The International Simutrans Forum

 

Author Topic: Questions for players: inefficient rail and 'bus networks  (Read 1844 times)

0 Members and 1 Guest are viewing this topic.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Questions for players: inefficient rail and 'bus networks
« on: September 03, 2018, 11:54:31 PM »
When undertaking some tests recently into the reasons behind the relatively low proportion of passengers transported by player trains/'buses/ships/aircraft on the Bridgewater-Brunel server game, one thing that struck me when looking at datasets in the debugger was the very low average speed of a high proportion of medium and medium/long distance journeys, with average door to door as the crow flies speeds of often not much more than 35km/h.

Looking at the structure of player networks in the Bridgewater-Brunel game, this appears to be likely to be caused by two things: (1) relatively inefficient rail networks; and (2) a relative lack of intensive 'bus/tram networks in larger towns.

What I am interested to find out is whether this is because there is some in-game economic incentive of which I am not aware for doing things this way, whether there is some practical in-game difficulty in creating more efficient networks, or whether the inefficient networks are as a result of a lack of understanding of the importance of creating or how to create more efficient networks (in which case, I will need to consider how best to communicate this).

In terms of the nature of the inefficiencies, I deal with each in turn. I should note that I am looking at the current Bridgewater-Brunel saved game: my initial explorations were undertaken in the in-game year 1923, and I am currently looking at a saved game from 1928 for data.

Rail networks

The main cause of inefficiency in the rail networks appears to be the absence of limited stop express trains. The paradigm example of this is the western mainline, operated by Far Eastern & Western Railways. Its southern terminus is in Bugleigh, the Bugleigh Ves station, which serves trains travelling north along the western mainline, as well as suburban trains travelling west to Manly St. Martin and beyond, and inner suburban underground trains. The station is actually owned by Bay Transport, which runs many of the suburban services.

Its northern terminus is at Polditch, trains on the western mainline terminating at the somewhat monumental Polditch Ves Memorial Port Junction station, which has a passenger ferry port and links to the Wimham/Flaxdale branch line and the north-western secondary lines. Along that route, I have counted a total of 28 stations, including 5 inner suburban stations on the four track Bugleigh metropolitan section.

The diagram serving this line is the POD / J 21X (Polditch - Bugleigh) service run by Far Eastern & Western. This shows a total of 15 non-waypoint stops for this diagram between Polditch and Bugleigh. That is, out of the 23 stations that are not purely inner suburban stations in the Bugleigh metropolitan area, this diagram skips only 8 of them. The greatest distance between any two stops on the route is the 33.6km between Budpike and Great Trunkvale; the smallest distance is the 4.6km between Willingpool Herring and Stoney Forney. The average speed of a randomly selected train on that line (convoy no. 6608) is just 58km/h. These 8 carriage trains hauled by the powerful LMS "Royal Scot" class are capable of travelling up to 155km/h, but, being steam trains, they take a considerable time to accelerate to full speed. The trains seem to spend much of their time travelling at relatively low speeds (~70-80km/h), although I note that, in the saved game from which I am working, a train just approaching Willingpool Herring (not from the direction of Stoney Forney) is travelling at 97km/h.

Another issue with the speeds on this line is the somewhat indifferent quality of the civil engineering. The main line is recorded as having been built in the 1840s (I think originally by Ves before that company went bankrupt in the mid 19th century and some of its main lines were taken over) and has numerous sharp curves of the sort that were rather less of an issue in 1845 than they are in 1928: in my saved game, one particular train in the woods at around 1087,1899 appears to have had to slow to 46km/h to take the 90 degree turn in just 5 tiles in this location. A similarly slow part of the network is just north of Bracklesingham, where the bridge over the stream is on a very sharp corner. In reality, streams this small can simply be put into a culvert (i.e. deleted), so it is not necessary to build a bridge over them at all. This section of line seems to have been built in 1846.

The result of all this is that the journey to Polditch, a distance of just over 182km, takes 3 hours and 50 minutes to complete: an average speed of just 47km/h. To put that in perspective, that is only 15km/h faster than the top speed of a tram in this era.

With some careful timetabling and an increase in the number of stations that have more than two tracks, a train that makes perhaps 3-4 stops rather than 15 could make much better time, even without improving the civil engineering, as the train would be able to remain at a higher speed for much longer rather than spending most of its time slowly accelerating from a standstill.

The Bugleigh - Polditch line is the one that I have analysed in detail, but it is by no means unique: the Olderpool to Selford service operated by Crandon and Lakes Railway, for example, also only has services that call at every station, with the result that a 108km journey takes 2 hours and 17 minutes, giving rise to the same 47km/h average speed as the Bugleigh-Polditch line. Again, the main line trains on this line, even though not quite as impressive in their traction as those on the Bugleigh - Polditch line, can nonetheless manage a respectable 128km/h if given the chance to accelerate over a long distance.

