• Welcome to SC4 Devotion Forum Archives.

BAT Showcase

Started by mattb325, February 01, 2007, 04:07:43 PM

Previous topic - Next topic

0 Members and 4 Guests are viewing this topic.

bravesirkevin

Might be a result of me not using the subway feature.... Hadn't yet developed that infrastructure in the parts of town I plopped it in, and was expecting the passengers to make use of the extensive bus network I'd provided but they promptly ignored it. :) Perhaps with some subway lines they'll choose to use those instead of cars.

art128

Thank you for the recent uploads. I will definitely be using the Ultra-modern L1 7.5m E-RRW station.

Looking forward for the BOB.

That apartment building is also very nice.
I'll take a quiet life... A handshake of carbon monoxide.

Props & Texture Catalog

bravesirkevin

#1422
Done a little more testing on my side. It might be the NAM or some other factor that I haven't considered, but just for reference, here's the suburban one:


And zoomed in:


There are lots of things arriving at the station... We've got trains, buses, cars and there are a ton of sims who are walking there. There are no trains or buses leaving the station though.There are a couple pedestrians walking across to the offices across the road, but all of the other traffic arriving at the station has been converted into cars. The biggest problem is that there are a several thousand cars leaving the station and jumping on the highway, which is not so much of a problem in this city, but in the adjacent one they all get off and congest the streets refusing to convert to bus or subway traffic.



Here's a different screenshot of a different part of town where I've added a subway line and have swapped in the Modern station instead. Some of the arriving subway traffic has continued on, but again, almost all passengers arriving at the station have hopped into cars. It's not clear from the picture, but there are almost 12000 cars leaving the station. That TLA-7 is red on the congestion map.

mattb325

It would appear that your scenario is forcing many passengers to switch to cars.

When it's a little tricky, the best thing is to test in vanilla. It has to be done in order to rule out other possibilities.

There is absolutely nothing in my plugins - no NAM, nada, zip, zilch... nothing but the station prop pack and strickland station itself. (Yes, these stations work in vanilla). I am using the vanilla cheats of 'howdryiam','fightthepower' and 'weakness pays'. This is simply to get the city off the ground and it's safe to say that these cheats have nothing to do with transport.

1) I then created a brand new region, and a blank tile. I have tried to set it up with just one or two main commercial roads and res zones feeding into these, just like your set up. Here it is after a bit of zoning growth:


2) Here's strickland station. There are no other train stations connected at all on the tile at this point, so strickland is just acting as a transit jump from one area to the next. You can see that it's not exclusively converting an arriving passenger type from one type to a car, as pedestrians are going through the station, as are cars.


3) I have now added a Maxis station at one end of the line and zoned res around it:


4) As expected, strickland now starts to work as a station. However, from these numbers, the incoming trips aren't being converted to cars on the way out (ie the little green arrows are 99% pointing towards the station):

Congestion on this set up is low, too:


5) Next, another Maxis station is added at the opposite end of the line:


6) Growth and congestion are picking up, but as the congestion is at I-D and I-Ag in a vanilla setting, this isn't attributable to the station:


7) I'm now adding bus stops at other points on the map to activate the bus-stop feature in the rail hub (you said this was in your set-up). Again, all the green arrows appear to be pointing in the right direction and the incoming numbers don't show that they necessarily being turned into outgoing car trips:

Here is the associated congestion map:


8) Just looking at the numbers, everything appears normal when compared to the other 2 Maxis stations' behaviour - which only have minimal trip types:




So from this, unfortunately all I can say definitively is one of four things. 1) The vanilla example needs more testing in possibly a full region; OR 2) My NAM & traffic controller set up is causing my stations to behave exactly as I expect; OR 3) Your Nam & traffic controller set up is causing my stations to behave contrary to how I expect; OR 4) You have something else, such as a mod or ordinance that is causing Sims to seek preference for a vehicle.

The only one I can safely eliminate is option 1. It doesn't help your situation, but all I can offer is that my nam/traffic controller is set up on the default.


bravesirkevin

Thanks for taking the time to investigate, Matt! I really do appreciate that. :)

Having played around a bit more, the station does seem to work well as long as I avoid placing it where a very high volume of passengers need to disembark, and I'm happy with that solution and am willing to leave things there.

If you are interested in examining things further though, one thing that is worth noting about my case studies, is that the roads flanking the station in the last image were not main roads and only had traffic from the station on them, so I could tell exactly how many cars were coming out of the station, and could rule out other causes of the congestion. In your image 7, you can't pin the congestion of the road to the industrial sector on the station because that road is the only route to that area, and is being fed by the entire residential neighbourhood surrounding the station. There's no bus stop in the industrial zone, so you can't tell if passengers arriving at the station are willing to get on buses to go there, or if they'll insist on taking a taxi. You also only have 96 trains arriving, while I had something in the region of 12000 trains, buses and subway trains disembarking and that makes quite a significant difference.

mattb325

#1425
^^ No problem  :), it's easy to test out in the vanilla mode and it's worth it as well, given that there are very few multi-function stations around.

Also, there is another issue with stations in game that hasn't really been adequately addressed, IMHO.

