But, where the nodes are in the roundabout, a car could be counted twice, three times or four times. The nodes should have been just outside the roundabout.
@@BiffaPlaysCitiesSkylines The nodes prior to the round-a-bout not the nodes actually on the round-a-bout. This way, like the other set ups, you only count each vehicle as it makes it's approach and leaving rather than triggering potentially three or four times as it runs around the round-a-bout. Edit: At any rate, it's been obvious for a while that round-a-bouts in game are not good high volume traffic solutions, but for neighborhoods, they're pretty awesome decorative pieces that raise land value and reduce noise if you detail them, so knowing how to set one up for good low volume flow is still somewhat advantageous over a generic 4 way intersection.
@Biffa, I've always thought that the game lacks the realism of traffic when all vehicles travel the same speed and all ages of drivers travel at the same speed. We know that in real life , this is not the way....!
@@sane9875 TM:PE has a setting to vary driver behaviors, and I've noticed cars driving much slower on highways then surrounding traffic. Though I can't rule out the car just had a lower max speed.
comparing the results would have been most interesting if each of them had been pushed to the saturation point, which only happened in the small intersections and small roundabout. The rest just showed the average number of vehicles spawned in 10 simulated hours... I would have loved to see a real efficiency comparison, but I guess that 2 days filming would not quite cover it. More like 2 *months* hahaha
@@lukearts2954 All of the interchanges between 7:00 and 13:00 are the same interchange. So probably the same volume. And because none of them have weaving (lane is picked before entering and no changing on the interchange) you wouldn't see them saturated, any congestion that occurs would occur somewhere before the interchange where people pick the correct lane. For each of these single lane non-weaving interchanges the capacity would be the same because it's basically just determined by how much lane-changing happens right before it.
There's a problem with your methodology on measuring the traffic flow: The mod seems to count vehicles that pass through one of the monitored nodes. For most of your setups, this means counting every car twice, once as it comes into the interchange, and once as it comes out. However, if you place your monitored nodes on the nodes of a roundabout, all left turning cars pass through all 4 of those nodes, and get counted 4 times. And similarly, all cars that go straight get counted 3 times. Similarly, if you place your counting nodes on the central node of a 4-way intersection, ALL cars get counted only once. To count exactly all cars coming into plus all cars leaving these junctions, you'd need to place the counting nodes on the approaches to the junction, just as you did for the highway junctions. I.e. the "4 Way 4 Lane; No traffic lights" would actually have a traffic flow of 2*1243=2486 The "4 Way 4 Lane; Vanilla traffic lights" would actually have a traffic flow of 2*1508=3016 The "4 Way 4 Lane; Timed Traffic Lights + Asymmetric Roads + Dedicated lanes" would have 1834*2=3668 Which would put all of them well ahead of the "Large Roundabout (Vanilla)" which has a flow of somewhere around 2383*2/3=1589, And in one league with the "Large Roundabout (vanilla); Asymmetric Lanes + Dedicated Lanes + Give Way" which is around 5335*2/3=3557.
obviously we don't know the details of how the mod works, but it would be good to have an option for UNIQUE vehicles so each car only gets counted once.
Alternatively, to get the traffic flow of how many vehicles pass through the junction, you'd have to divide the "total volume" by the number of nodes a vehicle passes through on average. Assuming each of the three turning directions is equally likely for any vehicle, this means you'd have to divide the larger, more complicated intersections by two (vehicles are counted going in and out of the junction), the roundabouts by three (each vehicle passes through two, three or four "counting nodes", each equally likely, so three on average), and the four-way intersections not at all (technically, you'd divide them by one).
Thanks for writing this! I was going to mention the same thing lol. There is an older video from 2017 or something that tests a lot of junctions as well. That does it pretty well.
Yeah it does seem if a vehicle passes through two or more junctions it gets counted more than once, thus skewing the accuracy of the mod. As is, it is a neat mod that can help provide more information about the junctions and interchanges, with a grain or two of salt.
This is a good start to actually measure how good any given traffic solution (TS) can handle traffic. The next step would be to include a traffic spawner type of building where you can send out specific volumes of vehicles to really test the limits of each TS. Say you want to test a 4 way kind of situation so you build each a traffic spawner and despawner at each of the 4 ends of the road and configure the spawners to spawn say 500 entities each. Each spawner will then send out 500 vehicles heading for a despawner. You now know that there are 4 x 500 = 2000 vehicles your TS needs to handle. This way you can test which solution will take the least amount of time to clear the traffic, when it will collapse due to too much traffic and how long it'll take to sort itself out afterwards. You could even go further if the spawner allows for more options. Say which kind of vehicle spawns how often (to test if longer vehicles like in an industrial area affect the TS differently than only private cars and taxis). Or how much volume should be send to each despawner etc. Hopefully the explanation is somewhat comprehensible.