I know that the difference in size between the different towns is far less than it should be: this is planned to be addressed when I redesign town growth. However, there are still clearly some towns that are more important than others (clusters of towns forming a metropolitan area (e.g. Selford) as opposed to isolated towns (e.g. Hadnell), for instance), and it should be possible in principle to deduce a sensible service pattern.

'Bus and tram networks
These vary somewhat, and are better in some towns than others. However, all towns seem to have a much more sparse network than one might find in reality.

One common pattern of 'bus transport is a single 'bus line from the main railway station heading into the centre of town in a straight line and back again (see, for example, Staffwick or Netstoke). Many other towns have two such lines in a V shape (e.g. Inglewick or Bumbleingworth). Ther eare many variations on this theme, but what many local transport networks seem to have in common is a propensity to do nothing other than serve the main railway station. Often, for many passengers, a long walk is needed to the 'bus stop. Take Newingate, for example: passengers in the houses at 819,1183  have to walk 37 minutes to get to the nearest 'bus stop, Newingate Green, wait 9 minutes for a 'bus, which takes 11 minutes to get to the railway station, on the north-west main line to Polditch, meaning that they will have spent nearly an hour travelling to the station before they even start on their rail journey. For passengers with a 3 hour journey time tolerance,  by the time that they have waited 11 minutes for the train to Polditch and reached that destination, they will have little over an hour left to complete their journey, including the 'bus ride and walk at the other end.

In reality, most towns had far more than one or two 'bus routes and a very high density network of 'bus stops; people will, in reality, rarely spend half an hour walking just to catch the 'bus.

Adding 'bus networks is perhaps less glamorous than main line railways, but adding efficient networks can make a real difference to passenger numbers (and will hopefully be able to make a real difference to town growth when that change has been implemented, albeit that is a while off yet).



I write the above not to be critical, but rather to try to see whether I can understand whether these inefficient networks are a result of a perverse incentive somewhere which I need to fix, or whether a better solution would be to educate people as to the value of efficient networks.

One pattern that I do suspect that I can see in these data is that many of the network choices which are time inefficient for passengers are very close to some of the network design styles that are optimum for an environment in which journey time does not matter, but whether it is possible for passengers to get from one place to another at all does matter - i.e., are closer to optimum strategies for Simutrans-Standard than Simutrans-Extended.

Are people able to tell me how much that they are thinking about journey times when they are planing their networks, as opposed to thinking about whether people can get from one place to another at all? If you are thinking more about the latter, might you design your networks differently if you realised that passengers' choice as to which destination to go to is heavily influenced by how long that it takes passengers to get there such that passenger ridership on a route can be increased as much by reducing journey times as by serving additional destinations?

Offline DrSuperGood

  • Dev Team
  • Devotee
  • *
  • Posts: 2727
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #1 on: September 04, 2018, 02:34:38 AM »
Currently a big problem is that one really cannot build stuff underground. One can, but it is really painful and slow to do so. As such few underground railways are being built even in areas that it would make sense for them. Obviously this will not be a problem once the standard fix for fast underground changes is merged in.

Busses are also quite slow at 20km/h. Only in the last in game year odd have faster busses been added one can expect generally faster bus routes.

I think the main problem is that people are just expanding rather than building up. Why spend 2 hours adding bus lines to a few cities when you could add a dozen new cities to your network with quickly constructed (not very straight) rails.

Offline Ves

  • Devotee
  • *
  • Posts: 1677
  • Languages: EN, SV, DK
Re: Questions for players: inefficient rail and 'bus networks
« Reply #2 on: September 04, 2018, 09:10:55 AM »
Yeah, thats my old line  :P
When I built it, it was not so much important how straight the line was, it was more important to not have too many hills and also the forge cost was an issue. I made sure not to have any 90 degrees, so thats someone else.
Also I was inspired by real world lines that (in Sweden at least) sometimes not goes straight at all, but follows the nature, which was a very big incentive to develop tilted trains. My plan was to perhaps straighten out some quirks down as the years passed. My schedules had to some extent different qualities of services, and also some freight running part of it.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #3 on: September 04, 2018, 10:14:40 AM »
Thank you for your replies. The civil engineering issues are to a large extent the expected result of the economic constraints of early railways, although some of the corners on the western mainline are tighter than on equivalent mainlines (e.g. the London & Birmingham Railway), requiring a speed limit of <50km/h: although there are no 90 degree turns in a minimum number of tiles (3), there are 90 degree turns in only a few more tiles than this (5-6), resulting in very slow maximum speeds around these corners - slow even by 1840s standards. Contour railways are definitely a thing, but one would not normally expect quite such slow speeds on a main line. I imagine that straightening the line as time passed as originally planned might well have produced a line capable of higher speed running.

I do not think that difficulties in building underground is really relevant to this issue, as, in reality, building underground was generally only done where absolutely unavoidable owing to the cost. People have in fact built substantial tunnels and underground urban railway networks. Building underground is not necessary for the purpose of having a relatively smooth, flat mainline, nor for having the infrastructure necessary for limited stop trains.

