Zipper Merge News

[kphoger]

If the volume exceeds the capacity for a single lane, jamming and delays will occur in the continuing lane as people try to change out of the terminating lane--in this case, early merging would only worsen the jam (case 2 as described above).

Wrong.

Excellent refutation 10/10       

I wrote full explanation at least twice before. Try looking things up on a previous page.

And ya still didn't learn anything.  From me, from jake, from any of the people who have made pretty much the same point as I did.  So we'll just agree to disagree I guess.

And vdeane summed up the general conclusion pretty well in the post above.

Well, point is that zipper merge elsewhere, as vdeane called it, is effectively an early merge. As long as you agree that is an equally good (or bad - since there is no truly good option here) alternative, we're on the same page.

I'm not convinced that a single merge point–whether early or late–is actually any better than scattered merge points along the path.  The only thing that seems to be universally agreed upon is that coming to a dead stop at the cones and then merging at 4 mph is bad.  Everything else so far has been argued to death but not really supported by actual data.

[paulthemapguy]

Okay, I see the point about a single merge point even if it is early though.  Ideally, if everyone merges at the same point, the effect will be exactly the same as if everyone waited until the last second.  But good luck trying to make everyone agree on the same arbitrary point, when you can just do the normal expected thing and wait until the point that's geographically defined!  It's an impossible proposition, getting everyone to think exactly like you.  And furthermore, shifting the merge to a single earlier point will cause a backward shift of the front and back endpoints of the resulting queue.  The jam will end earlier and also start earlier, causing a greater probability of blocking intersections and driveways upstream of the merge.

[kalvado]

If the volume exceeds the capacity for a single lane, jamming and delays will occur in the continuing lane as people try to change out of the terminating lane--in this case, early merging would only worsen the jam (case 2 as described above).

Wrong.

Excellent refutation 10/10       

I wrote full explanation at least twice before. Try looking things up on a previous page.

And ya still didn't learn anything.  From me, from jake, from any of the people who have made pretty much the same point as I did.  So we'll just agree to disagree I guess.

And vdeane summed up the general conclusion pretty well in the post above.

Well, point is that zipper merge elsewhere, as vdeane called it, is effectively an early merge. As long as you agree that is an equally good (or bad - since there is no truly good option here) alternative, we're on the same page.

I'm not convinced that a single merge point–whether early or late–is actually any better than scattered merge points along the path.  The only thing that seems to be universally agreed upon is that coming to a dead stop at the cones and then merging at 4 mph is bad.  Everything else so far has been argued to death but not really supported by actual data.

Let people settle with idea that merging at last instant (with coming to complete stop and forcing way into a formed line)  is not the best option. Then we may talk about distributed, early, late and what not.
And actually there are 2 groups here.

"The agreed upon merge location"? Says who? I see NO agreement on the roads, period. People get over at all sorts of different points and then they all seem to say, in essence, "Thou shall not use this lane past this point because I said so."

So you are that asshole, who drives in an empty lane past patiently waiting cars and demands an entry at the head of the queue?

No, he's the smart one using the full capacity of the road like everyone else should be doing.

[kalvado]

Okay, I see the point about a single merge point even if it is early though.  Ideally, if everyone merges at the same point, the effect will be exactly the same as if everyone waited until the last second.  But good luck trying to make everyone agree on the same arbitrary point, when you can just do the normal expected thing and wait until the point that's geographically defined!  It's an impossible proposition, getting everyone to think exactly like you.  And furthermore, shifting the merge to a single earlier point will cause a backward shift of the front and back endpoints of the resulting queue.  The jam will end earlier and also start earlier, causing a greater probability of blocking intersections and driveways upstream of the merge.

Single point early merge, from my experience,  is usually "once you see a line waiting - join the end of the line". Not conceptually different from zipper, first in-first out.
Usually once the reason for the merge (accident, stopped vehicle, "lane end" sign) is no longer in sight from points at the end of the queue, people no longer know where to merge, and single point merge (usually several car lengths - consistent with openings in continuing lane)  moves to the point where sign/accident is visible. Traffic is moving in congestion mode at that point anyway.

What I see as a wrong behavior is driving  in already vacated lane past a formed line and jumping the queue at the cones. But some see that as "using all available resources".

