The International Simutrans Forum

PakSets and Customization => Pak128.Britain => Topic started by: precor on April 08, 2013, 09:31:19 PM

Title: comments on playing simutrans
Post by: precor on April 08, 2013, 09:31:19 PM
So this is actually my 2nd post, but hey ho, I’ll make it my first in many ways.

First off, my massive congratulations to the thought, effort and work that has gone into creating both Simutrans and the fantastic Pak GB.  I really am immensely impressed at what has resulted, which is a game playable over nearly 300 years of transport evolution, is hugely playable, and which has superb graphics.

I’ve now nearly finished my 1830 start to 2020 and I loved it, my wife loved it quite a lot less though, mainly because I was on it for so long always modifying, extending and so on.

I particularly enjoyed the feeling that many of the challenges I faced were “realistic”, in that I had legacy lines, routes, stations and junctions which were in full operation whenever I had to look to the future for remodelling. I enjoy looking at my network and seeing the roots of my very first hesitant single and double track lines in the octuple & Maglev routes that now ply between parts of the city conurbations that were once <10 building hamlets. I can see in my routing how geography early on led me to various line shapes, that more or less, survive 150 years later. I can see how my first stations have evolved, extended sidewards, upwards and downwards. I can see that very early decision on what stations a line would serve led to how the network evolved for capacity, and how changes in those lines (offering longer through services, or vica versa chopping and pruning) radically affected usage of those and others. All with a very pleasant-to-the-eye changing landscape of buildings that nicely reflected the Britain through those ages.

I found over time that the optimum was separated twin track for every route and train speed (express, suburban, freight – I didn’t use mail as I found the admin just about bearable for the previous), and if I could, I’d probably have separated freight types too. This seemed to produce a) the most profitable lines and b) lines which actually could shift the numbers of pax/freight that were being generated. In hindsight, early on I put too much effort into flying/grade separated junctions which due to the steep up/down gradients of say a flyover bridge, were probably more damaging to flow than a flat junction (as in mid 19th Century timeframe !), and into providing “all directions” options rather than being focussed on providing just what was needed.

Now of course, I do have a list of observations, mainly because by trade I am a design Engineer and my natural focus is to look for weeds in the garden. Please see these just as observations on what are generally minor issues.

So, anyway, less eulogising. The issues in general...

1)   Out of Service Dates. Is the concept that these reflect real world end of build, end of service or are they to be adjusted for playability dates ? There seems to be quite a lot of variation in numbers that I can see in many places have been set so there is some kind of linear replacement option, but other times are quite limited to when something was in real world usage (e.g. APT). Often you find yourself technologically going backwards which doesn’t seem realistic (other than early BR diesels underperformed existing premier express locos which was true).

2)   Usability of some equipment. The use of gearing, and the values set, often strongly disincentivise against using what in reality, were heavily used locos. The modern examples would be the Class 86/87/90, all of which are markedly inferior to the Class 81. Similarly the Class 56 & 58 are very underpowered compared to the older Class 37, and yet are ferociously expensive (Class 60/66 cost but vastly less ability to recoup that although the 66 fails to make a profit on any service I run). Other than a desire for some variation and real world linking, there is no point in using any of these BR locos – which hardly seems fair given the effort to create them in the pak.

3)   Bus service bunching. I found that no amount of tinkering with the delay functionality allowed me to put a consistent spacing between busses, and even that assumed that you could hold each bus at at least one stop for some time (when on a busy route that is quite destructive). No matter how equidistant they were when the service started, fairly soon they were all bunched behind each other, half of them empty yet stops with huge overcrowding. Due to this, and their general unprofitably, I switched to trams for nearly every single local network since the ability to use signals enforces that distribution along the line.

I appreciate “balancing” is a horribly complex (and no doubt tedious) task, and one which every addition of a new train or bus must screw up the efforts to date, and in general, I found the game very playable, and although post 1930 it was impossible for a bus route to be in profit, the vast profits made prior to that, and the profits from the rest of the network (trams being very profitable) didn’t affect my game playing – which was much more about “building lines to transport everything that can and wants to be” rather than an obsession with the bottom line (that being enforced in the modern “SimCity approach” and which frankly, is dull).