I note that neither of the replies addressed the relative absence of limited stop expresses - may I ask what the reason is that players are not using these, given how long that it takes steam trains to accelerate? Had you just not thought of it (or realised the importance of this to passenger numbers on your network), or had you thought of doing this but rejected it? If the latter, it would be very useful to know the reasons.

As to 'buses, incidentally, the speeds are entirely realistic: until the later 1920s, trams are far more capable, but require fixed infrastructure.

In relation to expanding, I can see the incentive to expand rather than to improve in the early days, but people have not really been expanding very much recently. Is that because there are not a great many players (perhaps related to the high computational demand making the game slow on many people's computers) - or is it for economic reasons?

Offline Ves

  • Devotee
  • *
  • Posts: 1677
  • Languages: EN, SV, DK
Re: Questions for players: inefficient rail and 'bus networks
« Reply #4 on: September 04, 2018, 12:05:32 PM »
I did have a little mix of slow and direct lines, although not so much on that particular one, but on my other big line going from polditch if I dont misremember. Also I had on other of my main lines in the "Center Island" (cant remember the city names currently). It worked ok as to what I remember, it was just a little bit of a hazzle to make them depart and overtake when I wanted, but that was also a little bit due to the signals that where not behaving completely as they should leading to the ultimate bankrucy...
I did an experiment with some "luxuary trains", where I made only a few trains run direct between hubs on most of my lines with no intermediate stops on the highest class, but I had troubble filling them up to make them profitable. However the mail pickup on those where huge (priority mail), as most mail wanted to go via those direct lines.

Offline DrSuperGood

  • Dev Team
  • Devotee
  • *
  • Posts: 2727
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #5 on: September 04, 2018, 12:41:59 PM »
Quote
I note that neither of the replies addressed the relative absence of limited stop expresses - may I ask what the reason is that players are not using these, given how long that it takes steam trains to accelerate? Had you just not thought of it (or realised the importance of this to passenger numbers on your network), or had you thought of doing this but rejected it? If the latter, it would be very useful to know the reasons.
No idea how one would create such a thing quickly... I struggle getting my lines to run on schedule, with the process taking months to find the right frequency for number of trains.

Quote
In relation to expanding, I can see the incentive to expand rather than to improve in the early days, but people have not really been expanding very much recently. Is that because there are not a great many players (perhaps related to the high computational demand making the game slow on many people's computers) - or is it for economic reasons?
Because keeping my infrastructure working at a profit replacing the 50-100 year old vehicles on it is taking all my playing time.

I cannot even use replace on my trains, a mistake I learnt the hard way yesterday. I have to withdrawal them all and then start new ones as otherwise the line will likely get stuck due to a train trying to enter the depot as one tries to leave.

Offline Junna

  • Devotee
  • *
  • Posts: 1082
Re: Questions for players: inefficient rail and 'bus networks
« Reply #6 on: September 05, 2018, 05:12:14 AM »
I have slowly been working to realign the old Ves line Bugleigh-Pendale and also the main line from Selditch, but it's crashed and resetted my work several times... I do run one limited stop Bugleigh-Polditch but, due to the suburban services and the three sectioned local services, with the old signalling it was not efficient to run more trains, and I would have to rework the stations with more passing possibilities or it would just be pointless to add more, and it's a bother to work realigning lines with services running, so it takes a while... Also, one reason I don't run more limited long-distance services is because if I do, too many will simply use these rather than the local and you would end up with queues of passengers waiting for only long-distance trains...

I considered for a while a number of direct express services from Selford/Amditch to points East, including limited stop to Roxingstoke, but I feared that adding them would simply result in massive queues of passengers requiring a ridiculously frequent service to keep up with, quickly causing congestion in the lines. Generally I made most of the U-and-V shaped passenger networks, I've just used this same set-up since Experimental version 5 or 6 when I first played it (I should note I did never play Standard before Experimental, but after, though I originally am familiar with OpenTTD, though I always played that with a more 'realistic' style as opposed to the type who design networks that look like rail motorways with cloverleaf rail interchanges).
« Last Edit: September 05, 2018, 05:27:25 AM by Junna »

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #7 on: September 05, 2018, 09:19:41 PM »
Thank you for your replies: this is helpful.

A few further queries, if I may. Dr. Supergood - you write that replacing the older vehicles is taking all of your time. Presumably, this is not a reason to expand rather than to improve your existing network, but rather a reason to do neither? Incidentally, I am hoping that the forthcoming schedule and maintenance features will address the long-standing problem of congestion and service disruption when vehicles are replaced, along with other replacer related problems.

