Has the flashing yellow left turn signal made it to your state?

[mrsman]

Interesting right turb FYA in NYC. The FY arrow is on the bottom and has both a LPI and lagging protected right. Sequence (I believe) is

Green ball/red arrow (LPI)
Green ball/FYA
Red ball/green arrow (lagging protected right)

This is a wide intersection with a lot of pedestrian traffic, so the FYA helps here. Strangely enough the left turns on the cross street were not converted to FYA

https://maps.app.goo.gl/qZmda1UTjMExgAmF6

There is no reason to convert the left turn to FYA.  Only the green right arrow phase (and the solid yellow arrow phase that follows) will mimic the green arrow phase of the left turn signal.  If the permitted phase of the left turn were FYA, the right turn would still be a red arrow, as there is no turn on red in NYC.

Another issue is that the FYA sign being used to yield to pedestrians is so common that in NYC I'm not sure if it's understood that a FYA is used to yield to oncoming traffic.  While there are a few examples like that in Staten Island (I believe), using the FYA in its normal context is rare for NYC and non-existent in the pedestrian heavy areas.

NYC has been doing some wonderful things lately in the name of pedestrian and bicycle safety.  I like what they did here, providing protected phases for the right turns as well as for the pedestrians.  This should improve safety here.

[kphoger]

All signals should start with the least permissive phasing possible, and only progress beyond that as absolutely necessary.

Is that a mis-type?

[roadfro]

All signals should start with the least permissive phasing possible, and only progress beyond that as absolutely necessary.

Is that a mis-type?

I'm assuming Jake meant "least restrictive phasing possible"

[jeffandnicole]

All signals should start with the least permissive phasing possible, and only progress beyond that as absolutely necessary.

Is that a mis-type?

I'm assuming Jake meant "least restrictive phasing possible"

Both phrases will work.

Basically, I believe he's saying is start with a typical RYG light.  Then, does it require permissive/protected phases?  Does it require a FYA?  Does it require protected only phasing?  Does it require off-set phasing? 

[jakeroot]

All signals should start with the least permissive phasing possible, and only progress beyond that as absolutely necessary.

Is that a mis-type?

I'm assuming Jake meant "least restrictive phasing possible"

That's exactly what I meant, yeah. I updated my post to reflect that.

Basically, I believe he's saying is start with a typical RYG light.  Then, does it require permissive/protected phases?  Does it require a FYA?  Does it require protected only phasing?  Does it require off-set phasing? 

That's what I'm getting at. It really seems to me that a typical intersection, with single turn lanes (if any), should just start off with the most "permissive" (aka least restrictive) phasing possible, with modifications made beyond that as necessary. Even double left turns may not need them immediately.

For the longest time, my understanding was that this was WSDOT's policy (effectively, all intersections should be operated starting with the least restrictive phasing, and then modified accordingly), but it's extremely rare to actually see them do this. New intersections rarely forgo a separate left turn signal (be it an FYA or protected-only turn).

One of the only examples of them following this policy (sort of) was the rebuild of WA-161 (Enchanted Pkwy) north of Edgewood, in the mid-2000s, when the three main signals (19th Way S, 28th Ave S, and Military Road) were installed without any separate left turn control. All signals are operated (and continue to be operated) as permissive-only "LEFT TURN MUST YIELD" type intersections, despite WA-161 being a divided expressway-type road with a 45 mph speed limit. On paper, policy might dictate that the higher speed limit result in protected-only turns, but they were not installed. As a result, at least from my personal experience, this stretch of WA-161 is by far the best-flowing stretch. The lights have only two phases (apart of 28th Ave S because the offset requiring split-phasing), and through traffic does not need to wait for left turns to go.

Now yes, it helps that the left turns in this example are not busy anyways, but how many left turns are truly so busy that gaps in traffic aren't enough to clear them? After all, a gap in through traffic is no different than a protected left turn.

[jakeroot]

When designing roadways, it's always important to look at various options.  Having a FYA at every signal isn't necessary.  Heck, having a left turn arrow at every intersection isn't necessary.

I would say that left turn arrows are arguably not necessary at most intersections. Traffic turning left can turn behind through traffic, and at least two cars should be able to turn at the end of a fully-saturated cycle. Left turn signals can be added in this case, at least when there is a side-street with enough traffic to not allow the signal to remain green to clear all traffic. The idea of adding left turn signals when there are too many oncoming lanes, or a speed limit that's too high, or just as part of policy, seems (as much as I hate the term) like policy overreach. All signals should start with the least permissive restrictive phasing possible, and only progress beyond that as absolutely necessary.

