Why did doghouses take so long to gain popularity

[traffic light guy]

Anyone notice how there are rarely any doghouses that are over 30 years old. I made a topic a couple years ago about the First Doghouse, and I've came to the conclusion that the earliest possible doghouses were 8-inch heads from the 1960s. A few collectors owned 8-in Marbelite Doghouses, while the earliest recorded doghouse was documented by Bailey (Known as Signals Unlimited on Flickr), which was a long-groove Econolite Doghouse, with a square-door bull's eye red section, which has the earliest possible date of 1963. As for me, the oldest Doghouses in my area are 12" Eagle Flatbacks, dating from the early-1970s. However, finding doghouses that were installed prior to 1980 is extremely uncommon. Yet, Flashing Yellow Arrow signals were invited roughly a decade ago, and are already common, depending on where you live. However, Flashing Yellow Arrow signals are a lot more common today, than doghouses were 40 years ago. Why did it take so long for doghouses to gain popularity, although they were invited around the early-1960s, they didn't become common until at least the mid-to-late 1980s, early-1990's at the very latest. In my area, and based on photos that I've seen from other states, finding doghouses that are advanced in age is extremely difficult to find. Especially in my area, over 90% signal installs that predate the 1980s, lack doghouses, although plenty of them have protected left turn 3M signals for left turn signals. Out of hundreds of 40+ old signal installs in my area, I've only found five of them that have doghouses. Even before PennDOT did upgrades, Doghouses that were installed during the 1970s and earlier, even 1980s, were extremely uncommon. Any logical reason why Doghouses took so long to gain popularity.

[Brandon]

The 5-lamp towers go back to the 1960s in Illinois, and were very popular starting in the 1970s.  That's going back well over 40 years, almost 50 years.

Remember, there's more than one way to have five lamps in a signal head (tower and doghouse).

Here's an example of one (now replaced with newer towers, same trusses) in Illinois from the mid-1970s:
https://goo.gl/maps/uBaTBph6HhZJBeDx5

[traffic light guy]

The 5-lamp towers go back to the 1960s in Illinois, and were very popular starting in the 1970s.  That's going back well over 40 years, almost 50 years.

Remember, there's more than one way to have five lamps in a signal head (tower and doghouse).

Here's an example of one (now replaced with newer towers, same trusses) in Illinois from the mid-1970s:
https://goo.gl/maps/uBaTBph6HhZJBeDx5

I guess it does depend on where you live. Do you know of any other places where 5-section signals were common over 40 years ago

[Rick1962]

Tulsa began using 5-lamp towers in the late '50s, didn't adopt the doghouse until the late '80s. First doghouses I recall seeing were in Springfield, Missouri in the mid '70s.

SM-T580

[jeffandnicole]

The 5-lamp towers go back to the 1960s in Illinois, and were very popular starting in the 1970s.  That's going back well over 40 years, almost 50 years.

Remember, there's more than one way to have five lamps in a signal head (tower and doghouse).

Here's an example of one (now replaced with newer towers, same trusses) in Illinois from the mid-1970s:
https://goo.gl/maps/uBaTBph6HhZJBeDx5

I guess it does depend on where you live. Do you know of any other places where 5-section signals were common over 40 years ago

I would say Philadelphia was a late-comer to the 5-section dog houses. Very similar to how the city will be a latecomer to the 4-section FYAs.

Much of the country adapted to dog houses fairly quickly, including PennDOT.

[roadfro]

Any logical reason why Doghouses took so long to gain popularity.

My guess would be it has to do with technological advances.

When the 5-section displays were first invented in the 60's or so, I think it's a safe bet that many more traffic signals were still operating off electro-mechanical controllers running fixed-time cycles. In that kind of setup, running a protected/permitted display, especially one that was responsive to traffic demand, is highly unlikely. Even if an agency was using computerized signal controllers in that timeframe, that was likely newer technology that either wasn't widely available or was fairly expensive. So in the early days, it might have involved a significant monetary and time investment, including possibly having to replace the signal controller, to install a 5-section display.

