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
For example
If you set up a train route that runs A to B and the round trip is 20 minutes. (That is A to B to A again). If you buy 4 trains for the route, the depot will only release one every five minutes
I would suggest at least one thing. When you build a line and trains and when you want to send them on the track, make them leave the depot in the set intervals. They could be calculated like based on the frequency of the line by the game. Even that would help a lot.
Edit: Dave beat me to it. :)
I think it would not be that easy to implement that since there can be many lines on the same track and tracks have different speeds, but probably the most feasible way would be to make trains wait at station a little longer or run a little slower if they come too close to the train before them until they come approximately at the right distance (e.g. in terms of frequency measured at several chekpoints on the track e.g. after a certain distance).
I thought I read somewhere that this was already implemented in train fever, but it did not seem to really work even if the track had only one line... If it is the case it should be a little more aggressive in spacing out trains.
This is how it works:
- the first vehicle is timed and the frequency is adapted to it's time
- the next vehicle will wait at all stations untill it's (roundtrip time / # vehicles) behind the previous one, BUT it can only wait for a max of 10 seconds extra at a station.
The system works well for buses and trams because they generaly have much stops, so they can wait for an extra 10 seconds often. (a bus route with 5 stops that takes 5 minutes for the lead vehicle, allows the following vehicles
It works the worsed for long routes with slow vehicles (early truck lines); because the amount of time the vehicles can wait is nothing compared to the time they are driving.
The system has a couple of problems though:
- if you cross multiple lines and send them through the stame platform, the waiting vehicles of one line, can disturbe vehicles on another line.
- as mentioned before, it doesn't work for long routes with few stops, especially when the vehicles are also slow
- the player can't match frequencies of different lines well enough (this goes together with the first problem)
- when you have at least 3 vehicles, the 2e will wait 10 seconds extra and the 3e will do the same. At the next stop however, the 2e will wait 10 seonds again, which will bring the 3e vehicle back together with the 2e vehicle. This means that it will take a very long time to space a lot of vehicles out.
Therefore I propose these changes:
1) Vehicles should only be allowed to wait if the platform they are on is exlusive for their line
2) The time a vehicle is allowed to wait, should be longer for slower vehicles (maybe let them wait x% of the roundtrip time)
3) A 3e vehicle should wait double the amount of time, a 4e should wait triple the amount etc...
4) If a train has to stop for a signal, it should also start the autospacing routine
5) The player needs better information about frequencies (5 minutes is not accurate enough, we need to know if it's 4,8 minutes or 5,2 minutes so we have a better idea if we can mix the routes)
6) To prevent clogging at stations, a depot should also be considered a waiting spot, so that when starting the line, the vehicles should already be spaced a bit
7) Allow the player to adjust the frequency time upwards if needed..F.e. a line with a frequency of 6,5 minutes has to be mixed with a line of 6 minutes. Adding an extra vehicle to one of the lines, will not solve the mismatch in frequencies, so we should be able to increase the frequency of the 6min line to 6,5 minutes. The best way to do this would be an automatch feature, so that when something happens to the largest frequency, the other frequency is also adjusted. (we don't want to babysit all lines to see if the frequencies are still matching)
Which is what signal are for. Signals wil work regardless of the line a train is on. Just put the number of signals equal to (the max number of trains minus one) halved, assuming you use double lines. Signals are the quickest way already available to space out vehicles.
Would cause a new line not to space out if it shares a platform with an exsisiting spaced out line. This is not an issue for trams, but would trouble truck lines.
Nitpick here, but from your earlier statement, should this not count for longer lines, instead of slower vehicles?
How can a line tell what the first, second, third or fourth vehicle is? All a computer sees is multiple vehicles on a line with no beginning or end. This might cause the 'last' vehicle to back up sofar that the first vehicle catches up, which makes it wait for.... how long?
Signals already space out vehicles. Without addressing the issues above, this causes more problems than it solves.
Not in combination with the current autospace system. There is a risk that vehicles of the slightly faster line will lose some speed because they have to deal with vehicles of a slower line, but as long as the vehicle speeds are equal, the problems are neclegible.
Good idea, regardless of any change to the system or not.
Again, not really necessary if vehicles are of equal speed, and share little infrastructure. Systemic irregularities are caused by unequal vehicle speeds, not my unequal ferequencies as much.