@@BiffaPlaysCitiesSkylines To add to this idea: if the mod could slowly increase the traffic spawn rate, you could slowly ramp up the traffic until traffic gets stuck on the intersection. Start at 100 cars per hour, 200, 300 etc... The breakpoint of the intersection is the most interesting.
Yup, I was about to post the same. Every one above 5k could handle much more. They need to be brought down to grid lock and then end count. That is where you can see the maximin load.
@@danielgaron1908Hard to do. For all the freeway interchanges that were on here except the first, all the lane changing happens before the interchange, there is no lane interaction on the interchange, so any capacity limit is determined somewhere upstream where people pick their lane. The interchange itself will always have free flow (assuming there's no interchanges downstream). The only gridlock that can occur is *before* the interchange in a test like this.
Next mod to be commissioned by Biffa, one that generates a s**tload of traffic at the entrances to the map, way, way more traffic than would be spawned in vanilla
This video needed more Biffa, I missed your voice sir ☺ Cool concept though and that last intersection/interchange/junction was crazy! Question about your traffic counter mod, it appeared to count some cars twice (I'd guess if they changed lanes at the node? Can't think why else it would be) so maybe that's something to take into account. I might be wrong though I was trying to spot how the totals changed when a single car went through and I may have just misread it. Very cool experiment, must have taken ages to set it all up, record and edit! Merry Christmas to you, the Biffam (that's a word now) and your editor 🥳🥳
Was just thinking that, yep needs more Biffa and a sip of tea at the half way mark 😉 But very mesmerizing video, very obviously involved a lot of work, good job! I'm motivated now to try building that beast at the end!
Your traffic management videos are what prompted me to try Citirs: Skylines in the first place. While I've grown fond of most aspects of the game as a world-builder, I always enjoy treating it as a traffic simulator.
Interesting, but all it really showed was that the maximum load of cars supplied by the environment was around 7000 in 10:00 hours, regardless of interchange. To really test the capacity of an intersection/interchange, you‘d need traffic build-up on the incoming roads. In the video the incoming roads sometimes looked empty. But as always: Love your videos.
Surely on the highway measurement you should only measure all the entry ramps or all the exit ramps but never both as this would be counting the same vehicle twice? (Once while entering the junction and another count leaving the junction.)
@@BiffaPlaysCitiesSkylines Its not thar easy. If a roundabout also counts „cars staying on roundabout“ when you selected the inner nodes, you create number inflation. So a car would be counted in the incoming node, the passing node and the outgoing node. If the mod remembers which car was counted already then it would be fine.
The only suggestion I would have for using this tool is selecting inbound or outbound nodes only, not both. The way you have it set up, you are counting each car twice. :)
Two comments and a thoughtful suggestion: 1) AWESOME potential for that mod! Want to see more of it, for sure! 2) For all of the big interchanges, the mod was clearly just measuring how many vehicles were being sent and not how many they could handle. This is quite evident in how close those mid results were (almost all in the mid 6000s), and none of those systems in question ever became overloaded or stressed. Anyone who has used many of those types of interchanges knows quite well that their *max* capacities while still maintaining adequate flow are in reality quite different - some maintain flow rather admirably when overloaded and some *completely* lock up. I think any intersection/interchange tested with this mod should have enough vehicles thrown at it that at the very least it *starts* to struggle with maintaining good traffic flow - only then will we be able to see what the REAL results would be 3) Perhaps a future update of the mod could count not just the pure number of vehicles passing through each node, but maybe also count the "max time between two set nodes" (or max time IN a set node if that's possible) to better gauge the true capabilities of the intersection. For example, the last version of the roundabout outperformed the first version of the three-way interchange in this video. This being despite the roundabout having cars backed up and waiting for quite some distance in every single direction - it was definitely at it's limit of the volume it could handle. Conversely, the three-way in quesiton had absolutely no backup at any of it's nodes and clearly could have handled a lot more vehicles before starting to fail - which to me is a much better indicator of true flow. So maybe with my suggestion for a mod enhancement the mod might output something like: (v) 5335 & (tmax) 115 seconds for the roundabout and (v) 5156 & (tmax) 3 seconds for the three-way interchange. Yes, I just made up numbers for max time in those scenarios for illustrative purposes, but when running comparisons it could give a better feel for the true capability and "flow" of the intersection
Wait, is Total Volume just the sum of the individual nodes? No deduplication? If I'm trying to compare different intersection designs, I really need a count of *unique* vehicles.