Fast-foward to now, and most traffic signal controllers in use nowadays are basically computers that can be programmed for a variety of operations. Many signal controllers manufactured since the mid 2000s have a variety of phasing schemes built in, including PPLT FYA operations, or can be reprogrammed to implement such operations relatively easily. So nowadays, to implement an FYA display where a 5-section or standard left turn display was used before, may in some cases just involve the amount of time it takes to change out the signal heads and enter the new phasing parameters in the controller. (I'm oversimplifying, but I think you get the point.)

[traffic light guy]

The 5-lamp towers go back to the 1960s in Illinois, and were very popular starting in the 1970s.  That's going back well over 40 years, almost 50 years.

Remember, there's more than one way to have five lamps in a signal head (tower and doghouse).

Here's an example of one (now replaced with newer towers, same trusses) in Illinois from the mid-1970s:
https://goo.gl/maps/uBaTBph6HhZJBeDx5

I guess it does depend on where you live. Do you know of any other places where 5-section signals were common over 40 years ago

I would say Philadelphia was a late-comer to the 5-section dog houses. Very similar to how the city will be a latecomer to the 4-section FYAs.

Much of the country adapted to dog houses fairly quickly, including PennDOT.

There are no FYA signals in the city itself, but theres a few in the area:

https://www.google.com/maps/@39.8873185,-75.4025985,3a,15y,7.88h,100.82t/data=!3m6!1e1!3m4!1sdiec4w2kVpSSfMk3v7JDEg!2e0!7i16384!8i8192

https://www.google.com/maps/place/PA-52+%26+Pocopson+Rd,+Pocopson+Township,+PA+19382/@39.9128489,-75.6316243,3a,15y,217h,107.76t/data=!3m6!1e1!3m4!1s3xSJPERmSLxfyu7-WvsL7Q!2e0!7i13312!8i6656!4m5!3m4!1s0x89c6f74f44b9c8f7:0x357fd95ab5716f4e!8m2!3d39.9127443!4d-75.6317011

New, late coming, and not very common, just like how Doghouses were in PA, 45 years ago.

[mrsman]

In some ways, it takes time to adjust the long-standing practice.

As far as changing current signals to FYA, I don't see the point in most cases.  FYA does have a purpose to allow both sides of a lead-lag to operate protective/permissive while avoiding yellow trap.  But in many cases where there is already a safely operating doghouse signal, is there any good reason to replace the doghouse with a FYA?  If there is no reason, then why bother?

With regard to the adaptation of doghouses 30 years ago, the qn is what was there before?  In some cases it was a standard RYG.  Some traffic departments, very famously Los Angeles, were of the mindset to not provide arrows except in extreme cases of high left turn demand so as not to take away signal time from through traffic, so very few left arrows were installed.  the exceptions were cases with visibiltiy issues (where a protected only left would be more appropriate) and some types of lagging left signal (that used to be shown with a 4 aspect tower not a doghouse until quite recently).  Doghouses were virtually non-existent in L.A. until the late 1980's because of this mindset.

In other cases there were singals with both green and yellow arrows, just not displayed in a doghouse setting.  In some areas, the five aspect tower was just more popular.  In other areas, a four aspect tower with a bimodal arrow was also used.  I remember an intersection in Beverly Hills, CA that utilized four aspect towers, 12-12-12-12 on the overhead signal and 8-8-8-12 on the supplemental left pole signal.  Bu the intersting thing about it is that the arrow only on both signal faces was a 3M signal.  You had to really look carefully to determine if this was a green or yellow arrow being displayed.  (The signal has since been coverted many years ago to 5 aspect 12-12-12-12-12 tower on the pole and a doghouse overhead.  A very recent GSV shows that the doghouse is now a horizontal signal head.  Very weird for CA.)

https://www.google.com/maps/@34.068526,-118.4074386,3a,75y,58.73h,82.64t/data=!3m6!1e1!3m4!1sDq1fMHdKP4cIvpNla6HL3w!2e0!7i13312!8i6656

[fwydriver405]

In some ways, it takes time to adjust the long-standing practice.

As far as changing current signals to FYA, I don't see the point in most cases.  FYA does have a purpose to allow both sides of a lead-lag to operate protective/permissive while avoiding yellow trap.  But in many cases where there is already a safely operating doghouse signal, is there any good reason to replace the doghouse with a FYA?  If there is no reason, then why bother?

