Safe Speed Forums

Due to the (almost) completely random relationship between travelling speed and impact speed, if you take a sufficiently large number of samples (several billion, say) you will see a shift in the classification of between one and two percent.
But the number of collisions per annum in the UK is comparatively tiny, albeit a large number, and therefore the amount of randomness will completely swamp the trend.
And in a single collision, the resulting injury classification is a lottery. The trick is to avoid the collision in the first place.

Regards
Peter

Sorry - that's rubbish. You could say "impact speed is determined by a number of factors of which pre-impact free travelling speed is one but (depending on the circumstances) not the most important".

If you're saying what I think you are, then we're actually in complete agreement.
Unless, of course, you're saying something else, in which case we're not.

Regards
Peter

I agree with the logic. I disagree with the assumption that driver behaviour would be unalterered. Don't forget we've already got altered behaviour - they are driving 1mph slower.

There has to be a means of altering their behaviour. Does this means distract? Does this means alter safety priorities? Does this means mislead?

Let's say they are goin 1mph slower in 30mph zones. That's about 3%. In order to "succeed" to the same standards as before, around 3% less attention is required. Is 3% less attention provided? Or 2% less or 5% less?

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

Just how much attention will need to be diverted from watching the road to make your car drive "imperceptively" slower?

This is where the the whole Speed Kills argument unravels:

You don't have to pay any attention to your speed if you are driving at what you feel is the appropriate speed for the road and conditions.

You can concentrate all of your attention to spotting when the appropriate speed is about to become lower, and you can then reduce your speed immediately it becomes necessary.

If you are trying to driver slower than the speed you feel is the appropriate speed for the road and conditions you will be concentrating all of your attention on maintaining an unnatural speed, miss the hazard and hit the pedestrian.

Or am I missing something here?

_________________
http://www.itsyourduty.org.uk

Exactly, well, almost. I'd go further:

You can concentrate all of your attention to spotting when the appropriate speed is about to become lower, and you can then have already reduced your speed so that further adjustment is not necessary.

Lol

Don't all tread on each other in the rush to 'prove' me wrong will you

I'll agree that this notion has an extremely compelling 'hook' to it; hardly surprising that so many want to connect with it.

But you're not right.

When you are discussing a subject that is full of so many what if's and variables, is anyone absolutley right.

That by itself, does not make you right either

_________________
Steve

"Random" might be too strong a word as it implies unpredictability, and that's not quite true. Hugely variable might be a better way of putting it.
What's very difficult is nailing down all the factors, establishing how they change the initial-impact speed relationship and then quantifying them all. The idea that all else being equal, higher initial speed means higher impact speed which means more severe crash results, is over simplified but does at least show a relationship. But in the real world everything else is never equal so all those other factors must be considered. Taken in total they may make no change between initial and impact speed (total faliure to observe, set off with no brake fluid, etc), reduce it to zero (alert driver sees hazard early, experienced driver takes avoiding action etc) or rarely the impact speed might even be higher than the initial speed (crap driver hits the gas instead of the brake). To me that says that the relationship is not a hugely important one in reality except in circumstances where all the other factors either cancel each other out or are strangely absent.

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

Well I seem to have stirred the pot enough to get things bubbling. There's some interesting stuff coming out. It needs coherent presentation with step by step analysis, imo, before you can say "Yes of course there is a relationship between free travelling speed and impact speed and between impact speed and crash/injury severity but the input of initial free travelling speed in determining impact speed is outweighed by [other factors] because [analysis]"

It doesn't seem as though it would be too difficult to model this in the way that Pete has described but adding some additional input variables. So what would the inputs (fixed and variable) be:

initial free travelling speed
reaction time (clearer to have reaction time as a constant but to have a variable 'inattention gap'?)
braking performance (fixed at 20mph/s deceleration effect?)
road positioning (see below)
concentration factor (see below)
inattention gap
braking performance (need to take this as a constant I think?)
[others]

Expanding on some of the above:

- How can we model the effect of 'good road positioning? Can we (say)treat it as a reduction of the standard reaction time?

- by concentration factor, I mean the enhanced level of concentration which (I think) "better" drivers switch on where there is a high hazard density. This will feed through as reduced reaction time.

Can we assign some values to the above?

I'm not sure that this road positioning entirely removes the risk of a pedestrian ending up underneath our wheels if he's determined to do so, unless we reduce speed to almost zero. I thought it is more to avoid hitting a pedestrian who steps off the kerb without looking (or is pushed off or whatever). I'd think it's very hard indeed to drive defensively enough to avoid hitting somebody or something who is determined to hit you (e.g. if on a major road and approaching a junction with a vehicle waiting to turn right across your path, you can do all the usual stuff but if that vehicle pulls out at the critical moment, there's not much you can do about it).