@@BiffaPlaysCitiesSkylines Wouldn't you want it to do something like, okay Vehicle ID 999 has passed one of the entry nodes, it is now inside the interchange/intersection boundaries .... then when it passes an exit node, leaving the interchange, okay vehicle ID 999 has left the interchange system, incrementing traffic flow count and so on and so fourth for a vehicles per minute rating?
Yeah. Either only entry nodes should be counted. Or only unique vehicles should be counted when entering and exiting through nodes. Otherwise if you have only one node for a 4 way intersection it will count way less than a roundabout with 4 nodes all counting as each car passes through each node.
Currently a 4 way crossing has 1 node to count: Fine. Highway system: (on/off way) Divide by 2 or only count incoming / outgoing. Roundabout is trickier, as you need to avoid counting passing vehicles within the roundabout. When there are no dedicated on/off ways (single 2way roads) then… ?
I am pretty sure this video confirms what I always suspected, all free-flowing interactions with no points of conflict and proper lane mathematics (tm biffa) perform within a percentage point of each other (stack vs turbine, vs infinity loop, vs double diamond, vs inverted cloverleaf etc, etc) the next step is to include the cost of building and maintaining the road and total space taken up, that is the next dimension of efficiency! Thanks for putting all the work into this both of you,
I think there is a bit of a flaw in the test setup. The numbers say nothing about the interchanges, because the results are almost the same. Every car goes through two junctions and in some cases even three. This is different from the junctions, where sometimes they only go through one. Also, the junctions suffer more, while the interchanges do not. If you want to judge the quality of an intersection, the load should be so high that just about any congestion occurs. The fact that the results are very close suggests only one thing - they only give information about the load or the maximum number of cars that the game spawns. I suspect that this tool is not very helpful in this test setup. But I can see why you're so excited, Biffa. It can help to find out the load of an intersection and thus the choke points.
This DEFINITELY needs an in depth explanation. For example, this mod rewards/indicates good traffic layout, but a REALLY busy traffic intersection that doesn't move will show similar counts as small, or EMPTY, intersections with low car counts but are capable of moving. Maybe that could be in the next revision of the mod? How many vehicles are queue'd up for a given intersection, as an indication of the POTENTIAL traffic going through an intersection, and used as a gauge of efficiency? EG 1000 vehicles in queue but 0 moving indicates a blocked junction 1000 vehicles in queue and 1000 moving vehicles is a healthy interchange. Finally, that 16-way shouldn't be called "complex," that should be called an unholy traffic nightmare LOL!
It seems like roundabouts really perform best with moderate traffic, once you get traffic jams it performs only marginally better than a 4 lane junction. I recently started playing this game again and built a roundabout with an underpass down the middle so traffic going straight doesn't have to go on the roundabout, with sliplanes down the side for coming on. I don't know if it's difficult to visualise but you should test it, it worked wonderful in my city. Also a tunnel for pedestrians seems to improve it even more, mostly stuff I learned from this channel.
I like that most of those custom interchanges don't go much further than 6500, meaning that whichever you pick would work the same, and it's all up to how you like it visually looking :d Great video and mod
I thought the same but then I saw another comment about the amount of traffic being generated. Is 6500 the upper limit for all those interchanges, or is that just the max traffic this map was generating to put through the interchanges in the first place??? I'm not sure how you would stress-test all of those and put the max amount of traffic through each one but that would probably be the next step. Regardless this is a cool concept (and looks like a great mod - hope we see a lot more of that one yet!) and thoroughly enjoyed the video. :)
I think some additional stats could be quite helpful, especially if this were taken to be useful for real life: - total road length - build area (like a bounding box or a bounding circle or something) - number of layers (particularly relevant for that 16w ay monster) - ratio of raised or tunneled to flat pieces - total construction cost - number or merges (how often do you have to look out for other traffic to find your spot) - number of forks (how often do you have to decide what way to go) - number of forced interleaves (particularly relevant for cloverleaves) and the ratios of some of those various things compared to the resulting volume I know in the game the only two things that actually matter for these interchanges are volume and aesthetics, but it would be fun to see whether any patterns emerge from this.