Stations are generally listed as taxi makers, but because of  a lack of a passenger > car switch (they only have car > passenger switch as Andreas said), the taxis (which are a car) only go one way per morning or evening commute: from the sims house to the station, or in the evening from the station to the sims house, which is unrealistic.
So in cities like London and Sydney where I've lived and worked for many decades, If I were to do a pitch to a client in area where I had to take the train from my house (and being dressed for work and carrying a brief case, folders, large tubes of paper/lap-top etc, I wouldn't really want to walk), I would take a taxi to the station closest to my house - sim city can cope with this - and then once I got off at the station near my client's address, I would take a taxi to their offices/house to do the pitch. This 3rd part of the journey is not simulated in SC4, but is very common in real life. All you have to do is stand anywhere near a busy railway station to see how chaotic the roads are and how many people are coming and going on taxis, cars and buses.

Additionally, that extra congestion (within reason of course) around the station hub is good for med-high density commercial zoning, and also bear in mind, that these trains stations provide employment (strickland = 140 jobs) so sims are actually drawn into the station area because of the civic jobs. This is possibly partly an effect of what you are seeing in your set up, but to my mind something else in your plugins (or congestion in other parts of the network is high enough for the NAM to make sims go to cars from these stations) must be forcing them to be switched to cars in your scenario, because in my full plugins test pictures on the previous page, we are in the 000's as well, and the green car arrows are all going the right way.

Soon, I intend to make some non-transport hub stations for the 7.5m viaduct rail, which can be used for the quieter suburban settings, but the stations in this set were really designed for high volume settings and - because 7.5m is so new - it was crucial for them to be able to interchange with other existing networks rather than being used in isolation on the 7.5m network only, just in case I don't get around to making additional 7.5m station-types.

bombardiere

Quote from: mattb325 on March 25, 2018, 01:37:11 AM






Matt, in this your picture you can see the same effect than Bravesirkevin is describing. There are green arrows coming out off the station.

(I can't see what is the source. Train, pedestrian or subway. Also left lower corner, next to large police station, is another source for outwards green arrows.)

Your traffic seems to be modest, but perhaps if one is using car orientated traffic simulation setup, US style setup, then SC4 may want to overemphatise car use and not to convert the traffic into pedestrian or buses.

I guess that you have made it on purpose and if I understand your post, I think you want to simulate taxi traffic.

Quote
Stations are generally listed as taxi makers, but because of  a lack of a passenger > car switch (they only have car > passenger switch as Andreas said), the taxis (which are a car) only go one way per morning or evening commute: from the sims house to the station, or in the evening from the station to the sims house, which is unrealistic.
So in cities like London and Sydney where I've lived and worked for many decades, If I were to do a pitch to a client in area where I had to take the train from my house (and being dressed for work and carrying a brief case, folders, large tubes of paper/lap-top etc, I wouldn't really want to walk), I would take a taxi to the station closest to my house - sim city can cope with this - and then once I got off at the station near my client's address, I would take a taxi to their offices/house to do the pitch. This 3rd part of the journey is not simulated in SC4, but is very common in real life. All you have to do is stand anywhere near a busy railway station to see how chaotic the roads are and how many people are coming and going on taxis, cars and buses.

As we know, the SC4 doesn't do taxi traffic. So you may have wanted cars. However, in my personal opinion the buses are better representative for the taxis. In fact the SC4 buses behave as taxis. The buses are single passenger only and do not follow routes, but go where ever the commuter wants. As long as there are bus stop nearby.

art128

Quote from: bombardiere on March 26, 2018, 08:49:57 AMMatt, in this your picture you can see the same effect than Bravesirkevin is describing. There are green arrows coming out off the station.
(I can't see what is the source. Train, pedestrian or subway. Also left lower corner, next to large police station, is another source for outwards green arrows.)

Green is car traffic. Aren't Matt's station modded with some parking capacities?
I'll take a quiet life... A handshake of carbon monoxide.

Props & Texture Catalog

mattb325

Yes, the car trips are green.

Whether buses are more suited than cars as taxis is personal preference, I guess. I'm fairly certain that a car trip is simulated more quickly than a bus trip in SC4 (as it is in real life), so a car is generally preferable for me.
Also, I don't know what is causing the issue that I can see with sirbravekevins setup, which is why I have said that it could be a mod or congestion some where else on his network, I asked him what his settings were, and also offered the settings that I have used with the nam. So, in that instance, as you surmised, it could be that he has his nam settings skewed to car based ones, or he could have another mod at play, or it could be something else entirely.

As for the passenger trip types template that I have copied over to these 7.5m stations, it was provided by a NAM team member and it's most certainly not just something I have made up out of thin air. The template was provided for this lot: https://community.simtropolis.com/files/file/32073-city-link-transport-hub/ which was based on z's work and figures around capacities. That station and these 7.5m stations in question have car parking, bus, rail and subway facilities.

Now as to the picture, what you see is the morning commute taken by all sims in that station.

There are a number of other stations interacting within the spaghetti squiggles (including ones that I haven't made or modded) so what you are looking at is any number of sims going to the station and then continuing onto their commute not just along the 7.5m network but also interacting with the HSR (Monorail) hub next to the police station and the train station (by someone else) on the opposite side of the picture and subways, and other stations not visible.
My understanding of the purpose of the nam is fundamentally to improve the pathfinding capabilities. It also provides a ton of cool stuff, but that is secondary to the pathfinding. Point being, if in a city a car is quickest trip, then a car is what is given preference on the trip out. But that same logic is also evident in reverse when stations simply don't get used because they're not the quickest trip.

