Safe Speed Forums

Let's not get the over inflated self opinion types mixed up with Mr Perfect, eh. He really is perfect. If he has a fault about the only one it could be is that he's not real.

_________________
Make everything as simple as possible, but not simpler - Einstein

That’s why I take a mean – it cuts a lot of fluff from the discussion. We know that, theoretically, MDBDF (mean distance between Disastrous failure) could be very long if we all went very slow. But we don’t want to do that. We know that MDBDF is very short if we all drive around like lunatics. Is it linear between? Perhaps, almost - who knows?

But once we agree that MDBDF should be as short as possible, without going at a snail’s pace, then the goal is no longer just safety – we want to get somewhere too! So the optimum place for MDBDF is not zero, but something more, but not so much more that we all suffer too much. So the bottom line is that we can set MDBDF almost anywhere, but we choose to set it as short as we can politically afford to, because we want to get somewhere quick! Or, in other words, we set as long as we can, because we don’t want to die!



I’d imagine that, on average, there is a correlation between close shaves and crashes. As close shaves are not measurable, and are a factor of total crashes, it seems logical to trend (and discuss) physical crashes, rather than phantom ones (although I agree we should learn what we can from phantom crashes).

_________________
I stole this .sig

Its a pity we can't learn enough from 'phantom' crashes. The aviation sector has a mandatory 'near miss' reporting process designed to investigate and improve circumstances which lead to near collisions. Then again, there are relatively few pilots and the industry is highly regulated. Getting 30 odd million motorists to file near-accident reports wouldn't work in a zillion years.

Oooo, a new term to have fun and games with and arguments over the definition of . Seriosuly, how short or long MDBDF is must depend on what we call disastrous. Deaths? SI? Walking wounded? Vehicles written off but people unharmed? Minor damage only? Contact but no damage? Wherever the line is drawn there will be a case for including those incidents that nearly made it, and if you get down to swapping paint then you might as well start including near misses. Why? Because the difference between death and SI, or write off and minor damage, or between any of them and any of the others, may well boil down to nothing more than providence. I can't see much value in discounting a potentially serious driver error because they were lucky. Also, a potential crash situation caused by one driver's error may well be avoided only by another driver's skill. Can we really not give a metaphorical black mark to a driver who has only avoided a crash through luck or by relying on the reactions of another? I don't think so, and so we come back to counting driver errors rather than crashes.


That rather depends on whether the correlation between near misses and crashes can be determined. If we knew what it was then we could easily extrapolate MTBF from MTBC. But that's very difficult and probably varies from one driver to another, and certainly varies for individual drivers throughout their driving careers.

I don't think it's logical to consider physical crashes so much as easy. It's MacNamara all over again, measuring what is easy to measure and then coming up with reasons why it's important. I'm not saying it's unimportant or that it has no value at all. Just that as far as crashes are concerned luck plays a factor, and I think that needs to be taken out. The only way to do that is to consider all errors that can lead to a crash, regardless of whether or not they actually do. In practical terms that's incredibly hard to do independently, so the only real solution is for the idea to get the majority to be honest with themselves about near misses and to then try to learn from it.

What would be great is if there was an emergency manouevre/emergency stop counter on the dashboard. If that read three when the odometer read 120,000 you could feel pretty pleased with yourself, but if it read 3 a month after you take delivery of the car then you should ask yourself what's going wrong. And if it reads three by the time you get the car home for the first time you probably shouldn't start reading any long books. No idea if such a device is feasible (God knows what it would read if it's owner took it for a lap of the Nurburgring unless it was very clever at distinguishing between hard driving on a track and real emergencies on a road), and it'll probably never happen anyway. But I think it might tell drivers something about the quality of their driving.

_________________
Make everything as simple as possible, but not simpler - Einstein

Oh dear, I must be dead then. I make a point of knowing how a car handles in varying conditions, and will always include an emergency stop from 60+ in a test drive. Obviously I take care to ensure that I am the only car on the road first, but the suggested counter wouldn't know that.

