Safe Speed home
About Safe Speed
The Big Lies they tell to support speed cameras and speed reductions
If the policy really worked, why would they need to keep embellishing the truth?

Fiddle, fabricate, fudge and fib.


We are well and truly sick and tired of hearing the same old lies and misleading statements trotted out again and again and again. The Scamera Partnerships apparently love the lies, and we'd be surprised to find a Scamera Partnership site that didn't contain a few of them. Ask yourself: Why do they need to fiddle, fabricate, fudge and fib? 

The one third lie


  • One third of accidents are caused by speeding
  • Speeding kills 1,100 every year
  • One third of accidents have excessive speed as a cause or contributory factor
  • Speed is a factor in one third of crashes

Not known - it's probably just spin. We've never seen any research or scientific evidence to support it.


There's a shortage of actual UK data published about "excessive speed as an accident causation factor". In the UK and in other countries, it has long been the convention to gather data for accidents where the speed was excessive for the conditions without reference to the speed limit. Many Police forces gather accident causation factors, but few publish them in public.

West Midlands Accident Review

A notable exception is West Midlands Police who publish an annual Accident Review. West Midlands figures for excessive speed accidents are as follows:

non-pedestrian causes pedestrian causes
excessive speed fatals serious overall fatals serious overall
1999 9.30% 6.25% 0.46% 2.94% 0.72% 0.46%
2000 15.79% 4.94% 0.11% 0.00% 0.76% 0.11%

excessive speed definite factor any factor
2001 3.27% 4.05%



TRL 323 was an in-depth investigation by the Transport Research laboratory into a "New system for recording contributory factors in road accidents". It involved 8 UK Police forces on a trial basis. One test of effectiveness was to be: "Would the 8 forces yield similar results?". Similar results were obtained, indicating that the system was operated correctly and consistently.

There's much talk about TRL323 and we're planning a detailed page discussing the report. We have a copy (which we are not allowed to post for download) and it's very clear indeed to us:

  • that the report is good and clear
  • that the system was defined properly and worked correctly
  • that skilled police officers made valid and valuable judgements about each accident
  • that consistency across the 8 regions means that the results were true and unambiguous
excessive speed number of factors percentage of factors against 2,795 total accident reports

[these percentages are not calculated or included in the report]

definite factors 126 6.0% 4.5%
all factors 424 7.3% 15.2%

TRL323 allowed factors to be classified as "definite", "probable" and "possible". Where the table above lists "all factors", that includes excessive speed as a definite, probable and possible factor. It follows then that the "all factors" figures include cases where the excessive speed was only a possible factor.

These TRL323 factors are contributory factors not causation factors. In fact, TRL323 recognises that most accidents have complex causation and cleverly avoided the pitfall of trying to make reporting police officers identify the primary causation factor from a number of contributory factors. This approach gives freedom for the reporting police officers to record observable details without have to make fine value judgements about which factor actually caused the accident.

TRL323 general methods are widely respected as the best way to gather contributory factor information, except perhaps by those who have an agenda.

Which brings us to the last column in the table above. The calculated percentages shown in the last column do not appear anywhere in TRL323, but latterly those with an agenda have suggested that the true excessive speed value from TRL323 should have been 15.2% based on a calculation that the impartial authors considered unnecessary. If anyone quotes the 15.2% figure from TRL323, you will know to point out that it includes "possible and probable" excessive speed accidents as well as definite ones.

new We would like to present the split in TRL 323 between "possible, probable and definite" factors in the "excessive speed" catagory:

new table

excessive speed factors - TRL323 number of times recorded percentage of excessive speed factors percentage of 2,795 accident reports
possible 68 16% 2.4%
probable 163 38% 5.8%
definite 126 30% 4.5%
confidence not reported 67 16% 2.4%
all 424 100% 15.2%

We might make some guesses about how many of those factors represent a genuine excessive speed accident. We suggest a reasonable guess could be based on: 

30% of the possible factors (20) + 75% of the probable factors (122) + 75% of the confidence not reported factors (50) + 100% of the definite factors (126)  = 318.
318 factors from 2,795 reports suggests a new percentage of excessive speed accidents of: 11.3%. We're still a very long way indeed from "one third".


