Bike lights when on the pavement.

[mjr]

....As I understand it, it's assuming no focusing optics or even reflector - definitely not anything as fancy as white-vs-red lumens - and that you want 4cd showing to the side as well as "to the rear" as required by the current regs. You could maybe halve that with a reflector, depending on any case law about interpreting "to the rear", ....

Well The 4cd in BS6102/3 only seems to the brightest 'central' point A, [...] I would interpret the 4cd law on sole rear flashers as meaning a lamps minimum centralish intensity is 4cd and relying on the light source not being a laser for the surrounding smaller values specified in BS

Good for you! But I stand by my advice because while we might hope that the 4cd added to the regulations (in 2005 I think) might be interpreted like the measurement in the British Standard, we don't yet know that (and I suspect we won't for a long time, with the state of roads policing), so I'd play it safe and go for 50 lumens if you're buying the sort of rear light whose strength is stated in lumens (which might well suffer from overstating like those "10,000 lumen" front lights), or I guess maybe 25 if you're reasonably confident there's some sort of half-decent reflector or lens on it.

[SA_SA_SA]

....

As the law accepts BS6102/3 as an acceptable sole rear lamp it would seem odd if the direct rule was intended to mean 4cd over a hemisphere (I was a bit surprised it didn't include the intensity around points B and C etc but I suppose that would be too much to expect of a pre-existing non-BS/K lamp whereas knowing the minimum for the brightest central-ish bit probably isn't too much effort). Also, at the time that law(SI)) was introduced weren't LEDs still in the 'five directional epoxy 5mm through-hole red LED on 2AA Cells era' ? --- so the super-wide beam of your pessimistic? interpretation seems unlikely to me: plus I think CJ the then CTC Technical Officer would have commented on such a very large jump in light requirements over BS6102/3. It would destroy any battery life advantage over as BS steady lamp and ... it would also be a much higher requirement than for ECE car tails lamps ....

I thought CJ had suggested to the DFT copying essentials of BS6102/3 intensity levels directly into law, but that they did it minimally by only copying the 4cd central bit...

Lastly, I found that in the Dec/Jan edition of CTC Cycle mag, on page 74, CJ described
the VIS180 as claiming 19 red lumens and that by assuming a uniform 90degress spread* that would give 19candelas in that cone. *but said he expected it to be a bit brighter in the middle.
NB the ECE max for motorcycle tails is 12cd and the max of all a cars tail lamps should be 17cd or less (ie max 8.5cd each for 2)...