The traffic in my picture is reasonably modest ~7K but bravesirkevin's had ~12k so neither has hit congestion (the station has ~115K capacity), which means that something else is at play to cause an overwhelming preference to cars.

mgb204

The behaviour, passengers from trains leaving the station in cars, it because of the transit switches which allow for them just as Andreas noted. It's not a bug per-se, even if you might think the behaviour looks odd. If you have a transit switch that allows for cars to leave the lot (Passenger -> Car), then it is just one valid option for traffic using the station. If a disproportionate number leaving are cars, that's because the quickest routes for onward travel to jobs is by car. Which as postulated here, can be dramatically altered depending on your NAM TSCT settings.

mattb325

Just to recap, so nothing gets confused with these stations:

1) bravesirkevin initially stated that all train trips were being converted to cars on the way out;
2) Andreas stated that it's possibly because of the passenger > car switch;
3) I posted pictures showing that I could not replicate all train trips being switched to cars;
4) I suggested it was to do with the individual Nam settings, and asked what they were on bravesirkevins PC. I still don't know the answer to that (even though I can guess with reasonable accuracy that they might favor the car). I advised that my nam settings are default;
5) bravesirkevin showed that adding subway lines didn't necessarily fix the problem with most passengers leaving as cars;
6) I tested the station without the nam to show that in vanilla mode, the stations don't throw all passengers into cars on the way out; therefore, it is something else in bravesirkevins plugins that is causing this behavior.




There are mods available (I think it's called the radical ordinance mod, and is HUGELY popular) that among other things such as making commercial demand skyrocket, have settings that can either force all sims to drive or force all sims to take PT. He could be using that mod. Or he could be using another mod that does something similar (there are plenty available) or he could have the NAM and traffic controller settings skewed to vehicle preference. None of us, apart from bravesirkevin, knows what is in his plugins.

That is why I've posted the pictures and done the testing to show that under typical, default settings and circumstances, these stations will not convert all incoming passenger trips to outgoing cars.
It is also why I have left my conclusions open-ended, as I simply cannot state one way or another what is causing this behavior.



As an aside however, if you stop and think about what the obvious cause is (ie bravesirkevin is using the traffic controller settings that favor cars over PT) then apart from the obvious inconsistency of running train lines through a car-centric setup, having passenger > car switch actually stops the network from being broken as it takes the secondary option in the settings (public transport) and converts it to the primary option in the settings (car). That congestion is experienced as a result simply means that the roads around the station need upgrading to handle the extra volume being forced on them.

Andreas

#1431
Maybe I should add that congestion itself isn't necessarily a bad thing. If you'd see the congestion map of some of my downtown areas (all roads and such blood red), you might shake your head. :D We have to keep in mind that SC4 is still a game, and sometimes, we have to work around certain phenomenons, and make use of them.

As stated somewhere above, congested roads are actually attractive for commercial development (while it's bad for residential areas). You can use this effect for deloping a busy CBD, or some commercial hot spot in a suburban setting. In case you prefer to use the station as a "park & ride" one, just make sure that such "transit hopping" doesn't happen.

Avoid roads (highways, etc.) that provide the shortest possible connection in favor of public transportation. I know that the usual grid development encourage such layouts, so just go ahead and be more creative with your traffic layout, remove short-cuts, or add "penalties" to such usage (a toll booth on a congested highway works magic if there's a nearby railway that supplies the necessary transit switches) etc.

It's not easy to find the proper balance, and experimenting obviously takes some time. Fine-tuning the NAM traffic controller might be necessary, but even with a "moderate" default one, most things should be fine, even though your maps show lots of red (hey, this is realistic after all, isn't it? ;) ). If your traffic really gets out of hand, you'll notice it otherwise (residential buildings getting abandoned etc.).
Andreas

bravesirkevin

With regards to my settings I don't really use any game-changing mods apart from NAM and IRM. My NAM settings are pretty much a default NAM installation with the Medium Traffic controller as that was the default option. Didn't play around with any of those settings at all. Doubt that IRM would have any effect on traffic as it's primarily about where industrial lots build.

Quote from: Andreas on March 26, 2018, 01:54:09 PM
Maybe I should add that congestion itself isn't necessarily a bad thing. If you'd see the congestion map of some of my downtown areas (all roads and such blood red), you might shake your head. :D We have to keep in mind that SC4 is still a game, and sometimes, we have to work around certain phenomenons, and make use of them.

Not really bothered by congestion in general, and I'm not really aiming to eliminate it either, because it is realistic and it is useful in its own way, as you pointed out. I do aim for some degree of town-planning though and I generally use public transport to provide an alternative to cars, and with the intention of reducing car traffic. Once I swapped out the station for a standard 15m NAM viaduct station the congestion returned to normal levels as everyone reverted back to using buses, subways and their feet to get from the station to work, instead of all jumping on cars.

Quote from: mattb325 on March 26, 2018, 11:16:12 AM
The traffic in my picture is reasonably modest ~7K but bravesirkevin's had ~12k so neither has hit congestion (the station has ~115K capacity), which means that something else is at play to cause an overwhelming preference to cars.