The DfT have accident causation data which they are not publishing. We have written to request information but had no reply. This document contains: "Fifteen police forces are currently supplying contributory factor information to DfT, using the proposed national form, on a voluntary basis."

We need that data! It's of enormous public interest. The DfT committee is called "Standing Committee on Road Accident Statistics (SCRAS)" The former senior person (Peter Wilding) still mentioned in the documents has left. The senior person now is Alan Oliver. If you feel like chasing this up, please do so.

new. Finally some of this data has been published on the 30th March 2004. It is contained in a report called "review of the contributory factors system" and can be downloaded from the DfT (here). This is by far the largest published survey of accident causation data in the UK in recent decades and the data was compiled with the excellent "TRL323" system. The conclusion is that excessive speed was coded as "possible,  probable or definite" for 7,600 out of 60,797 accidents. That's 12.5%, which is no where near "one third". As usual this catagory includes inappropriate speed within the speed limit - see below.

The top 7 contributory factors as reported were:

  • Inattention:  25.8%
  • Failure to judge other person's path or speed:  22.6%
  • Looked but did not see: 19.7%
  • Behaviour: careless/thoughtless/reckless: 18.4%
  • Failed to look: 16.3%
  • Lack of judgement of own path: 13.7%
  • Excessive speed: 12.5%

Hampshire County Council

Hampshire County Council helpfully publish this pie chart: (click here)

And this table:
Factors Number of casualties Calculated percentage
Driver error 5704 65.6%
Excessive speed 993 11.4%
Driver impairment 415 4.8%
Actions of pedestrians 161 1.8%
Road conditions 320 3.7%
Vehicle defects 112 1.3%
Weather conditions 382 4.4%
Other 605 7.0%
Total 8692 100.0%

With such a large "driver error" category, it's somewhat hard to see how excessive speed would not be a "driver error", and if it's worthy to separate out excessive speed from other drivers errors, then why not separate out other factors too? Still good on Hampshire - they are doing more with the info than most. 

DfT Sleep Related Accident Study

This document  (dft_rdsafety_pdf_504599.pdf) examines accident causation with particular reference to sleep related accidents. However, accident causation data is included and 19 out of 556 accident reports were due to excessive speed. That's 3.4%. (see table 2.16). Once again the low incidence of excessive speed as an accident causation factor is highlighted.

Durham causation figures from press reports

This Daily Telegraph article (click here) (or here)contains the following words attributed to the Chief Constable of Durham - Paul Garvin:

Mr Garvin explained: "I actually believe in casualty reduction and trying to make the roads safer but, having looked at the accident statistics in this area, we find that if you break down the 1,900 collisions we have each year only three per cent involve cars that are exceeding the speed limit. Just 60 accidents per year involve vehicles exceeding the speed limit.

"You then need to look at causes of these 60 accidents. Speed may be a factor in the background but the actual cause of the accident invariably is drink-driving or drug-driving. Drug-taking is becoming more of a problem. In 40 per cent of fatal road accidents in this area one or more of the people involved have drugs in their system."

Many accidents were caused by fatigue, although one of the most common causes of crashes was the failure of drivers to watch out for oncoming vehicles when turning right. "The cause of accidents is clearly something different than exceeding the speed limit and we ought to be looking at those other factors," Mr Garvin said.

Summary of the truth

Every single authoritative reference makes excessive speed a small accident causation factor. (If you think you know otherwise, please respond to our statistics challenge on the front page of the web site) The highest of all the figures is 15.79% (West Midlands, 2000, fatals only). The highest figure quoted in TRL323 is 9.2% (in Staffordshire, table 4). Clearly none of the figures is anywhere near "one third". So that's a lie then.

The Secondary One Third Lie


We're expected to believe that "one third" of accidents will be affected by enforcement.


Not known - it's an ongoing implication.


The vast majority of excessive speed accidents are entirely out of reach of speed enforcement. There are three important ways that the accidents are out of reach.