I also agree that the cars entering each junction should steadily increase until the traffic comes to a near stand still. Then you will actually know how well each interchange actually deals with pulses of incoming vehicles. Like when you put in a big warehouse and a million delivery vehicles start showing up!
nice vid Biffa I also saw some people wanting a summary of it all, so: Small Roundabout (vanilla) 908 Small Roundabout (vanilla AL+DL+GV) 977 4 Way 4 Lane Junction (No TL) 1243 4 Way 4 Lane Junction (Vanilla TL) 1508 4 Way 4 Lane Junction (TTL+AR+DL) 1834 Large Roundabout (vanilla) 2383 3 Way Y-Pretzel Interchange 4763 3 Way Interchange (vanilla LM) 5156 Large Roundabout (vanilla AL+DL+GV) 5335 3 Way Interchange (vanilla LM+DL) 5407 3 Way Interchange (vanilla) 5625 4 Way Turbine 5978 4 Way Pinava 6358 4 Way Full Diamond 6418 4 Way Parsio Bypass 6549 4 Way Quadfractal 6572 4 Way Turbine (large) 6563 4 Way Diverging Windmill 6562 4 Way Twisted Cloverleaf 6565 4 Way Windmill 6581 4 Way Twisted Full Diamond 6596 4 Way Infinty 6627 4 Way Pinava Condenesed 6646 4 Way Stack Interchange 6649 4 Way Stack 6652 4 Way DCMI 6655 4 Way Twin Trumpet 6680 4 Way Inverted Cloverleaf 6697 4 Way Double Trumpet 6765 16 Way Complex Interchange 7149
So many of these intersections won’t definitely not work irl, it would confuse the hell out of everyone xD! Also the one who build that monstrosity of a 16 way complex intersection has way too much time on his hands, props for making it so nice symmetrical
The number of interesting junctions out there is amazing. I think that there needs to be at least twice as many cars running through each to check if there is any build up. Most junctions seem to be capping out at 6500, and there is no pile up. Speed lost in changing lanes, or tight bends would affect the flow if there were more cars.
This is great, but it looked like it wasn't stress testing all of the interchanges, so I'm not sure that it really provides a comparison of the maximum flow rate before they degrade? I think that you might get more interesting numbers if you stress test by ramping up the traffic until it starts to degrade in throughput, and then backing it off to see how quickly it recovers.
I find the best fix for roundabouts in vanilla is to separate the entrances and exits, even with small roundabouts or single lane traffic. It's not perfect, but it does help reduce congestion from the default roundabout setup
This was already reasearched on reddit a couple years ago (search for "Traffic flow measured on 30 different 4-way junctions"). For sure the different intersections differ in throughput, so my guess here is that the traffic is probably too small to actually measure the real throughput (so it measures only part of it).
That was relaxing video and the mod is already very useful as is! I'd love to see the tool get some upgrades though, so we could see some more information. Like how big portion of cars go from direction A to direction B since with wild interchanges directions often are not balanced. It would make it extra easy to see between what the most traffic goes to especially if the traffic is smooth but heavy to all directions at the moment of inspection and before disaster. Additionally, removing the double counting from the mod would be nice. I'm lazy especially during holidays and don't do maths even if it is as simple as division by 2🙈
I am a little disappointed, as the fact that many of these intersections had comparable throughput shows that the stress testing is insufficient. It is hard to tell which of these interchanges would result in genuine improvements over others in a high-traffic situation. I would recommend comparing cut-down versions of these (reduced lanes where possible, shortened sections for increased merging), and it would be good to see side-by-side comparisons for scale. For those horrifying situations where you have *wayyy* too much traffic and nowhere to put a nice freeway interchange (because actually there's a train line running underneath and a metro line on top). Biffa also missed some of the best uses of this as it currently stands: service interchanges and complex intersections. - The SPUPclo (a 2-phase light cycle cross between the parclo and the single point interchange) - The Michigan Left (a 2 phase, 3-light linear intersection system where a left is composed of a right and a U-turn or vice versa, each movement at lights) - The at-grade DDI - The classic diamond, 2-level roundabout, and single-point (SPUI) - Swindon Magic Roundabout and variants and last but not least, the traffic-light riddled spaghetti junction. Whose traffic fails at system interchanges? If that's your problem, this is not the tool you need.
