The Alliance of British Drivers (ABD) has labelled ACPO’s new speed limit enforcement guidelines as “commendable”, but has expressed doubts about how they will be interpreted and implemented by local police forces.

The ABD highlights a number of extracts from the ACPO guidelines, including: “Where limits are not clear (that is they don’t feel/look like the limit or are on inappropriate roads) they will not be routinely enforced.

“The gist of these guidelines is that limits should suit the road and that police should use their discretion to enforce only where the limit is obvious to drivers and would be what they would expect from what they can see through the windscreen. Limits should not be set well below what drivers would expect to see and then be enforced to achieve high rates of offender punishment.

“Unfortunately many local authorities and police chiefs appear not to understand this simple concept, believing that just sticking up a sign and prosecuting thousands of safe drivers somehow improves safety.”

The ABD goes on to say that it is “disappointed to see ACPO … supporting setting of speed limits at the average speed of traffic”, suggesting that this will lead to the “excessive prosecution of safe drivers”.

The ABD goes on to call for the reintroduction of “setting of limits at the 85th percentile”.

Click here to read the full ABD press release.

Hugh

As a way to measure pre and post 20mph speed you could measure the mean speed over time at a particular point. But the mean speeed of the “road” for assessing whether you need additional measures beyond signage for a 20mph limit ahould be estimated or measured for the whole road according to DfT. Of course that would still enable you to put in additional calming measures at any particular pont on the road such as additional roundels, markings, engoneering, etc.

Rod King, 20’s Plenty for Us0

Thanks all – an interesting debate about the 85%ile which I shall close off at this point.

Nick Rawlings, editor, Road Safety News0

Eric: No space in previous response for this bit, but hopefully the editor will allow me to indulge you so, out of interest, if in your hypothetical example, only 80% all did 40 and 20% did 120, the 85th%ile would then shoot up to 120 ..agreed? (i.e 85% did not exceed 120mph), but hopefully you would you still rightly be saying “I would hope we all agree that 40mph would be an appropriate speed limit” sensibly based on what the majority did, or would you then have to conclude that 120 is the right limit just because it’s the 85thille speed? By the way, the mean speed (56) has now dropped to considerably less than the 85th. See the conundrum? Real world scenarios are much better to work with!

Hugh Jones, Cheshire0

Yes Eric, in a fantasy road scenario, you’d probably be right but if I could just drag you back into the real world for a moment. As you know quite well, speeds are not spread like that – if they were, we wouldn’t be having this exchange and anyway you said previously that “the number travelling at precisely the 85%ile speed is irrelevant – it is the number travelling above/below it which defines it”, but in your hypothetical and impossible example, nobody seems to be travelling below it and I’m not sure you can usefully apply percentiles to a ‘range’ of two values anyway. So whilst it makes for an interesting brain-teaser, in practice it doesn’t help in the slightest. To avoid prolonged exchanges – now and in the future – I can only suggest you conduct a proper speed survey (or two) of your own and all will become clear.

Hugh Jones, Cheshire0

Hugh’s view is easily proved wrong.

Lets say that 85% of drivers just happen to drive at exactly 40mph on a stretch of road and the other 15% of drivers manage to hit 120mph. Hugh would calculate just over 50mph as the mean speed, but the 85%ile speed is 40mph. I would hope we all agree that 40mph would be an appropriate speed limit.

Eric Bridgstock, Independent Road Safety Research, St Albans0

But if the majority of walkers were uneasy about the 15m line and strayed no further than say 30m, you could say that that was the optimum point by consensus and if you worked out the mean distance chosen from the edge by all the walkers (including the foolhardy few who strayed up to and beyond the 15m line and those who stayed even further back than 30m) it would be around 30m. The 15m line might then still represent the 85th%ile chosen distance to/from the edge, but not the most popular and not that chosen by the majority.

Hugh Jones, Cheshire0

The cliff analogy is used to express the fact that people can wander around where they like, but it’s only when they get within 15 metres of the edge that they feel they have gone far enough and that to go further would significantly elevate their risk.

The mean distance from the edge is a bit of a red herring in that statistically there is no difference in actual safety risk (falling off the edge) if you were to walk at the mean distance or to walk at the optimum distance.

All that setting the limit at the mean achieves is to reduce the benefits of the view with no appreciable increase in safety.

Duncan MacKillop, Stratford on Avon0

Duncan: It’s just that in the last para of your first posting, from the phrase “the majority of drivers will use to maintain the optimum balance”, I took it that you were concluding that whatever speed the majority tend to adopt, we should presume to be what that majority therefore consider to be the optimum speed i.e progress balanced with safety? If so, then as the majority tend to cluster around the mean speed – more than any other – then this should be taken to be the ‘optimum’ speed by consensus – not a higher speed adopted by a significant minority.

Hugh Jones, Cheshire0

Not so Hugh. The mean is an entirely diifferent number to the optimum. In the example the 85/15 metre line marks the extent of the best balance between view and safety and even though there may be a spread of distances from the edge which can be averaged, the optimum remains the optimum. Whether people choose to take the optimum path or the mean path there is no discernable difference in their safety, but there is a difference in the benefits (the view) received.

The rapid fall off in the number of peeople chosing to go over the line (only 15 out of 100) is an indication that the limit is a good one for the majority of people for the majority of the time.

Duncan MacKillop, Stratford on Avon0

I think Duncan’s admirable explanation is that of the mean speed, rather than the 85th, as surveys always show that the greater number of drivers – presumably balancing reasonable progress with a reasonable safety margin – are driving at, or very close to, the mean.

Hugh Jones, Cheshire0

The 85th percentile has always caused a great deal of confusion, so here is a good explanation.

