Centralised discussion space for TfL plans and cycling in London

I won't give you a full analysis of this, but overall, like most of the 'low traffic neighbourhoods', which as mentioned before, is a very silly name, these are not well-conceived. The officers mostly designing the schemes have very little experience of modal filtering and fail to understand some key principles of how these schemes should be done.

(1) Define cell boundary streets. Ideally, cells shouldn't be too large, or you get aggressive driving when people have to cover quite a lot of distance through the cell. 'Area 2' is much too large. This may be because none of the streets inside it are suitable as cell boundary streets, but it's (counter-intuitively) better to break very large cells up into smaller ones. It leads to much better driver behaviour.

(2) One-way streets are a 100% no-no in any filtered area. There are two main purposes to filtering: increase cycle permeability, i.e. remove one-way streets in the cell, and remove crashes caused by rat-running drivers turning injudiciously on streets bounding a potentially filtered cell. The vast majority of turning crashes, and the vast majority of serious and fatal cycling casualties are caused by such turning behaviour. When you filter, these largely disappear, as people accessing cells turn much more carefully.

If, therefore, one-way streets persist in any 'filtered' area, the filtering is nonsensical. I'm not hugely familiar with the area but rode around there just recently, and all of these one-way streets could, and should, easily returned to two-way operation with the right filtering scheme. This needs to be rethought entirely.

(3) One of the most persistent mistakes people make when filtering is that they filter at the edges of cells. You can see this here in both yellow areas. It's bad for a number of reasons: (a) It increases distances to be driven inside the cell, with the consequences for driver behaviour mentioned under (1) (and actually doesn't do that much to improve turning behaviour where people have to drive into a cell from the other side to get to, say, the streets immediately adjacent to Leytonstone Road); (b) it causes much greater priority to the main street, which has suddenly 'lost' a junction with a side street that drivers would otherwise have to pay attention to, and this is not at all good--except for footway level entry tables, any measures that de-prioritise side street junctions on cell boundary streets, such as the modish 'continuous footways' are very much to be avoided; and (c) it removes a healthy level of side-street interaction and often certain design possibilities, such as the possibility of constructing better-quality pedestrian crossings. Designers are often worried about side-street interaction, but if an adjacent cell is filtered, this is typically very low. Hence: NEVER filter at the edge of cells. It is always avoidable. Filter as deep inside the cell as possible instead (with the above caveat not to make cells too large).

(4) Following on from the above point, the main constraint that local authorities typically encounter is the operation of refuse vehicles. This is a very reasonable point, and must be taken into account. In most cells of an acceptable size, refuse vehicle paths can be defined through openable gates. It should usually be possible to limit their number to one or two for a whole cell. It's actually something that refuse services should welcome, as good filtering causes far less conflict with other drivers. Needless to say, there can also be local buses that run along streets an authority may not wish to consider boundaries. This is the most difficult problem to solve when filtering and will usually require bus re-routing if possible.

(5) Loops are also to be avoided, but can't always be avoided (e.g. for access to public buildings). The most problematic loops are where drivers on a cell boundary street for some reason persistently turn around by driving a loop through adjacent local streets. This is often caused by banned turns. Deciding what to do about these requires local knowledge and observation. Generally, the effectiveness of a filtered area is greatly increased if drivers have to turn around a lot. Designers often try to prevent this by leaving lots of loops in, but that's not a good idea. It can in due course lead to conflict between drivers going in different directions and the usual impetus to make said streets one-way.

(6) You should use the fewest possible filters. I favour four-way filters at junctions like the filter at Northchurch Terrace/Northchurch Road/Culford Road/Lawford Road in De Beauvoir.

Anyway, the most important rule is to filter in the centre, not at the edges.

I've seen one scheme so far, in Southwark, that seems well-thought out, but this one here is basically a full card at bad filtering bingo. All the mistakes that can be made are present.