One thing that makes me less quick to jump on FYAs is that, unlike left turns where there is no left turn arrow at all, traffic seems slightly less apt to pull forward. Even in areas that I deem relatively conservative, drivers will always pull forward to turn on solid green lights where there is no protected phase (how else can you guarantee you'll make it?), but the split more like 75 (forward)/25 (behind the line) at FYAs. There have been more than a few situations where I missed a left turn because the driver in front refused to enter the junction, where I know they would have, had the light not had a green arrow phase. Some drivers fully embrace the concept of "I'll just wait for the green arrow if there's no gap" style of driving that seems to pop-up when I see an FYA installed.

That problem also exists at doghouse signals as well.

Its this phenomena that makes me prefer lagging lefts over leading lefts, provided that the yellow trap problem is addressed.  Its the "natural" phase for turning left, particularly in saturated intersections.  Another plus is that with the right sensors, the left turn phase can be totally skipped altogether if the left turn lane clears during the green orb phase.

The downside of course is yellow trap and that generally the lagging lefts have to be simultaneous.  This is why lagging lefts are more common where the opposing left is prohibited.

Lagging lefts do have a tendency to result in more drivers pulling forward, yes. I think this is because they think that the green arrow was skipped, so they pull forward to avoid missing the light entirely. That said, even at the left turns that I go through regularly where lagging left turns are normal, drivers do not always pull forward. Eventually, there is usually a gap, but then the green arrow activates, so the "pulling forward" bit only helps those of us trying to complete the left turn quicker. At that point, especially at those intersections where lagging green arrows are employed habitually (side-streets approaching a major road in this area), drivers will sometimes wait behind the line (unless there is a gap) because they become accustomed to the lagging green arrow ("I don't need to worry about pulling forward because there is a green arrow at the end"). Most drivers to creep forward, yes, but there are still some that don't. This is compared to those intersections where there is no green arrow at all, where pulling-forward compliance is near 100%.

This talk of lagging left turns and creeping forward does remind me why I tend to prefer advanced left turns: the capacity seems to be a bit higher at intersections where a left turn will absolutely be needed: you get all of the cars at the beginning of the phase, and then at least one or two (sometimes more) at the end during the clearance phase. It's a relatively minor increase in capacity, but certainly noticeable, especially in everyday driving. In comparison, there is no "clearance" phase to accept a couple more cars at lagging lefts.

[RestrictOnTheHanger]

Interesting right turb FYA in NYC. The FY arrow is on the bottom and has both a LPI and lagging protected right. Sequence (I believe) is

Green ball/red arrow (LPI)
Green ball/FYA
Red ball/green arrow (lagging protected right)

This is a wide intersection with a lot of pedestrian traffic, so the FYA helps here. Strangely enough the left turns on the cross street were not converted to FYA

https://maps.app.goo.gl/qZmda1UTjMExgAmF6

There is no reason to convert the left turn to FYA.  Only the green right arrow phase (and the solid yellow arrow phase that follows) will mimic the green arrow phase of the left turn signal.  If the permitted phase of the left turn were FYA, the right turn would still be a red arrow, as there is no turn on red in NYC.

Another issue is that the FYA sign being used to yield to pedestrians is so common that in NYC I'm not sure if it's understood that a FYA is used to yield to oncoming traffic.  While there are a few examples like that in Staten Island (I believe), using the FYA in its normal context is rare for NYC and non-existent in the pedestrian heavy areas.

NYC has been doing some wonderful things lately in the name of pedestrian and bicycle safety.  I like what they did here, providing protected phases for the right turns as well as for the pedestrians.  This should improve safety here.

RTOR notwithstanding, i know that the right turn signal would still be red since the Northern thru traffic would have a green.

The comment about converting the left turn signal on Northern was moreso about the fact that there were some signal upgrades on Woodhaven Blvd, 108th st/LIE, and other places in Queens and NYC that have seen normal FYA installation.

I think NYC drivers do in fact understand that FYA equals yield, both to cars and peds, and installation should continue wherever it would be beneficial.

[mrsman]

When designing roadways, it's always important to look at various options.  Having a FYA at every signal isn't necessary.  Heck, having a left turn arrow at every intersection isn't necessary.

I would say that left turn arrows are arguably not necessary at most intersections. Traffic turning left can turn behind through traffic, and at least two cars should be able to turn at the end of a fully-saturated cycle. Left turn signals can be added in this case, at least when there is a side-street with enough traffic to not allow the signal to remain green to clear all traffic. The idea of adding left turn signals when there are too many oncoming lanes, or a speed limit that's too high, or just as part of policy, seems (as much as I hate the term) like policy overreach. All signals should start with the least permissive restrictive phasing possible, and only progress beyond that as absolutely necessary.

One thing that makes me less quick to jump on FYAs is that, unlike left turns where there is no left turn arrow at all, traffic seems slightly less apt to pull forward. Even in areas that I deem relatively conservative, drivers will always pull forward to turn on solid green lights where there is no protected phase (how else can you guarantee you'll make it?), but the split more like 75 (forward)/25 (behind the line) at FYAs. There have been more than a few situations where I missed a left turn because the driver in front refused to enter the junction, where I know they would have, had the light not had a green arrow phase. Some drivers fully embrace the concept of "I'll just wait for the green arrow if there's no gap" style of driving that seems to pop-up when I see an FYA installed.