This will only happen if the new line only uses platforms that are already used by other lines. This is rarely the case and bad design in all cases anyway. (notice I said platforms, not stations. you can have two truck lines in a small truck station because it has two platforms)
If the vehicles are slower or the route is longer, the roundtrip time will be larger. ;)
This problem only happens when putting a bunch of vehicles on the same line from the same depot. So the order should be the order they left the depot. The game already keeps track of a first vehicle too. (I think it's always the first vehicle shown in the line manager)
This should indeed only be implemented if point 1 is finished.
But it would help to keep vehicles seperated and will allow for new seperations to be completed quicker (f.e. when you add a new train to an older line)
What's wrong with more information? The game already updates frequencies, we just can't see the number because it is rounded. The game even shows frequencies up to seconds if you go below 2 minutes... The problem I currently have is when you try to mix two lines that both have 5 min frequency, they should be able to alternate nicely. But if the first line is actually 4,8min and the second 5,2 minutes, they won't alternate nicely, and it will cause the vehicles of both lines to start grouping up instead of staying nicely spread out.
The effect of a mismatch in frequencies is much bigger than a mismatch in speed. If you have two lines alternating perfectly with a frequency of exactly 5 minutes and they share track that's not too long, a mismatch in speed doesn't matter, because they will never be at the same place together.
If however, the speed is the same, but they arrive at the same time (where the lines merge together) because of mismatched frequencies, one of them has to stop.
Correct me if I am wrong, but it seems that your premise for solving vehicle traffic issues should be good engineering. To use some of your words; to make good networks, players need to use signals, bridges and tunnels, the right roads and tracks to generate an optimum frequency for most or all of your vehicles. Please correct me if I am wrong, or incomplete in this statement.
I think I understand you here, but I'll talk about this at point 4.
No, after a 10 second delay, a faster vehicle will lose a greater amount of distance to the vehicle in front of it than a slower vehicle, regardless of the distance vehicles need to travel. Therefor, a fixed delay for all vehicles causes fatser vehicles to space out over a line quicker.
To quote Arkady Darrell, a circle has no end. That is how a computer looks at a line. It sees a number of vehicles in a circle, going from waypoint to waypoint. A computer can see how far apart the vehicles are in either time or space, but it will make no call on which vehicle is first.
Now you can make a computer look up a pair of vehicles whose time-space difference is greatest, and assign the last of the two vehicles the honour of being first, but by the increasing the difference in time-space between the next vehicles, the last vehicle may well back up into the first, changing the order between vehicles again, and setting any spacing routine on a rabbit's chase.
The advantage of the current system is that because it is so simple, it cannot be broken in any way. It may perform poorly at times, but it will never fail.
Here we go. I don't understand this. If I understand your premise correctly, then players create a good network that maximizes frequency on their own, and they do it by adding an extra pair of signals before adding another train to a line. If so, no modification to automatic spacing is necessary.
Nothing is wrong with a little extra information, but you might not need it. I'll explain at the next point.
This is a misconception of what frequency is, and is in fact a mixup of two separate problems; the mixup is partially my fault. Thinking caps on (for me too) and here we go:
Frequency mismatch isn't the problem you make it to be. Lets work this out with an example.
Assume Line A has 4 trucks and visits a certain depot (Y) every 30 seconds. At depot Y, we would see trucks from line A arive as follows:
30 sec
60 sec
90 sec
120 sec and so on.
However, lets asume that Line B as a single truck, which happens to arive at 120 seconds, just before the truck from line A does. Lets update the arivals:
A 30 sec
A 60 sec
A 90 sec
B 120 sec
A 130 sec This vehicle is now 10 seconds late, which saves me a lot of time in my explanation later on. Each of the other trucks from line A will now wait 10 seconds at Y because they have a truck 20 seconds in front of them, and another truck 30 seconds behind them. Trucks could also wait at depot Z (where they came from). Eventually, autospacing will correct the frequency of the vehicles on line A.
This is what the timetable will then look like:
A 30 sec
A 60 sec
A 90 sec
B 120 sec
A 130 sec
A 150 sec
A 180 sec
A 210 sec
B 240 sec
A 250 sec
A 280 sec
A 310 sec
A 340 sec
B 360 sec
A 370 sec And from here on, line A and B will no longer interfere.
In terms of physics, the frequency of line A has corrected itself to remain at 30 seconds. The only permanent change in line A is its PHASE; it is now 10 seconds out of phase, compared to its original schedule from before the truck from line B interfered.