Real problem was that the roads can't handle 12000 cars coming from a single point. It's not actually true that all the traffic was being converted into cars... occasionally, they'd stay on the train and continue to the next station, and once the subway network was in place, a few would use it,  but they still had a massive preference for cars, and something like 98% of disembarking passengers would get straight into one, in spite of the other options being available. Not particularly opposed to the idea of sims hopping on a car when leaving a train station... just ran into a lot of issues because the sims plainly refused to take any option that involved walking. There were even sims who would get in a taxi just to cross the street to the offices opposite the station. A literal 16m drive.

There were a lot of other oddities too, like buses from the neighbouring town coming in and using the station to convert to cars so that they could get to work at the power plant all the way at the other side of the map, which I'd forgotten to add a bus stop to (basically adding more than 5km to their trip), as well as residents walking to the station, but jumping into a car there, instead of getting on the train. Do absolutely love the stations though, but I've taken to keeping them in residential areas, far away from any commerce or industry, and in that context they work perfectly!

fantozzi

IRM changes zone density and purpose type of lots, shouldn't affect traffic. If I remember right, there are option in the NAM traffic controller setup tool to alter also preferences - if sims walk longer distances etc. - that's a setting apart from the capacities.

But imho - I have seen this and other stations already in different pictures, they are popular downloads like the NAM. I'm pretty shure, if it was the NAM other players had confirmed the bug.

Basically, even if not likely, it can be any plugin related to traffic. Even a plugin we wouldn't expect.

At this point it's very difficult to drag down the bug - the database is too slim. If there were other users confirming, we could ask: what do they have in common and so find some trace.

But in this classical situation - for one player it is working fine, for one player not - it's much effort for both sides to narrow the search. bravesirkevin cf.e. had to go through all his plugins, check many NAM settings etc. He could spent weeks with this. The same mattb325.

So me, personally, I would suggest to wait a bit and experiment and hope other players share their experience. It's fishing in the dark atm.     

mattb325

#1434
Thanks, I would like to get to the bottom of this for you (and me). When I say 'all' I mean that more than 90% of incoming trips being converted to car traffic as being 'all' for the sake of ease.

As fantozzi said, it is going to be very difficult to determine why your set up is behaving like this, but something must be causing everything to deviate from the norm, and I'd like to help figure it out, as when I play the game, I don't get any of the congestion/abandonment issues that I have seen from your pictures.

When I look at your city 'the plain' - the map doesn't seem to be full and the pictures that you show have low density houses (13-20 occupants each) but it is giving 12,000 cars on the road. Of course they could be coming from somewhere else on the tile or an adjacent tile, but for all train journeys (ie more than 90%) to be switched to cars is just extraordinary.

The first pictures that I showed had city population of approx 120K (region = 1M), and the stations were only used at half the volume you are experiencing, and they are not converting everyone to a car. The Maxis city also showed (albeit in a limited population) that train journeys aren't being switched to a car, so for me the next logical test is in a big city in a big region with lots of congestion.

Luckily given all the testing I normally do I have a city with 500K on a full region (>6M). This particular city is designed primarily for cars and has highways and avenues etc. There is a train line loop that draws other sims into this city for work. The city has been going for hundreds of years.
It doesn't have subway or bus networks.



It is a busy city - existing train stations on the line are congested and over capacity:



So I have dropped strickland station into this and run the simulation for many years. Again, I just can't get it to convert all the train trips to cars:



Admittedly, it's not 12,000 cars, but I am just struggling to see the difference between the two cities and the traffic volumes that my skyscraper city is making vs your low density city.



And that's not to imply that I don't believe you; it's just that I am at a loss as to how two NAM set-ups on medium could deliver such vastly different results when there is supposedly nothing else at play (the IRM won't impact traffic). This is a copy of my controller settings that have been used when I am showing the pictures from a full plugins (obviously this wasn't the case with the vanilla Maxis only test and no NAM):







bravesirkevin

#1435
Really appreciate your commitment, Matt! My NAM settings are an exact match of those you showed in your screenshot.

Quote from: mattb325 on March 27, 2018, 12:22:50 AM
When I look at your city 'the plain' - the map doesn't seem to be full and the pictures that you show have low density houses (13-20 occupants each) but it is giving 12,000 cars on the road. Of course they could be coming from somewhere else on the tile or an adjacent tile, but for all train journeys (ie more than 90%) to be switched to cars is just extraordinary.

The city is a brand new development in the region that I've been working on for a few months. It's got low density residentials at the moment, but I'm developing the infrastructure for medium density. Most of those houses are not 13-20 occupants variety... they're mostly your mansions and similar kinds of houses, so there are generally around 40-50 occupants per building. The actual city has a population of about 40 000 and it has very large city to the north, and a pretty large one to the west. A fair chunk of the traffic is coming from the western city.

Quote

So I have dropped strickland station into this and run the simulation for many years. Again, I just can't get it to convert all the train trips to cars:

Admittedly, it's not 12,000 cars, but I am just struggling to see the difference between the two cities and the traffic volumes that my skyscraper city is making vs your low density city.

As I'm getting more data from you, I'm becoming less convinced that our respective stations do have different behaviour... Merely different manifestations of that behaviour based on our different approaches to town planning. My siims will use the train station as a train station. They will get on trains and shuffle off to the next train station if the next train station is where they want to go. In my tests, it's not uncommon for the majority of passengers to just stay on the train, provided that their ultimate destination is closer to the next station. The problem I'm facing is not one of sims not wanting to be on a trains, it's that the ones that get off at the station will always get on cars, ignoring all other forms of mass transit, and that non-car traffic that is passing by the station, but not headed to a destination near the train line will use the station to convert to car traffic to complete the trip.

While you keep looking at your screenshots noting, quite rightly, that not all the trains are turning into cars, you're not noticing that in all your screenshots, pretty much every arrow leaving the station that isn't orange is green.

With regards to the town planning:

I had placed one station right next to the busiest road in my city, so it was expected that a lot of sims would be getting off the train to get to their jobs, rather than continuing on to the next station, so it's not a surprise that they got off the trains there. The problem was that they all got off as cars. A further problem was that, being right next to the main road, all the buses passing it (and there were very many buses) immediately turned into cars to finish their journey to work (presumably because cars are faster.)

The second station was placed in a primarily residential district with the purposes of loading sims there to send them north-west to the business district, or south to dirty industry district... Hadn't developed anything the to east yet, so the station was effectively at the end of the line at the time of the test. It's generally my practice to use viaducts where the rail would cross a major arterial, so that I don't have a level crossing there, and I generally line my major arterials with commerce, even in a primarily residential districts. The effect of this was that the station wasn't popular as a point to get on trains, but it was a popular place to get off of one, and all the folks getting off were getting off in cars. Again, being a major arterial, there were a lot of buses going past, and those buses all stopped at the station and turned into cars to complete their journeys.

Thereafter, I tried putting the station in the middle of a viaduct in the middle of a residential area, far from any commerce or industry or main roads, and had replaced the two stations with the NAM 15m viaduct stations. In this third case people were primarily using Strickland station to get on trains, journeying to the other stations and getting off on buses or foot at those destinations. There would still be the occasional passenger getting off the train at the New Strickland (and those passengers would still get into cars rather than catching buses or walking,) but the congestion problems had vanished completely.

Suppose the real lesson from all of this is that the transport hub stations need to be placed with care.

mgb204

Again I must stress there is no bug, the behaviour described is totally logical and makes complete sense. I think it's best to point out that I worked on the Transit Switches, or at least the template of them, for these multi-network stations.

When sims get to the station, they have many options to complete their journey, they will always use the quickest one of those, to get where they want to go. But, that includes the prior rail part of their trip, in addition to travel to the station of embarkation. I.e., for the entire combined journey, use of the station in this manner, makes for the quickest trip from start to finish. Provided the route is also valid, based on other factors, sims will use the quickest route every time, that's a fundamental concept of how SC4's traffic simulator works.

For example, a sim might walk to a bus stop, take a bus to a train station, travel a few stops, then get in a car, drive to another bus stop and then take a bus to work. That might sound crazy, but the traffic simulator doesn't stop to consider if such changes fit some perceived real world logic, it simply looks for the quickest valid route. The game doesn't stop to consider how that sim can find a car at the station, or why they wouldn't have one at home to use from the outset, it simply was never coded to take that into account. Since the routes that utilise the station are so hugely varied between one user and another, this is simply down to the particular design of a given city/region. Of course the preference to use certain types of transport is also dictated by other factors, like TSCT settings, but in this case that would seem not to apply. But due to the zoning/buildings and placement of everything, using the station in this manner makes for the quickest transport routes. If that's causing congestion, you have to look at why so many people want to use the same route. It may be you need to alter your transport networks to prevent this happening.

The way I look at it, which Matt stated before, this station is designed to be a large commuter hub. Placed in a location with plenty of jobs nearby, commuters will use the full range of transport options to switch between. Essentially medium-long distance lines connect to local ones and traffic may flow between them in both directions. But if you intend to use this as a more typical station, then it's probably not ideal for that.

When using other stations, you state this behaviour changes, but that again is completely explainable. Unless those stations have a Passenger -> Car transit switch, it's impossible for rail passengers to leave the station in a car. This switch was included because the parking facility at the station, means it's a valid switch to have. But you will not find such a switch commonly used and I'm pretty sure the NAM stations do not have it. I did wrestle with whether or not to include such a switch, since it's well known that such effects can occur (referred to as "car stealing" in this post). You have to consider too, how without the ability to switch to a car, that may alter the overall commute time, such that it no longer makes using the station the quickest solution for sims.

However in most user-cases, which includes my comprehensive testing too, sim's will not use the station in the manner described. That doesn't mean any city/region design where it does happen is a problem, any more than the transit switches are faulty. Everything is doing exactly what it's been designed to do, in the specific setup of bravesirkevin though, that leads to an odd but explainable phenomenon. But if anything, I think perhaps this station is simply not the best one to use in this instance.

bravesirkevin

Quote from: mgb204 on March 27, 2018, 05:00:46 AM
Again I must stress there is no bug, the behaviour described is totally logical and makes complete sense. I think it's best to point out that I worked on the Transit Switches, or at least the template of them, for these multi-network stations.