Imagine a path that meanders along the edge of a high cliff. 15 metres from the edge of the cliff the local authority has painted a nice white line and and regularly spaced along it are signs which say “danger cliff edge”.

The lie of the land is such that if you walk 100 metres from the edge, you don’t get to see the magnificent views, but these views will open up the closer you get to the edge.

The whole purpose of walking the path is to see the views, so you evaluate the situation and quickly work out that you get the best ratio of view to safety if you stay just inside the white line.

As you walk you notice that the odd person is walking beyond the line so that they can get an even better view, but they seem to be getting away with it even though you might fear for their safety they seem to suffer no ill effects.

The next year you return to the cliff to enjoy another walk, but this time you notice that a terrible winter storm has eroded the cliff by 20 metres and the line and signs are no longer there.

The question is do you still walk a distance from the edge that gives you the best balance of view and risk which experience has shown to be at least 15 metres, or do you blunder off the edge of the cliff and die horribly?

Instead of the best balance between view and risk as in the cliff example, people take the best balance between progress and risk when driving on the roads. There may be nothing as obvious as a cliff edge to act as a guide, but there are many less obvious clues which by far the majority of drivers will use to maintain the optimum balance.

Duncan MacKillop, Stratford on Avon0

Rod: To be of any use and so one can compare like with like, you have to measure at the most free-flowing part of the road to establish the highest mean speed of that road. Logically and fairly obviously speeds past any obstacles are inevitably going to be lower but this would be taken as read anyway. The exception would obviously be if you specifically wanted to know the speeds at these points as a post-traffic calming exercise for example.

Hugh Jones, Cheshire0

Eric, my understanding is the same as yours, but all too often I have seen it wrongly interpreted as the speed that 85% of vehicles are travelling at, which it definitely isn’t, which is why it’s worth clarifying every now and again. Yes it’s the speed that 85% do not exceed, but what is often overlooked is that of that 85%, most do not come anywhere near it as most travel significantly below it, clustering around the mean, which is why I mentioned the 1 in 20 ratio.

Once again, speed limits (in the UK anyway) are not ‘set’ according to any percentile speed, but if they were, why do you think the 85th is a ‘suitable value’ anyway if so few do it? Why not just above the mean? The 85th%ile seems to have erroneously been set in stone a few decades ago and only recently have the authorities re-examined its supposed significance.

Hugh Jones, Cheshire0

Hugh

I never meant “sample 20 vehicles”, just using 17 in 20 as another way of expressing 85%.

But you clearly do not understand the 85%ile speed which, by definition, is the speed that 85% of drivers DO NOT EXCEED (or, if you like, 15% of drivers exceed).

I would like to know what your precise definition of the 85%ile speed is, if it is not the same as mine. The number travelling at precisely the 85%ile speed is irrelevant – it is the number travelling above/below it which defines it and, by convention, considers it as a suitable value for a speed limit.

Eric Bridgstock, Independent Road Safety Research, St Albans0

Also remember that when mean speeds are measured then they also apply to the whole length of road. Simply measuring the average speed of vehicles at the most free flowing point does not accurately produce the mean speed for the road. You need to factor in the deviation from this that you would expect where the road has obstacles, natural calming points, junctions, crossing, etc.

Rod King, 20’s Plenty for Us0

Correction Eric:

Your rather naïve calculation simply can’t apply to 20 vehicles as it is too low a sample to give the full range of actual speeds on a road and you’d be lucky to capture anyone travelling above the 85th%ile speed within that small sample. They simply wouldn’t show up at all, so you see how they couldn’t figure in your 17/3 split.

If however, you were to measure the free-flowing speeds of 100 vehicles, you would see what I mean and see also how small the percentage of vehicles travelling at the 85th actually is, (typically 5%, hence 1 in 20) compared to those travelling at the average, which is why the latter is the better indicator of optimum speed. I think you’re misunderstanding my use of the 85th%ile as a spot-speed, rather than the 85th when used as an indicator of a range of speeds, below this spot-speed.

Hugh Jones, Cheshire0

Correction Hugh.

85% is 17 in 20, leaving 3 (three) in 20 exceeding it. And the 85%ile speed has long been recognised by road safety engineers as the optimal speed for setting the limit.

Eric Bridgstock, Independent Road Safety Research, St Albans0

People often drive the same road time and again, often twice every day (the commute). They obviously will not select the exact same speed every time (if they did they would not be adjusting for pedestrians, cyclists etc).

Therefore, when the average (50%ile) is, say, 31mph, most drivers will have safely driven both above and below that speed, making the average 31 mph.

Speed limits set at 50%ile therefore makes the safe actions of the responsible majority illegal. The issue (as an engineer) though, is “What effect does this have on road safety?” There are good arguments that deaths and serious injuries will reduce and good arguments that they will increase. To find the overall effect, we need to implement speed limit changes in scientific trials.

Enforcement adds another confounder therefore, in the absence of scientific trials, we can never be certain what factor is causing what effect.

Dave Finney, Slough0

Re-the last two paras: Speed limits are not ‘set’ at the prevailing speeds of traffic and never have been. We have national speed limits which are applicable by default, dependent on the road type. The limits should influence the speeds – not the other way around.

If a Highway Authority is considering subsequently applying a limit other than the default limit, it will certainly take note of the existing speeds, but if we are to assume that the speeds ‘chosen’ by the majority of drivers represents a good balance between reasonable progress and an acceptable safety margin, then this is invariably centred around the average speed anyway and definitely not the 85th%ile which is typically the speed of only 1 in 20 vehicles and the safety margin then becomes compromised.

I recall the ABD once pushing for speed limits to be set by Police ‘experts’ rather than the HAs – does this mean they have changed their mind?

Hugh Jones, Cheshire0