This is why a difference infrequency isn't the biggest problem; a difference in vehicle velocity over a significant stretch of road or rail is. Don'tt ask me to explain right now, my head is swimming a bit.
My premise is that I want to improve the current autospacing system so that vehicles get spaced out quicker on their line.
This can be done in two ways (premise 1 and 2):
1) increase the time a vehicle is allowed to wait at a platform
2) increase the number of waiting spaces on a line
This spacing however, should also be stable, otherwise you'l see weird oscillations in the frequencies of lines. Let's call this premise 3
This has a lot to do with what you mentioned in your previous post. As you can see, if both lines are a multitude of eachother (like 120sec is a multitude of 30sec), the lines will be able to alternate between eachother without causing interuptions. However, if the lines' frequencies are not a multitude of eachother, you'l never be able to get a stable alternating phase.
example: line A arrives every 30 sec, but instead of 120sec, line B only takes 110 sec.
A: 30, 60 and 90
B: 110
A: 120 (everything seems fine)
A: 150, 180 and 210
B: 220 (everything still seems fine)
A: 240, 270 and 300
B: 330
A: 330 (suddenly, there is a problem, where line A and B arrive at the same time, causing one of them to be late)
Now imagine that you have another line that has to alternate with A, and one that has to alternate with B. Now one of those lines will get out of sync too and the problem translates through your whole network. (this is especially a problem for a train network, because they don't stop and accelerate as quickly as trucks)
My 1e point is to improve premise 3.
By disallowing a vehicle from line A to wait at a platform where line B also passes, a problem in line A, won't cause as much of a problem for line B. (If a vehicle from line A waits 10 seconds while a vehicle from line B wants to enter the platform, you get an unnecessary delay for line B)
Point 2 is to improve premise 1
If you have a line with a long roundtrip time, it will take longer for vehicles to spread out, because the time between the vehicles should be larger. This is directily proportional. f.e. a line with a roundtrip of 6min and 3 vehicles, needs 2 minutes between the vehicles, so vehicle two will need to wait about 120 seconds before it is spaced and vehicle three will have to wait about 240 seconds. If the vehicles are slower or the line is longer (or a combination of both), you could have a rountrip time of 12 minutes, which would require vehicle two and three to wait double the amount to achieve the right spacing.
By directly linking the amount a vehicle is allowed to wait with the roundtrip time, the time to achieve the right spacing becomes independent of the lenght of the line and the speed of the vehicles. 5% seems like a good number as a 5min line (roundtrip, not frequency), will allow up to 15sec wait, while a 10min line will allow up to 30sec wait. A minimum time of 10seconds for short routes can be kept.
Point 3 is to improve premise 1 again.
As mentioned in the previous point, a 3e vehicle will have to wait double the amount of time to be spaced than a 2e vehicle. So allowing the second and third vehicle to wait longer, it will drasticly improve the speed at which lines autoseperate when a lot of vehicles are put on a line at once. Ofcourse, if it can't be coded, it can't be implemented, but I think it should be possible.
Point 4 is to improve premise 2
Currently trains are only allowed to wait at a station platform. By allowing a train to wait at a signal (if the track is exclusive to that line as mentioned in point 1), the number of places trains can stop is increased, which will improve the speed at which vehicles get seperated.
After thinking, trains should probably also be allowed to wait, even if the signal is green.
Point 5 is to improve premise 3
If the player knows two lines are not in sync (have a frequency that is a multitude of eachoter), he/she has the option of not combining those lines, or adding/removing vehicles to make them more in sync. That is currently not possible. If you mix two lines with a frequency of 5 minutes, you don't know if they will keep alternating nicely or if you'l end up with 2 trains of line A behind eachother and a hopelessly late train of line B behind them.
Point 6 is to improve premise 2 (but you already agreed)
Point 7 is to improve premise 3
It will not be possible to get frequencies in sync nicely most of the time. To imrpove this, we should be able to add a bit of waiting time to a line (this can be limited to the percentage mentioned in point 2 to prevent a line trying to reach an unobtainable frequency). The game can decide where (only on platforms that are exlusive as mentioned in point 1 ofcourse).