A couple observations: 1. Most of the later intersecting could have handled a lot more traffic than what was shown. 2. Some intersecting seem to have a slightly higher throughput, but would collapse much more completely over that limit. A good example is the built in 3-way intersection. Adding dedicated turning planes and using lane mathematics made the volume lower, but both were running at less than max capacity. Overload the vanilla 3-way and you get gridlock,but the one with lane mathematics and dedicated lanes would continue to flow, even though they would continue to back up. This leads me to want to see a brain of the mood that enables you to overload the intersecting with super dense traffic and compare the flows then.
It's interesting how a lot of the fancier 4-way highway interchanges are roughly the same flow (c.6500-7000). Also, had no idea how BEAUTIFUL some of them could be!
Hopefully we will see some of the fancy intersections in your next builds. It seems like that your traffic creater mod is limited by around 6500 to 6700. it would be interesting to see how much the intersection can really take. May you find a way to increase the traffic and check what is possible and compare it against necessary space per intersection. From this comparison you can make an overview what is the most efficient setup to take a huge amount of traffic. Because it seems like in the beginning the cross section with traffic lights can take more or less the same traffic like a smaller round about, which need much more space. So from a space and a build costs point of view an intersection with traffic lights is more efficient then a smaller roundabout. Maybe there are other surprises we don’t see directly on the first view. Maybe it’s a bit to scientific but it would be fun too.
Thanks for the interchange ideas. Other than that the differences are so minimal I don’t really see a need to monitor efficiency. It’s a cool mod regardless
Ah, I see we finally find out what the question on the discord about people's favorite steam workshop interchanges was about! On another note, I'm surprised at how well the vanilla large roundabout did. Especially with how it appeared to be at a total standstill in the footage. It still moved a surprisingly large volume of cars!
I use the vanilla cloverleaf to make a windmill, but it isn't like the one you posted. I like to put the "left turn" into the center, giving an extra node for traffic to pick a lane before the next merge. Lane math brings it from a 3-2-3-4-3. Since I use the vanilla clover as a base, the footprint is exactly the same size. I have never seen it fail.
So as long as your building a 4 way intersection with lane mathematics (no crossroads or roundabouts) your going to get the same steady traffic flow though, so it's really about which intersection fits best or looks best for your city, cause basically they all pretty much let the same traffic through.
Except in cases where a single lane turning ramp can't handle all the traffic, but firstly, the traffic flows stay the same for all interchanges in this test and secondly, that just means that you have to upgrade the number of lanes.
This mod looks really cool! It would be nice to see the average flow rate or even how long cars are stopped in each lane before they enter the intersection. Definitely a very useful tool!
Hi Biffa... Another good vídeo. I think the "4way full diamond" and the "4way stack" will be the most effective junction, because the changing speeds. To much corners and turns, will slowing they're down, and create Traffic a bit easier.
Wonderful video! The only omissions I'm really missing is the continuous-flow intersection and diverging diamond interchange. We went quite rapidly from some pretty basic street intersections into a long series of really really good no-conflict highway interchanges. It's a bit thin on the middle ground: really beefy street intersections, and service interchanges.
This was quite interesting. But I think, the differences in the results of all the 4-Way Intersections come mainly from randomness. To get an intersection fast, you have to think of a few important things: 1. lane mathematics, 2. dedicated lanes and 3. don't put merging before spliting. The later concept is the problem of a regular clover-leaf intersection where the streams of cars cross each other. As far as I can see, every 4-Way intersection presented here (apart from the mere crossroads) pays attention to those three points. From a mathematical point of view one would say they are all topologically similar to each other. But I would like to know whether and how one can use this mod to analyze the traffic in our cities.
Like the idea of this mod and it was relaxing to watch let alone great music with it! I look forward to seeing more about it. I'm not a techy person so if it needs other things or adjustments, I leave that to you techy people 😁 Missed hearing you though!!!
Actually your readings are off, since you counted traffic double or even worse on some of the interchanges falsifying the total amount a intersection can handle. i.e. 5:11 913 at the bottom and 937 on top right now the cars driving straight up make both of those counters increase basically counting double. Besides that thanks for sharing them mod with us.
It lulled me to deep sleep. Not because boring ; it was very informative on the contrary for the part I was mesmerized by. But they hypnotized me into slumber midway. I'll have to watch again...AWAKE.