Firstly, a large proportion of excessive speed accidents take place entirely within the speed limit. We've never seen UK figures for this proportion, but research in Canada put the proportion at 2/3rds. For want of UK based figures we'll assume that the same applies here for the purpose of estimation. (update: see Avon and Somerset figures - we found some!)

Secondly, there are "lawless" drivers who are functioning outside of the law for one or more of the following reasons:

  • Joyriders in stolen cars
  • Emergency services drivers on emergency call
  • Reckless drivers who have abandoned all caution
  • Reckless drunk drivers
  • Drivers attempting suicide by deliberately crashing
  • Drivers with false number plates or incorrect registration details who know that they cannot be traced
  • Criminals escaping the scene of a crime
These drivers are very highly represented in the records of excessive speed accidents - they are very likely to represent more than half of all excessive speed accidents, and much more than half of all excessive speed accidents that take place over the speed limit. We've called it half as a conservative estimate. (We're excluding normal responsible motorists simply exceeding a speed limit from the group we have termed "lawless". Some might wish to term them lawless but we don't.) 

Thirdly, no system of speed enforcement is going to prove 100% effective at preventing speeding. The authorities clearly understand this because they plan to issue more speeding tickets every year. They pay lip service to the claim that they intend to change behaviour, but the reality is that there will always be speeding drivers - especially in safe circumstances. It's highly unlikely that blanket enforcement with draconian penalties would prevent anywhere near half of the cases of speeding. So we've called this half as well as another conservative estimate.

So it follows that we should calculate the size of the group potentially affected by speed enforcement as follows:

[number of excessive speed accidents] x 0.333 [over the speed limit] x 0.5 [law affected group] x 0.5 [speed enforcement affected group]
Here's an example calculation: 3,400 fatalities x 11.4% x 0.333 x 0.5 x 0.5 = 32 potentially affected fatalities.

By our estimation, just 8% of excessive speed accident are potentially within reach of speed enforcement. But it doesn't stop there - speed enforcement itself brings dangers which far more than offset the possible benefits. We maintain a list of the possible and probable dangers of speed enforcement. (click here) Clearly there is an urgent need to improve the quality of the figures used here - they are intelligent and conservative estimates - but in the complete absence of applicable UK research they will have to do for now. Update: Figures from Avon and Somerset tend to confirm our suspicions. (click here)

newSection summary

So just to clarify this 10% or so of accidents are due to excessive speed, and within that 10%, some 8% are potentially within reach of speed enforcement.  8% of 10% is 0.8% of the total of accidents. With 3,400 fatals in the UK annually, we're talking about just 27 fatalities.

There's one more important effect that hinders the potential benefit of speed enforcement in reducing excessive speed accidents. Speed cameras have their greatest effects where road dangers are smallest. It's easy to find people exceeding the speed limit on the open road away from hazards - but exceeding the speed limit in the absence of hazards cannot be dangerous. Where hazards are present, drivers slow down anyway - this is the main way they avoid having accidents. 

We need drivers who slow down where the danger is greatest, not where the danger is smallest. This is where emphasis on speed limits misses the point, because such emphasis tends to affect areas where the danger is least. On the other hand publicity about using speed wisely and safely would emphasize reducing speed where danger is greatest. 

Further Reading

OneThirdLie web site (click here)
SafeSpeed: Emails exchanged with the DfT and the TRL (click here)
SafeSpeed: More emails (click here)
SafeSpeed: Letters to Robert Gifford of PACTS (click here)
SafeSpeed: Accident causation in Grampian (click here)
RingRoad web site: The West Midlands Accident Review (click here)
SafeSpeed: Strange TRL claims: (click here)
SafeSpeed: Avon and Somerset figures (click here)
SafeSpeed: Tayside causation figures (click here)
Also note our "statistics challenge" displayed on the front page of the Safe Speed web site since 30th September 2002.

The one mph lie


For every one mph reduction in speed accidents reduce by 5%. Etc.


Finch 1994 (not available on the web)
TRL421 (copyright report, not available for public download)
TRL511 (download pdf)


It's utter fabrication, made of out bizarre manipulations of statistical models which start off proving that faster roads are safer and are then bent to show the opposite. Once the purported link between speed and accidents has been concocted, the reports authors make an enormous leap of faith by assuming that the relationship is causal, then declare that a reduction in speed of 1 mph will reduce accidents by 5%. 

We know the authors of TRL421 and TRL511 cannot be trusted because they lied about the meaning of their research - within the research itself. It's completely amazing. But if you check the references listed below under further reading it's completely obvious.

TRL: we are so ashamed of you!

Further Reading

ABD : About Finch 1994
SafeSpeed: TRL421
SafeSpeed: TRL letters

clarified The thirty-five mph lie


At 35mph you are twice as likely to kill someone as you are at 30mph. (official example)

At 20 mph 10% are killed
At 30 mph 50% are killed
At 40 mph 90% are killed
At 35 mph it takes an extra 21 feet to stop
Kill your speed or kill a child


Ashton and Mackay 1979


It's a funny sort of lie this one. The figures and the claim are broadly true, it's just that free travelling speed and impact speed are almost unrelated to one another. Ashton's figures relate to vehicle speed at impact - but we're expected to assume that they refer to free travelling speeds.

We know from official figures from 2001 (for example) that 0.7% of child pedestrians injured in accidents with cars died. And we know that in 2001, 65% of cars at sample sites were exceeding the 30 mph speed limit. But the 0.7% of child pedestrians killed were in all speed limits. Clearly we would have expected more than half to be killed if the implied claim were true.

Further Reading

SafeSpeed: Kill your speed
SafeSpeed: Proof
SafeSpeed: That Ad

The thirty-five percent lie 


According to government research accidents at speed camera sites were reduced by 35%.


Department for Transport:  "A cost recovery system for speed and red-light cameras ~ two year pilot evaluation" (click here) Report prepared by: Adrian Gains and Richard Humble, PA Consulting Group; Professor Benjamin Heydecker and Dr Sandy Robertson, University College London.


The document is a travesty of science. See our letter to Professor Heydecker. The headline claim is completely worthless. No conclusion about camera effectiveness can be drawn from the report because insufficient data is available to eliminate massive error sources. The only useful conclusion that can be drawn is that the authors of the report should not be trusted.

But they hired skilled folk to write the report. It doesn't contain any lies. Nevertheless, the headline conclusions are entirely misleading. It's very likely true that 35% fewer accidents occurred at speed camera sites. The question is: Why?

Was it because of the beneficial effects of speed cameras or or because of something else? On proper investigation, it is completely clear that all of the following are included within the 35% claim: 

  • County level regression to the mean benefit illusion 
  • Site level regression to the mean benefit illusion
  • Benefit of other (i.e. non camera) safety treatments 
  • Traffic reductions at camera sites 
  • Accidents that were too distant from the camera to be affected 
  • The actual effect of the cameras 
Until someone properly puts percentages to all these effects, nothing whatsoever can be concluded about the effectiveness of the cameras from the report. It's probable that site level regression to the mean effect is strong enough by itself to produce the entire benefit. The others are likely to be worth just a few percent each. The actual benefit of the cameras is probably negative even at speed camera sites.

new: Professor Heydecker admitted on the BBC radio show "More or Less" that no regression to the mean compensation had been applied.

Further Reading

SafeSpeed: Gambling with Cameras
SafeSpeed: Regression to the mean
SafeSpeed: Letter to Professor Heydecker.
new SafeSpeed: More or Less

Speed cameras save lives

It's everywhere. "Speed cameras save lives". Perhaps they think if they say it often enough it will become true. Perhaps they know if they say it often enough that people will believe them. Perhaps they think it's "obvious" that speed cameras save lives.

None of that helps. There's no evidence anywhere that speed cameras save lives. We have a statistics challenge on the front page of the web site inviting evidence to support the claim. We have not been offered any. We wrote to Richard Brunstrom asking the question and had no useful reply. We wrote to the spokesperson for "National Safety Camera Liaison" (Susan Beck) and had no reply at all.

Oh, and our national road safety performance is much worse since we started installing cameras.

