[mapt]

Sure, but most people do not walk ten times as fast (or even twice as fast) as their escalators lift them.

Visual inspection of https://www.youtube.com/watch?v=7fNDgwp_w_s suggests that when two people enter the escalator side by side, the walker exits the escalator at about the same time as the stander is 75% of the way up it (~=133% of the rate of speed). The density of the walking side is surprisingly high to me (maybe accounted for by the short span relative to the monsters I'm used to on WMATA MetroRail), about 60% of the standing side. This leaves 80% throughput for the walking side of short-span split escalators relative to the standing side, only a ten percent overall cut in capacity for this one escalator in Singapore.

That number will probably go down the taller the escalator is, as the walking side slows down due to personal stamina (adjusted walking speed and percent willing to walk), and bunching, a phenomena we're familiar with from busses.

[kazinator]

If walking is just 50% faster than riding (i.e. walkers go at half the speed of the escalator), but requires 50% more space, then it's a break-even situation for capacity. And a win for latency.

[csours]

In traffic, it is not about the latency, it is about the jitter.

[PhasmaFelis]

It required quite a bit more than 50% more space, assuming you want to avoid people being constantly kicked in the shins.

[mapt]

Edited post with math.

[kazinator]

It doesn't look like a very busy time, in that video. I count about 10 people standing on one of the escalators while there are 8 walkers going by them. There are some irregular gaps among the walkers; more walkers could easily be accomodated; it's just not busy enough; I don't really see a backlog of people waiting at the bottom; everyone is just marching naturally toward the escalator and getting on.