That problem also exists at doghouse signals as well.

Its this phenomena that makes me prefer lagging lefts over leading lefts, provided that the yellow trap problem is addressed.  Its the "natural" phase for turning left, particularly in saturated intersections.  Another plus is that with the right sensors, the left turn phase can be totally skipped altogether if the left turn lane clears during the green orb phase.

The downside of course is yellow trap and that generally the lagging lefts have to be simultaneous.  This is why lagging lefts are more common where the opposing left is prohibited.

Lagging lefts do have a tendency to result in more drivers pulling forward, yes. I think this is because they think that the green arrow was skipped, so they pull forward to avoid missing the light entirely. That said, even at the left turns that I go through regularly where lagging left turns are normal, drivers do not always pull forward. Eventually, there is usually a gap, but then the green arrow activates, so the "pulling forward" bit only helps those of us trying to complete the left turn quicker. At that point, especially at those intersections where lagging green arrows are employed habitually (side-streets approaching a major road in this area), drivers will sometimes wait behind the line (unless there is a gap) because they become accustomed to the lagging green arrow ("I don't need to worry about pulling forward because there is a green arrow at the end"). Most drivers to creep forward, yes, but there are still some that don't. This is compared to those intersections where there is no green arrow at all, where pulling-forward compliance is near 100%.

This talk of lagging left turns and creeping forward does remind me why I tend to prefer advanced left turns: the capacity seems to be a bit higher at intersections where a left turn will absolutely be needed: you get all of the cars at the beginning of the phase, and then at least one or two (sometimes more) at the end during the clearance phase. It's a relatively minor increase in capacity, but certainly noticeable, especially in everyday driving. In comparison, there is no "clearance" phase to accept a couple more cars at lagging lefts.

Herein lies the difference between leading and lagging lefts.  For a leading left, I get a protected turn at the beginning of the cycle and I get a chance for one or two cars to turn at the end of the cycle.  From the point of view of how many cars can turn left for a given amount of protected time, I agree you get an extra one or two cars during the clearance phase.  A lagging left does not have the clearance phase, so if the amount of time for the green arrow in both cases is equivalent, more cars will be able to turn in the leading scenario.

However, when taken from the point of view of how much time does servicing lefts take away from the through traffic phases, the lagging left is better.  [This is addressed more to Jake's comment 1579 not 1580.]  As you said in Reply 1579, through traffic is better served to the extent that we do not break away from the through traffic to service a protected left turn.  This is true.  Assuming good sensors on the pavement, you are far more likely to be able to skip the protected left turn phase entirely with a lagging setting  than you are with a leading setting.

Let's take an example.  Suppose the signal is set up to activate the left turn arrow if there are at least three cars sitting in the left turn lane.  In a leading setting, it will notice three cars and provide a protected left.  This is true even if there will likely be gaps in the through phase that follows.  But in a lagging setting, the green arrow follows the through phase.  If enough left turning cars are serviced in the through phase (by finding natural gaps in traffic), then it is possible (even likely) that the left turn phase is skipped entirely.  There are more likely to be cars waiting to turn in the leading scenario since the left turn lane traffic can build up while the light is red.  But in the lagging scenario, we provide the maximum opportunity for cars to clear during the green orb phase, and only if it is still necessary do we light the green arrow to provide a protected left to prevent backups in the left turn lane.

So, IMO, leading lefts better serve the left turners, for the reasons you stated, but lagging lefts better serve the intersection as a whole.  Of course, this assumes that all the yellow trap issues are properly addressed.

[Amtrakprod]

Sometimes lights can even have both a lead and lagging cycle. This is best to be used at an intersection with a shared turn lane that has heavy left turn usage. IMO T intersections should utilize lagging lefts way more than they do.

[thenetwork]

When designing roadways, it's always important to look at various options.  Having a FYA at every signal isn't necessary.  Heck, having a left turn arrow at every intersection isn't necessary.

I would say that left turn arrows are arguably not necessary at most intersections. Traffic turning left can turn behind through traffic, and at least two cars should be able to turn at the end of a fully-saturated cycle. Left turn signals can be added in this case, at least when there is a side-street with enough traffic to not allow the signal to remain green to clear all traffic. The idea of adding left turn signals when there are too many oncoming lanes, or a speed limit that's too high, or just as part of policy, seems (as much as I hate the term) like policy overreach. All signals should start with the least restrictive phasing possible, and only progress beyond that as absolutely necessary.

