Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Discussions Rules and Guidelines
Report this post
2. This is an occasional persistent problem. I have sent some savegames of this via email ages ago, but never made a discovery of consistent steps to reproduce. Can be quite frustrating. The track looks and behaves normally right up till you try to build a switch.
An hour to upgrade a station? How exhausting! Pretty sure you can trim that down a lot, more so if you dispense a little realism. How complicated are your routes? I'm curious to see a screenshot or two. Have you forced specific platform usage for each train? How long are your games then?
It looks like you are sticking with the approach of connecting everything to all of the Platforms at all Stations. I can't imagine doing that, but it would explain how bringing in trains and making every incoming line have access to 8 Platforms per City would need an hour of switch placing. That would certainly be frustrating. Point to point is just so fast, easy, cheap, efficient. I like your desire for a more realistic looking setup, but it really doesn't play well in this game, sorry :(
When you lay tracks (...) they won't be necessarily a single piece but a set of tracks. You can see this by hovering and thus lighten them up while still in Track construction mode. This includes straight flat tracks as there is no such thing unless you have an underlying grid you can snap to.
Every time you add a switch (= two new track points) these block(s) are broken up again and new sets of coordinates and vectors (one for both ends) have to be calculated. Since floating-point numbers are real numbers they are an approximation and not exact numbers while the track between those two sets are a function.
You may think that adding a new set would work perfectly and it would if the new set would be used for both new segments, left and right of the new added track point. But there are two new sets generated, one for each new segment.
You can spot this during construction by adding a track point preferably when flat terrain switches to steep slope. You'll end up with a knee instead of a curve (vertically). You may not see all these knees on flat terrain but they are still there.
You can even find horizontal shifts on straight tracks, most likely on long stretches you've constructed at once.
While I can understand the knees - it would be a more complex track building to prevent them. But when a fast solution seems to work why invest more time into a better, more realistic one... but that's a game design decision and time is money, except when you're transporting freight ;) - someone seems to have messed around with integers or unnecessary rounding functions. At least that's what come into my mind when you see a straight track jumping to the left or right before proceeding straight. That's a huge difference, too big to be based solely on a new calculations using floating points only.
Another easy to spot adding of new segments is when you add and remove a Supply tower and than add new Signals pressing Ctrl thus switch into Multi-signal mode. The new Signals will be at both sides of the former Supply tower and not side-by-side as usual.
So all these new segments you're creating with each switch you're laying shifts the tracks more and more apart til the point where they are not considered to be parallel anymore even when your eyes tells you otherwise.
I'm not sure if your explanation fits what I experienced. I don't remember doing much track work in those areas beforehand. The question would be, can you take a "good" track and make it misbehave? In my case I couldn't make a switch down an entire section. These sections were, if I recall correctly, at least a couple of train lengths long.
You can find those horizontal shifts of straight tracks using Track construction mode. While you barely spot the rift otherwise you'll easily can spot the black bar between the two segments when they are highlighted. Since it's a 3D map the vectors can be off on any axis while not every shifted axis has to be visually spotable - the eye works analog and has physical limitations, like floating point numbers, and the monitor works digital, more like integers even so higher resolutions are better than the human eye. In the end it's all about precision and when or how fast rounding adds up to become visible. Unless they use very complex functions with variables differing in multiple magnitudes there should be no such thing as adding up to visible "glitches". But that's my personal interpretation based on my experience and not the sole eligible one.
Sometimes you can even spot tracks moving slightly when adding switches. So yes, you can take a "good" track and make it misbehave.
Sections don't have fixed length. When you lay a (long) track it will be broken up into a variable number of sections of very different lengths. Bridges and tunnels are obvious. The section adjacent to a bridge will often be short. But don't ask me about the underlying rules, that would involve too much speculation.
Worst case for me so far is when I have to remove a whole sections of once parallel tracks after adding and removing a switch that didn't allow to add the second X-part. It's not the normal behaviour but it happens much too often to be a simple glitch.
As for the replication, that's called testing and should be a part of any QA-department. And yes, GMS, KMG, etc., they all have their names listed under Contributors, eight QA Managers, plus a bunch of externals.