This means that for the example I made in premise 3 (with line A 30sec and line B 110sec), we should have to option to tell the game we want to sync lines A and B by increasing the frequency (time between vehicles) to match a multitude of line A. The game can than very easily adjust the frequency of line B to 120sec. If something might happen to line A, which changes the frequency to 35sec f.e., line B can automaticly adjust to the new fequency: 140sec. (or when line A gets to 37sec, line B can go to 111sec)
I like the effort you put in your suggestions but I feel like it is not really a solution to the Problem. It'll all work fine in a very small rail system with two or three routes but I feel like most people will want to do more complicated stuff.
Considering that with your solution the system works best if the trains have the freedom to wait at a station and/or a signal that will lead to the conclusion that the more complicated the train lines get and the more traffic you have at your train stations the worse the spacing system will work. And a complicated route plan is where you'll need it the most.
I definately don't want a system as complicated as A-Train for Transport Fever but I would suggest the option of either departure times or a fixed frequency that you can set for a route. You could give it as an option of either that or the automatic spacing system (for simpler lines). Of course that provokes the risk that with dep times or fixed freq the trains will stack up behind one another if one train misses its departure.
I think you're right that we may wel agree on a lot, and I think that my own statements cause some confusion here. By premise I mean the original outlook or idea which shapes a solution. What you state above are solutions. Let me state the two contradictionary premises I was thinking about out loud you you can quickly (dis)agree with them:
1) Smart players should use dumb tools like signals and autospacing to solve line issues and maximizing frequency if they want to create complex networks.
2) Dumb tools should be replaced with smart tools to make the creating of more complex netowrks easier. Whether or not the player needs to deliver some brainpower can be debated.
I understand what you're getting at, but there's a lot more going on here. In my example, I used rounded figures to make it clear that the frequencies remain the same, but differernt lines will change phase which in the end will avoid interference at stations or depots.
These easy to read numbers however aren't common numbers, one might as easily have two lines with frequencies of 59 seconds and 127 seconds. Notice that I picked prime numbers here. It becomes much harder to guestimate when these lines will interfere.
Now my argument regarding phase stands regardless of the numbers, but if two lines meet up in a station with the frequencies of 59 seconds and 127 seconds, the lines will interfere once so many years, and auto spacing will induce an average loss of frequency that may not even be a second per year.
Which means that while line A may effect B and line B might affect C, I think the actual time lost is negligible.
As for trains and stations; good point, but this is where the need for a premise shows its face.
A smart tool might help out players here, but shouldn't a smart player avoid this issue altogether by not having vehicles that accelerate slowly share platforms?
Premises...
Ok, I am on board with this. You seem careful about stating why you came up with these numbers (5%) and I agree here. Note by the way that this game isn't making it easy to debate this. The game states frequency in seconds or minutes; what it is talking about in fact is roundtrip time. What one ought to want is a low roundtrip time anf/or a high frequency, because a high frequency actually means, more vehicles pass by in a certain amount of time.
Which is what our disagreement is about. I am trying to tell you that this is not how a computer thinks. Point 3 may have some merit in an intuitive sense, but it is prone to failure. Auto-spacing as it is, is stupid but robust.
When you say "only allowed to wait' you are talking about the auto-spacing feature, right? But I am trying to tell you that signals work as an auto-spacing tool anyway because trains wait before signals anyway. A signal will turn red if another train occupies a piece of track in front of it up to the next signal or station. Because of this, if signals divide the track into equal pieces of length of a number equal to the number of trains on a track, you got your improved auto-spacing tool righ there!
I am combining point 5 and 7 here. There is no problem here if as you say, the vehicles accelerate quickly anyway, the lines will just get out of phase. As for trains, let's talk premises then.
It remains the job of the player to do something usefull with the given tools ofcourse, but the tools currently in the game are not sufficiënt, even for smart players, to create a good working network.
Most of my idea's are to improve the autospacing tool itself, without the player having to give any input. (point 1, 2, 3, 4 and 6)
Point 5 is to give better information to the player
Point 7 is to give a simple tool to the player because it would be very hard for the game to do this on it's own.
Maybe I should add an extra point 8 (another simple tool for the player):
add some amount of time to reach a given roundtrip time, warn the player if a synced vehicle goes more than ...sec above the given time.
f.e. a line with a roundtrip time of 3 minutes 53 seconds, can be rounded to 4 minutes.
If the roundtrip time would exceed 4 minutes and 5sec(?), the player should receive a warning.