[Mr_Northside]

PennDOT has signs in some work zones where they want to have a zipper merge...one stating "USE BOTH LANES TO MERGE POINT" and then at the merge point, "MERGE HERE - TAKE YOUR TURN".  THIS would clear things up quite nicely if used more widespread where engineers determined a zipper merge should be used.

Yes.  In addition to the signs, they tried to get the word out in print and TV media (when doing press releases about the construction projects) on how to properly merge as well.  I gotta give PennDOT some credit there.

I'm not convinced that a single merge point–whether early or late–is actually any better than scattered merge points along the path.

Having been in zones with no indication on "how" / where to merge, and thru the explicit "Zipper Merge" that PennDOT has been pushing - and almost everyone actually doing so - I can say that I am convinced that it is better.  (And I have at least 2 friends who now think so, and at least 1 thought it was stupid when they started popping up a few years ago)

[kphoger]

The only thing that seems to be universally agreed upon is that coming to a dead stop at the cones and then merging at 4 mph is bad. 

Let people settle with idea that merging at last instant (with coming to complete stop and forcing way into a formed line) is not the best option.

Merging at the last instant with coming to complete stop and forcing way into a formed line is not the same thing as merging at the last instant without doing so.  You seem to be missing the fact that a zipper merge explicitly expects people in both lanes to take turns at the merge point.  There's no such thing as "forcing your way in" at a true zipper merge, because there is no "right" and "wrong" lane to be in.  It's essentially a "form one lane" situation.  This notion of zooming past traffic or coming to a stop in the "ending lane" is by nature contrary to what a zipper merge is intended to do, which is to make both lanes the "ending lane" or, if you prefer, both lanes the "continuing lane."

[kalvado]

The only thing that seems to be universally agreed upon is that coming to a dead stop at the cones and then merging at 4 mph is bad. 

Let people settle with idea that merging at last instant (with coming to complete stop and forcing way into a formed line) is not the best option.

Merging at the last instant with coming to complete stop and forcing way into a formed line is not the same thing as merging at the last instant without doing so.  You seem to be missing the fact that a zipper merge explicitly expects people in both lanes to take turns at the merge point.  There's no such thing as "forcing your way in" at a true zipper merge, because there is no "right" and "wrong" lane to be in.  It's essentially a "form one lane" situation.  This notion of zooming past traffic or coming to a stop in the "ending lane" is by nature contrary to what a zipper merge is intended to do, which is to make both lanes the "ending lane" or, if you prefer, both lanes the "continuing lane."

you're 100% right, we're not only on the same page, we're on the same line.
Earlier in this thread there was a great idea - which I totally like: do not go against common trend. Everyone merges at last point - let that go smoothly and merge there. Everyone merges early - let that go smoothly.  Outcome is the same, give or take a tiny bit.

Look again at Rothman's and 1995hoo's comments: their point is if everyone is doing it wrong, I will still do it right by merging at the end. And that is the problem from my perspective.

[Rothman]

Look again at Rothman's and 1995hoo's comments: their point is if everyone is doing it wrong, I will still do it right by merging at the end. And that is the problem from my perspective.

Glad to know that we're doing it right and you're doing it wrong, then.

[kphoger]

My main problem with early merging is that it the expectation can lead to road rage.  A driver zooming past everyone in the near-vacant lane and then butting in makes the other drivers angry and encourages them to tail-gate in order to box him out; then, when he does find a spot, people have to brake.  Also, if early merging is the expectation, then there are the vigilantes who think it's their job to ride the center stripe and block people from using the near-vacant lane, which encourages potentially unsafe maneuvers on the shoulder.

[kalvado]

Look again at Rothman's and 1995hoo's comments: their point is if everyone is doing it wrong, I will still do it right by merging at the end. And that is the problem from my perspective.

Glad to know that we're doing it right and you're doing it wrong, then.

Well, First Amendment says you are always entitled for your wrong opinion

[kphoger]

The only thing that seems to be universally agreed upon is that coming to a dead stop at the cones and then merging at 4 mph is bad. 

Let people settle with idea that merging at last instant (with coming to complete stop and forcing way into a formed line) is not the best option.