In my area, as part of the US BUILD GRANT, preemption sensors are being installed at some locations where previously there were none. From what I've been told by one of the signal designers, an emergency vehicle who requests ROW can cause yellow trap if turns are allowed on circular green if the intersection is not set to an all red clear. Of course, some of the newly configured doghouses in Maine still cause yellow trap via phase skip...

In other cases there were singals with both green and yellow arrows, just not displayed in a doghouse setting.  In some areas, the five aspect tower was just more popular.  In other areas, a four aspect tower with a bimodal arrow was also used.  I remember an intersection in Beverly Hills, CA that utilized four aspect towers, 12-12-12-12 on the overhead signal and 8-8-8-12 on the supplemental left pole signal.  Bu the intersting thing about it is that the arrow only on both signal faces was a 3M signal.  You had to really look carefully to determine if this was a green or yellow arrow being displayed.  (The signal has since been coverted many years ago to 5 aspect 12-12-12-12-12 tower on the pole and a doghouse overhead.  A very recent GSV shows that the doghouse is now a horizontal signal head.  Very weird for CA.)

There's this interesting example in Old Town ME where a 5-section tower is used for the left pole signal and 4-section for the overhead signal.

From what I have been told after going to an NHDOT job shadow, NHDOT-installed and maintained traffic signals have always used 4-section bimodal signals for their PPLT signals and PPRT signals, although they are switching to FYA for PPLT. The doghouse signals you may see in NH are usually installed by another contractor, usually inside cities and towns.

Maine and Massachusetts are a bit mixed on what they use for their PPLT and RT signals. Sometimes they use 4-section bimodal and other times doghouse on new or retrofit installations, although Maine has started to use the doghouse signal more on newer installations.

[traffic light guy]

There were 7 intersections (Currently only 3) within the Philadelphia area that have Doghouses that predated 1980. These doghouses were installed during the early-to-mid seventies:

Four in Lower Merion:
https://www.google.com/maps/@40.015839,-75.2293386,3a,30y,109.74h,91.14t/data=!3m6!1e1!3m4!1sI-0kg0eR_4H9gdPgG3fs0Q!2e0!7i16384!8i8192

https://www.google.com/maps/@40.0182688,-75.2401715,3a,15y,243.43h,96.5t/data=!3m6!1e1!3m4!1sWR5DHGkyq2_rJJl7eH5_Tw!2e0!7i16384!8i8192

https://www.google.com/maps/@40.0064873,-75.2320016,3a,30y,260.18h,89.39t/data=!3m6!1e1!3m4!1sAw_dJB8K4ZNFyZOYWYR2bw!2e0!7i3328!8i1664 (replaced in 2009)

https://www.google.com/maps/@40.0033207,-75.2280129,3a,15y,266.74h,92.46t/data=!3m7!1e1!3m5!1sHYPd71lf7g4wu-A5s7uIWQ!2e0!5s20091001T000000!7i13312!8i6656 (Replaced in 2009)

And three more in Bristol Township:

https://www.google.com/maps/@40.1423541,-74.8460826,3a,29.7y,228.06h,95.82t/data=!3m6!1e1!3m4!1s5JPPj0DGPZDqLRVM76QhPA!2e0!7i13312!8i6656

This one is one of my favorites, but got replaced in 2018:
https://www.google.com/maps/@40.1188158,-74.864391,3a,55.3y,70.74h,90.46t/data=!3m6!1e1!3m4!1sjf1Q2E-0nieSXMI2wQ_eLA!2e0!7i13312!8i6656

For some reason, it seems as though, back in those days, doghouses were more likely to be used for right turn signals, rather than left turns.

[mrsman]

I definitely can see that a five aspect signal is more used for right turns than left, yet the decision to use a doghouse over a 5 tower is one of personal preference of the engineer.

5 aspects on left is the standard for pplt, but not for protected only left, or split phasing, or ( until recently) lagging lefts.  Yet, the corresponding right turns would likely use 5 aspects in all of those cases, especially if there is a pedestrian crossing.

Nexus 5X

[traffic light guy]

I definitely can see that a five aspect signal is more used for right turns than left, yet the decision to use a doghouse over a 5 tower is one of personal preference of the engineer.