One thing that makes me less quick to jump on FYAs is that, unlike left turns where there is no left turn arrow at all, traffic seems slightly less apt to pull forward. Even in areas that I deem relatively conservative, drivers will always pull forward to turn on solid green lights where there is no protected phase (how else can you guarantee you'll make it?), but the split more like 75 (forward)/25 (behind the line) at FYAs. There have been more than a few situations where I missed a left turn because the driver in front refused to enter the junction, where I know they would have, had the light not had a green arrow phase. Some drivers fully embrace the concept of "I'll just wait for the green arrow if there's no gap" style of driving that seems to pop-up when I see an FYA installed.

edit: word choice

With the FYA intersections I see in my area (CO/UT), if the opposing left turn lane gets a leading green arrow, in some cases, I see the left turn lane  in my direction will get a FYA, in case there was little to no through traffic one can turn left without waiting for a green arrow/green ball.

[mrsman]

When designing roadways, it's always important to look at various options.  Having a FYA at every signal isn't necessary.  Heck, having a left turn arrow at every intersection isn't necessary.

I would say that left turn arrows are arguably not necessary at most intersections. Traffic turning left can turn behind through traffic, and at least two cars should be able to turn at the end of a fully-saturated cycle. Left turn signals can be added in this case, at least when there is a side-street with enough traffic to not allow the signal to remain green to clear all traffic. The idea of adding left turn signals when there are too many oncoming lanes, or a speed limit that's too high, or just as part of policy, seems (as much as I hate the term) like policy overreach. All signals should start with the least restrictive phasing possible, and only progress beyond that as absolutely necessary.

One thing that makes me less quick to jump on FYAs is that, unlike left turns where there is no left turn arrow at all, traffic seems slightly less apt to pull forward. Even in areas that I deem relatively conservative, drivers will always pull forward to turn on solid green lights where there is no protected phase (how else can you guarantee you'll make it?), but the split more like 75 (forward)/25 (behind the line) at FYAs. There have been more than a few situations where I missed a left turn because the driver in front refused to enter the junction, where I know they would have, had the light not had a green arrow phase. Some drivers fully embrace the concept of "I'll just wait for the green arrow if there's no gap" style of driving that seems to pop-up when I see an FYA installed.

edit: word choice

With the FYA intersections I see in my area (CO/UT), if the opposing left turn lane gets a leading green arrow, in some cases, I see the left turn lane  in my direction will get a FYA, in case there was little to no through traffic one can turn left without waiting for a green arrow/green ball.

One of the great features of FYAs, giving more time for a potential left turn.  While it isn't too common to allow the FYA to be displayed opposite a leading green arrow - there is no safety issue to do so (other than possible protection for pedestrians ala leading pedestrian interval).

The opposite situation, FYA in the opposite direction of a lagging green arrow, is absolutely required to avoid yellow trap.  The only other ways to avoid yellow trap with a lagging left is to totally prohibit the left opposite a lagging green arrow, restrict the opposite left turn by making it protected only [red arrow],  or by making totally simultaneous lagging lefts.

[roadfro]

Let's take an example.  Suppose the signal is set up to activate the left turn arrow if there are at least three cars sitting in the left turn lane.  In a leading setting, it will notice three cars and provide a protected left.  This is true even if there will likely be gaps in the through phase that follows.  But in a lagging setting, the green arrow follows the through phase.  If enough left turning cars are serviced in the through phase (by finding natural gaps in traffic), then it is possible (even likely) that the left turn phase is skipped entirely.  There are more likely to be cars waiting to turn in the leading scenario since the left turn lane traffic can build up while the light is red.  But in the lagging scenario, we provide the maximum opportunity for cars to clear during the green orb phase, and only if it is still necessary do we light the green arrow to provide a protected left to prevent backups in the left turn lane.

So, IMO, leading lefts better serve the left turners, for the reasons you stated, but lagging lefts better serve the intersection as a whole.  Of course, this assumes that all the yellow trap issues are properly addressed.

A technical note: The detection used with most signals does not count vehicles, especially inductive loops. It works based on presence detection–is there a vehicle within the detection zone or not? Detection zones are usually only a car length or two, but it's usually a "yes/no" input to the controller as opposed to a "0/1/2/...".

And given the random arrival nature of traffic, unless the signal is in a coordinated system where arrival platoons are frequent, you're just as likely to have vehicles stack up in the left turn pocket during red (leading left scenario) as you are during FYA/permitted left (lagging left scenario). When it's time for the controller to address the protected left turn in the cycle, it's just going to look and see whether any left turn vehicles are still detected to determine to activate the green arrow.

[SignBridge]

We have an intersection in my neighborhood where the induction loop was placed one car length back from the stop line so it would only activate the arrow if there were at least two cars in the left-turn lane. But LOL, if I was the only car in the lane, I would stop well back from the stop-line, right over the loop so I would get the arrow!