Taking a close look at track Utilization at Stations, you quickly discover that point-to-point architecture is NOT all that efficient. Even on a 4-track Large Station Track #1 shoots past 100% _real quick_. So spreading out incoming trains by assigning specific tracks/platforms lightens the load -- AND speeds up train turnaround time. Which is VERY important for making Express trains. It also _should_ make it easier for an incoming train to dodge around a Random Breakdown that didn't quite make it to the Maintenance Shed in time. (I seem to get A LOT of those.) But I doubt that the program is sophisticated enough to allow a routed train to go to any other empty track/platform rather than the (blocked) one to which it is assigned.
Anyway, visualize this as my standard switchyard for a Large Station:
Terminal building at the top. Four tracks extending out left, and four tracks extending out right.. All four tracks on both sides are precisely adjacent to one another.
Starting at the bottom, lay in a switch going from #4 to #3. Then immediately from #3 to #2, and then again from #2 to #1.
The 'yards are now transitional from #1 to #4. Buuuttt a train coming in "below" the switches can't navigate from #4 to #1.
At this point, a 4-track Supply Depot can be placed. Or place it later when the switches are complete.
Now, going further outward. lay in a switch that goes from #1 downward to #2. Then another from #2 to #3, and finally from #3 to #4. Then, if one hasn't already been placed in the middle, place a 4-track Supply Depot.
Further out, the tracks then can head off in four different destination left, and four different destinations right.
Now, what this pattern accomplishes is that any train coming in from ANYWHERE, on ANY track can reach ANY track/platform. And likewise depart the station on ANY track to end up on ANY track outbound.
I tried to find a more compact pattern by executing multiple switches at the same point, but none of those made that ANY solution possible.
Now if you want a REAL headache, consider what is involved with tearing up an existing 2-track Station pattern and rework it into a 4-track Large Station pattern. And a PHENOMENAL headache converting to an 8-track double Large Station pattern. Making a line of of switches that go up from #8 to #1, _and then down again_ from #1 to #8 takes a phenomenal amount of uncluttered real estate.
^That is why "more than an hour per" comes into play.
[BTW, take into account the _scale_ of the map. For the scale used by the game, ONE of my double Large Station switchyards would be the width of most States.]
https://steamcommunity.com/sharedfiles/filedetails/?id=2471016015
I tried to "simulate" your solution. If I am wrong, I apologize in advance. I added a signal-control one below it for comparison. If you want to keep "manual" control, you could just do a similar solution by hand.
Notice that I am able to use 3 up-down pairs instead of 2.
Service tower anywhere near a station is a bad idea. In my view it's a bottleneck just waiting to form.
Also, I want to tell you that the utilization figure for a platform is a simplistic tool. I believe it mainly considers how many trains use that platform, and NOT how long trips those trains are taking. I'm not inclined to believe that maintenance time for long vs. short trips will equalize this properly. I don't pay attention to this figure, but for a rough idea: longer trips 150%+. Local trips may be at 50-75%.
It's a game. I like efficiency. I can describe what seems to be best for that. I want arriving trains to wait just slightly before using a platform, but have no delays when departing. But, don't think you should play a certain way based on something I said. Games should be for fun. :)
Looking at the screenshot though, I see the Large Station has 4 tracks inbound, but _8_ tracks outbound. On my double Large Station configuration, that would turn into _16_ outbound tracks. That's a bit much, even for me.
Four switches (middle configuration) or 8 switches out (end point configuration) breaks up the train traffic across four tracks, so backups don't occur unless two bumper-to-bumper trains are assigned to the same track/platform. (In which case, delay was inevitable.)
I am ABSOLUTELY certain that the Utilization figure reflects the number of trains using that particular piece of track... even if they are only passing through. I can see this when I work my way through the "Trains using this Station" train list. I might, for instance, account for, say 52% on track/platform #1, but after doing assignments, I look again and track/platform #1 is showing 108%. The difference are those trains taking the shortest path between two other destinations. Because of this, I rarely assign any trains to the track/platform showing the highest Utilization value.
