At the start of the month, when 20 ticked over to 70, I suggested three numberplate spotting games. Three different sequences to try to follow on new-style registration plates, for no particularly good reason other than to pass the time while out and about.
» Alphabetical By Area Code: A~, B~, C~, D~, E~, F~, G~, etc
» Sequential London Identifiers: LA, LB, LC, LD, LE, LF, LG, etc
» Reverse Chronological: 70, 20, 69, 19, 68, 18, 67, etc
I've now tried playing all three and can confirm that one of the games works really well and the other two very much don't. I was intrigued as to why, so this week have undertaken mass collection of numberplate data to try to understand.
Alphabetical By Area Code
This is the game where you try to spot a plate beginning with A, then a plate beginning with B, then C and so on. Each letter represents a different part of the country so A is for cars registered in East Anglia, B for Birmingham, C for Cymru and so on. You can see a full list of area codes here.
Alas this is one of the games that doesn't really work. Only 19 of the 26 letters of the alphabet are in regular use as area codes, so if you try working your way through this particular alphabet you will get repeatedly stuck.
• Codes I and Z are never used (because they look took much like 1 and 2).
• Codes Q and X are only used on vehicles purchased tax free for export, hence particularly rare.
• Codes J, T and U have no meaning whatsoever, but can still be purchased as part of a personalised plate, so are very rare too.
Experiment 1 - first letter
I stood beside Stratford High Street long enough to watch 571 vehicles go by. I tallied the lot, I didn't wait to do them in order. I wondered what the overall distribution of first letters might tell me. Here are my results as a graph.
There's a lot to unpack here. Numberplates starting with L were a lot more common than any other letter - 22% of the overall total - because L is for London. In second place was E for Essex with 11%, and in third place K for Herts/Beds/Bucks/Northants. I suspect if I'd stood in a different part of London I might have seen a lot more G for Kent or R for Reading. The least common regional plates were C for Wales and V for Worcester because they're a long way away, and O for Oxford because it's only small. I saw more Ss than expected because a lot of our local double decker buses were registered in Scotland.
At the lower end I was amazed to see a Q among my 571 vehicles because they're ridiculously uncommon, and far less surprised not to see an X. Of the three letters used only on personalised plates I only spotted a J, this on a white van owned by J Kent Roofing Specialists from Basingstoke. I could probably have stood there for most of the morning and not seen a T or a U. And this is why Alphabetical By Area Code doesn't really work, the letters are very much not evenly spread.
Sequential London Identifiers
This is the game where you have to work your way through all the initial two-letter pairs in your local area, which in my case means starting with LA and ending with LY. It's crucial to know which pairs were never released, which in London's case is LI, LQ and LZ, and not to waste your time waiting for those.
I know from my previous experiment that around 20% of vehicles in London have L codes of some kind, so there ought to be plenty of spotting opportunities. But when I tried playing the game I got very stuck around LE to LH, wasting days looking without getting any further forward, so something untoward was evidently going on.
Experiment 2 - first two letters
While walking around London I spotted 636 vehicles whose registrations started with L. I tallied the lot, I didn't wait to do them in order. It took a couple of days. I wondered what the overall distribution of area codes might tell me. Here are my results as a graph.
The two-letter codes are a lot more equitably spread this time. Most of the 23 possible codes appeared a reasonable number of times, so if you were hunting for (say) LA then LB then LC then LD you ought to be in luck. But two codes appeared a lot more often than the others, namely LV and LX, with LV easily the champion with 12% of the total. Meanwhile two codes, namely LE and LU, hardly appeared at all. I was so surprised when a single LU finally turned up that I went over and took a photo of it. But I didn't spot LH at all, indeed still haven't, despite the fact it has allegedly been released.
When the new numberplate system was launched in 2001 London had three separate DVLA offices. Wimbledon issued codes LA to LJ, Stanmore used LK to LT and Sidcup took LU to LY (although these offices closed in 2013 and everything's now run from Swansea). These sub-areas may help explain the overabundance of LV and LX in east London, but not the mysterious scarcity of LE, LH and LU. All I can say is that Sequential London Identifiers doesn't really work because there are too many exceptionally rare occurrences along the way.
Which brings me to the game that does work. Counting back from 70 (September 2020) to 51 (September 2001) works rather well, so long as you manage to keep remembering which number comes next in the sequence. There are just enough 70s out there now to start you off, and still just enough 19-year-old 51s to allow you to finish.
Experiment 3 - age identifier
I stood beside Ruckholt Road long enough to watch 417 vehicles go by. I tallied the lot, I didn't wait to do them in order. I wondered what the overall distribution of age identifiers might tell me. Here are my results as a graph.
At last a genuine pattern, and one that makes economic sense. Most vehicles on the road are fairly recent, hence the hump at the beginning, while there are fewer older vehicles the further back you go. I saw more 68s than anything else but the other codes from 2016 to 2020 weren't far behind... indeed 43% of the overall total came from the last five years. Meanwhile a quarter of the total came from 2011-2015, and this dropped to just 5% by the time I got back to 2001-2005.
The surplus bar at the end of the graph includes all the vehicles registered before the current system was introduced, plus all the personalised numberplates that sit outside the system. Altogether they made up about 7% of the total, or one in every 15 vehicles. On this particular dual carriageway they split roughly fifty-fifty between old and personalised, but you'd get a very different split if you surveyed traffic in Barking or Belgravia.
And this is why, if you want to play a sequential numberplate spotting game, it's much better to focus on ages than area codes. Always do your research before undertaking a potentially pointless activity. I'm sure you'll all be rushing to have a go this weekend.