However most drivers did not notice the placement of the loop and thought the arrow was failing to work. I assume the County got too many complaints about it because they finally dug up the loop and re-installed it in the usual spot behind the stop-line.

[roadfro]

^ I've seen installations in Nevada where the through lanes have detector loops just right at the stop bar, but left turn lanes often have two sets of loops so that two cars in the pocket can set off detection. I've also seen installations where there are some detection loops beyond the stop line, in order to detect left turning vehicles that are waiting for a gap during permissive phasing.

[jakeroot]

But LOL, if I was the first car in the lane, I would stop well back from the stop-line, right over the loop so I would get the arrow!

Why? Since when have arrows been required to make a turn? If you're the only one there, just turn in a gap or at the end. Don't be selfish.

^ I've seen installations in Nevada where the through lanes have detector loops just right at the stop bar, but left turn lanes often have two sets of loops so that two cars in the pocket can set off detection. I've also seen installations where there are some detection loops beyond the stop line, in order to detect left turning vehicles that are waiting for a gap during permissive phasing.

Shit, I've seen intersections in parts of WA where video detection systems can see well beyond the stop line. I've had lights that trip from my car just barely nosing into the left turn lane.

Definitely haven't seen one activate when I was waiting to turn, although I always wait in the middle of the junction, well beyond the typical detection points.

[fwydriver405]

Never noticed this until now. FYA left turn signal is installed on a shared left-thru lane. Lee NH.

But LOL, if I was the first car in the lane, I would stop well back from the stop-line, right over the loop so I would get the arrow!

Why? Since when have arrows been required to make a turn? If you're the only one there, just turn in a gap or at the end. Don't be selfish.

^ I've seen installations in Nevada where the through lanes have detector loops just right at the stop bar, but left turn lanes often have two sets of loops so that two cars in the pocket can set off detection. I've also seen installations where there are some detection loops beyond the stop line, in order to detect left turning vehicles that are waiting for a gap during permissive phasing.

Shit, I've seen intersections in parts of WA where video detection systems can see well beyond the stop line. I've had lights that trip from my car just barely nosing into the left turn lane.

Definitely haven't seen one activate when I was waiting to turn, although I always wait in the middle of the junction, well beyond the typical detection points.

There's an intersection in my hometown where the sensor is also placed farther back. It's placed where the left painted arrow is beneath the ONLY marking. Come to think of it, my town is also replacing loop detected with 360 degree cameras for the sensors when new signals will be installed next year. With the limited field of view, would it be possible for the detection zone on these kind of sensors to be placed back farther than at the line? I just wonder if those sensors would be confused if the detection zone is placed farther back with other traffic.

[Revive 755]

One of the great features of FYAs, giving more time for a potential left turn.  While it isn't too common to allow the FYA to be displayed opposite a leading green arrow - there is no safety issue to do so (other than possible protection for pedestrians ala leading pedestrian interval).

At least in the Midwest, not having a FYA active when the opposing left turn gets a green arrow would be the exception.  Almost all the FYA installations I have seen and can recall in Illinois, Missouri, Iowa, and Nebraska will bring up the FYA when there is a not a vehicle present to call the leading green arrow.  The main variable seems to be how long after the opposing through movement turns green does the FYA come up.

[SignBridge]

Jakeroot, to answer your question above: It's faster, easier and safer to make a left turn on a leading green-arrow than to wait for a gap. Also visibility of oncoming traffic is an issue at this intersection as the road takes a right bend as it crosses over, so it can be difficult to see oncoming traffic if there are vehicles in the opposing direction waiting to turn left.

But I also have to confess the mischievous satisfaction of beating the County's dumb set-up to make the arrow come on. LOL And as I stated above, the County has since changed the installation to make the arrow work the way most drivers would expect it to.

[mrsman]

Let's take an example.  Suppose the signal is set up to activate the left turn arrow if there are at least three cars sitting in the left turn lane.  In a leading setting, it will notice three cars and provide a protected left.  This is true even if there will likely be gaps in the through phase that follows.  But in a lagging setting, the green arrow follows the through phase.  If enough left turning cars are serviced in the through phase (by finding natural gaps in traffic), then it is possible (even likely) that the left turn phase is skipped entirely.  There are more likely to be cars waiting to turn in the leading scenario since the left turn lane traffic can build up while the light is red.  But in the lagging scenario, we provide the maximum opportunity for cars to clear during the green orb phase, and only if it is still necessary do we light the green arrow to provide a protected left to prevent backups in the left turn lane.

So, IMO, leading lefts better serve the left turners, for the reasons you stated, but lagging lefts better serve the intersection as a whole.  Of course, this assumes that all the yellow trap issues are properly addressed.

A technical note: The detection used with most signals does not count vehicles, especially inductive loops. It works based on presence detection–is there a vehicle within the detection zone or not? Detection zones are usually only a car length or two, but it's usually a "yes/no" input to the controller as opposed to a "0/1/2/...".