I'm not normally one for just watching traffic flow, but this was really relaxing, and your mod is a really interesting one. I think it would be really interesting to see a follow up video that uses the mod to do some data analysis on the interchanges showcased to get a better idea of just how well they each perform when pushed to their limits. I appreciate that would be a very time consuming video to make though, so I understand if that can't happen. Side note; I loved the infinity interchange and I'd love to see it in Biffalo or Blightea at some point!
Apart from the standard roundabout, 4 way Pinavia Condensed and 3 way Y-Pretzel seem like the only interchanges that allows a vehicle to take a u-turn and head back to the direction it was coming from. Which in my view is a better option for real world application
What I learned from this video: (This is just my opinion) - Roundabouts suck - TM:PE adjustments always help - If the traffic from every direction has a dedicated slip lane to every other direction, the flow figure is always around 6500 - That 16 way interchange at the end fills half of your map 🤣
Towards the end, some of the best intersections are limited by the small amount of traffic coming in to them so it would be cool to see the maximum number of cars that could flow through each intersection (as we are not seeng how good they really are)
The fact that a majority of the interchanges had nearly identical scores seems to indicate there were not enough cars being spawned. Admittedly, watching the traffic move smoothly was awesome, but these don't accurately represent the maximum throughput of most of them.
In my latest city, I've been noticing that lane math doesn't seem to be that helpful without TPME lanes. And still, there can be some pretty silly lane changes as cars approach junctions. I'd be interested in seeing more about how you set up this test. Is this a map with only outside connections? (what it looks like) The next test could be to build 4 similar/equal "towns" and see how the intersections handle flow as the population increases. One strategy that I've started to try out is to avoid or eliminate left-hand turns (with overpasses that circle back around) It's also been fun going through some of Biffa's older videos!
Thanks for sharing the link to the interchanges on Steam. Now you need to apply lane mathematics on them and rerun them with the mod updates for unique vehicle counts and have results ordered in a nice-looking table at the end of the video. Make sure you package up the new interchanges with lane mathematics and make it available as a download as well. And I want it by Friday. 🤪 I'm only joking!...only...joking...😵💫
I don't think this mod measures traffic volume as well as you may hope for. For starters on every highway interchange you count vehicles when they enter the interchange and again when they leave it. So every vehicle is accounted for twice. On roundabouts it's even worse. If a car leaves the roundabout on the last exit before the point of entry, the car is accounted for four times - point of entry, two nodes where it stays on the roundabout and then the exit. The only correctly counted junctions are the basic crossroads. For the mod to work properly it would need to allow to select single lanes, not the entire nodes. In which case you should only select either incoming or outgoing lanes.
Something to think about, but it still gives an accurate flow indication, rather than an individual vehicle count, which is what I was going for and about as accurate as you can get in Cities Skylines 👍
For those situations where traffic jams never formed, your counter was basically counting every single car that spawned. The study doesn't mean anything.
Kinda reminds me of Real Civil Engineer playing "free ways", but in that game the cost of concrete is also included in your final score. Also feel like you could pump up the amount of vehicles and then some junctions would struggle more (like some clover designs)
little clips of your intersections would be perfect for low budget movie productions who couldn't afford the vehicles or stuntmen for an actual accident scene.
This is interesting, but there is a couple of things I would like to know/include in those stats. One is how long it take to traverse the junction/interchange and to be honest most of those aren't used inside a city, more on the outskirts or between 2 regions. In one of them junctions there was a broken node that was holding up traffic (slowing it) that would have affected the numbers leaving the junction over time. I am pretty useless at building intersection/junctions and can't use many on steam as they are specific for left hand drive - where I build right hand drive as that is what I'm used to, so would love a builder to create a few of those bigger interchanges with thoughts of those who use right hand drive. (hint hint to the builders out there)😁 The other thing I really would like to be highlighted here is that there is a lot of talk about the junction/interchange but I haven't seen anyone explaining and using examples of where the different intersections can be the most effective in what circumstances. As an example I have never seen a diverging diamond interchange and it makes no sense to me so I never use it as it's counter-intuitive to my experiences. Also the testing used is based on a high traffic dual carriageway highway which a lot of intersections aren't designed to cope with which gives a false perception on where/when to place or use these. Enjoyed the basic idea as an interesting concept but not so sure of it's usefulness, sorry Biffa.
And when the last interchange is not handling the traffic, just add more lanes, bridges and tunnels; that's what you're supposed to do in a car-friendly society, not reducing lanes and providing buses or trams, or God forbid, pedestrian areas. Long live cars!