So in detail, tweaks that I think would add playability (and which may be easy for me to do myself, although I don’t know how):
•   GNR C1 underpowered, only alternative to LNWR/GER/GWR fast express, in reality, was a success and set pattern for big engines.
•   GWR Bulldog, doesn’t last in service until new GWR Castle, or the intermediate dated Star needs more usable power to match/exceed and fill the gap.
•   LNWR G1, reality stayed around for much longer, leaves big gap in freight locos until LMSs available, yet the earlier (and inferior) A continues.
•   5MT (Black 5) not much better than the G1, but much more expensive, 8F less powerful but more expensive !
•   Need some LNER freight locos (Ks ?). This builds into a theme that post 1900 I tried to “Group” my lines into pseudo companies, e.g. LNWR to LMS, GNR/GER to LNER, LSWR/LBSCR to SR and so on. Thus having a triad of options in both express, suburban and freight to give variety, and of course, easing the replacement burden by spreading it out over time.
•   LNER P2 to stay on until late 50s/60s as offers a high power loco for hillier routes, no replacement.
•   EE1 to stay until EM2 as a more logical pax electric loco replacement.
•   Class 123 seems to have a large gap between 1st and other 3 cars (middle 2 offset to rear, as overlap 4th). A graphics glitch I assume.
•   Running cost jump LUL 23 to 39 stock and subsequent.
•   LUL '67 100kmh, yet tube tunnels limited to 80kmh (also hits '59s ability to make a profit), this affects all subsequent LUL stock in the same way that city road limits destroy the ability to use the higher speed of later busses to pay for their running costs. Fair enough in cities (50kph), but either tube tunnel limits could be raised, or a second tunnel type available say from the mid/late 60s which offers a higher speed.
•   Livestock wagons end '73, what next ? do we stop transporting livestock by rail (is that what happened ?).
•   Blue Pullman last until HST?
•   Also keep EM1/2  longer than reality to give a DC electric option and a sensibly priced Electric loco for freight (if kept in reality, i.e. Woodhead line surviving, I assume they would have lasted like 73s probably). This looks at the “what if element of the game”, allowing deviation from reality. In the game, there is no real electric freight options (I used Class 73s which visually with “Intercity” branding looks a little odd !), until the Class 92 which is very expensive.
•   Get 200kmh HST in '75, but not tunnels capable of this until 1980...
•   Exacerbated with 240kmh APT in ’78, but no track, tunnels or bridges able to support this higher speed until 1980. Thus impossible effectively to use with any hope of profit.
•   Extend Class 50 through to late 80s for secondary duties as was the case in reality. Could be retired coincident with Class 40s or in both cases, retained until Sprinter generation DMUs available.
•   Class 86, 87 & 90 all underpowered, significantly poorer than CL81 !
•   Class 56 (which very expensive, twice a Class 37) & 58 underpowered, worse than a Class 37.
•   Mk1 Suburban ends in 80, yet nothing in 70s/early 80s to replace (wait for Sprinter generation DMUs to turn up?)
•   APT to be kept longer to simulate squadron service, Pendolino a logical follow on (tilt and 24 years later), although only Eurostar in ’94 actually provides something of equivalent/greater speed (perhaps the most relevant replacement since you’ve already built 320kmh infrastructure, which isn’t needed for any other trains – so I waited until this was available to re-equip my APT lines).
•   Class 317 has different entry dates for different parts !, Jan 1981 (front) and Mar 1982 (rest).
•   Ability to make up MUs of different lengths rather than pre-formed 3/4/5 cars (esp 5 car as would need very long plats for doubled, which OTT). This seems to be the case with some MUs (e.g. Class 170 turbostar), but not others. I prefer the flexibility, which I note is available with the LUL stock (my tube routes were around from the 1860s I think as I used tunnels and normal platforms / locos since stagecoaches simply could not handle the local traffic – but the result was I ended up with 2tile tube trains forever and growing “overlap” of stations. In truth, I should have blitzed my local networks and started again with fewer, larger stations – but then again, this is part of the interesting “working with legacy” aspect I enjoyed.
•   Sprinter generation DMUs are very short lived (unlike reality !), and replaced older Loco hauled and 1st gen DMUs – tying the replacement times together would help playability.
•   Continuation of some of the more modern stuff, Sprinter DMUs, HSTs, Class 60, Class 86/90 all still serving (2013) in fleet strength yet all vanish circa 2000. The Mk3 could benefit from extending, as otherwise the Class 67 has to run with Mk4s which isn’t realistic.
•   Class 66 seems very, very expensive for what it offers - unless I should be using it's power to pull much longer freight trains, i.e. 8-10tiles ?, 8 is doable (I use Class 57s in a "Freightliner" sector on all piece/cooled goods which work well) but 10 starts to really cause problems on track layouts and signalling, plus preceding locos (Class 37 with the 56/58 not really usable) aren't really any good above 6tile trains so it's a hell of a jump if that is what is needed to make the 66 profitable.
•   Desiro 450 seems very underpowered, far worse than predecessor (4VEP in my usage). Struggles to get much above 100kmh and not near it's max.

[when I say underpowered, I mean the combination of power/gear. I am not 100% on how these work, but I did have a test track set up with parallel identical lines to test how different locos/formations performed and I understand from reading the forum it is effectively power x gear = game performance, which fitted my test track observations]

I don’t want the above to be taken as some kind of criticism, they were just the points I noticed when trying to play the game through the period (using timeline). The ability to use obsolete stuff makes the game still perfectly playable, but I would suggest tying replacement dates together better perhaps – especially as it seems the vast bulk of the trains are now modelled, certainly in the BR  modern era if not also the Grouping era onwards as well. I do have an excel file I created with all the trains and their dates, mainly so I could plan what I’d do with the various lines and companies (and later, BR regions and sectors !).