And now that I'm busy creating more points ;)
point 9:
Show the player the roundtrip time for a line seperate of the frequency (that would make talking about it also less confusing ;) )
Well, point 4 isn't that important anyway and you are right it wouldn't make a big difference on most lines. But there is a scenario where it might be usefull:
Two lines A and B, both with double track.
A goes station 1 <> station 2 (platform 1)
B goes station 3 <> station 2 (platform 2)
they arrive at station 2 from the same side and share a decent amount of track (f.e. 1000m).
the track from station 3 to the merging point is shorter than the shared track.
Normally line A and B should arrive alternating, but when f.e. line A is build and you just start line B, they are not in phase yet. Currently the new trains for line B can only wait in station 3 and before they merge with line A.
However in point 6, the depot is added as a waiting spot, so line B can increase the time between the trains in two stops. (so there is already an improvement)
By adding signals as waiting spots, there is one place where it can make a difference and that is the signal before the merger of both lines.
The reason is that the signals on the shared track need to be closer together because you need more trains there. If then the distance between station 3 and the merging point is too short, you can't space out the trains with signals. Together with the trains being not spaced out yet, this means that it is likely that you get two trains from line B after eachother on the shared track, which in turn will hinder line A (bcause the second train from B will wait in front of the station preventing a train from line A to enter the station). The extra signal increases the waiting spots from 2 to 3, reducing the likelyhood of that problem occuring.
Thanks for clarifying your point of view. I understand now, so lets wrap this one up then.
And most of my comments on these ideas caution against these because they cause problems elsewhere. Unfortunately you haven't adressed these.
And why not. What is the harm here.
If you adjust the frequency without buying new vehicles, you can only decrease the frequency. And you can only do this by reducing the average speed of vehicles, which, as I explained before, cause more problems.
If you adjust the frequency by buying vehicles AND adjusting the frequency, you might have to go through several cycles of buying vehicles and decreasing their average speed before you hit the sweet spot, but the speed of your vehicles will suffer more.
Frequency isn't some number that you can dial down without it affecting something else. It is a combination of two factors in particular; the average velocity (not the maximum speed) and the distance covered. This is the problem I have with this one in particular; it don't consider the system as a whole, and/or could be overkill for a small problem.
This idea I like very much, particularly if you can set it as a passive warning. You open the line manager and see a red light or something next to the line to indicate its frequency or average vehicle velocity is fluctuating or falling. Or maybe having it pop up after all is a good idea regardless. I needed this idea in my game today, the Cobra-trams were getting stuck on the tramstations.
That is the weird part about Train Fever. They state the frequency in seconds or minutes. What they state in fact is the roundtrip time. To us, intuitively, a low roundtrip time is good. So we can easly pick up this information and put it to use. More vehicles = lower triptime. Train Fever would be a lot harder if they actually stated the frequency:
A roundtrip time of 100 seconds for a four-vehicle line would amount to the following frequency; 1 divided by 100 seconds times 4 = 0,04 Hertz. That's not an easy number to work with. UG did well by just stating the roundtrip time instead of the frequency.
I stopped reading here because I think you're on to something, but I can'tt quite follow you. I won't ask you to explain your idea further. I am on the fence about this suggestion; one part of me says players should suck it up and separate-double-track everything; on the other hand, I like a challenge too. Soooo, ok, why not.
When I talked about roundtrip time, I meant the time a vehicle needs to complete a lines full order list. So the frequency (game meaning) becomes the same as the roundtrip time if there is only one vehicle on a line.
If you re-read my comments now, maybe it all becomes clear? ;)
What problems do you foresee by changing the autospacing? I know point 3 might be a problem to code, but I think the others: 1, 2, 4 and 6 should work, not?
problems I foresee:
1) if all stops for a line are shared with other lines, there is indeed a problem. Two fixes:
- inform the player that vehicles only space out if they have a spot they can wait (maybe even show a message to the player when a line doesn't have a waiting spot for vehicles)
- leave a minimum time (maybe 5 sec?) that vehicles are always allowed to wait
2) I don't see any possible problems if you include a min and max waiting time. (mine could be the previously mentioned 5 sec and max could be 60 sec?)
3) This depends on the possibility to have the game know which vehicle is 2e, 3e, etc...
4) as long as point 1 is implemented, I don't see a gameplay problem here. Maybe let the train show:"waiting for autospacing" or something so that the player knows why that train is standing still.
6) the same as in point 4, let the vehicles show:"waiting for autospacing" so that the player knows why the vehicles aren't leaving the depot.