The same goes for relatively violent direction changes - I want to know how the car will react; will it turn, or slide, or like the new Renault Logan (and Merc A Class before it) will it roll over? (Renault Logan to be sold as Dacia in UK failed ADAC tests when it rolled during a sudden avoidance manoeuvre, or Elk Test for short, at only 40 mph).

Knowing this helps me react better should something unexpected happen while driving the car, and gives me an idea of how much latitude I need for safe driving. Unfortunately the counter you suggest would have me down as a very dangerous driver...

You're right if a 'near miss report' is paperwork.

I believe you're wrong if a 'near miss report' is an individual mental process.

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

Yep, that's one major downside, which is what I was trying to say with that Nurburgring remark. It'd probably go up a hell of a lot in only 13 miles or so .


Not really since you'd be the one looking at it, and you'd know better. Obviously the 'stopometer' record would need to be interpreted unless it could be made very very clever at telling the difference between a real emergency and other circumstances where you'd stop that hard, or alternatively have an off switch. The stopometer wouldn't be there for the scammers and tptb, it would be there to provide information for the driver alone just like the odometer. You know when you've made it go up by deliberately chucking the car about, so you just ignore those. Actually, it'd probably be better to make it like the trip counter, and have a reset button.

Of course, this is all irrelevant since the biggest disadvantage of it is that the chances of manufacturers ever bothering to fit 'em is probably near zero. I can think of a few people who could do with one though. Two members of my family who scare the hell out of me by not seeing other cars until the last minute and then flooring the brake pedal almost hard enough to stand the f er on its nose, but who both bitch at me for driving too fast even though I can barely remember what the ABS feels like.

_________________
Make everything as simple as possible, but not simpler - Einstein

Of course they judge your driving on the basis of the nasty surprises they keep getting...

Fundamental problem alert! We've got politicians, research scientists, 'road safety campaigners' and god knows who else making the same basic error.

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

I'm not discounting it - I'm saying it cannot be captured on record today. Crashes, on the other hand, are rarer, and more serious, they can (and are) assessed and logged.



That is a benefit of MDBDF. Because it (statistically) includes the effects of errors, it tells us something of our expected safe-driving life span (so to speak), warts and all.



Now you are talking MY language! Especially if it was hooked up to an RF Sensor and a database backend so we could ALL tell when a crappy guy is driving along.

_________________
I stole this .sig

I know, but if we do no more than that we're failing to take the next step just because it looks difficult.


Nope, can't see that at all I'm afraid. It just includes the effects of errors where luck ran out (or whatever). We're still no nearer to knowing whether the errors that lead up to a certain crash are made by the driver(s) involved one a week, once a month, once a year, or whether the situation is so unique it can't reasonably be expected to recur. Nor does it tell us a thing about whether one driver involved in a crash is making other errors that might lead them to crash again in a couple of weeks if their luck remains bad.


I should have known you'd come out with something like that. How can I put this? No. Chance. In. My. Car.

Anything that can be used as a driver aid, fine provided we don't get to the point that the dashboard looks like it belongs to Concorde. Any tale telling devices I am resolutely opposed to. We should be encouraging drivers to take an interest in self improvement, not spying on them. Leaving aside civil liberties, any monitoring device is likely to change the way drivers behave. You may (probably will, in fact) say that that's the whole point of it, but if it turns out to distract them then the change will be for the worse. They should be concentrating on driving safely, not worrying about whether their own car is grassing them up for failing to achieve perfection.

_________________
Make everything as simple as possible, but not simpler - Einstein

If we need to understand why motorists become involved in near-misses, they would have to report the circumstances. That wouldn't happen voluntarily.

I also believe that it is verging on fantasy to even imagine that the majority of average drivers would engage in any sort of self assessment that would amount to a near-miss report. In an alter-Britain perhaps, but not the one I know. Sorry if that sounds defeatist, but I also think I'm right in being realistic.

If we're talking about forensic investigation, then of course you're right. But that's not what I meant - although I appreciate it looks like that.

If the average driver now spends 5 seconds thinking about his near miss, and we took action that raised that to 6 seconds, I believe we'd see a real long term benefit. But I also believe we could do far far better than that.