Yeah, I initially thought that it was a bug because I'm not used to seeing trains generate cars at all, let alone generate them exclusively. I quickly realised that the fact that the train station generated cars was actually an intended feature.

Quote from: mgb204 on March 27, 2018, 05:00:46 AM
The way I look at it, which Matt stated before, this station is designed to be a large commuter hub. Placed in a location with plenty of jobs nearby, commuters will use the full range of transport options to switch between. Essentially medium-long distance lines connect to local ones and traffic may flow between them in both directions. But if you intend to use this as a more typical station, then it's probably not ideal for that.

While I have no doubt that this is the theory, in practice I've found that the station causes far more problems than it solves when placed in a busy town centre, precisely because all the other options for transportation are unable to compete with cars in the race to get to work. It actually breaks all of the other commute options because it gives the option to get off of mass transit and get back into a car, and that option is always taken because sims think buses suck. While I have no doubt that sims may use the full range of options available to them upon exiting the station, I'm less convinced that it'll be an even spread. In my estimation, based on what I've seen while playing around with the station, it's far more likely to be something like up to 1% foot, up to 1% bus, maybe up to 5% subway/light rail and the other 95% being cars... a very uneven spread!

On the transport map it is quite difficult to tell how the disembarking traffic is being distributed while testing because the route query tool is quite imprecise and only tells you how many of each type of vehicle are on a given portion of road/rail/whatever, and where they came from, but doesn't tell you what proportion of that tally came from each source, and a situation where 95% of the disembarking passengers are cars and the remainder is spread over all the other options, it could easily give one the false impression that it is creating a more-or-less even spread even though it is not. If you see 10 000 cars on a main road near the station, it's tempting to say "obviously it has 10 000 cars on it... it's a main road after all!" and come to the conclusion that the station is not contributing to that in any significant way, but in one of my tests all of the train passengers were exiting the station on to a side road that was used almost exclusively by the station, and that meant that I could get a particularly accurate count of how many cars were being generated by it... That number was pretty close to being equal to the sum of the number of buses, trains, pedestrians, subways and other cars arriving at the station, minus the small count of subway passengers that stayed on the subway to complete their journey.

To actually get them to favour using the other transport methods over cars, leading to a more even spread, you'd have to make it incredibly inconvenient to use a car so that the other options are able to compete. Using bus-only lanes to forces cars to take a long detour or surrounding the station with pedmalls to encourage people to walk or get on mass transit instead of simply getting into a car, but if you're going to go to those lengths, it's probably simpler to just use a station that doesn't generate cars.

You're quite right that changing the station for one that doesn't generate cars will change the entire city's transportation dynamic and will mean that the train station is no longer the quickest route for some sims to get to work, but that's fine. There might be a lot fewer people using the station now, and there might be just as many cars travelling around the city, but the difference is now that the car traffic is spread over the entire network, rather than a significant portion of it being concentrated on just one stretch of road.

mattb325

#1438
I don't know how much further we can get without new information (I'm not trying to wrap the discussion up or anything, but I fear that pushing opposing points will soon run it's course, and we know where that leads  $%Grinno$%).

Just to answer a few points from bravesirkevin raised. First off, all of this testing is de riguer for releasing this sort of content, so it isn't especially onerous; there are plenty more cities that I haven't shown to demonstrate how this station functions.

1) I'm not ignoring the green arrows in my screenshots leaving the station at all. However, I'm not giving singular credence to the arrows as they are simply not granular enough to be meaningful. By that I mean they can represent one single trip, joined trips along the same route, a million trips or everything in-between (or, as you put it, imprecise).
What I am using is the numbers on the train station query in conjunction with the arrows AND the congestion data to draw any meaningful conclusion(s).

I'm usually loathe for large walls of text, but let me explain the contents of one of these pictures in further detail:



What you see in this picture is the morning rush at the station without any other PT connections around it (and there is also no other neighboring city development). The station fronts an avenue and has roads hugging it on the other 3 sides. There is commercial development on the avenue and residential behind the station. The plugin folder is empty apart from the train station. There is no NAM pathfinding, as there is no NAM present. However, Maxis' pathfinding also seeks the quickest route: the NAM merely corrects errors and greatly expands functionality when cities get larger, but in this little setup, Maxis' pathfinding behaves very similarly to the NAM.

The station has 140 civic jobs (nothing radical, even the Maxis station has 4) and there are a total of 360 'journeys' to, from and through the station on the morning commute. The 360 journeys are 163 ped (45%) and 197 (55%) car. It makes sense as nothing as is connected, so there are only two options.

Assuming that 140 of those journeys are for sims simply working at the station, this leaves 220 other 'journeys' in and out of the station, which, based on the percentages leave 99 ped both in and out vs 121 car both in and out. The picture then shows that 3 purple pedestrian arrows are leaving the station to go and work at the commercial buildings to the left of the station. It also shows one purple pedestrian arrow on the right hand side getting into a car and crossing the avenue.

How many are doing this? It actually isn't possible to determine by looking at the arrows (as I said, these can show 1 trip or a million trips), but in that small scenario the train is divvying up journeys out  between the two in types and is definitely not favoring a car. Not 95% car, 1% ped + 4% other as you have experienced. In this test, it is just the station behaving as modded. There are no plugins & mods (either my plugins or your plugins) influencing the journeys.