Merging at the last instant with coming to complete stop and forcing way into a formed line is not the same thing as merging at the last instant without doing so.  You seem to be missing the fact that a zipper merge explicitly expects people in both lanes to take turns at the merge point.  There's no such thing as "forcing your way in" at a true zipper merge, because there is no "right" and "wrong" lane to be in.  It's essentially a "form one lane" situation.  This notion of zooming past traffic or coming to a stop in the "ending lane" is by nature contrary to what a zipper merge is intended to do, which is to make both lanes the "ending lane" or, if you prefer, both lanes the "continuing lane."

you're 100% right, we're not only on the same page, we're on the same line.
Earlier in this thread there was a great idea - which I totally like: do not go against common trend. Everyone merges at last point - let that go smoothly and merge there. Everyone merges early - let that go smoothly.  Outcome is the same, give or take a tiny bit.

Look again at Rothman's and 1995hoo's comments: their point is if everyone is doing it wrong, I will still do it right by merging at the end. And that is the problem from my perspective.

There is room in this world for a campaign to change the common trend.  In my opinion, the zipper merge wastes less pavement for vehicle storage in the event of jams, encourages fair respect by/for drivers in both lanes, gives a clear expectation of behavior, and offers no significant disadvantage over early merging.  So I'm in favor of campaigns to implement the zipper merge, even if it doesn't actually improve the delay impact of jams to any degree whatsoever.

[kalvado]

There is room in this world for a campaign to change the common trend.  In my opinion, the zipper merge wastes less pavement for vehicle storage in the event of jams, encourages fair respect by/for drivers in both lanes, gives a clear expectation of behavior, and offers no significant disadvantage over early merging.  So I'm in favor of campaigns to implement the zipper merge, even if it doesn't actually improve the delay impact of jams to any degree whatsoever.

There is room in this world for a lot of different things - but we cannot afford all of them at the same time.
Question is if such campaign would be cost effective; and where it has to start.
One of interesting questions is legal meaning of this sign:

I believe (maybe I couldn't find appropriate paragraph) NY doesn't define driver behavior. To the point of authorities interpreting things differently.
I believe I have seen "ending lane yields" and "both lanes equal" in other state's laws.
Which basically means zipper campaign has to start with states legislation harmonization - and that is EXPENSIVE.

[kphoger]

There is room in this world for a campaign to change the common trend.  In my opinion, the zipper merge wastes less pavement for vehicle storage in the event of jams, encourages fair respect by/for drivers in both lanes, gives a clear expectation of behavior, and offers no significant disadvantage over early merging.  So I'm in favor of campaigns to implement the zipper merge, even if it doesn't actually improve the delay impact of jams to any degree whatsoever.

There is room in this world for a lot of different things - but we cannot afford all of them at the same time.
Question is if such campaign would be cost effective; and where it has to start.
One of interesting questions is legal meaning of this sign:

I believe (maybe I couldn't find appropriate paragraph) NY doesn't define driver behavior. To the point of authorities interpreting things differently.
I believe I have seen "ending lane yields" and "both lanes equal" in other state's laws.
Which basically means zipper campaign has to start with states legislation harmonization - and that is EXPENSIVE.

I have never seen this signage at a zipper merge.

[kalvado]

There is room in this world for a campaign to change the common trend.  In my opinion, the zipper merge wastes less pavement for vehicle storage in the event of jams, encourages fair respect by/for drivers in both lanes, gives a clear expectation of behavior, and offers no significant disadvantage over early merging.  So I'm in favor of campaigns to implement the zipper merge, even if it doesn't actually improve the delay impact of jams to any degree whatsoever.

There is room in this world for a lot of different things - but we cannot afford all of them at the same time.
Question is if such campaign would be cost effective; and where it has to start.
One of interesting questions is legal meaning of this sign:

I believe (maybe I couldn't find appropriate paragraph) NY doesn't define driver behavior. To the point of authorities interpreting things differently.
I believe I have seen "ending lane yields" and "both lanes equal" in other state's laws.
Which basically means zipper campaign has to start with states legislation harmonization - and that is EXPENSIVE.

I have never seen this signage at a zipper merge.

Yet it is default sign for lane end - and a point where lanes have to merge. Defining default right of way at merge - each and every merge - is a must if you want to achieve uniform motorist compliance. Interestingly enough, that may include merge of highway acceleration lanes. 