And given the random arrival nature of traffic, unless the signal is in a coordinated system where arrival platoons are frequent, you're just as likely to have vehicles stack up in the left turn pocket during red (leading left scenario) as you are during FYA/permitted left (lagging left scenario). When it's time for the controller to address the protected left turn in the cycle, it's just going to look and see whether any left turn vehicles are still detected to determine to activate the green arrow.

With random traffic patterns, yes the left turn lane will stack up at an even rate.  But with a lagging left turn, there is more of a likelihood that whatever cars are there would go during the green phase.  At times of heavy traffic, yes, you will have a left turn signal during every traffic phase.  But if there are only one or two cars waiting to turn during moderate traffic, for the leading left the left turn signal will be triggered -- but under similar conditions the lagging left will not be triggered, becuase the left turn lane will be emptied before the end of the green orb phase as the left turners will go during the natural gaps in traffic.

While it is true that most induction loops are designed merely to detect presence, they can be aligned in such a way as to indirectly "count" the number of vehicles.  Look at this intersection La Cienega/Olympic in Los Angeles, near Beverly Hills.  (View satellite and hide labels).  Do you see all those circles in the left turn lane?  They are induction loops.  It seems that one lane has as many as five loops.  Indirectly, this is used to "count" the number of vehicles.  [By needing more than one of the loops to trigger the signal change.]  My recollection is that Los Angeles DOT traditionally favored thru traffic over left turners, which explains why they were late adopters of left turn signals at many intersections.  And to the extent they allowed left turn signals, which were usually leading lefts unless the opposing left was somehow restricted, you would typically need at least three cars waiting in the lane to activate the signal.

https://www.google.com/maps/place/Beverly+Hills,+CA/@34.0590798,-118.3762021,38m/data=!3m1!1e3!4m5!3m4!1s0x80c2bc04d6d147ab:0xd6c7c379fd081ed1!8m2!3d34.0736204!4d-118.4003563

[roadfro]

A technical note: The detection used with most signals does not count vehicles, especially inductive loops. It works based on presence detection–is there a vehicle within the detection zone or not? Detection zones are usually only a car length or two, but it's usually a "yes/no" input to the controller as opposed to a "0/1/2/...".

While it is true that most induction loops are designed merely to detect presence, they can be aligned in such a way as to indirectly "count" the number of vehicles.  Look at this intersection La Cienega/Olympic in Los Angeles, near Beverly Hills.  (View satellite and hide labels).  Do you see all those circles in the left turn lane?  They are induction loops.  It seems that one lane has as many as five loops.  Indirectly, this is used to "count" the number of vehicles.  [By needing more than one of the loops to trigger the signal change.]  My recollection is that Los Angeles DOT traditionally favored thru traffic over left turners, which explains why they were late adopters of left turn signals at many intersections.  And to the extent they allowed left turn signals, which were usually leading lefts unless the opposing left was somehow restricted, you would typically need at least three cars waiting in the lane to activate the signal.

https://www.google.com/maps/place/Beverly+Hills,+CA/@34.0590798,-118.3762021,38m/data=!3m1!1e3!4m5!3m4!1s0x80c2bc04d6d147ab:0xd6c7c379fd081ed1!8m2!3d34.0736204!4d-118.4003563

Well, I admit that I haven't seen that kind of detector layout before. It would effectively operate as a counter then given the distance from the stop line.

For the example you linked, it seems as though the three loops in the center of the lane are likely the typical presence detection. That lone loop near the start of the turn pocket might actually be a passage gap detector, although passage gap operation is more typical of through lanes. Most setups I'm aware of will use multiple loops for a presence detection zone–in Nevada, it's usually two loops (or sometimes four loops in turn lanes) near the stop line (two loops covers one typical car length), with a pair of loops being wired to the same circuit–but passage gap detection is usually only one loop much further back from the stop line.

[mrsman]

I know I linked this document before, probably in the Pac SW boards, but here is a great history of L.A. Traffic innovations.  Relevant to the conversations in this thread, I have eluded to the Los Angeles DOT philosophy that was traditionally hostile to protected left turns, because it would take time away from through traffic.  While all of this document is interesting, the most pertinent discussion can be located in the section discussed as "The Evolution of Left-Turn Phasing in the  Los Angeles Area " (starting on page 91)

As discussed in Part 1, left-turn phasing in Los Angeles was a novelty, in the 1950's and 1960's, due to signal equipment restrictions and the limited number of continuous raised median islands. By the early
1970's, left-turn phasing became a routine design on State highways and suburban boulevards in areas adjacent
to Los Angeles. However, it would not become a standard feature on Los Angeles' more urban streets due to
the traffic signal operating philosophy that prevailed throughout most of the 1970's.