Further Reading

SafeSpeed: Letters to Richard Brunstrom
SafeSpeed: Letter to Susan Beck
Safespeed: Statistics Challenge II

Speed Kills

It's everywhere. Like all the best lies, there's an element of truth to lend plausibility. On the road driving too fast for the conditions is dangerous and can kill. Exceeding a speed limit in itself is never dangerous and cannot kill. It may seem obvious that the faster you go the greater the danger, but that's simply not true either - faster roads are safer and the fastest roads are the safest of all. 

But no road is faster than any other road - not until a driver decides it is safe to accelerate to a faster speed. This is a critically important point - there is always a driver in the loop setting a speed he feels comfortable with. If he perceives a danger he will slow down.

Never forget that all a driver controls is course and speed. To suggest that speed is dangerous threatens to remove and restrict drivers control of speed, while speed is the drivers primary risk management tool. In fact speed is so many different things that it's hardly surprising that "speed kills" means one thing but is easily stretched to appear to fit others.

The main problem with the phrase lies in the precise definition of the word "speed" - some sorts of speed can kill, but that does not mean that exceeding a speed limit is dangerous.

When someone says: "speed kills" the correct response is: "What sort of speed is that then, exactly?"

SafeSpeed: Speed
SafeSpeed: Speeding

The public supports speed cameras

There are two parts to this lie.

In the first place the government and various supporters and vested interests have put huge resources into trying to convince the public that speed cameras are a good idea. It not at all surprising that the non-driving population accepts the claims at face value - why would a non driver question the claims at all? Obviously they have convinced some drivers as well. What a shame they didn't put those resources into a real method for improving road safety. It is highly illogical to suggest that road safety can be improved by lying to the public about specific road dangers.

Part two of the lie involves the survey methods (and particularly the survey questions) which have been designated by the Department for Transport. The questions are quite obviously loaded to elicit the desired response.

With loaded questions and a misled public it should come as no surprise that they are able to claim that public opinion supports speed cameras.

SafeSpeed: analysis of official surveys of speed camera approval (click here)
Admiral Insurance: (independent survey)

new Truth and delusion

So is it a massive conspiracy? How has this come about? Why are so many important people lying about road safety?

In the large part, sincere people with insufficient knowledge of the process of driving have been seduced by three key factors:

The "obvious" idea that slower is safer.

It might be "obvious" that slower is safer, but for slower to be safer, one has to be going slower in absolutely critical places, at absolutely critical times - like immediately before an accident. Going slower on a clear road has no positive safety implication. There's mounting evidence that modern policy is actually leading to more excessive speed accidents as drivers fail to slow when necessary more often than they used to. (click here)

The promise of something that is easy to measure. A road safety "quick fix".

The idea of an automated road safety system has considerable attractions. It's cheap, potentially reliable and can detect and punish dangerous offences 24 hours per day. If only it were so simple. But it's so attractive that we seem to have accepted a poor standard of supporting evidence.

Failing to distinguish between normal responsible drivers and reckless behaviours.

The are appalling high speed crashes caused by reckless drivers sometimes unlicenced, or unregistered or driving stolen cars. Far too often such accidents are equated to the behaviour of a normal responsible motorist exceeding a speed limit. But failing to distinguish is contributing strongly to the "speed kills" obsession.

SafeSpeed: conspiracy


It's absolutely staggering in 21st century Britain that we're being fed lies and propaganda about road safety. Mostly it's well meaning but misguided people who sincerely believe that they can save lives by convincing us not to drive too fast. If the emphasis was on responsible driving and safe speeds for the prevailing conditions it might just work. But while it's all directed at speed limit compliance it doesn't stand a snowflake's chance in hell of working.

Please can we have sensible road safety policies based on truth?


Safe Speed encourages comments, further information and participation from our visitors. See our (forums).

Read about our comments policy (here).

Many pages (including this one) have a specific associated forum topic. You can (view) or (add comment) to the forum topic for this page. Posting in the forum requires simple registration.

Calling for real road safety, based on truth

We have a strict editorial policy regarding factual content. If any fact anywhere on this web site can be shown to be incorrect we promise to remove it or correct it as soon as possible.
Copyright © SafeSpeed 2003, 2004
Created 31/08/2003. Last update 12/05/2004