[kphoger]

Defining default right of way at merge - each and every merge - is a must if you want to achieve uniform motorist compliance.

I do not agree with that statement.  "Take turns" or "Form one lane" is an appropriate indication.

[kalvado]

Defining default right of way at merge - each and every merge - is a must if you want to achieve uniform motorist compliance.

I do not agree with that statement.  "Take turns" or "Form one lane" is an appropriate indication.

Well, then you need to make sure drivers post appropriate sign before they crash and close that damn lane.. 
Or you still need a law (52 of thm, including DC and PR) defining default behavior and mechanisms to override the default....

[jakeroot]

Defining default right of way at merge - each and every merge - is a must if you want to achieve uniform motorist compliance.

I do not agree with that statement.  "Take turns" or "Form one lane" is an appropriate indication.

Well, then you need to make sure drivers post appropriate sign before they crash and close that damn lane..
Or you still need a law (52 of thm, including DC and PR) defining default behavior and mechanisms to override the default....

Why can't the default law be, "take turns when forming single-file line"? If a crash occurs at the merge point, the fault is 50/50. Having one lane with absolute priority promotes road rage, because those drivers rightfully have the right of way, and can single-handedly ruin a zipper merge. Many drivers merge in advance because they know the drivers in the thru-lane don't have to let them in. If the law designated both lanes as equal, and any collision at the merge point as having 50/50 fault, drivers might be more likely to treat a merge with capitulation.

Then again, I'm not a law-specialist; maybe this type of law isn't allowed to exist.

[kalvado]

Defining default right of way at merge - each and every merge - is a must if you want to achieve uniform motorist compliance.

I do not agree with that statement.  "Take turns" or "Form one lane" is an appropriate indication.

Well, then you need to make sure drivers post appropriate sign before they crash and close that damn lane..
Or you still need a law (52 of thm, including DC and PR) defining default behavior and mechanisms to override the default....

Why can't the default law be, "take turns when forming single-file line"? If a crash occurs at the merge point, the fault is 50/50. Having one lane with absolute priority promotes road rage, because those drivers rightfully have the right of way, and can single-handedly ruin a zipper merge. Many drivers merge in advance because they know the drivers in the thru-lane don't have to let them in. If the law designated both lanes as equal, and any collision at the merge point as having 50/50 fault, drivers might be more likely to treat a merge with capitulation.

Then again, I'm not a law-specialist; maybe this type of law isn't allowed to exist.

Next thing you have to explain how such merge is different from highway acceleration lane merge. Or - as we have here - a merge with an extra lane going through traffic light. merge after passing.
I am sure if I make a post on "weird lane ends" there will be a bunch of examples from all over the place. Problem is that the law must address ALL of them in some sensible way.
A lot of these things actually hang on common courtesy - like truck drivers moving over to let traffic from the ramp into the flow.  If you will, that is sort of why i emphasize "common" way of doing things...

[jakeroot]

A lot of these things actually hang on common courtesy - like truck drivers moving over to let traffic from the ramp into the flow.  If you will, that is sort of why i emphasize "common" way of doing things...

On this last point. Washington's "Keep right except to pass" law (RCW 46.61.100) has a section on moving left to allow people to merge:

Upon all roadways having two or more lanes for traffic moving in the same direction, all vehicles shall be driven in the right-hand lane then available for traffic, except...(c) when moving left to allow traffic to merge...

[mrsman]

As far as I'm aware, on a regular lane drop, the lane that is ending must yield to the lane that is continuing and that somebody in that lane should merge when it is prudent to do so and not wait to the last minute.

There are places, though, where a zipper merge is preferred.  For example, where construction normally closes a lane, this is the preferred method.  And generally there will be some signs to state that, although the signs aren't very clear.

I know of a permanent zipper merge in the Baltimore area on I-70 approaching the north I-695 ramp.  They have some very good signage there.

FORM SINGLE LANE AHEAD:
https://www.google.com/maps/@39.3053042,-76.7492325,3a,75y,113.46h,81.94t/data=!3m6!1e1!3m4!1sUKj_IxMECx2XNI5Ih0EuGw!2e0!7i13312!8i6656

ALTERNATE RIGHT OF WAY, with yellow diamond sign showing both lanes merging to the middle.  It's a little different than the sign posted earlier in the thread:
https://www.google.com/maps/@39.3047953,-76.7475411,3a,75y,113.46h,81.94t/data=!3m6!1e1!3m4!1sllG0iY8qvW7-6aG1xit-TQ!2e0!7i13312!8i6656

FORM SINGLE LANE NOW:
https://www.google.com/maps/@39.3048744,-76.7460204,3a,75y,52.67h,78.61t/data=!3m6!1e1!3m4!1s3futPYb7D5RVhCC_M7cvCw!2e0!7i13312!8i6656

So here, there is a purposeful zipper merge.  The three lanes of I-70 essentially come to an end at I-695.  Originally, the right lane led to I-695 south, the middle lane to I-695 north, and the left lane led to the park and ride.  Now, the right lane leads to I-695 south and the left two lanes lead to I-695 north, with a left exit to the park and ride.  [Very little traffic heads that way.]  Even though two lanes of traffic go to I-70, the ramp narrows at a point to one lane and there used to be tremendous backups here.  By encouraging the zipper merge, the backup is allowed to spread over both lanes, so it is not as severe.

[doorknob60]

In Idaho, the drivers manual states that "merging traffic must yield" on a normal lane drop (eg the following sign).


The Ada County Highway Distric Road Wizard has talked about merging a fair amount in the newspaper column (which by the way is quite interesting to read for road geeks).

Here's one example, this one on a surface street, but same idea as a freeway.

Dear Road Wizard: With abundant two-thru-lane intersections that merge down to one lane past the intersection in Boise, and especially Meridian, why don't Idaho drivers happily alternately merge (like most other states) for faster movement of traffic? Here drivers line up for blocks in the left-most thru lane instead of also using the right-most thru lane and merging down to one for fear of road rage and people saying they "cut them off." The intersection of Eagle Road and Chinden Boulevard is a prime example. Don't people realize that everyone will move faster if they allow merging? Merging is not "cutting someone off,” it's a rule of the road. Paula


The classic thru-traffic showdown. In one corner of the ring, a driver in the left-most thru lane who has waited out multiple signal cycles during rush hour traffic to finally make the green light. In the other, a motorist who has used the right-most thru lane to zip past drivers on the left, demanding to merge where the right lane ends past the signal. Who deserves the title of “First Down the Road”?

Preventing someone from merging may feel like winning the battle, but it’s really losing the war, which is otherwise known as the struggle to get everyone into the one lane in the most efficient way. The better approach is for drivers in the left-most lane to leave gaps for merging traffic while maintaining a steady speed. People in the right-most lane should look for those gaps and try to merge at about the same speed as traffic to the left. Motorists should use—but not abuse—either lane. Don’t be shy (or rude)!

Source: http://www.achdidaho.org/Departments/PR/RoadWizard/Entries.aspx?MID=451

And 2 weeks later:

Dear Road Wizard: I'm glad “Paula” brought up the two thru lanes on Chinden Boulevard at Eagle Road. You see, I'm one of those guys who wait in line on the left for traffic to proceed past the light. Why do I do this? It’s because I'm following the rules of the road. There is only one westbound thru lane on Chinden past Eagle Road. The right-hand lane past the signal is designated as a turn lane into the Banbury subdivision. Problem is, folks like Paula think it’s a merge lane and use it to cut off drivers in the correct thru lane. By the time four or five cycles pass and a person in the left lane finally makes it to the light, most have no compassion for the "cheater" who is in the right lane. Does anyone share my view? Dan


I decided to be one of “those” people who use the right-most thru lane to pass drivers on the left at this intersection, then cut into traffic at the last minute where the right lane ends just beyond the signal.

Drivers in the left-most thru lane made room for those of us merging from the right, and everyone got down the road smoothly without any noticeable anger or inappropriate gesturing.

People don’t always behave this way, but the right lane actually is a designated merge lane. There is a sign posted and painted arrows (somewhat faded) on the lane that direct drivers to merge. The lane then becomes a right-turn-only entrance to the Banbury subdivision. A source of confusion, perhaps, and maybe “cheaters” are going too far past the merge point, but Paula had it right. Both thru lanes at the intersection can be used by drivers continuing westward.

Source: http://www.achdidaho.org/Departments/PR/RoadWizard/Entries.aspx?MID=452

I go through the I-184/I-84 westbound interchange during rush hour, where it's basically 6 lanes of traffic (from left to right: 3 lanes I-84, 2 lanes I-184, 1 lane Franklin Rd.) merge into 4 lanes, where the far two right lanes end. The traffic in the farther left lanes (I-84) is much lighter than the traffic coming from I-184, which all has to merge. The farthest left I-184 lane (the one that does not end) backs up way farther than the other I-184 lane. I always take the right-most of those lanes. I cruise in that lane until a large gap to my left appears (usually happens after the ramp from Franklin merges in, and where it's 5 lanes, but before the farthest right lane ends). Only a small handful of times I've had to ride it out all the way to the end of that lane. When I did, I had no trouble merging in, nobody "blocked me out". But people here do tend to merge really early, often clogging up the non-merging lane.

I feel like that interchange might be better if one lane on the left ended (instead of a second right lane ending), but that could slow things down outside of commuting hours (where I-184 traffic is light), and increase weaving of I-84 drivers getting into the far right lane to exit at Eagle Rd./ID-55. So I don't know if it would be that much better.

[mrsman]

In many cases, the road markings are designed for the busiest times, so it appears that it would be most helpful if lanes on both the left and the right merged and not put the full onus of merging on I-184 traffic.

So generally speaking, if 6 lanes merge rapidly to 4 lanes, it seems it would make the most sense to have the left most lane and the right most lane end, instead of quickly ending two lanes on one side or the other.  But the highway authorities apparently disagree.

[jeffandnicole]

In many cases, the road markings are designed for the busiest times, so it appears that it would be most helpful if lanes on both the left and the right merged and not put the full onus of merging on I-184 traffic.

So generally speaking, if 6 lanes merge rapidly to 4 lanes, it seems it would make the most sense to have the left most lane and the right most lane end, instead of quickly ending two lanes on one side or the other.  But the highway authorities apparently disagree.

Traffic merging from both sides is generally a disaster.

[1995hoo]

I said I wasn't going to waste my time with this thread any more given the troll.

But then this afternoon I thought of a semi-related question. It's not quite the same issue, but I think it's pretty close. See Google Maps satellite view linked below of I-66 near Fair Oaks Mall in Fairfax. I was getting onto the highway via the ramp that curves in from the top right of the picture (I had been at an all-day meeting near Fair Oaks and for some stupid reason I decided to take I-66 west to the Fairfax County Parkway, not thinking about what the rain would do to the traffic). Traffic on I-66 was at pretty much a complete standstill, or at best what the radio in Raleigh used to call "stop-and-roll" where you use the clutch more than the accelerator.

I think most of us would agree that it is not appropriate for someone who has nobody ahead of him in the onramp merge lane (I won't say "acceleration lane" given how slow traffic was) to come to a stop right at the point where the shoulder changes color, well before even the end of the solid single white line. Yet that is what some guy in a minivan did, putting his blinker on and sitting there. Why in the world would you not continue down the merge lane in this situation?

https://www.google.com/maps/@38.8612864,-77.3531256,179m/data=!3m1!1e3

[UCFKnights]

For those early mergers, I'm curious what you all think of the frequent expansion of roads at "choke points" only to return to the smaller size right after them. Should we not use the expansions designed to alleviate traffic?

For example, when too many people are trying to turn left onto a road or highway, they frequently install a second left turn lane and a few hundred feet of a second lane on the road to the left to allow people to merge without requiring the left turn signal to be green and impeding the opposing thru traffic for so long?

Also, I know in Palm Beach County, they frequently install a 2nd or 3rd lane for a few hundred feet before a busy intersection only for it to end a few hundred feet after, so people can "skip the line" and not get stuck missing a light cycle. Doing this seems to successfully allow the road to handle much more traffic without the expense of expanding the entire road to 3 lanes, such as if you look at eastbound Clint Moore here: https://www.google.com/maps/@26.4077098,-80.1464466,195m/data=!3m1!1e3

They usually seem to try to expand all the intersections first to alleviate traffic, and only after that begins to fail, connect the 3rd lanes at the intersections (which also makes that eventual 3rd lane construction much less disruptive, as they don't need to rebuild the intersections at that point). If we're all supposed to early merge, wouldn't we never use these lanes specifically designed for us to enter so we can pass the choke point?