Would seem this your memory is correct - 4 year experiment which for first 2 years they thought was working far better, but then they realised it was mostly just because people were no longer driving drunk in the afternoons... Main benefit in southern scotland and northern england, worst impact norther scotland:
There have been frequent unsuccessful attempts in the UK parliament over the past few decades to make changes to the current system.
Older readers may remember the three-year experiment launched in 1968, when British Standard Time (GMT+1) was employed all year round.
The clocks were put forward as usual in March 1968 and not put back until October 1971.
The Department for Transport's initial analysis of road casualty data during the experiment suggested more people were injured in the darker mornings, but fewer people were injured in the lighter afternoons.
It estimated a net reduction of 2,700 people killed or seriously injured during the first two years of the experiment.
However, it was recognised at the time that the calculations did not take into account drink-driving legislation passed in 1967, so the Department for Transport eventually re-analysed the data and factored that in. In 1989, it found the data agreed broadly with the earlier estimates.
It suggested that those living in central England and southern Scotland benefited most from the experiment. However, northern Scotland saw a net increase in the number of people killed or seriously injured.
Linkee