RogerBW's Blog

There's one bit of the canonical route to Cambridge where I have a choice: from the M25, should I leave at junction 21a, 22 or 23 to reach the A1(M) northbound? If in doubt: data! (Images follow.)

First of all let's zoom in on the red/black speed map to show what I'm talking about. The approach is from the left (with the little-used M1 option turning away off the map). 21a is the first exit from the M25 (and you can see that I've sometimes diverted to refuel there); 22 is nearer the centre; 23 is at the bottom right and obviously much longer.

The traces vanish at the top right because I enter the Hatfield Tunnel (under "Comet Way" on the map).

By taking each outbound trip and selecting the segments that are not common to all three routes, i.e. from where they start to split just before J21a until they merge again just north of A1(M) J4, I can extract speeds and distances. The distances are, as one would expect, fairly consistent; the times are rather less so.

Granted, some of the long times at 21a are caused by my stopping for fuel. But not all of them. Sometimes the traffic's just heavy, and sometimes I've been annoyed enough to bail out of the queue and head round by another path. (Obviously I choose whatever I think is most likely to be the best route based on what I know of road conditions; sometimes I've come off at 21a because there were major road-works before 22, but chances are that lots of other people did too, and that'll have slowed things down.)

The J23 route is certainly the longest, but also the most consistent; it's the fastest on average, and clearly the best choice for speed. On the other hand, the J21a route is shortest (not surprising as it's the most direct), and occasionally the fastest, but the most inconsistent, being prone to major delays on the approach to the Park Street roundabout from the south and on the approach to the London Colney roundabout from the west.

The J22 route is on average 110 seconds slower than J23, but also more than three miles shorter. It's probably a good compromise.

What I find particularly interesting in this is how much more variation the shorter routes show than the longer ones.