Speed does not kill. It never has, and never will.
What causes collisions is a failure to stop or manoeuvre in time to avoid another object - be that moving or stationary, another car, pedestrian or a tree. This is true of 100% of impacts.
Excessive speed is a risk factor for this, but actually a relatively small one - some studies publish statistics for speed being a 'significant factor' in a collision, but oddly there are no statistics for accidents where the speed in question was actually above the limit. (As opposed to 'it's foggy, should have slowed down)
As to your premise:
"If the cameras could be hidden but everybody knew there were ten of them then only a complete fucktard would break the limit on that stretch of road."
How would 'everybody know there were ten of them'? Visitors wouldn't know about *any* of them, but only see the (one or two) white/black traffic enforcement camera signs. (Maybe - there's no 'end of zone' marker, so...)
The 'bright yellow' colour of the cameras is quite amazing - now you only need a bright yellow box on a stick to get all the sudden braking and speedometer-watching that a genuine camera would have created.
As for a coherent argument:
54% of scenes of collision gave inattention/distraction as a significant cause
28% of collisions gave speed, tailgating, being in a hurry and aggressive driving as significant.
- Note the deliberate clustering of Speed with Aggression, tailgating etc.
Only 13% of collisions considered 'Excessive Speed' significant, and as mentioned above this includes driving below the posted speed limit in bad conditions - which no camera could ever note.
I could not find any notation of which collisions involved actual illegal speeds, and the study even states that there is almost no data given on the estimated speed so it's impossible to find out how many collisions partially caused by excessive speed involved illegal speeds.
The only speed data given was the posted speed limit.
Source: Pages 30-36: http://www.dft.gov.uk/pgr/roadsafety/research/rsrr/theme5/roadaccidentdatabase.pdf (February 2008)
A quick analysis of the above study implies that the maximum for collisions partially caused by traffic enforcement cameras is more than those caused by Alcohol, so it should be worthy of study.
However, I couldn't find any research at all into the numbers of collisions likely to have been caused by the distraction of speed cameras. I found lots of claims of this, and quite a bit of circumstantial and anecdotal evidence, but no real studies.
So, why have there been no studies? I'll leave that one for you guys...