[If you try to lay down track that would bypass the city Station, it doesn't always work because the program will ALWAYS use the shortest path, and if the shortest path passes through a Station, that ramps up the Utilization percentage on whichever track/platform acted as a conduit.]
I can only see p2p being viable for city-to-city traffic. And if all the city stations are tied up with city-to-city tracks, where are the Resources trains coming in from the countryside supposed to unload? [This is probably where Warehouses become a significant factor, with all products from rural businesses going to the Warehouses and all city production (Passengers, Mail, and processed goods actually going INTO a city.] But doesn't the system fall apart with city-produced tertiary goods? For example, Wood goes to a Warehouse (outside), gets transformed into Planks. Now all those Planks need to go to cities that produce Furniture. Again, it gets routed to the destination city's Warehouse. (Or not. It may go direct.) Then comes the Furniture... which can only be routed to cities with large enough populations. How do you accomplish p2p tracking between cities that may be 2-3-4 cities apart from one another?
I've never had a situation so far where eight cities were connected to a single one. The standard model is four.
Basic city-to-city-connections are between those producing different primary food, like Beer and Meat. There are very rare circumstances when you have to link up Beer with Beer or Meat with Meat. Trains are using Automatic and after the first exchange of Beer and Meat most cars are express cars and most trains end up as express trains automatically. As long as the utilization stays in the two digits there is no waiting time unless a random breakdown happens which is unavoidable. After that they simply sort it out by their own as they split out evenly again.
That's a working concept without interference by other connections.
It also deals with any industrial good that is produced in the other city.
Keep in mind that swapping Beer and Meat covers ~40% of early game demand. As a rule of thumb the first five consumer goods are demanded about twice as much as the rest. It shifts, some scenarios have more detailed formulas but you wanna push growth as early as possible to get the money you need.
Basic rural-to-city-connections provide goods like Cattle for Meat production, which are only demanded in certain cities while other cities don't want them at all. In most scenarios this includes Cotton and other non edible goods.
Depending on your city growth you can steer new industries to some degree. In most cases you can avoid to haul goods across the map.
At this point you're not even close to eight tracks in most cases.
Nearby food can be served with refrigerator car adding another 5% to their value.
At some point the incoming goods become too much to be handled by the other tracks alone. That's when warehouses may show up to add another set of tracks. Depending on the map, how far away common goods are, how sparse common goods are, how clogged the map is with cities, connecting warehouses and swapping certain goods between them can help. speeding up the distribution process.
There is no reason for city-to-far-away-city-connections except when a task demands it. Any express good can and will be take the next train in the next city that goes to their destination as long as the whole chain of trains stays under 150% of the direct wagon route.
That's why you won't see express goods going from Sacramento via Redding to Carson City, neither by pass through nor bypass. The bypass has to be over or under the Rockies not coming near Redding at all.
That's the perfect argument against passing though :)
With p2p you know exactly what train is using which track. In late game there are often more express goods than your network can handle. So what? You have more money than you can handle, too ;) We're in the steam era, there are always horses and wagons just waiting to earn some money, too. Don't starve them to death, give them something to earn so they may take your train for their holiday ;)
So we agree on this... why don't you just do it the way you've described it? Adding a waypoint when necessary shouldn't be an issue, they are there to steer trains they way you want them to go.
Btw the devs have taken this into account a long while ago after reports by players and since than bypasses are favored over shortest routes. Technically it looks like they've taken into account that trains are not supposed to run through a station on high speed in reality thus increasing the calculated length for tracks within stations when the shortest route is searched for.
When the city bypass feature got added, I tested it out. This is an exception to the shortest path rule. Passing through a Station and not stopping seems to add a virtual 17 miles or so to the calculated path length, making the bypass automatically selected as the shortest (effective) route most of the time. A waypoint will take care of the cases when the bypass is too much longer.
Point to point architecture works well for me. It is what I use 90% of the time for gold checkmarks. By the time the game is over I have typically used one or two Warehouses attached to Cities, maybe three. Until the City gets really large, two Large Stations with p2p can handle all the express and freight with minimal traffic jams. You don't need to be absolutely strict with isolated track and only one Platform connected at each end, but that dominates the setup for me. Sure a few exceptions are needed.