5 aspects on left is the standard for pplt, but not for protected only left, or split phasing, or ( until recently) lagging lefts.  Yet, the corresponding right turns would likely use 5 aspects in all of those cases, especially if there is a pedestrian crossing.

Nexus 5X

How come 4-section signals are never used for right turn signals?

[fwydriver405]

I definitely can see that a five aspect signal is more used for right turns than left, yet the decision to use a doghouse over a 5 tower is one of personal preference of the engineer.

5 aspects on left is the standard for pplt, but not for protected only left, or split phasing, or ( until recently) lagging lefts.  Yet, the corresponding right turns would likely use 5 aspects in all of those cases, especially if there is a pedestrian crossing.

Nexus 5X

How come 4-section signals are never used for right turn signals?

My guess is that with most 4-section PPRT signals, there is no yellow clearance when the protected green interval expires. AFAIK, 2009 MUTCD (4d-24-01a) requires the green arrow to terminate with the green ball - so the green arrow can't just disappear without a yellow arrow.

Of course, you can use a bimodal yellow/green right turn arrow, like what Maine, NH and MA have done for years and the functionality is exactly the same as if you were to install a 5-section signal. Except for NHDOT signals which only uses 4-section signals and never 5-section signals, in the case of Maine and Massachusetts, I believe it is the preference of the state agency and/or the engineer who decides what kind of signals are to be installed...

[traffic light guy]

I definitely can see that a five aspect signal is more used for right turns than left, yet the decision to use a doghouse over a 5 tower is one of personal preference of the engineer.

5 aspects on left is the standard for pplt, but not for protected only left, or split phasing, or ( until recently) lagging lefts.  Yet, the corresponding right turns would likely use 5 aspects in all of those cases, especially if there is a pedestrian crossing.

Nexus 5X

How come 4-section signals are never used for right turn signals?

My guess is that with most 4-section PPRT signals, there is no yellow clearance when the protected green interval expires. AFAIK, 2009 MUTCD (4d-24-01a) requires the green arrow to terminate with the green ball - so the green arrow can't just disappear without a yellow arrow.

Of course, you can use a bimodal yellow/green right turn arrow, like what Maine, NH and MA have done for years and the functionality is exactly the same as if you were to install a 5-section signal. Except for NHDOT signals which only uses 4-section signals and never 5-section signals, in the case of Maine and Massachusetts, I believe it is the preference of the state agency and/or the engineer who decides what kind of signals are to be installed...

What makes more right turns more dangerous than left turns?

[jakeroot]

4-section right turn signals are basically the norm in the South Seattle area. Particularly Puyallup, but also other municipalities: https://goo.gl/maps/uA7TbPnnp262ir2u8

They are all bimodal, although this one was, until quite recently, only a green arrow that simply disappeared.

[DevalDragon]

I remember driving with some friends on one of my first visits to Illinois. We were stopped at a red light (with 5 lamp towers) in Naperville. The left turn arrow turned on and I tried to go straight. I stopped quickly - no damage but a lot of honking. My friend commented "Don't they have traffic lights like this where you're from?"

I thought for a second and said no. I drove in Texas for 14 years where they have sideways (horizontal) traffic signals and (at that time) lived in Ohio where they had dog houses.  It was the first time I'd ever seen a vertical 5 lamp tower.

Going back to Doghouses, that's all I remember as a kid when we visited Ohio. And they still have a lot of them there, even with special adapters for the different size lenses for turn arrows.

[mrsman]

I definitely can see that a five aspect signal is more used for right turns than left, yet the decision to use a doghouse over a 5 tower is one of personal preference of the engineer.

5 aspects on left is the standard for pplt, but not for protected only left, or split phasing, or ( until recently) lagging lefts.  Yet, the corresponding right turns would likely use 5 aspects in all of those cases, especially if there is a pedestrian crossing.

Nexus 5X

How come 4-section signals are never used for right turn signals?

Not never, but rarely.

I won't discuss the bimodal yellow/green arrows, since as others suggested, they are functionally equivalent to 5 aspect signals, except that the yellow and green share the aspect.  (I'm personally not a fan, although I know there are areas like NJ and BC where these are common place).