In relation to it being difficult to set up slow and fast train paths on the same line, this is an interesting point that I had not considered. I am hoping to resume work soon on the schedule features (mostly relating to dividing/combining/shunting), and I wonder whether any of those (conditional depart, perhaps?) might assist in making it easier to set up fast and slow train paths on a double track main line with passing points at stations. May I ask what the principal difficulty in doing this is? It may be worthwhile seeing whether any other schedule features might assist in addressing this since schedule features are a current focus.

Junna - I am curious as to your response about too many people using the long distance services. Why is this a problem - surely more passengers means more profit? If demand exceeds supply, you can either increase supply (by having longer trains - it was not uncommon to have express trains of up to 16 carriages by the 1930s; I notice that many trains on the Bridgewater-Brunel server seem to be no more than about 8 carriages long) or increase prices (by uprating the classes). I know that the current locomotive hauled carriages lack first class provision, which I plan to address eventually, but you can still, for example, use "medium" rather than "low" as the base class for your long distance services if demand really does exceed supply even if you lengthen your trains. In reality, express trains of this period were much, much longer than local trains (the latter of which tended to be no more than 3-6 carriages long for the most part, express trains tending to be 7-16 carriages long).

In any event, thank you all for your feedback so far.

Offline Rollmaterial fi

  • Devotee
  • *
  • Posts: 572
  • Languages: EN, FR, DE, FI, SE
Re: Questions for players: inefficient rail and 'bus networks
« Reply #8 on: September 05, 2018, 09:37:38 PM »
One feature that would assist with scheduling overtakings would be allowing convoys to depart late by only making them wait for the next departure slot once a certain specified time has elapsed from the previous one. That would considerably reduce the risk of obstruction by convoys arriving shortly after a slot and taking up the platform for a long period of time.

Offline Junna

  • Devotee
  • *
  • Posts: 1082
Re: Questions for players: inefficient rail and 'bus networks
« Reply #9 on: September 05, 2018, 09:45:32 PM »
Some kind of priority signalling method or wait until overtaken would be good-- in standard, I used to set it to make short-time waiting (using the 1/64 or so time spacing) at some stations to allow overtaking by non-stopping services, but using this in Experimental after the timetabling changes, just ends with the train once again trying to space out to the original, unless I bother with complicated things that usually doesn't work as intended anyway, which makes it more of a bother than is worth.

Mostly the number of trains would just seem silly-- would there be 40 trips daily on any service but a most generous commuter operation? And the reason I don't set anything but very low (I only set medium if there are some other compartment classes) is because if the entire train is set to low, medium or so on, the number of passengers simply is so small it makes little to no money, and most annoying of all, if you do set the passenger class to only medium, you just get passengers queuing at stations with no way of getting elsewhere. I experimented some months ago with running services of only one class on the line but the passengers do not make a choice depending on the availability of accommodation (or at least they did not at the time, I don't know if it has changed since), so they queue for services they do not afford, so I just set everything very low unless it has secondary accommodation carriages.

The reason for the lack of length on the trains is that it is too much work to lengthen platforms and realign track and station approaches and points extensively, and even if one were to only lengthened stations called at by express trains, the work would still be extensive, and the cost of double-heading the expresses to get any acceptable acceleration with so many carriages--and reaching only a fraction of the maximum speed, just makes it not quite worth it either, and it adds quite a lot to the transfer times when the stations are long. I generally rarely lengthen any stations beyond 7 tiles, and try to make most 6 initially. Sometimes things may warrant 8 or 9 tiles, but the terribly slow acceleration of such a long train if not double-headed makes it kind of a cause of congestion on a line with many services running.

Another issue right now is that I have to manually replace endless city road sections not yet upgraded if I want to replace the now expired earlier buses, which is an enormously time consuming and difficult task, compounded by the unresponsive controls on the server game that makes road selections delayed. I thus cannot automatically replace all old road vehicles, but have to look at each city first, and fix all the roads, before they can be replaced, because so many roads remain cobblestone.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #10 on: September 05, 2018, 10:51:02 PM »
Rollmaterial - I am having some difficulty following your suggestion, I am afraid. Can you elaborate at all on precisely what you imagine would happen in the case of departure slots?

Junna - thank you for the reply. I will have to consider whether and how the conditional depart feature could be used (or be modified to be able to be used) as a "wait until overtaken" feature.

In relation to the number and length of trains, the two are, of course, closely linked. Because we have no day/night cycle, it is more sensible to think of trains per hour rather than trains per day. 40 trains per day would equate to a little under 2 train per hour. 1 train per hour was actually not uncommon during this time period: the London to Southampton/Bournemouth service, for example, was approximately hourly in 1938, and ran with up to 12 carriages (depending on demand). If your trains are 8 carriages long running half hourly, then this is not far off the capacity of a 12 carriage train running hourly, which is reasonable for the inter-war period. I imagine that London to Birmingham might have been even more frequent.

May I ask what in particular makes extending platforms difficult?

In relation to acceleration and top speed, the game code is designed to give realistic performance for any given loading - are you aware of any reason to believe that performance of any given locomotive with any given set of carriages is below a realistic performance for that combination? In this period in reality, long trains of up to 16 carriages pulled by a single steam locomotive was not uncommon (12 was more common than 16), and these trains could reach a significant maximum speed, albeit took a long time to get there (which is why they stopped so infrequently). Generally, one would time the local service to depart just after the express service so that the express passengers can transfer to the local train and continue their journey and so that the local train would not get in the way of the express. Even with the slow acceleration with a 12 carriage load, the local train would, stopping at many stations, not conflict with the path of the express.

In relation to passenger numbers and classes, what sort of patronage figures are you seeing for services set entirely to low rather than very low? Had you considered a combination of faster trains and increased prices? Remember, faster services will draw more passengers to your network, making up for the increase in price.

As to very low passengers accumulating at stations for which there is no service, this would, if correctly understood, amount to a bug. If you can find a reliable reproduction case for this, I should be very grateful if you could post a bug report thread to deal with this.



Edit: In relation to the cobblestone roads, I have just increased the weight limit for these, as, being a hard surface, they should in principle be able to accommodate a reasonably high axle load, even though the uneven surface would require a lower maximum speed. I wonder whether I should also increase the wear capacity of these roads to above those of tarmacadam roads - does anyone have any data on this?
In any event, the increase in the maximum weight of the cobblestone roads should eliminate the need manually to replace these in the 1920s/1930s just to be able to run the latest 'buses: this was not done in reality, and many roads remained cobbled right into the 1960s.
« Last Edit: September 05, 2018, 11:12:01 PM by jamespetts »

Offline ACarlotti

  • *
  • Posts: 483
Re: Questions for players: inefficient rail and 'bus networks
« Reply #11 on: September 05, 2018, 11:55:02 PM »
f you do set the passenger class to only medium, you just get passengers queuing at stations with no way of getting elsewhere.
As James mentioned (while I was writing this post) this would be a bug. There was some discussion of either this bug or a similar bug back in about February, and I think at the time I suggested I would look into the routing system to investigate at some point. That is still my intention, but my priority at the moment is merging changes from Standard, as I think that is a more productive and/or higher-priority use of my time.

Mostly the number of trains would just seem silly-- would there be 40 trips daily on any service but a most generous commuter operation?
Would this level of service still be profitable? If so (and you're using sufficiently long trains), then it might suggest that the pakset and/or code might not be properly balanced. If not, then running the services unprofitably is unrealistic.
However, I would note that 40 trips daily is roughly the service level that exists on many limited stop intercity routes in the UK.


Based on the above discussion it seems that a lot of the issues arise from player effort being insufficient to manage the level of public transport that would be expected in a similar size/density map in real life. Part of that can probably improve by adding more features to facilitate management and by improving the stability of existing features. A lot of effort seems to have been lost due to short-term bugs or loss of the server saves. However, the main things that I think could be done to overcome this defecit are:
1) Reduce the map size. I think you already suggested doing this, and it would have the added benefit of reducing the memory/computational requirements.
2) Adjust the speed at which the game runs. That is, make months take twice as long as they do at present, so that one month would last approximately 13 in-game hours (if you're currently using the default month length) and players can spend more time developing/managing their networks stably due to a lower rate of external change.
(Disclaimer: I have never played/investigated the server game due to the limitations of my current computer, so these observations are almost entirely second-hand.)

Offline Rollmaterial fi

  • Devotee
  • *
  • Posts: 572
  • Languages: EN, FR, DE, FI, SE
Re: Questions for players: inefficient rail and 'bus networks
« Reply #12 on: September 06, 2018, 12:13:40 AM »
Currently, if a convoy arrives to a stop with "Wait for time" enabled right after a departure slot, it will always wait for the next one. This is fine for the ends of the line where longer dwelling times or several convoys waiting at the same time are less of an issue, but poses some challenges on intermediate stops where a late convoy would have its journey time from a previous stop to a following one considerably lengthened and would possibly block the line for other convoys. What I am proposing is the ability to specify a time after a departure slot - either as a percentage of the departure interval or simply as a time shift - before which an arriving convoy would depart immediately instead of waiting for the next slot. Here is an example:

The month length is the default 6:24:00. A line has a stop with "Wait for time" enabled with 6 departures per month (departure interval 1:04:00) as well as a new "Do not wait until" setting set to 60% or 0:38:24. If a convoy arrives to that stop less than 0:38:24 after a departure slot, that convoy will depart immediately. An arriving convoy will only wait for the next slot if that time has elapsed since the last slot.

Offline DrSuperGood

  • Dev Team
  • Devotee
  • *
  • Posts: 2727
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #13 on: September 06, 2018, 12:44:48 AM »
Quote
A few further queries, if I may. Dr. Supergood - you write that replacing the older vehicles is taking all of your time. Presumably, this is not a reason to expand rather than to improve your existing network, but rather a reason to do neither? Incidentally, I am hoping that the forthcoming schedule and maintenance features will address the long-standing problem of congestion and service disruption when vehicles are replaced, along with other replacer related problems.
The problem is that depots have no signals built into them so cause OpenTTD style collisions.

For example at a station place a depot behind a choose signal that faces the platforms. The intention is that trains that are being replaced will unload at the station and then go into the depot and be replaced. The choose signal allows them out to go to the platforms if they were replaced before (eg empty at last slot). The reality is that a train will go to be replaced, another will go to be replaced and the previous now replaced train will depart the depot and run into the second replacing train blocking up the track to/from the depot and eventually the line due to it blocking up the intersection. Block signals are being used so one would imagine it reserving to the depot before allowing the train to go to it however that does not seem to stop a train from departing the depot itself and blocking it. One has to manually intervene and order the previously replaced train back to depot so the other train can come in and be replaced.

Road vehicles do not have this problem due to road mechanics being bi-directional, but instead suffer from the "sell immediately not working" bug. So for example moving from 20 horse drawn coaches to 4-6 busses requires manually selecting 14 coaches and ordering them to be withdrawn otherwise the coaches will sit in a depot costing you money unnecessarily until manually sold.
Quote
In relation to it being difficult to set up slow and fast train paths on the same line, this is an interesting point that I had not considered. I am hoping to resume work soon on the schedule features (mostly relating to dividing/combining/shunting), and I wonder whether any of those (conditional depart, perhaps?) might assist in making it easier to set up fast and slow train paths on a double track main line with passing points at stations. May I ask what the principal difficulty in doing this is? It may be worthwhile seeing whether any other schedule features might assist in addressing this since schedule features are a current focus.
Honestly I have no idea how I would set such a line up without spending hours of time fine tuning it and a least 2-3 months staring at it to make sure everything works in sync. Then every time rolling stock is changed repeat that to get timing right.

In real life this is exactly what happens, but that is ok because in real life peoples jobs consist entirely of doing this. Now days I am sure there are dedicated software packages to aid with this which likely are made of more than several dozen or hundred times the programming code than Simutrans Extended is made out of. Even if creating the schedule for a single line takes 1-2 man weeks it is fine. However in a game where you are one person trying to run 1 company you just do not have the time to do all of that. By the time you finish tuning a line it will already require vehicle replacement...
Quote
(by having longer trains - it was not uncommon to have express trains of up to 16 carriages by the 1930s; I notice that many trains on the Bridgewater-Brunel server seem to be no more than about 8 carriages long)
You can fit over 4,000 people into a 2 unit (6 coach), 4 tile "LMS suburban EMU" with top speed 110 km/h before the 1930s. How 4,000 people in real life can fit in that I have no idea and I just hope everyone is alright. In retrospect this is a bug as the assembly is meant to have a maximum capacity of only 800.

The problem is that having a 1-2km long platform is itself not very realistic. Where as with old coaches it was easy to have 12-16 or more in a 4 tile long stop, with the 1900s ones you can get between 1 and 2 per tile as they are so massive. They also weigh a lot more than the older coaches so the per ton of coach carrying capacity is a lot worse.
« Last Edit: September 06, 2018, 01:49:02 AM by DrSuperGood »

Offline Junna

  • Devotee
  • *
  • Posts: 1082
Re: Questions for players: inefficient rail and 'bus networks
« Reply #14 on: September 06, 2018, 12:58:16 AM »

In relation to passenger numbers and classes, what sort of patronage figures are you seeing for services set entirely to low rather than very low? Had you considered a combination of faster trains and increased prices? Remember, faster services will draw more passengers to your network, making up for the increase in price.

Something like going from 40-70% loading of capacity filled to under 10-20%.

40 trains per day would equate to a little under 2 train per hour. 1 train per hour was actually not uncommon during this time period: the London to Southampton/Bournemouth service, for example, was approximately hourly in 1938, and ran with up to 12 carriages (depending on demand). If your trains are 8 carriages long running half hourly, then this is not far off the capacity of a 12 carriage train running hourly, which is reasonable for the inter-war period. I imagine that London to Birmingham might have been even more frequent. 
When I say "day" I mean the in-game month, which I consider roughly to be a day or so. Out of Roxingstoke I currently run something like 90 trains per in-game month, but most lines are 7, 9, 10 or 12 trips per in-game month in frequency.
Quote
I do not see the point of longer trains. You can fit over 4,000 people into a 4 unit (8 coach), 4 tile EMU with top speed 110 km/h before the 1930s. How 4,000 people in real life can fit in that I have no idea and I just hope everyone is alright.
I haven't noticed that myself, if that happens it seems that the overcrowding not respecting the overcrowded capacity bug that used to occur still occurs. I noticed it frequently when there were stage coaches, because sometimes stagecoaches would randomly get stuck, and I found it was because they had so many people in them that their weight exceeded the limit (in one case, 241 people in one 4( 8) limited one. A commuter EMU could well handle an overcrowding of 100-250% but, 4,000 in a eight-carriage set is obviously bugged, and I don't remember the actual numbers of the LMS EMU being wrong.