I think we could get 10% of them thinking about their near miss for at least an hour. And I think the benefits are huge. For a start I believe that there are about 100 near misses for every injury crash.

I apologise for misleading you about what I meant.

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

It make no odds – the thing is that a UNIT for road safety is established. Right now, actuaries in insurance companies are really making these calculations of risk and they set their rates based on that, so that they get more in premiums than they are likely to pay out in costs. That means they have to assess each drivers MDBDF (or something similar) and work out the costs based on that, and add profits to arrive at the figure. You would think that people with high insurance costs would twig that the reason behind it is their short MDBDF, but they don’t. They put it down to a rip off. On the other hand, people (like me) who get full comp with two drivers for under £150 do recognise that it is due to their long MDBDF! One explanation is that people with long MDBDFs are smarter than people with short MDBDFs (!), which is no surprise if you think about it!!!



That’s right, but you are confused because you think I mean that MDBDF relates to real-world accidents, but it doesn’t. MDBDF relates to the chance of a real-world accident, which is proportional (I expect) to the number of near misses – they are different units for the same thing (in statistical terms). It is a notional value – after you die, the actual number of crashes you had in your life could be used to compute your Actual Distance Between Disastrous Failures, but a fat lot of good that would do for you!



The assumption is that a driver with a short MDBDF is more likely to crash soon than a driver with a long MDBDF. The fact that the driver has observed ‘characteristics’ that gives him a short MDBDF (convictions, age, locality, sex, past history, car-type, income, emergency-stop-count, etc.) predict that he won’t last long. If the MDBDF turns out (after you are dead) to be inaccurate, then the inferences drawn from the observed ‘characteristics’ were false.

But of course, a coarse MDBDF for the whole driving population is easy to find – take the number of miles driven by us all, and divide it by the number of serious injuries. It is exactly what SafeSpeed has been doing with his graphs all along to determine the safety of our roads – he just has omitted to give it a Unit (MDBDF). So we all have a global MDBDF right now anyway – it is 10 million miles! So, on average, we could each drive for many lifetimes without seriously hurting anyone, yet collectively kill or badly injure 50,000 people a year!

The idea behind MDBDF is to work it out for smaller groups, though. That is more difficult, and needs a lot of data.



Then we will never get the data to work out MDBDFs for various types of driver, because people like you won’t let the data flow. Perhaps you are right, and we are stuck with coarser measures, like cameras, humps etc!

_________________
I stole this .sig

By way of a point of order, that's not right at all.

They assess average risks across wide population groups, and apply a suitable premium to the whole group. If you look at the young male problems, there are indications that risk is not spread anything like equally across the group, with 1 in 5 or 1 in 10 really causing the lion's share of the risk applied to the whole group.

Of course most imaginable schemes can't recognise the high risk sub-group because there's no history to go on for a new driver. But I bet the warnings are there in the near misses...

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

No need for a point of order – although I might have phrased it better as “The companies actually attempt to assess what risk group each driver fits into”. In any case, I acknowledge your point when I say that ‘The idea behind MDBDF is to work it out for smaller groups”, i.e. insurance people divide the entire driving population into smaller groups, but very small groups or individual calculations need more data and time, and, below certain thresholds, may not be viable at all, for political reasons as much as anything.

As you say, groups are coarse, and one solution is to improve the model with specific, high-resolution data. As you know, I believe that is absolutely inevitable. But you are opposed to it, for some reason, aren’t you?

PS: MEAN can only apply to groups of various sizes, so the insurance grouping is inevitable too.

_________________
I stole this .sig

My wife has no complaints!



Nope - large no- nonsense traffic cop with a lot of acid in lecture usually usually works!

_________________
Take with a chuckle or a grain of salt
Drive without COAST and it's all your own fault!

A SMILE is a curve that sets everything straight (P Diller).

A Smiley Per post
FINES USfor our COAST!

Approach love and cooking with reckless abandon - but driving with a smile and a COAST calm mind.