2) Apart from the last test in the big city, which I have done to try and 'break' the station by forcing car usage, I have taken no care or caution when placing these stations, I have simply placed them where ever I felt like and in multiple scenarios. Can I predict all scenarios? Of course not. The ones I have tested  have been in a new region, new tile and let development spawn around it. An existing region, existing tile connected to other existing tiles and rail, an existing region, new tile with multiple other transit types interacting in the same tile, an existing region with multiple other transit types on adjacent tiles, a vanilla region and a big city region.
In every iteration, I have not seen 95% car conversion on the journeys. I wouldn't have released it otherwise, or I would have taken that transit switch off (I have manipulated the template that mgb204 gave me for things like adding monorail, HSR, removing freight, etc depending on the building type and the function I wanted it to have, and I wouldn't have hesitated to remove the passenger > car switch if I didn't agree with it).

I certainly would hate people to think that it somehow needs to be handled with 'great care' or that is highly unsuited to certain applications as the last few posts imply. My tests indicate that the opposite is true.

3) To answer the point that it is unsuited to a 'busy town centre', again this picture is showing what is happening:

This city is in a full plugins. It is 500K inhabitants and acts to draw in surrounding tiles to work by a) highway, b) road and lastly c) rail. It has no other PT to speak of. The other railway stations are congested. The default NAM will absolutely therefore 100% show a preference for sending train journeys out via a car because of the other congestion all along the railway line.



But again, it doesn't do this. There is a purple pedestrian line running left to right leaving the station after getting off a train. How many in this purple line? Again, it is not possible to say. It is somewhere between 1 and 851. The percentage split of the station that I am absolutely trying to force car usage on is ~1% ped, ~23% car and ~75% train.

And because it is impossible to be precise with the arrows, that's why I am showing congestion data:



You can see that the station is in green just under the 'n' in 'congestion' on the data map. It is on a road, and it is all green immediately around the station (apart from one little yellow corner, but those, as you can see, are trips in). The query says that a total of 6599 cars come and go. But how many cars are leaving? It isn't possible to determine, but if I use the NAM settings that tell me a road can handle 4000 cars before being congested it is very clear that none of the roads adjacent to the station is handling anywhere near this number of cars.
To be able to get 12,000 in such a low density setting something else is influencing this (be a plugin, mod, eternal commuters due to plopping a skyscraper district on an adjacent tile, or NAM settings).

3) As to whether the trips out being a car is a good thing, please consider what is happening in that same picture, and contemplate this scenario: by leaving the station as a car, the sims can economically travel great distances to find work on the tile. This means that the station can be placed in a big commercial-only tile and draw residential sims from surrounding residential only tiles to get work without the commercial tile abandoning or needing a bus stop/subway on every single street. This is one of the many reasons I am only to happy to have this switch feature as part of the station. But regardless, the station works for me in every other scenario.

bravesirkevin

#1439
Quote from: mattb325 on March 27, 2018, 02:00:36 PM
I don't know how much further we can get without new information (I'm not trying to wrap the discussion up or anything, but I fear that pushing opposing points will soon run it's course, and we know where that leads  $%Grinno$%).

I was keen to wrap up the discussion because I'm not interested in coming into conflict with any of you guys because you all do such an amazing job with your mods, I use them all and I love them, and am so deeply grateful to all of you for the hard work you've put in. As this conversation has progressed though, I feel like there is more to be said and I'm going to attempt to provide some new information.

QuoteJust to answer a few points from bravesirkevin raised. First off, all of this testing is de riguer for releasing this sort of content, so it isn't especially onerous; there are plenty more cities that I haven't shown to demonstrate how this station functions.

I absolutely believe that you do test things quite thoroughly. Can't say I've ever run into any trouble with any of your lots, and I've pretty much got every single one that you've ever released installed. I do believe that there are limits to how far one person's testing can go though, and it is possible that I've stumbled across something that your testing, and the testing that the others who worked on the base template might not have revealed.


Quote1) I'm not ignoring the green arrows in my screenshots leaving the station at all. However, I'm not giving singular credence to the arrows as they are simply not granular enough to be meaningful. By that I mean they can represent one single trip, joined trips along the same route, a million trips or everything in-between (or, as you put it, imprecise).
What I am using is the numbers on the train station query in conjunction with the arrows AND the congestion data to draw any meaningful conclusion(s).

Unfortunately, those numbers are just as lacking in granularity, because they don't distinguish between incoming and outgoing traffic... it also includes every single component of every single trip. Because of the nature of the station, you're going to have very high pedestrian numbers even if no single sim leaves the station on foot. Congestion data is also sketchy because it only considers how clogged the roadway/railway/whatever is, but doesn't provide a lot of information about what's causing the clogging.



QuoteI'm usually loathe for large walls of text, but let me explain the contents of one of these pictures in further detail:

Walls of text do suck, but they're an unfortunately necessary evil when one is trying to communicate complex ideas. Will refrain from commenting on each thing below individually to avoid contributing to the wall-of-text epidemic more than I have to. :) I will, however, directly point out the things that I feel don't quite add up.