Offline DrSuperGood

  • Dev Team
  • Devotee
  • *
  • Posts: 2727
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #15 on: September 06, 2018, 01:45:57 AM »
Quote
I haven't noticed that myself, if that happens it seems that the overcrowding not respecting the overcrowded capacity bug that used to occur still occurs. I noticed it frequently when there were stage coaches, because sometimes stagecoaches would randomly get stuck, and I found it was because they had so many people in them that their weight exceeded the limit (in one case, 241 people in one 4(  limited one. A commuter EMU could well handle an overcrowding of 100-250% but, 4,000 in a eight-carriage set is obviously bugged, and I don't remember the actual numbers of the LMS EMU being wrong.
Yeh must be a bug.

It was actually a 6 unit "LMS suburban EMU" (2 sets). I have fixed this in the original post. The assembled thing is listed as having a capacity of 560 + 240 overcrowded. The break down is 80 high, 480 low and 240 low overcrowded. I changed it to 80 medium and 480 very low. Due to the above mentioned problem with replacing trains, my line got stuck for several months accumulating several thousand passengers at every stop. The EMUs made short work of it moving over 4,000 passengers at a time, which clearly is more than their technical maximum of 800.

Offline SuperTimo gb

  • *
  • Posts: 62
  • Languages: English, French
Re: Questions for players: inefficient rail and 'bus networks
« Reply #16 on: September 10, 2018, 06:39:54 PM »
I attempted to make a route with fast and slow services on the server but it turned out to be a massive waste of time and money. I had to build additional platforms for the slow services, which required building elevated track and replacing small signal boxes with larger ones. I also bought new rolling stock with express tender locos rather than the tank engines I was using for slow services.

I scheduled the fast trains to run from Belley to Strawburgh and then on to Green Quantinglow non-stop, whilst the slow services were to travel between Strawburgh and Quantinglow calling at St. Luke Lydock and Wime Downs. The logic behind this was that Belley - Strawburgh is much further than the other legs so it made sense to have faster more powerful locos. The journey time was approx 1 hour as opposed to 1 1/2 hours calling at all stops. Unfortunately all this has halved my income, as well as giving me a lot of additional debt. The new higher frequency in the local area doesn't seem to drum up any additional traffic and the express trains cost more to run then the tank engines whilst not generating any more money. An issue with the signalling at Strawburgh caused one of the express trains to get stuck for no discernible reason at Quantinglow and I wasn't able to fix this for many in-game months. (The train's schedule wanted it to go to a platform I had made one way and wouldn't use the adjacent platform that was accessible. Neither would it do what trains normally do in that situation and take a ridiculous route to the terminus before turning 180 and travelling to the specific platform.)

So I have deleted the additional platforms to save on maintenance and put all the trains on the same schedule. I would replace the express locos with tank locos and remove their dinning cars but due to the massive loss of income from my attempt to incorporate faster services I am likely to be unable to build any infrastructure or new rolling stock for a long time.

I think it may be possible to get a system to work with different speed services but I think the level of work necessary to do this, and the amount of in game capital required, will be grossly disproportionate to any benefit. (I would like to do it to have a realistic system, even if it lost money but I cannot afford that at the moment).
« Last Edit: September 10, 2018, 06:50:58 PM by SuperTimo »

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #17 on: September 10, 2018, 08:18:50 PM »
Thank you for your replies. Please remember to file bug reports in their own threads with a saved game in which the issue can reliably be reproduced at a specific time and place, or else I will have virtually no chance of being able to fix them.

Bug reports aside, Super Timo - can you elaborate on why your income fell when you added express trains? Also, may I ask why you had to add extra platforms rather than just using platformless bypass lines for the express trains (as was common in reality)?

May I ask whether you noticed any difference in patronage when you reduced the journey time by a third? I am trying to understand precisely how the incentives in Simutrans-Extended to build express trains differ from those in reality such that I can make the former more like the latter.

Offline SuperTimo gb

  • *
  • Posts: 62
  • Languages: English, French
Re: Questions for players: inefficient rail and 'bus networks
« Reply #18 on: September 10, 2018, 10:05:18 PM »

Bug reports aside, Super Timo - can you elaborate on why your income fell when you added express trains? Also, may I ask why you had to add extra platforms rather than just using platformless bypass lines for the express trains (as was common in reality)?

May I ask whether you noticed any difference in patronage when you reduced the journey time by a third?

I added the extra platforms as I wanted to maintain flexibility, the cost of adding extra platforms was not much compared to the cost of signalling and P-Way, so I could add additional stops to the fast trains or for other future services. At Strawburgh I added two bay platforms to allow the local trains to terminate without blocking the through platforms.