Ok, with that out of the way, let's discuss the purpose of the right turn signal.  In most cases, it is a corresponding signal to the concurrent left turn.  Assuming u-turns are prohibited (or u-turns yield to the right turns), the right turn at the same time as the corresponding left is totally protected from cross traffic and pedestrians.  The green right arrow will end when that protection ends.

Let's say you're at the intersection of First and Main, a left from First to Main with a corresponding protected right from Main to First.

If the left were a leading left, then when the leading left ends, opposing traffic on First Street begins.  From the point of view of Main St, the protected right ends and will then begin having to yield to cross traffic.  A yellow arrow is appropriate.

If the left were a lagging left, then when the lagging left ends, crossing traffic on Main Street begins.  From the point of view of Main St, the protected right ends and will then begin having to yield to parallel pedestrian traffic.  A yellow arrow is appropriate.

However, if the parallel pedestrian traffic (in the lagging left situation) were somehow nonexistent or prohibited, then when the lagging left ends and Main St gets the green, the right turn from Main to First continues to be protected.  The right turn arrow will remain green while the green orb is lit.  The protected right will end when Main St's phase will end.  Pre-2009 MUTCD, it would be OK to have the yellow ball signal the end of both the Main Street green orb and the Main Street green arrow.  Yet, this is rare because it only works when the pedestrians aren't present.  So it would be rare to see the 4 aspect signal for right turns.

[fwydriver405]

However, if the parallel pedestrian traffic (in the lagging left situation) were somehow nonexistent or prohibited, then when the lagging left ends and Main St gets the green, the right turn from Main to First continues to be protected.  The right turn arrow will remain green while the green orb is lit.  The protected right will end when Main St's phase will end.  Pre-2009 MUTCD, it would be OK to have the yellow ball signal the end of both the Main Street green orb and the Main Street green arrow.  Yet, this is rare because it only works when the pedestrians aren't present.  So it would be rare to see the 4 aspect signal for right turns.

While looking for other examples in the "double permissive left" thread in San Francisco, I stumbled across something that is related to the 4-aspect signals here. Does anyone know why San Francisco is now installing new 4 aspect signals for right turns (old, current) (and left from one way to one way situations) across pedestrian paths?

These ones first light the green ball during the pedestrian phase, then green ball + green arrow when the pedestrian phase is finished, then the green ball + green arrow terminates when the signal goes to yellow ball.

[mrsman]

However, if the parallel pedestrian traffic (in the lagging left situation) were somehow nonexistent or prohibited, then when the lagging left ends and Main St gets the green, the right turn from Main to First continues to be protected.  The right turn arrow will remain green while the green orb is lit.  The protected right will end when Main St's phase will end.  Pre-2009 MUTCD, it would be OK to have the yellow ball signal the end of both the Main Street green orb and the Main Street green arrow.  Yet, this is rare because it only works when the pedestrians aren't present.  So it would be rare to see the 4 aspect signal for right turns.

While looking for other examples in the "double permissive left" thread in San Francisco, I stumbled across something that is related to the 4-aspect signals here. Does anyone know why San Francisco is now installing new 4 aspect signals for right turns (old, current) (and left from one way to one way situations) across pedestrian paths?

These ones first light the green ball during the pedestrian phase, then green ball + green arrow when the pedestrian phase is finished, then the green ball + green arrow terminates when the signal goes to yellow ball.

This is meant to address pedestrian safety and blocked traffic in busy areas.

There are some intersections in other cities that divide up the green orb phase the following ways:

1) red right arrow with pedestrian "walk" - no right turn allowed, protected pedestrian phase, (leading pedestrian interval)
2) shared right turns and pedestrians, right turns allowed but must yield to pedestrians, pedestrians may either have walk or FDW, cars may have either no arrow or FYA
3) green right arrow with pedestrian "don't walk" - pedestrians should not be in the intersection at all, right turning traffic has right of way (but obviously please don't run down the slow grandmas or jaywalkers).

Some cities implement:  1,2,3; or 1 and 2; or 1 and 3; or 2 and 3.  (sometimes the order also changes having 3 before 2 for instance). 

San Francisco implements 2 and 3.  Without 3, very few right turning vehicles would be able to make the turn in one cycle.  So they shorten the pedestrian phase to allow some dedicated time for right turns that are protected from pedestrians that may block traffic flow for part of the phase.