The huge interference of massive switching yards means more effort is required to try and jam more trains in to handle the demand, which makes the idea of making even more switches necessary seem appealing. It just makes the problem worse. It doesn't matter if you give access to more Platforms for each train. With p2p (strict) the trains taking roughly equal loads (enforced by setting Min Cars) keep spaced apart so there can not be much of a cue apart from maintenance and breakdowns. This makes access to more Platforms unnecessary. This lets you use a single switch, improving swap time and overall throughput. It really does work about 90% of the time.
I can't see accomplishing ^that on a p2p system.
As to the bypass feature (of which this is the first I heard of it), how does it work? If this is something that REQUIRES the signaling system, I can't use it, as I ALWAYS set rail traffic to Normal (trains can pass in opposite directions without interference).
Normal track use does explain a lot of our different perspectives. I'm embarrassed the that didn't cross my mind! Thineboot and I will generally use Realistic track. With Normal, I suppose swap times will be completely unaffected by the switch networks, since the trains can pull up right next to the Platform and sit on top of each other. That means that the advantage of p2p are not realized, as the interference it reduces wasn't happening in the first place. I think Normal mode only allows as many trains to load/unload at one time as there are Platforms at a Station, but I thought that they could all sit on top of each other at the same Platform. I may not remember that right (didn't play much Normal) but if that is the case, I don't think there is any value in distributing trains to different Platforms.
The many switches and Supply tower setup do require a length of path stick out fairly straight from the Station. For trains that want to veer off either way, that does add to the length of their paths unnecessarily, and of course adds to cost.
I'm still at a loss as to why so many trains and years are required. I typically remember finishing scenarios in something like 3 years with not that many trains. I would have a couple of larger Cities (>> 100,000), maybe 7 more over 100,000 anyway.
We're talking about this now for decades - Covid-time - in multiple threads and your premise was Realistic and not Fantasy and now that?, Sorry, but I'm baffled :(
As for the city sizes at the end of my scenarios they are what they are. Unless there is a task forcing me to reach a specific size I grow them as fast as possible with the goods that are available. Which means normally they end up at 90K+ for Museums or Universities, whatever fits better to reach the goals.
Since I finish all tasks with golden ticks - the only exception so far is the Atlanta-tick in Gone with the headwind - and the slowest time is the early game the rest of the tasks are done much faster. I rarely develop industries other than the ones really necessary to finish tasks as they just don't pay back before the last task is done.
We're playing two completely different styles, I got it. And that's fine as long as everybody has fun with their own game play. Since you're complaining about problems you have with endless switches we've tried hard to show you another approach which seems to make no sense as there is... yeah, what?
With trains running through each other Is signalling even necessary? A rail to reach them all? Now I have nightmares ahead ;)
It seems to me that what the program does that makes a Large Station more efficient than a Station which is more efficient than a Small Station is that it reduces the Utilization percentage Cost by the number of tracks/platforms available, even if ALL of the traffic move track/platform #1. Regardless, what I have noticed is that with dispersed routing across a greater number of tracks/platforms DOES get in and out of stations MUCH quicker.
Ah, but you see, we are ALL playing a Fantasy! Given the scale on the maps of the various regions and nations, the scale of the train you see on the screen is about as big as a State _county_. And the tracks are to the same scale: HUGE.
In case it hasn't been noted yet, I'm something of a History buff. I know A LOT about the workings of historical RRs. (Helped along considerably by my grandfather who was a major shareholder in a third-tier RR in Northern WI and Upper MI. Learned A LOT about RRs from him.)
Everywhere you go between cities, you commonly see doubled-track, which obviously allows trains to get by one another without colliding. Less frequently, you may come across a siding that only runs for about a mile long. <Those are what's missing in RE. But when the Realism is set to Normal, the _effect_ of built-in sidings appears. Which is why I do it that way. {Plus I really have no patience for debugging a switching system across an entire network.) [Plus I dislike the concept of making trains sit for _days_ for the tracks ahead of them to clear. Never happened. Hours, maybe, but never days. (Unless there was a weather or collision ahead that needs to first be cleared.)]