I think despite the journey time being quite a bit faster the lower service frequency counteracted this (I only had enough money to buy 3 express locos, I am trying to use LMS stock but I may switch as I don't think their large locos are particularly good for my route). I think on my route, despite a lot of efforts to reduce corners and gradients, the terrain necessitates fairly tight corners and gradients so the extra speed from the express trains doesn't really help very much. With all the trains stopping my average speed is 63km/h which could be better. I think once I get some more money in I will look at electrifying the line with DC OLE in order to have faster accelerating trains and get away with some steep gradients in places.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #19 on: September 10, 2018, 10:44:15 PM »
I should be interested to know how you get on with that.

Offline DrSuperGood

  • Dev Team
  • Devotee
  • *
  • Posts: 2727
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #20 on: September 11, 2018, 03:32:58 AM »
I did not report the EMU carrying 4,000 people in a separate thread because it is completely unreproducible at the moment. Since their initial deployment I have not seen them violating their capacity anymore. My only guess as to the cause is that it is something to do with replacing an existing engine (a feature that will be removed by the new scheduler) or changing passenger classes while passengers are in them (something that should happen infrequently).

I am currently working on modernizing the central island chain. This includes many new boat lines as well as some pretty silly/spectacular bridges. As express trains are not an option, air travel was going to be used to move the passengers great distance quickly however currently that is not quite working as it should.

Offline Jando

  • *
  • Posts: 504
Re: Questions for players: inefficient rail and 'bus networks
« Reply #21 on: September 30, 2018, 04:45:45 PM »
Interesting thread here. Just noticed it, coming back to Simutrans after a break.

As to express trains: I tried it a few times with passengers on offline maps and never could make it work well. It's a huge headache getting the schedule right so that faster trains don't get stuck behind slower ones - and there's little benefit to it anyway. I think I remember right that Extended passengers don't even care about travel time; instead they care about the sum of waiting time and travel time. And on my typical maps it's by far more efficient to reduce waiting time (easily done by more frequent service) than it is to reduce travel time by introducing express trains. And express trains would have to run as frequently as normal commuter trains, else the express trains would get no passengers because the more frequent commuter line often has a lower sum of waiting time and travel time combined.

As to city transport with bus or trams: My networks tended to be small as well, usually trying to have a max. 20-25 minute walk to the next bus station. I think there's some psychological barrier here as well: takes a lot of time to design a good city transport network, and just like in reality, many of the city lines will run at a loss, thus hard to judge what the in-game benefit is.

Offline waerth

  • *
  • Posts: 145
Re: Questions for players: inefficient rail and 'bus networks
« Reply #22 on: November 11, 2018, 05:00:04 PM »
JamesPetts said:
Quote
Generally, one would time the local service to depart just after the express service so that the express passengers can transfer to the local train and continue their journey and so that the local train would not get in the way of the express.
Hi I am not playing the saved game but I am doing express trains (IC) and stop trains (ST) on my game. In 1837 I have 6 ST routes and 1 IC route. The problem is the IC route crosses multiple ST routes. So although I can plan the start of the IC route to be after the ST route on the same section. Later when crossing onto track driven by other ST routes it is a lottery whether the IC will end up behind or in front. Which is why on some parts I will make dedicated IC track also to make short cuts.

IC 1 route shared with (all my lines start on rail yards so they don't block other trains at stations while waiting)

Rail yard ---- st1 st2 st3 ---- first city ---- st3 ---- other lines coming in ---- st3 st4 st5 st6 ---- second city ---- st 5 st6 ---- branching ---- st6 ---- third city ---- no st dedicated track ---- fourth city ---- Rail yard (the total distance is 400km)

My map is 1300km by 400km approx btw.

Offline jamespetts gb

  • Simutrans-Extended project coordinator
  • Moderator
  • *
  • Posts: 18745
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
Re: Questions for players: inefficient rail and 'bus networks
« Reply #23 on: November 11, 2018, 06:18:54 PM »
This is interesting - when the next major set of features relating to vehicle maintenance and convoy re-combination are added, I will have to look into making these features allow for this sort of timing pattern: there are already planned to be schedule triggers, and this may already suffice for the purpose.

Those features will have to wait until some very difficult critical bugs can be fixed, however.

Offline Junna

  • Devotee
  • *
  • Posts: 1082
Re: Questions for players: inefficient rail and 'bus networks
« Reply #24 on: November 17, 2018, 09:00:31 PM »
What I used to do in standard (and experimental early on) was stop locals at special platforms with a 1/xx wait for load time, to make it dwell longer at the station by just a little bit, to increase chances of it being overtaken if an express was running after it, but this is a bit of a bother to set up in experimental currently from my experience, because using timetabling and waiting easily results in waiting too long; it would be useful if instead, some easier way to just add a little bit of additional dwell time was possible.