Perhaps a Nor Cal local can let us in and tell us if they work as intended.

EDITED TO ADD:  While I've seen these new signals all over Downtown SF on GSV, there is an extra reason to see them on Market St.  New regulations basically force all cars except taxis and buses to turn off of Market and not drive on it.  All traffic is essentially forced to turn right at the next opportunity.  Since this is a heavy right turn, they need an exclusive phase (away from peds) in order for traffic not to be blocked.

[jakeroot]

Seattle has quite a lot of those right (or left) turn "filter" arrows for one-way to one-way movements (#2 and #3 in the list above). Here's an example at 4th & Pike. This one was necessary because the original double right turn was eliminated when the outer turn lane was replaced with a bus lane, forcing all traffic to turn + yield from one lane.

There is another at 5th & Spring that has a shared left/through lane. This one lags as can be seen in the image (FDW on left, solid WALK on right).

Unlike what I've seen elsewhere, Seattle uses both lagging and advanced green arrows (depending on each individual install). I think I prefer advanced green arrows with these signals, as you'll get the cars at the beginning, any during gaps in pedestrian flow, and then a couple at the end when the light turns red. Slightly higher capacity. But it might be more dangerous with cars pushing the end of the protected phase right as the WALK sign comes on.

[mrsman]

Most of the signals that provide an exclusive phase for right turning (or left turning one-way to one-way, functionally the same thing) do so as lagging green arrows.  I believe the thinking is that pedestrians are liable to jaywalk and start crossing when they see a break in cross traffic.  They are more liable to jaywalk before the walk phase then keep walking right at the end.

[jakeroot]

I did manage to recall where I'd seen at least one version in Seattle that had advanced phasing. The right turn from one-way Howell to the I-5 southbound on-ramp at Yale is a flashing yellow arrow. As can be seen in this GSV imagery, the protected phase is at the beginning, as the solid yellow at the end of the cycle corresponds perfectly with the "solid don't walk [0 seconds]" symbol on the right. If that "0" wasn't visible, it would have been possible to assume that the solid yellow arrow had been following a green arrow, but that wouldn't be possible if the FDW phase was still active.

This signal/intersection is actually doubly strange. The far left lane is a shared left turn & bus lane. At the beginning of the cycle, the through signals are red, but the right and left turn green arrows activate (for the left onto Yale and the right onto I-5 southbound). The transit signal also gets its "green" signal (vertical white line). The reason for this is that buses need to get all the way to the right to access the northbound 5 express lanes. This is effectively a queue-jump; the green left arrow is necessary to clear left-turning vehicles that might be in the way of those buses. It's possible that the advanced green for the FYA on the right was set up this way to mirror the advanced left onto Yale.

[mrsman]

I did manage to recall where I'd seen at least one version in Seattle that had advanced phasing. The right turn from one-way Howell to the I-5 southbound on-ramp at Yale is a flashing yellow arrow. As can be seen in this GSV imagery, the protected phase is at the beginning, as the solid yellow at the end of the cycle corresponds perfectly with the "solid don't walk [0 seconds]" symbol on the right. If that "0" wasn't visible, it would have been possible to assume that the solid yellow arrow had been following a green arrow, but that wouldn't be possible if the FDW phase was still active.

This signal/intersection is actually doubly strange. The far left lane is a shared left turn & bus lane. At the beginning of the cycle, the through signals are red, but the right and left turn green arrows activate (for the left onto Yale and the right onto I-5 southbound). The transit signal also gets its "green" signal (vertical white line). The reason for this is that buses need to get all the way to the right to access the northbound 5 express lanes. This is effectively a queue-jump; the green left arrow is necessary to clear left-turning vehicles that might be in the way of those buses. It's possible that the advanced green for the FYA on the right was set up this way to mirror the advanced left onto Yale.

I agree with your theory.  For queue jump operations, the bus signal must be in advance of the general traffic flow.  As left turners could block the bus, left turners must also be allowed unfettered turning access (i.e. a protected turn vis a vis pedestrians) at this time.  If signal time were already provided for the left turners, and there is really no reason why right turners should not also get a protected phase, both left turns and right turns are given the advance slot.

