David Cantrell wrote:
On Sun, Nov 29, 2015 at 01:00:51PM +0100, Eric wrote:

I wouldn't object, I believe the capacity argument

I believe it too, but I also recognise that capacity and throughput are
different things, and it's throughput that matters the most.

Consider a road. The M25, for example. Its capacity is highest when
traffic isn't moving, because the gap between vehicles is minimised.
Throughput is typically highest at a speed somewhere between 40 and
60mph. At higher speeds throughput decreases because the distance
between vehicles is too high.

Now, an escalator is a bit like a road which has a non-zero minimum
speed. If you ignore the people in the current "standing lane", then all
that matters is the speed that maximises throughput in a single lane. Is
it higher than the minimum or not? Is it a viable walking speed? Can the
traffic sustain that speed over an extended period? Remember, the
relationship between throughput and speed is non-linear and involves
lots of uncertainty and unknown parameters which make it hard to model.
It may even be discontinuous.

There's also the length (ie, rise) of the escalator to consider. If it's
very high, fewer people will choose to walk up, so the walking lane will be
under-used, with long gaps. In such cases, having two standing lanes will
maximise throughput. But with short escalators, lots of people will prefer
to walk, so it's better to have a walking lane.