Back to topic - as mentioned eleswhere DIS and Speed Awares (if done prpeerly) usually have the desired effect in tackling a problem and educating in right direction - based on reports

. Also - think lot of drivers do heed our friendly advice when we do pull them. Certainly - no record of further transgressions in our patch

Think we need to tighten L-test and have some further assessments every now and then to iron out bad habits and encourage at least a regular re-read of the Highway Code.

Also - new gadgets in the car - perhaps a marketing ploy could be a free lesson in using them properly to prove their value. This in turn would improve standards perhaps.

_________________
Take with a chuckle or a grain of salt
Drive without COAST and it's all your own fault!

A SMILE is a curve that sets everything straight (P Diller).

A Smiley Per post
FINES USfor our COAST!

Approach love and cooking with reckless abandon - but driving with a smile and a COAST calm mind.

And I think I know exactly what kind of specific data you're thinking of - I don't even need to say it, do I?

The problem is, unless they know exactly what data is important, they'll collect the wrong data - and that data, besides being useless for the purpose, is likely to be used for all sorts of other stuff.

I would have no objection to a 'black box' type data recorder, which recorded everything - including a video of the road ahead - for a period of, say, 10 seconds prior to a collision.

Cheers
Peter

_________________
Only when ideology, prejudice and dogma are set aside does the truth emerge - Kepler

No, they base their premiums on the risk of having to pay claims, any type of claim. Risk of collision is in there, but you're oversimplfying it by implying that it reflects crash risk. Premiums go up and down on the basis of postal code, where it's kept, or on extra equipment fitted to the car. One idiot company decided that the optional alloys the original owner of my car ordered made it non-standard and bumped up the premium from reasonable to outrageous. They're factory fitted Honda wheels that as far as I can tell are just slightly fatter at the hub and slightly narrower at the rim, but it would have cost me a load extra on my premium. Does that reflect crash risk or MTBF or whatever? No.

Where it probably is reflected is in the 3rd party only component of the premium, but I don't know about you but I have no idea what proportion of my premium that is. Just got the renewal quote in, but nowhere does it say that £X is for 3rd party £Y is for fire and theft, and £Z is for the remaining fully comp bit. I have noticed that it's slightly cheaper than last year despite the claim I had to make over that suicyclist last autumn. Sure as hell glad I had the NCB protected.


No confusion here mate. I realise exactly what you mean by it, though if any poor wording on my part has given that impression, sorry. However, I'm not convinced that it's a reasonable assumption to think that the proportion of errors that go on to cause crashes is so straightforward. As you say, we can apply it fairly coarsely to the driving population as a whole or perhaps large groups (which is presumably what insurers do for various age groups for example). But that's about all we can do with it. So good for insurance premiums but of limited use in telling an individual how much s/he needs to improve.


I notice you've put emergency stop count in there .

I notice you've also put car type in there. Going back to the use of premiums to determine crash risk, the component of the premium that is determined by car type really tells us little more than how much the insurer expects to have to cough up if it gets bent or written off. Small Fords are cheap and expensive Italian exotica are frighteningly dear. Does car type tell you anything about the driver? Only if you resort to stereotypes.


Cameras and humps don't change the way data flows. They change the data itself by making driver behace unnaturally. I'm all for driver aids but not for snooping devices. Snooping devices will just increase the fear level, and again, that will bugger up the data. How can data flow properly if the act of observation changes what is being observed?

_________________
Make everything as simple as possible, but not simpler - Einstein

Lift that goalpost a bit higher please BW, it's making an awful screeching noise as you drag it along to another position
You actually used the term "neanderthal", but failed to specify exactly what classed someone as such.


I think you will find that people who drive (or have driven) at competition level have absolutely no desire to drive the same way on a public road at all.


As I said before BW, the bloke (or bird) who is going to drive however they feel and without consideration for other road users doesn't give a stuff for the camera.
They are invariably in a car that is not registered to anybody, and when it does start to draw too much attention, it is simply sold on and another vehicle is purchased to enable them to start all over again.

_________________
What makes you think I'm drunk officer, have I got a fat bird with me?