• The 12000 thing isn't actually important. It was probably a mistake on my part to give that so much emphasis, when the key issue for me was actually that virtually all of the traffic on that very busy train line was being turned into cars. (The reason for the large volume of traffic was that the western neighbour is predominantly a high density residential city and there weren't enough jobs to go around there so they all flooded into the new city looking for work. I'm 99% sure that it wasn't an eternal commuters problem because I'd taken some pains to avoid such a thing, but it is possible.)

• Congestion isn't that big a deal to me either, and again it was a mistake to overemphasize that portion of it, because that's getting a lot of attention when I'm most concerned about the fact that it's turning everything into cars. Even a train station that has a 100% conversion rate of passengers into cars is not generally going to cause congestion at all, until it gets really busy, and it's difficult to get a train line that busy. That busy train line was just an extreme case, and an outlier. It did however make me very aware of the fact that the presence of the station was breaking my public transport network, rather than being a useful component of it, as I'd expect a commuter hub to be.

• In the first picture, you're coming to the conclusion that the traffic departing the station is sorta evenly split between pedestrians and cars based on the numbers and the pink arrows leaving the station to go to the small shops at the side. The problem is that you're extrapolating how many pedestrians are leaving the station based on the rough proportion of numbers. Those small shops occasionally do have as many as 55 workers, but if they're the ones I'm thinking they are, then they each only have about 9, and there's no guarantee that they're each receiving 9 pedestrians from the station. You could actually click the route query tool on each of those 3 buildings and see how many actual pedestrians are leaving the station, and from that you could get a slightly better estimate of how many of the pedestrians entering the station are being turned into cars or station workers. I actually think that the workers heading to the small shops didn't actually go into the station at all, and simply walked through it, (as opposed to having driven to the station and parked before walking to work) and that would mean that they're not actually a part of the equation.

Now for some attempted new information. What I did here is open up an existing city with some well established infrastructure and picked a central train station that's already in use and gleaned as much information as I could off of it before bulldozing it and replacing it with Strickland station, ran it at cheetah speed for a couple months, and then gathered all the same data. I did do it with a full plugins folder (because it was an already established city) and had NAM active with the default settings as always.


Here's the view of the train station as it was... As you can see from the news ticker and no-jobs zot, the city has a lot of problems anyway, but that's not the focus of this, so I'm just pre-emptively mentioning it to avoid having those conversations later.


Here's the train traffic on the eastern line.


Here's the train traffic on the western line.


Here's the traffic for the station in general. Unfortunately the Victorian Viaduct Station doesn't seem to have any useful tooltip for the Route Query tool, so I can't glean exact numbers of each kind of transport involved in the process, but there are no cars coming into or out of the station at all.


Here's the congestion map. Those intersections are already pretty busy... that avenue is a major arterial and it's a fairly dense neighbourhood.


Here's the volume of pedestrian traffic.


Here's the car traffic


Bus traffic


Subway Traffic


Train Traffic.


Here's the area with the station replaced. Not much to see here, including it simply to demonstrate that the only thing that changed was the station.


Train traffic on the eastbound line. Interestingly, it's quite a bit higher than it was with the old station.


Train traffic on the westbound line. Again, it's higher. In fact the dynamic has changed completely, but that's not unexpected.


Now here's the whole station, along with a helpful tool tip giving us lots of fun numbers to crunch. That's an awful lot of cars considering that the old station had no cars at all. Twice as many as there are trains. There are also a lot of subway trains going through there too.


Congestion map... It's about on par with the previous one, but we're not adding that many cars to the mix so I wasn't actually expecting everything to turn bright red.


Pedestrian map... Notable is the sharp drop in pedestrian activity in the commercial areas near the station.


Car traffic volume. This has jumped sharply around the station. Was previously marked as around 10% and now it's around 70%.


Bus traffic volume. Previously there was a ton of bus traffic around the commercials, but now it has all but vanished!


Subway traffic. Looks about the same.


Train traffic looks about the same too, but it is actually heavier.


Just another quick look back at the station stats now that we've examined all the other data. We've got this station and it's got lines coming off of it in pretty much every colour. Congestion certainly isn't worse than it was before and It seems like we can say that the station is working perfectly, doing everything that's intended... can call this test done, right? Alas, there's more to this story...

Let's look at the lines of traffic there.... There's blue approaching the station, but none leaving it. Buses are only arriving, not leaving. There's a lone pale pink arrow heading to the Eurolux boutiques where a glorious 7 people are employed. All of the other pink is people heading to the station. Quite a lot more orange that's heading away from the station than before, but they don't work in this part of town anyway, so they have no reason to get off here. That's a ton of yellow down there... the subways are booming! And then every other arrow headed away from the station is green. It's a safe bet that apart from 7 people who like exercise, all of the passengers who got off at this station either took a car or got on the subway.

So let's check out the subway:

Oh, damn! No one leaving the station got on the northbound subway line at all. They either stayed on the subway and passed through the station or they got off at the station and got onto the train or into cars!


Looks like no one leaving the station got onto the southbound subway line either.

Thing is, from the numbers, the congestion and the rainbow arrow of arrows surrounding the station it really did look like the station was functioning perfectly, but as has been clearly demonstrated here, that perfectly functioning station was converting ALL of the disembarking passengers into cars. There are 3 kinds of people leaving the station: those that got onto the train, those that passed through the station without interacting with it, and those that got onto cars.