Amazing. I've just finished doing it, and there's your post more or less describing what I've done.

I've uploaded it: http://www.safespeed.org.uk/pedinc.xls

It's a bit scruffy and undocumented so far.

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

Yeah well you can say that but until it's been shown to be so, your 'view' is no better than a hunch. Perhaps Paul's done it with his spreadsheet - I'm still trying to work out what it means.

Yeah, at the moment it's a hunch. I'm not in a position to work it out except by thought experiment, like my twins on the dual carriageway heading towards the downed tree. I reckon the way to go about it is to think up as many factors as possible, then apply them to the situation individually and see what affects what. Initial speed doesn't affect much beyond the actual distance covered during braking and the distance covered between ":shock: ****** tree!" and hitting the anchors. On the other hand, what we might call quality of observations will determine the point at which " ****** tree!" occurs, so it follows that it also determines when the brakes are applied and from that the speed at which the tree is hit (or whether the tree is avoided altogether). Far as I can see initial speed can't affect observation, though something like the physical condition of the driver could.

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

Why is it rubbish?

Impact speed is essentially: i = s - (d * t)

where s = travelling speed, d = deceleration and t = time spent decelerating.

As t is essentially random, what makes i less than random?

Regards
Peter

But it is really unpredictable, as it depends on factors such as, the exact time both parties started their journeys, their average speed over the journey, etc. etc.

The relative time that both parties arrive at a particular point in time can vary between 0 attoseconds and an infinite number of light years.

What was the exact time you last drove past the door of 33 75th Avenue, Manhattan?

To say that it's hugely variable is an understatement.

Regards
Peter

I think I see what you mean but I don't think we're talking about exactly the same thing. Correct me if I'm wrong but I think you're talking about the random nature of the exact time and place that a crash happens. Like I might hit a bus tomorrow if I leave at a certain time, but if I go a minute early so I can pick up a paper, or a few minutes late because the cat threw up I've avoided it without ever knowing how close it was. Maybe even a couple of seconds delay would mean that I don't end up where I'd need to be for the crash to happen. Something like that? If so, then yeah, I'm with that but it wasn't what I meant by things being variable.

I'm talking about after that. Just after the point at which a collision becomes unavoidable. Like drivers A and B, and that tree. They are both going to hit it (my imagination, I'm allowed to say it's too close to stop in time ). If we presume absolute identical vehicles, same quality road surface in each lane, identical reactions and so on then B will hit it slightly harder. Vary so much as tyre pressures or weight in the car and things start to change. Something major like how quickly they spot the tree, and we could have B hitting the brakes long before A, and so A ends up hitting it harder. Say the tree is right on the very edge of B's best possible braking distance and he spots it as early as possible and responds perfectly. Maybe he hits it at less than 5mph. At the same time say A's a smoker and dropped his fag in his lap, delaying him spotting the tree by a few seconds. He's going to hit it at a significantly higher speed. In other words, the instant after a collision becomes unavoidable the difference between initial speed and impact speed may vary between nothing and not quite 100%.

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

In 25% of cases (according to an Australian study SafeSpeed quoted), t is zero, so it is not essentially random.

_________________
I stole this .sig

Yes, I agree that things get somewhat more predictable once you reach the point where a collision is inevitable. But, even then, a small fraction of a seconds difference in response time makes a large difference to the impact speed. So, yes, it's then not quite random, but hugely variable.
But, my point is, when this is taken together with everything that's gone before, it is essentially random - especially seeing that the usual hazard (pedestrian, other car etc) is no longer going to be there a couple of seconds later, unlike the tree in your example.
Also, your example isn't very realistic in that it depends on several factors coming together at the same precise time - the two cars, despite doing different speeds, being alongside each other at the precise instant that they both happen to be a certain distance from the hazard, being the same instant that both drivers spot the hazard. I'm sure you'll agree that the chances of that happening are pretty remote.
Having said that, a very similar scenario is described in at least two pieces of government-sponsored research - you know, the ones which underpin the whole 'speed kills' policy. And those people really ought to know better.

Regards
Peter

In cases were braking does take place, t can be any value between zero and the time it takes for the car to stop completely (and beyond, but then no collision takes place). And it's that which determines the impact speed.
And even in that 25% of cases where t = 0, it's only zero because we cannot have negative values - that would mean going backwards in time.
Where t = 0, it means that either, a) the time taken to reach the hazard is shorter than that required for the driver to react (eg 0.25 seconds) or b) the driver doesn't see the hazard, therefore doesn't react at all.
In every case, t depends on the precise relationship in position and time between the car and the hazard. As this is random, so is t.

Regards
Peter