But as you point out, this is a unique circumstance.  The vast majority of one-way to one-way protected turns are lagging to better avoid conflict with jaywalkers.

[jakeroot]

The vast majority of one-way to one-way protected turns are lagging to better avoid conflict with jaywalkers.

This situation, at any rate, still strikes me as unusual given that most municipalities (outside of a few select areas) generally lead their turns, no matter which direction. It just seems to be the case that protected turns over crosswalks from one-way to one-way are almost always lagging. I can understand the jaywalking argument (although I hate that term, especially being someone who does it all the time to the chagrin of no one), but it seems that you could have the same issue at regular intersections. The left green arrow comes up, the car starts to turn, but has to proceed slowly because of someone who started crossing prematurely. I've seen a lot of municipalities around me address this issue by beginning a left turn phase with flashing yellow arrows instead, with the green arrow either at the end or not at all.

[mrsman]

The vast majority of one-way to one-way protected turns are lagging to better avoid conflict with jaywalkers.

This situation, at any rate, still strikes me as unusual given that most municipalities (outside of a few select areas) generally lead their turns, no matter which direction. It just seems to be the case that protected turns over crosswalks from one-way to one-way are almost always lagging. I can understand the jaywalking argument (although I hate that term, especially being someone who does it all the time to the chagrin of no one), but it seems that you could have the same issue at regular intersections. The left green arrow comes up, the car starts to turn, but has to proceed slowly because of someone who started crossing prematurely. I've seen a lot of municipalities around me address this issue by beginning a left turn phase with flashing yellow arrows instead, with the green arrow either at the end or not at all.

But in reality they are very different situations.

The standard left turn from a two-way street, the leading left is predominant.  Leading lefts are generally preferred since they can always be designed without worrying about the yellow trap problem.  This is true when only one side leads, both sides lead, both sides lead but one side has a heavier left turn load, etc.  it is very versatile.

Even though it is more rare, I prefer the lagging lefts, since they seem more efficient with respect to straight moving traffic.  If the traffic load is moderate, and there are good sensors, it is possible that a permissive left can be made during the green orb phase.  It would render the lagging left unnecessary and there won't even be a separate phase for it.  It improves the general signal operation.  For this reason it is preferred in situations where yellow trap is avoided (in situations where the opposing left is prohibited).  It is also useful (assuming yellow trap is avoided) in situations where there is an option lane present.  The lagging left will clear the traffic, where left turners have been blocking the intersection.  But given the problems with yellow trap, it is not used as often as leading lefts.

But the criteria for choosing a leading left in a two-way street is irrelevant for choosing a lagging signal for right turns or one-way to one-way left turns.  Right turns at a regular intersection will largely be dependant on the corresponding left turn and be lit during that period.  But if the right turn light were independent, and the sole reason for providing the right turn arrow is to have dedicated turning time in a heavy pedestrian area, then the criteria is based on how the turning traffic interacts with the pedestrian.

I declare that jaywalking is likely more common in areas with one-way street networks.  The pedestrian will likely look in the direction of the traffic he is crossing, determine that there is no cars and simply cross.  There are fewer turning conflicts in one way networks, plus many are in downtown areas with heavier pedestrian presence.  As it was determined that jaywalking is more common when a pedestrian gets to a corner, sees his light is red, watches cross traffic and crosses when there is a break in traffic even if his light is still red.  We do not want to put a protected arrow at this point in time, because it will be unexpected by the pedestrian*.  It is better at the end of the signal phase, where the pedestrian sees a FDW and runs knowing that his time is up and some conflicting traffic (usually cross traffic, but in this specific case right turning traffic) will likely start once the hand stops flashing.

So where the sole criteria of a protected arrow is the pedestrian interaction, the lagging signal is used.  But it will be leading to the extent that the turning signal interacts with another signal, like a corresponding (lagging) left turn or a bus queuing signal.

* You are correct that a leading left in a two-way intersection may also be a surprise to a jaywalking pedestrian, but the pedestrian interaction is simply not the criteria for choosing to use a left arrow in this instance.  The purpose is to provide protection against cross traffic, not pedestrians.  Also, jaywalking is less common as there are more moving parts to consider when crossing a busy two-way intersection.