The philosophy reflected the distinct signal system that
the City of Los Angeles operated. Unlike the Division
of Highways, which operated signals along a few, widely
spaced State Highways, the City of Los Angeles operated
signals throughout a network with signal spacing at approximately 1/4-mile intervals. This type of signal network allowed 30 mile-per-hour progression to be maintained in all directions with short (50 to 70 second) cycle lengths. However, the addition of left turn arrows
would require longer cycle lengths, which, in turn, would
severely compromise progression. This degradation in
progression was avoided by resisting requests to install
left turn arrows.

Part 2 explained how the City's approach to left-turn phasing started to change due to popular demand for
this feature. The turning point was when a judge of the Downtown federal court house could not convince
LADOT staff to approve a left turn arrow for eastbound Temple Street at Main Street. He appealed to Councilman Gilbert Lindsay who forced LADOT to install it by Council motion. Having to be compliant with the
mandate, but disliking being over-ruled, LADOT implemented a form of "occasional"  left-turn phasing at this
location in 1991 that differed in two ways from traditional left turn phasing. The first difference was that motorists could turn on the "protected"  green arrow after having an opportunity to turn during the "permissive"
circular green, upon yielding to opposing traffic. The second difference was that the arrow would be activated
only if the queue of left-turn traffic extended four car lengths, as monitored by in-pavement detectors. This
protected/permissive left turn operation with queue detection, would become known as the Los Angeles form
of left-turn phasing and the model for most subsequent installations in the City.

https://ladot.lacity.org/sites/default/files/documents/transportation-topics-and-tales-milestones-in-transportation-history-in-southern-california.pdf

So yes, while the leading left turn is more popular now in Los Angeles (and in most other places in the country), if the goal was to only allow for a protected green arrow when absolutley necessary, a lagging green would work better as it would only turn green if there were still cars to be serviced at the end of the cycle.

Main St Los Angeles, then and now, is one-way northbound.  So a lagging left turn onto it would not have a yellow trap problem.

Most other PPLTs in Los Angeles (like Olympic/La Cienega that I referred in my other post) are leading lefts.  But yet they are still designed to activate only after a certain number of cars are waiting in the lane.  As Jake pointed out earlier, even in saturated conditions, two cars can probably turn at the end of a cycle.  This is very ingrained in Los Angeles culture (at least when I lived there) and usually two cars were permitted to turn at the end.  So if two cars will turn anyway at then end, we shouldn't stop everyone to give priority to a left turn arrow unless there are a significant number (4 or more) waiting to turn.

And yes, based on my experience, very few other cities do this.  Most other places I've been with PPLT leading left will provide a left turn arrow even if only one car is waiting, and even if there is reasonable likelihood that there will be a gap in traffic during the green orb.

[jakeroot]

Most other PPLTs in Los Angeles (like Olympic/La Cienega that I referred in my other post) are leading lefts.  But yet they are still designed to activate only after a certain number of cars are waiting in the lane.  As Jake pointed out earlier, even in saturated conditions, two cars can probably turn at the end of a cycle.  This is very ingrained in Los Angeles culture (at least when I lived there) and usually two cars were permitted to turn at the end.  So if two cars will turn anyway at then end, we shouldn't stop everyone to give priority to a left turn arrow unless there are a significant number (4 or more) waiting to turn.

And yes, based on my experience, very few other cities do this.  Most other places I've been with PPLT leading left will provide a left turn arrow even if only one car is waiting, and even if there is reasonable likelihood that there will be a gap in traffic during the green orb.

First, thank you for linking that PDF. Absolutely brilliant write-up, and I can already see myself wasting the rest of the day reading it.

Second, yes, that is exactly what I'm getting at with the concept of green arrows. I can totally understand why they might need to exist, but the concept of activating them, potentially throwing off corridor progression and making through traffic wait for the green, yellow and all-red phases is simply a waste of time when that/those cars could either turn in gaps, or at the end. If cities made a concerted effort to coordinate their streets, with fewer dedicated left turn phases, there might be better throughput of traffic and thus more gaps for left turns anyway. Providing a green arrow when traffic in a left turn lane begins to build is more than logical, but providing for a green left turn arrow for a single car is just a waste, and frankly, from my experience, it teaches people to be lazy and too reliant on green arrows which you don't need to turn left.

Where I live in Tacoma, we've actually had quite a lot of influence from Los Angeles over the years. Although our signal placement is nowhere near as good as California, the way signals are operated around here is exactly how it's done in LA: signals are installed with no protected phasing, and are upgraded in time with protected phasing, generally at the beginning of the phase, and the green and yellow arrows are literally tacked onto the left-most signal head at the point where such signals become necessary. Fully-protected turns are relatively rare, and limited to only very busy intersections (many having been replaced with FYA signals over the last couple years anyway).

