume that two people receive training per accident (ok I know there's a fundamental flaw in the logic there due to the accident type, but I'm sure you get the drift anyway!) that's a total bill of (say) £1.3 million.

This would only need to save two fatalities in the following year to more than cover the costs, going by the Government's figures.

The expression "bolting the stable door..." comes to mind.

More seriously - this wouldn't do the job because there are too few fatals - the objective needs to be to pick up the dangerous 10% responsible for 45% of the accidents doesn't it?

Personally, I'd love to go in the other direction and pick up the near miss causers - this way we are truly into prevention. But it'd be a big task. I bet we could identify them if everyone had an emergency brake application counter fitted. More than 3 application in 12 months could well be an indication of a problem.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

Hi All,

I just got a report extract from the American part of VFRM: It contains:

===================================
Identifying the drivers most at risk

The first finding from the Napier University study data was that the system could be used to identify the most at risk drivers in an organisation. In the table below for example, the system identified the 5% of drivers responsible for 45% of crashes.

% of drivers - crashes - % of crashes
83% 0 0%
13% 1 55%
3% 2 25%
1% 3 10%
1% 4 10%
================================

So in this sample, 83% of drivers were crash free. We need to assume that this was a time limited sample and many of those 83% will have crashes in (say) the next decade.

So the curve isn't going to be as vicious as it first appeared, but it may well be vicious. In this sample we have 17% of drivers causing 100% of the accidents.

I'll what more and can find, and see if I can infer anything at all from the extra data received.

_________________
Paul Smith
Our scrap speed cameras petition got over 28,000 sigs
The Safe Speed campaign demands a return to intelligent road safety

Hope I'm not drifting off subject here with my 2-penny worth here.

I've long held the belief that there are people out there on the roads who are not temperamentally nor attitudinally suited to hold a driving license. And no amount of training is going to alter that.

Furthermore, the courts seem unable or unwilling to act decisively to get these people off the roads when they get the chance.

I once witnessed a crash involving a car whose driver had rounded both myself (on a motorbike) and another car ahead of me to jump a red traffic light at some roadworks. He raced through the lights and crashed into another car pulling out of a side street (also controlled by lights) around the corner ahead. He broke the leg of the young girl on the backseat - it could have been worse.
In court, the witness tesimonies of both myself and the car driver - stating that the guy had gone out of his way to jump the lights making a one finger gesture to us both as we pointed to the signal - were dismissed as understandable mistakes (very patronising). Anyway, he got away with a fine and my faith in the willingness of the court system to deal appropriately with an offence that I'd seen unfold in front of my eyes, was totally trashed.