This is an issue where constraints of physical topology meets constraints for the management of network cables, meets technical constraints on the use to which cables can be placed.
Where cabinets were put in many years ago, the whole point was to add a point of flexibility, where the E-side cables (up to 1,000 pairs, and pressurised to keep water out) were terminated, where the local distribution cables (10, 25 or 100 pairs) fan out to the surrounding streets, and where the two could be connected or (in the event of a fault) changed. It also marks a change in philosophy: There is very little desire to ever break into the larger E-side cables, or to add joints, for any reason. The smaller D-side cables don't show this, with plenty of joints existing.
In physical terms, coverage doesn't seem to be "circles" surrounding the cabinet. Instead, it seems the coverage area can be better thought of like a teardrop, with the sharp end pointing at the exchange, and the point where the cabinet is located; most houses connected to the cabinet will be further away from the exchange ... all the way up to the point where a new cabinet is sited, and takes over the coverage further out.
The natural consequence is that the houses at the far edge of the coverage for one cabinet will likely be closer to the cabinet at the start of the next. That was never an issue in voice days, of course - so no need to play with the layout of the network.
It is hard to draw this on a map, but the next best example I can think of comes from paving patterns. Something (but not exactly) like this:
Put the exchange at the bottom of the image, and a cabinet at the bottom of each fan, and you get to see something of the pattern.
The cabinets in question are new installations and properties connected to cabinet 2 follow the same copper route as mine from the exchange. I can see each connection as it winds its way down the road.
Cabinets installed now won't quite follow the same physical pattern, but are constrained instead by the layout of the existing cabling.
Both cabinets 1 and 2 will have respectively been sited at a location that all the existing cables (for the homes attached) already run through - whether they are the large (multi-hundred pairs) or smaller (10,25,100 pairs) cables. The cabinets will, as much as possible, just break into cables that run past that location, with as little new cabling needed as possible.
Your current cable will, before the cabinets were installed, already have physically existed where cabinet 1 was sited. Crucially, your line won't have run to the location where cabinet 2 is now sited.
To do what you ask, and re-route you through cabinet 2 needs a considerable amount of extra copper cabling added. Cabling that doesn't exist today, through ducts that probably aren't big enough. Quite a lot of rework.
It may have been a quicker design to meet connection targets.
The main constraints placed on BT, by the local BDUK projects, are financial: to offer as much value for money as possible (ie be as cheap as possible), and to get as many homes above the 25Mbps threshold as possible.
To do what you ask would take time, and would cost a lot of money, but wouldn't likely gain much additional coverage of superfast speed (especially if long-range VDSL (*) becomes an option on your cabinet).
Meanwhile, spending extra money on your line would be likely to deprive someone else of an upgrade of any form.
If you were spending your own money for this, that would be fine. But the council, and BT by extension, have other priorities.
Does not mean we have to accept it.
There is a technical reason too. It would be wrong to allow VDSL2 signals down one cable from FTTC cabinets that are located in different places, and are different distances away. Crosstalk could obliterate the DSL service (including ADSL service from the exchange).
This would become an issue for your line if you happened to share a cable (at some point) with a property that was closer to cab 1, and stayed connected there.
The likelihood of this happening depends on the layout of all cables in your locality - not just the one you are currently connected to, but any cable that you might swap to in future (for example, as part of a future repair).
I note that you make a regular reference to the fact that all the copper, for all houses, follows one route. You might be thinking that there would be little to prevent you from piggy-backing on the copper in the wrong direction out to cab 2, and then back toward the exchange again. Unfortunately, it is this technical restriction - the requirement to not interfere with other services sharing the copper cables - that makes this impossible. You would need BT to install 500m of brand new copper ... possible more ... and then keep track forevermore of the anomoly that this particular cable is being used "backwards".
(*) Long-range VDSL looks like it will change the range of superfast speed from today's 1.2km to perhaps 1.6-1.7km, and maybe further in the future.
This graph
(courtesy of TBB) shows what BT thinks right now for both the standard 17a profile and the (being trialled) LR-VDSL:
