Why can't people merge onto an interstate properly anymore?

[vdeane]

I saw someone end their life by an improper merge.  There is a large merge from TN 396 to I-65 north that goes on for a good distance.  A typical day consists of what I call right lane runners running the right lane all the way to the merge point.  Keep in mind there is like a good mile to make this movement. It starts out as two lanes going into two lanes until it is one lane going to two to the final two lanes.  The right lane runners keep running the right lane until they cannot any longer.  This causes traffic slow downs. 

I cannot believe that people just cannot see this for what it is.  If folks would merge as soon as possible then they would not choke up the highway such as they do.  Traffic can come to a standstill due to this.  Going back to the man that lost his life.  He decided to be a right lane run all the way to the must merge point and then tried slingshot across the "fast lane" to get ahead of the game like so many others.  Well needless to say that he miscalculated the space available and then proceeded to plant himself into his own car by the collision he just caused and he was gone.


https://www.google.com/maps/@35.7488561,-86.8707593,3a,75y,13.12h,80.49t/data=!3m7!1e1!3m5!1suQoYnNdzZQMuDhd26pRG1A!2e0!6shttps:%2F%2Fstreetviewpixels-pa.googleapis.com%2Fv1%2Fthumbnail%3Fpanoid%3DuQoYnNdzZQMuDhd26pRG1A%26cb_client%3Dmaps_sv.tactile.gps%26w%3D203%26h%3D100%26yaw%3D104.031876%26pitch%3D0%26thumbfov%3D100!7i13312!8i6656

Have you ever heard of a zipper merge? 

I cannot believe that people just cannot see this for what it is.  If folks would merge as soon as possible then they would not choke up the highway such as they do.  Traffic can come to a standstill due to this.

If folks would merge as soon as possible, the highway would be choked up more. Running until the end then zippering in lessens the congestion. They are doing it right. You're doing it wrong.

Additionally, you mention traffic congestion is an issue. Aren't you similarly opposed to highway widening on this segment of I-65 that would reduce that congestion? Kind of seems counterintuitive.

It appears I-65 needs to be widened from 4 to 8 lanes between I-840 and TN-396.

Of course, if you're the only person "doing a zipper merge", then it's not really a zipper merge, is it?  Especially if traffic has to slow down or stop to let you in.  A zipper merge is about two lines of cars taking turns to merge, not about one guy passing a bunch of people and then forcing them to stop to let him in.

My methodology is, if I can merge wherever without affecting traffic, then I will merge at the end.  If I can merge now without affecting traffic but have reason to believe that will change, I will merge now.  If traffic is so gunked up that traffic will have to let me in no matter what, I'll merge at the end.  It's people who can merge without affecting traffic but pass that opportunity by to pass the stopped traffic ahead (or because their preferred speed was simply faster than traffic was going in the first place) and then need to be let in that are annoying.

Now would things be more efficient if everyone uniformly merged in the end, in terms of vehicles per hour?  Yes.  But we drive in the real world where people don't do that in practice, and so need to devise our behavior accordingly.

[kphoger]

Now would things be more efficient if everyone uniformly merged in the end, in terms of vehicles per hour?  Yes.

Has this actually been scientifically shown to be true?  Despite all the conversations we've had on here about the zipper merge, and as much as I'm in favor of it, I can't recall having seen anyone link to a study showing it to be proven more efficient in any metric.

[sprjus4]

https://itre.ncsu.edu/itre-studying-how-zipper-merges-reduce-congestion-at-sites-across-north-carolina/
https://uknowledge.uky.edu/cgi/viewcontent.cgi?article=2602&context=ktc_researchreports

[jakeroot]

This definitely bears repeating for anyone who did not see it:

To the OP:

Think about this.  A forum of dozens of people from all over the country, of all different ages, with driving experience ranging from relatively little to very extensive, several of whom have studied the pertinent traffic laws in detail, who bring various personalities and temperaments and worldviews to the table–almost unanimously and unwaveringly describe the circumstances you capture on video and share with us as being normal, commonplace driving situations, the sort of thing a good driver should expect to see and account for;  and, as far as I can recall, have without exception described your reactions behind the wheel as the most hazardous factor in each one we've seen.

That should give you pause and prompt you to reevaluate your outlook.  This is not just a couple of guys chatting in the break room at work.  Rather, this forum is a cross-section of those most interested in roads and driving in America.  Take the opinions expressed here seriously.

This sums up perfectly what I have observed across |crash_it|'s thread littering: total disregard for any opinion that is not his. What makes an already bad driver even worse is failure to recognize when they have committed an error of their own. It may take some time, since humans are naturally stubborn, but eventually they must begin to wonder why they have so much content for their bad-driving YouTube channel (hint: I think he's making poor judgements on purpose!)

[jakeroot]

A typical day consists of what I call right lane runners running the right lane all the way to the merge point

If there's not already an existing name to describe an issue, I'd likely call it more a personal problem than a societal one.

[Ketchup99]

Ok, admittedly, I'm not really an experienced driver, I've only had a license for a little over a year. But I believe the technical term for people who refuse to move over to let someone merge on the highway is an asshole. If I'm driving and I see someone merging, if I can move over, I do. (It seems you had ample room to do so.) If I can't move over, I slow down (if the car is ahead of me) or speed up (if behind me) so that they can get in without an issue. Am I legally required to do this under Pennsylvania law? I have no idea. But it's called courtesy - I do it every time and I appreciate it when others do it for me - and it makes the road safer for everyone.

[kalvado]

Now would things be more efficient if everyone uniformly merged in the end, in terms of vehicles per hour?  Yes.

Has this actually been scientifically shown to be true?  Despite all the conversations we've had on here about the zipper merge, and as much as I'm in favor of it, I can't recall having seen anyone link to a study showing it to be proven more efficient in any metric.

Well, there is one reason to use wide backups - it reduces backup length possibly clearing upstream exits. Throughput of last second merge in congestion mode is worse, obviously

[ran4sh]

https://itre.ncsu.edu/itre-studying-how-zipper-merges-reduce-congestion-at-sites-across-north-carolina/
https://uknowledge.uky.edu/cgi/viewcontent.cgi?article=2602&context=ktc_researchreports

Usually when it is studied, it's because they are looking for ways to reduce already-existing congestion. Sometimes a formal research study will state that assumption, but a lot of times they won't.

But there are also merges that are not congested. Going all the way to the merge point on those would just cause congestion in what was otherwise free-flowing traffic. The correct move is to maintain speed and merge at the first safe opportunity.

[Avalanchez71]

A typical day consists of what I call right lane runners running the right lane all the way to the merge point

If there's not already an existing name to describe an issue, I'd likely call it more a personal problem than a societal one.

I say if one was to merge as soon as safely permitted it was ease traffic concerns. 

[sprjus4]

A typical day consists of what I call right lane runners running the right lane all the way to the merge point

If there's not already an existing name to describe an issue, I'd likely call it more a personal problem than a societal one.

I say if one was to merge as soon as safely permitted it was ease traffic concerns.

What situation are you describing? An on ramp lane coming into a mainline that's already backed up?

[Avalanchez71]

A typical day consists of what I call right lane runners running the right lane all the way to the merge point

If there's not already an existing name to describe an issue, I'd likely call it more a personal problem than a societal one.

I say if one was to merge as soon as safely permitted it was ease traffic concerns.

What situation are you describing? An on ramp lane coming into a mainline that's already backed up?

Negative.  This is the beginning of the back-up.  The traffic flow moving is free-flowing up to SR 396.  The problem is that the right lane runners wait until the last minute to merge causing the back-up.

[kphoger]

Ok, admittedly, I'm not really an experienced driver, I've only had a license for a little over a year. But I believe the technical term for people who refuse to move over to let someone merge on the highway is an asshole.

People who, when merging, fully expect everyone else to move out of their way to accommodate them:  those should be labeled the same.

[jakeroot]

A typical day consists of what I call right lane runners running the right lane all the way to the merge point

If there's not already an existing name to describe an issue, I'd likely call it more a personal problem than a societal one.

I say if one was to merge as soon as safely permitted it was ease traffic concerns.

So if we all merged when you deem appropriate, it would be fine then. All hail king Avalanchez71!

The problem with your approach is the vagueness of where "as soon as safely permitted" would be; the safest point literally changes from driver to driver, perspective to perspective. Some may just deem it inappropriate to ever use lanes that end under your guidance.

Luckily, our DOTs are aware of these issues, so they established a merge point for us: where the lane tapers and the arrows point left or right. Merging at this point maximizes roadway capacity, minimizes confusion around "the right place to merge", and maximizes safety as drivers know *this* point is where they need to watch for merging traffic.

[kalvado]

A typical day consists of what I call right lane runners running the right lane all the way to the merge point

If there's not already an existing name to describe an issue, I'd likely call it more a personal problem than a societal one.

I say if one was to merge as soon as safely permitted it was ease traffic concerns.

So if we all merged when you deem appropriate, it would be fine then. All hail king Avalanchez71!

The problem with your approach is the vagueness of where "as soon as safely permitted" would be; the safest point literally changes from driver to driver, perspective to perspective. Some may just deem it inappropriate to ever use lanes that end under your guidance.

Luckily, our DOTs are aware of these issues, so they established a merge point for us: where the lane tapers and the arrows point left or right. Merging at this point maximizes roadway capacity, minimizes confusion around "the right place to merge", and maximizes safety as drivers know *this* point is where they need to watch for merging traffic.

While a very well versed statement, one should take it with a grain - or rather teaspoon of salt and be prepared to dump a bucket of bullshit out of it. Merging at the last second may work fine while traffic is still free flowing throughout  - and then it really makes little difference;  but last second merge obviously makes things worse in congested flow.

[jakeroot]

So if we all merged when you deem appropriate, it would be fine then. All hail king Avalanchez71!

The problem with your approach is the vagueness of where "as soon as safely permitted" would be; the safest point literally changes from driver to driver, perspective to perspective. Some may just deem it inappropriate to ever use lanes that end under your guidance.

Luckily, our DOTs are aware of these issues, so they established a merge point for us: where the lane tapers and the arrows point left or right. Merging at this point maximizes roadway capacity, minimizes confusion around "the right place to merge", and maximizes safety as drivers know *this* point is where they need to watch for merging traffic.

While a very well versed statement, one should take it with a grain - or rather teaspoon of salt and be prepared to dump a bucket of bullshit out of it. Merging at the last second may work fine while traffic is still free flowing throughout  - and then it really makes little difference;  but last second merge obviously makes things worse in congested flow.

Speaking of grains: I've never seen a grain of evidence to support that statement.

I don't think there is anything "obvious" about which is more correct. I can see the logic behind merging when there is an "opportunity" (again -- very vague -- usually a bad thing when speaking about driving), but I can also see the bigger picture which is that congestion is not usually caused by the merge itself; at I-705 & I-5 in Tacoma, it's lane changing between auxiliary lanes post-merge that causes the backups, not the lane-drops on I-705. Drivers gain nothing by merging "early" because the lane-drop is not the source of congestion.

[kalvado]

So if we all merged when you deem appropriate, it would be fine then. All hail king Avalanchez71!

The problem with your approach is the vagueness of where "as soon as safely permitted" would be; the safest point literally changes from driver to driver, perspective to perspective. Some may just deem it inappropriate to ever use lanes that end under your guidance.

Luckily, our DOTs are aware of these issues, so they established a merge point for us: where the lane tapers and the arrows point left or right. Merging at this point maximizes roadway capacity, minimizes confusion around "the right place to merge", and maximizes safety as drivers know *this* point is where they need to watch for merging traffic.

While a very well versed statement, one should take it with a grain - or rather teaspoon of salt and be prepared to dump a bucket of bullshit out of it. Merging at the last second may work fine while traffic is still free flowing throughout  - and then it really makes little difference;  but last second merge obviously makes things worse in congested flow.

Speaking of grains: I've never seen a grain of evidence to support that statement.

I don't think there is anything "obvious" about which is more correct. I can see the logic behind merging when there is an "opportunity" (again -- very vague -- usually a bad thing when speaking about driving), but I can also see the bigger picture which is that congestion is not usually caused by the merge itself; at I-705 & I-5 in Tacoma, it's lane changing between auxiliary lanes post-merge that causes the backups, not the lane-drops on I-705. Drivers gain nothing by merging "early" because the lane-drop is not the source of congestion.

OK, so what is the mechanism of congestion from the bigger picture point of view?

[jakeroot]

So if we all merged when you deem appropriate, it would be fine then. All hail king Avalanchez71!

The problem with your approach is the vagueness of where "as soon as safely permitted" would be; the safest point literally changes from driver to driver, perspective to perspective. Some may just deem it inappropriate to ever use lanes that end under your guidance.

Luckily, our DOTs are aware of these issues, so they established a merge point for us: where the lane tapers and the arrows point left or right. Merging at this point maximizes roadway capacity, minimizes confusion around "the right place to merge", and maximizes safety as drivers know *this* point is where they need to watch for merging traffic.

While a very well versed statement, one should take it with a grain - or rather teaspoon of salt and be prepared to dump a bucket of bullshit out of it. Merging at the last second may work fine while traffic is still free flowing throughout  - and then it really makes little difference;  but last second merge obviously makes things worse in congested flow.

Speaking of grains: I've never seen a grain of evidence to support that statement.

I don't think there is anything "obvious" about which is more correct. I can see the logic behind merging when there is an "opportunity" (again -- very vague -- usually a bad thing when speaking about driving), but I can also see the bigger picture which is that congestion is not usually caused by the merge itself; at I-705 & I-5 in Tacoma, it's lane changing between auxiliary lanes post-merge that causes the backups, not the lane-drops on I-705. Drivers gain nothing by merging "early" because the lane-drop is not the source of congestion.

OK, so what is the mechanism of congestion from the bigger picture point of view?

Demand outstripping supply. That could mean two lanes instead of three, or a weave where there should be a flyover.

As to how ending a lane somehow doesn't create congestion (even if the congestion is masked by something worse down the road): lane drops clearly reduce roadway capacity. It can create a simpler environment with fewer lane changes, but that does not mean less congestion, just more cars in one lane. You could theoretically paint off the lane, eliminating the merge, but that does not increase capacity, and is likely to create a more congested environment somewhere else. Lanes comes from somewhere, and must go somewhere too; that 'somewhere' is usually that lane-drop we love to hate.

If you don't support the concept of merging at the point of the lane drop, you must then support merging before the lane drop, or eliminating the lane drop altogether. Assuming you intend to support the first (merging before), you are not providing any evidence of (a) where drivers should merge instead, (b) how merging there is superior to merging at the lane drop, and (c) how capacity is somehow diminished by drivers using all available lane space.

The problem I have with those who do not support zipper merging is that they seem to live in this fantasy world where everyone could merge at exactly the perfect time, creating no congestion... This is total and complete fantasy. The vast majority of drivers are spectacularly awful at doing even the most mundane things like maintaining a lane or using their blinker; you honestly expect them to be able to calculate, at-speed, the exact perfect time to merge relative to number of lanes, distance to lane drop, and number of cars? Come on. Zipper merging is the best we can do with what we were given; we were given a plate full of ass-tastic drivers who can barely understand the concept of taking turns to begin with -- it's literally the most we could ever hope for.

[kalvado]

So if we all merged when you deem appropriate, it would be fine then. All hail king Avalanchez71!

The problem with your approach is the vagueness of where "as soon as safely permitted" would be; the safest point literally changes from driver to driver, perspective to perspective. Some may just deem it inappropriate to ever use lanes that end under your guidance.

Luckily, our DOTs are aware of these issues, so they established a merge point for us: where the lane tapers and the arrows point left or right. Merging at this point maximizes roadway capacity, minimizes confusion around "the right place to merge", and maximizes safety as drivers know *this* point is where they need to watch for merging traffic.

While a very well versed statement, one should take it with a grain - or rather teaspoon of salt and be prepared to dump a bucket of bullshit out of it. Merging at the last second may work fine while traffic is still free flowing throughout  - and then it really makes little difference;  but last second merge obviously makes things worse in congested flow.

Speaking of grains: I've never seen a grain of evidence to support that statement.

I don't think there is anything "obvious" about which is more correct. I can see the logic behind merging when there is an "opportunity" (again -- very vague -- usually a bad thing when speaking about driving), but I can also see the bigger picture which is that congestion is not usually caused by the merge itself; at I-705 & I-5 in Tacoma, it's lane changing between auxiliary lanes post-merge that causes the backups, not the lane-drops on I-705. Drivers gain nothing by merging "early" because the lane-drop is not the source of congestion.

OK, so what is the mechanism of congestion from the bigger picture point of view?

Demand outstripping supply. That could mean two lanes instead of three, or a weave where there should be a flyover.

As to how ending a lane somehow doesn't create congestion (even if the congestion is masked by something worse down the road): lane drops clearly reduce roadway capacity. It can create a simpler environment with fewer lane changes, but that does not mean less congestion, just more cars in one lane. You could theoretically paint off the lane, eliminating the merge, but that does not increase capacity, and is likely to create a more congested environment somewhere else. Lanes comes from somewhere, and must go somewhere too; that 'somewhere' is usually that lane-drop we love to hate.

If you don't support the concept of merging at the point of the lane drop, you must then support merging before the lane drop, or eliminating the lane drop altogether. Assuming you intend to support the first (merging before), you are not providing any evidence of (a) where drivers should merge instead, (b) how merging there is superior to merging at the lane drop, and (c) how capacity is somehow diminished by drivers using all available lane space.

The problem I have with those who do not support zipper merging is that they seem to live in this fantasy world where everyone could merge at exactly the perfect time, creating no congestion... This is total and complete fantasy. The vast majority of drivers are spectacularly awful at doing even the most mundane things like maintaining a lane or using their blinker; you honestly expect them to be able to calculate, at-speed, the exact perfect time to merge relative to number of lanes, distance to lane drop, and number of cars? Come on. Zipper merging is the best we can do with what we were given; we were given a plate full of ass-tastic drivers who can barely understand the concept of taking turns to begin with -- it's literally the most we could ever hope for.

I thought you were studying urban design, not philosophy? Reducing an issue to a paradox was a fashionable thing in the ancient world, possibly even earning you a textbook mention of "Jakeroot's paradox". By now most of them are resolved by mathematical logic, though, so let's try doing that.

"Supply and demand" is a good answer. "momentary supply and demand" would be a great one. "momentary supply and demand in a situation of perishable supply and inelastic demand" for an extra credit.

So, why lane change affects "supply and demand" ratio?
PS: extra credit for considering supply and demand question in a context of  ">  like" traffic throughput curve

[hotdogPi]

I thought you were studying urban design, not philosophy? Reducing an issue to a paradox was a fashionable thing in the ancient world, possibly even earning you a textbook mention of "Jakeroot's paradox". By now most of them are resolved by mathematical logic, though, so let's try doing that.

"Supply and demand" is a good answer. "momentary supply and demand" would be a great one. "momentary supply and demand in a situation of perishable supply and inelastic demand" for an extra credit.

So, why lane change affects "supply and demand" ratio?
PS: extra credit for considering supply and demand question in a context of  ">  like" traffic throughput curve

1. Proof by contradiction isn't for the ancient world only, and it is mathematical logic.

2. Supply decreases when you lose a lane.

[Avalanchez71]

The opportunity to get ahead of the other guy causes congestion.  I have sat and watched this occur over and over.  The right lane runner runs the right lane as it is at that point the fast lane.  As soon as the lane runs out the right lane runner becomes a zipper and zips over to the left lane avoiding the middle lanes causes a sea of break lights.

[kalvado]

I thought you were studying urban design, not philosophy? Reducing an issue to a paradox was a fashionable thing in the ancient world, possibly even earning you a textbook mention of "Jakeroot's paradox". By now most of them are resolved by mathematical logic, though, so let's try doing that.

"Supply and demand" is a good answer. "momentary supply and demand" would be a great one. "momentary supply and demand in a situation of perishable supply and inelastic demand" for an extra credit.

So, why lane change affects "supply and demand" ratio?
PS: extra credit for considering supply and demand question in a context of  ">  like" traffic throughput curve

1. Proof by contradiction isn't for the ancient world only, and it is mathematical logic.

2. Supply decreases when you lose a lane.

1. if done properly.
2. Not the question I asked.
The question was specifically in response to

I can also see the bigger picture which is that congestion is not usually caused by the merge itself; at I-705 & I-5 in Tacoma, it's lane changing between auxiliary lanes post-merge that causes the backups, not the lane-drops on I-705.

Which is at least partially correct. Hence the question -

So, why lane change affects "supply and demand" ratio?

[kphoger]

If you don't support the concept of merging at the point of the lane drop, you must then support merging before the lane drop ... you are not providing any evidence of (a) where drivers should merge instead, (b) how merging there is superior to merging at the lane drop, and (c) how capacity is somehow diminished by drivers using all available lane space.

Very well stated.

(a)  If there are opportunities to merge earlier than the end of the line, then typically the highway isn't congested enough that this is even a problem to begin with.  If, on the other hand, the highway is congested enough to make this an issue, then it's likely there's no good spot to merge.

(b)  Merging late instead of early only causes problems if you have to slow way down in order to do it.  This commonly happens when the continuing lane is already backed up and drivers don't leave much space between them and the next car.  More on this later...

(c)  Does any of this actually relate to capacity?  I understand that people might get annoyed by someone using the "available lane space" and then merging over, when everyone else has chosen to merge over early.  But being annoyed does not equal reduced capacity.  Similarly:  trading a choke point at the merge taper for a longer tailback upstream isn't necessarily a win for "congestion".  We need to first define what our goal is before we can debate the best path to meeting that goal.

The right lane runner runs the right lane as it is at that point the fast lane.  As soon as the lane runs out the right lane runner becomes a zipper and zips over to the left lane avoiding the middle lanes causes a sea of break lights.

Now onto where I left off with (b).  This is precisely what a zipper merge is supposed to eliminate.  With a true zipper merge, there is no "continuing" lane and "ending" lane.  Rather, two lanes simply become one, and traffic in both lanes take turns to make that happen.  There's no "lane running", because the entire concept of "lane ends" doesn't apply to one specific lane.  This way, too, there shouldn't be a problem of a driving having to slow way down or stop before merging into the continuing lane, then entering the traffic stream at 0.8 mph, because–again–there's no such thing as "the continuing lane".

[kalvado]

If you don't support the concept of merging at the point of the lane drop, you must then support merging before the lane drop ... you are not providing any evidence of (a) where drivers should merge instead, (b) how merging there is superior to merging at the lane drop, and (c) how capacity is somehow diminished by drivers using all available lane space.

Very well stated.

(a)  If there are opportunities to merge earlier than the end of the line, then typically the highway isn't congested enough that this is even a problem to begin with.  If, on the other hand, the highway is congested enough to make this an issue, then it's likely there's no good spot to merge.

(b)  Merging late instead of early only causes problems if you have to slow way down in order to do it.  This commonly happens when the continuing lane is already backed up and drivers don't leave much space between them and the next car.  More on this later...

(c)  Does any of this actually relate to capacity?  I understand that people might get annoyed by someone using the "available lane space" and then merging over, when everyone else has chosen to merge over early.  But being annoyed does not equal reduced capacity.  Similarly:  trading a choke point at the merge taper for a longer tailback upstream isn't necessarily a win for "congestion".  We need to first define what our goal is before we can debate the best path to meeting that goal.

The right lane runner runs the right lane as it is at that point the fast lane.  As soon as the lane runs out the right lane runner becomes a zipper and zips over to the left lane avoiding the middle lanes causes a sea of break lights.

Now onto where I left off with (b).  This is precisely what a zipper merge is supposed to eliminate.  With a true zipper merge, there is no "continuing" lane and "ending" lane.  Rather, two lanes simply become one, and traffic in both lanes take turns to make that happen.  There's no "lane running", because the entire concept of "lane ends" doesn't apply to one specific lane.  This way, too, there shouldn't be a problem of a driving having to slow way down or stop before merging into the continuing lane, then entering the traffic stream at 0.8 mph, because–again–there's no such thing as "the continuing lane".

Again - if you think about it, instead of quoting some junky DOT writeups- there is a pretty narrow set of cases where zipper is better, a lot of cases where it is about the same, and many cases where it is worse. Last second merge when general traffic merges early generally makes things worse to much worse.

If Jake helps me with derivation, there will be a full set of reasoning - and of course, everyone is welcome to join discussion. But we'll be moving slowly...

[kphoger]

Last second merge when general traffic merges early generally makes things worse to much worse.

It sounds like you're still assuming there's one "ending" lane and one "continuing" lane.  Which means you still seem to misunderstand the concept of a zipper merge.



For what it's worth...





Oh, wait, that was some junky DOT writeup.

[jmacswimmer]

It sounds like you're still assuming there's one "ending" lane and one "continuing" lane.  Which means you still seem to misunderstand the concept of a zipper merge.

The above is a perfect introduction to me contributing the following:
There's unique signage at some merges in Maryland that are specifically geared towards encouraging zipper merging and making it clear that neither lane is "ending" versus "continuing".

The most prominent example is the I-70 EB ramp to I-695 NB, which has to condense into one lane to squeeze thru the existing stack interchange. (Click ahead in the streetview link to see the signs at the actual "zipper" point - this ramp backs up at peak times and will continue to do so until the interchange is reconstructed, but drivers are generally good at following the signs and alternating at the actual merge point.)

Have any other states tried signage similar to this?