[jakeroot]

Jakeroot, to answer your question above: It's faster, easier and safer to make a left turn on a leading green-arrow than to wait for a gap. Also visibility of oncoming traffic is an issue at this intersection as the road takes a right bend as it crosses over, so it can be difficult to see oncoming traffic if there are vehicles in the opposing direction waiting to turn left.

But I also have to confess the mischievous satisfaction of beating the County's dumb set-up to make the arrow come on. LOL And as I stated above, the County has since changed the installation to make the arrow work the way most drivers would expect it to.

But the "way most drivers would expect it to [operate]" is entirely symptomatic of how the County has set up their traffic signals in time, which is that drivers get a green arrow as long as they are at the stop line. This basically teaches drivers that you don't need a green arrow to turn. Which is obviously not the case. If it were, the left turns would have red arrows and no permissive phase. If the County is perfectly happy with a permissive phase, why do they go to such great lengths to make drivers interact with this phase as little as possible? As well, if that bend were truly an issue, it wouldn't be permissive.

It is faster for you to turn with a green arrow at the beginning, but all of the traffic coming towards you must suffer for you to gain this advantage. An advanced left for a single car throws off progression along any road, coordinated or not, because it forces traffic to wait 8-12 seconds for a single car to turn at the beginning of the phase, when that single car could easily turn at the end of the through phase during a two-second all-red phase (which could easily be implemented if not already); traffic, of course, can turn in gaps with both setups. By all accounts, such a maneuver is just as safe, because that maneuver takes place when everyone has a red light, and when left turning traffic would have excellent visibility of oncoming traffic (who could be running the red or "pushing" their yellow).

I want to make it very clear: I am absolutely in favor of left turn green arrows, but not at every intersection, and definitely not for a single car. Activating only for multiple cars, and preferably at the end, seems to be the most efficient setup.

In Vancouver, BC, where I spend a fair amount of time, many of their left turn phases activate only during peak hours, when left turning traffic is heavier, and sometimes still only every other phase. This obviously harms through traffic, but is obviously necessary when you have 378 cars trying to turn left, and only two or three are able to turn in gaps at the end. Providing an advanced left at 1130am for a single car with only a couple oncoming cars is a complete waste, and sets an unfortunate precedent for drivers that green arrows should be provided at every single intersection, at all times of day, at the beginning of every signal phase, irrespective of individual intersection characteristics (when no protected phase is more often than not perfectly fine).

[jakeroot]

I am bringing my comment into a second post so that others may consider this specific facet of operation more closely, and ideally, critique my thinking:

Consider that an unspecific intersection has a 120-second total cycle: assuming both roads are equally utilised, that's 60 seconds for each road. If a two-second all red phase is provided, that's 58 seconds for through traffic along both roads. If we implement green arrows at the beginning, that drops to about 50 seconds, with that number dropping further with each additional car turning left. Assuming there is only one or two cars waiting to turn, providing the full 58 seconds for through traffic, and leaving the two-second all-red phase for those two cars to finish their turns, we are providing for the most efficient setup physically possible at a signalized intersection.

Providing a left turn phase at the beginning could reduce the green time for through traffic by around 15 to 20%, if not more. Yet this reduction could potentially exist to allow only one or two cars to turn at the beginning, when they could just as easily turn at the end during an all-red phase (which should exist at all intersections anyways).

The only real issue with my thought process is that, in too many parts of the country, drivers absolutely refuse to enter the intersection during a permissive phase. This stubbornness absolutely harms traffic flow, because engineers basically become forced to provide green arrows for every phase, since drivers who don't enter the intersection will only be able to turn in gaps, at that is certainly no guarantee along busy roads.

In British Columbia, although I did not take a drivers licence exam there, I am quite familiar with ICBC's teachings. The most important thing is that, although it is not taught as a "must", drivers should wait in the intersection when turning left. This allow engineers across BC (and generally across Canada) to use less protected phases, because even a fully saturated phase will have time at the end for a couple cars to finish turning. This is compared to places like Oregon, where apparently it's illegal to be in the intersection on red, thus making it illegal to wait in the intersection unless someone is sure there will be a gap. It should be no surprise that driving in Vancouver is a hell of a lot more enjoyable than driving in Portland, since there are far less green arrows and thus much better throughput along major corridors.

[mrsman]

The real model for great signal progression is the Detroit area.  Tradephoric and others have posted videos showing drivers being able to driving distances of several miles only stopping at red lights a couple of times.  This is true along several corridors Woodward and telegraph being the most famous.  The secret in Detroit for their success are prohibiting most left turns entirely at Major intersections by implementing the so-called Michigan left.  Since you go beyond the intersection 2 then make a u-turn and then a right turn to complete a left turn, you only need to have two separate signal phases at the main intersection.  The proof is in the pudding because signal progression is great over there.

The classic Los Angeles model doesn't quite go this far.  In the busiest parts of the city it is too far built up to have the room to incorporate Michigan lefts. 

Nexus 5X