You don’t hear about it much, but the Highway Capacity Manual does have a definition for pedestrian LOS:
Pedestrian LOS is determined based on the average delay per pedestrian (i.e., wait time). Pedestrian delay is calculated using two parameters: cycle length and effective green time for pedestrians. The HCM 2000 recommends estimating effective green time for pedestrians by taking the walk interval and adding 4 seconds of the flashing DON’T WALK interval to account for pedestrians who depart the curb after the start of flashing DON’T WALK. The equation for calculating pedestrian delay based on equation 18-5 of the HCM 2000:
Where: dp = average pedestrian delay (s) C = cycle length (s) g = effective green time for pedestrians (s) Pedestrian LOS thresholds at signalized intersections.
LOS
Pedestrian Delay (sec/ped)
Likelihood of Noncompliance
A
Low
B
≥ 10-20
C
> 20-30
Moderate
D
> 30-40
E
> 40-60
High
F
> 60
Very High
Using this metric, one can predict the amount of jaywalking. Intersections with “A” or “B” pedestrian LOS will have a high degree of compliance. Intersections with “C” or “D” LOS will have a moderate amount of jaywalking. Intersections with “E” or “F” LOS, well, can expect to have lots of jaywalking.
Because I don’t have a social life, I spent the better part of last evening with a stopwatch timing the signalized intersections and observing pedestrians in my Berkeley neighborhood. Supposedly, this is a pedestrian-friendly city and yet pedestrian LOS were all within the “C” and “D” range — meaning one can expect a moderate amount of noncompliance. Jaywalking is fairly common, so the pedestrian LOS chart models reality pretty well.
As in most cities, the pedestrian signals in my neighborhood were upgraded to the beg-button type. Pedestrians must press a button, then wait (as long as a full cycle) to get the walk light. What if instead these beg-buttons were Pedestrian Priority buttons? The walk light would come much sooner, and LOS would move into the A” or “B” range. There would be less jaywalking, and a much improved pedestrian experience.
The Claremont/62nd intersection has pedestrian LOS of “E”. Everyone jaywalks here.
Systemic Failure 
Thank you for this formula. I will do some of my own timing. I live the suburbs so lots of beg buttons. Hopefully I can make the local transit groups aware of this.
“What if instead these beg-buttons were Pedestrian Priority buttons? The walk light would come much sooner, and LOS would move into the A” or “B” range.”
True, but wouldn’t that disrupt signal timing from intersection to intersection?
It’s not like signal coordination actually works when the blocks are as irregular in length as the ones in Oakland are. There is a very small range of block lengths and speeds where signal progression can work, and outside of that it’s delusional to think that it does any good for vehicles, especially at an intersection like this one where pedestrians are constantly present. The beg buttons are there to try to squeeze out a tiny percentage of increase in vehicle efficiency at the expense of massive inconvenience to pedestrians.
There are maximum phase lengths. So if the green duration max in a phase is 60 seconds and the walk phase is 30 seconds (walk-flashing don’t walk), there’s no reason if someone hit the button in second # 20 that the walk phase shouldn’t go on, especially for arterials. Instead, what you have happen is the green for cars go on and a ped hitting the button in second #5 and not getting to cross until second #120 (for a typical 120 second phase, as is often found in suburban areas). No one follows this and everyone crosses on the parallel green, although that is technically illegal in many states.
Sometimes you DO want cars to lead a few seconds before peds walk, usually in areas with high turning movements, and the ped head is don’t walk until that period ends. But without high pedestrian compliance, people start walking on the parallel green, and no one gets to make their right turn on the green ball. The counter is to have right arrows to grant this extra phase to signal that cars have an exclusive phase, but adding right arrows is costly.
What’s clear is that engineers are willing to do “substandard” facilities for walking and biking, while resisting “substandard” facilities for drivers
Interesting how the manual assumes that people will act in ways that are technically illegal, which could cost you, I believe, over $200.
I wonder– do they make the same assumptions about drivers? When they calculate the efficiency of a 35 mph road, do they assume that most drivers will be speeding?
Yes, they do make the same assumption about speed limits. Under the so-called 85% rule, the posted speed limit is dictated by the 15% fastest drivers. If more than 15% drivers are violating the posted speed limit, then the speed limit is unenforceable. Note that this is in contrast to jaywalking laws — if more than 15% of pedestrians are jaywalking, they can still get ticketed.