Anyway, regardless of what I've commented, please keep up the excellent work. I need to start a new map, which I think will be much larger, but with less cities, and non-beginner mode too, so hopefully I'll still make enough money to build stuff (less wastefully as I now understand how to do everything I think). Having AI companies will at least reduce the admin burden of trying to keep so many local networks optimised and uptodate - my interest is in the intercity/suburban stuff.

cheers  :)
Title: Re: comments on playing simutrans
Post by: sdog on April 08, 2013, 10:03:34 PM
Welcome to the forum precor.

Knowing people in this forum for a while, i should expect such extensive feedback will be very much appreciated by the team. There are always gaps and inconsistencies in such a vast pak-set, the most difficult is to spot it, while improving it often only needs a few lines to be changed.

Just one small advise, when you start a new, more extensive game, better test what the AI actually does in a test game, before relying on it. It is not very capable, and might spoil things. It might be much more effective to create huge stations, to cover a large part of a town, consider that some kind of community organised service. With this approach you only have to be careful with the latest versions of experimental (i guess you use it, since you mentioned having used convoy spacing), where this might cause pax to walk long distances and increases wait times.

With regard to convoy spacing: Perhaps you ought to try having only the terminal stops of a route set to 'wait', with enough room for buses to stage. Ie having buses unload at a stop, loop around (with plenty of space for some other buses), go to the same stop but different platform and have it wait there until full or on time.
Title: Re: comments on playing simutrans
Post by: jamespetts on April 08, 2013, 10:40:06 PM
Welcome to the forum! Constructive criticism is always welcomed. Development of this pakset is somewhat complicated by the fact that we have two versions of it, each maintained separately, albeit with close co-operation - one for Simutrans-Standard (which I think that you played), and one for Simutrans-Experimental (which I maintain). I can comment on your helpful observations from the perspective of the maintainer of the Simutrans-Experimental version (and as someone whose work in the production of graphics and data for vehicles, etc., is often equally applicable to both versions), and I imagine that, on return from his holiday, the maintainer of the Standard version, which you played, The Hood, will be able to comment more specifically in some areas. I should also note as an aside that, currently, the Experimental version as released is behind the Standard version as released as there are a number of things which need to be done with the Experimental version before I can release the latest version of that.

As to your specific observations, I shall reply in numbered order.

(1) On the Experimental version, I have for a while tried to set retirement dates to match real-life production dates, albeit often extended until a suitable replacement is available if there would otherwise be a gap. There was a time in the pakset when dates were partly based on service dates, which is incorrect, and I do not know whether the Standard pakset retains some of this anomaly.

(2) Gearing is one way in which the Standard and Experimental versions substantially differ. This is because the Experimental version has a different physics implementation, which has been further revised recently. This includes a value for tractive effort and brake force as well as power. In the currently released Experimental version (based on the previous revision of the physics), no vehicles have a gear value at all. In the next release of the Experimental version of the pakset, gear values will represent traction losses from engines in vehicles where the engine is a distinct unit separate from the drivetrain (i.e., not steam locomotives, horses, sailing ships or the like). In the Experimental version, these are based wherever possible on researched real life values, and otherwise extrapolated from those vehicles whose values are known. The gear factors are also based as far as possible on real live values for drivetrain losses (20% loss for electric, 50% for diesel). Power for steam locomotives is more complicated, but I have recently undertaken a large scale review of this ready for the next release, based on real life measured performances, and extrapolating others using mathematical formulae based on characteristics such as the heating surface area and firegreate area.

(3) This particular issue has been addressed by some of the enterprising contributors to Simutrans-Experimental using the spacing and scheduling ( feature.


This is indeed horribly complicated, and, in terms at least of cost balancing, work has not really started on this either in the Standard or Experimental versions. One of the main reasons that I have not yet started is that there is no point in cost balancing until the parameters of the simulation are sufficiently clearly set, and, since there are some significant changes planned ( for Experimental, many of which have the express intention of allowing cost balancing to work properly, it does not seem worthwhile trying to hit a moving target. During that time, I am endeavouring to gather real life pricing data ( from which I can extrapolate relative costs when the balancing process finally starts. That being so, however, there has been some work done on balancing the revenue side: from version 0.8.4 onwards, the relative cost of transporting different types of things is based on real life data, and work is ongoing to calibrate the way in which passengers are generated ( to increase the realism of the simulation of transport demand in that respect.

As to the availability of track to go with certain types of train, the issue of track has recently been addressed in preparation for the next release of the Experimental version, which ought to deal with these various issues.

Incidentally, the project takes a long time because there are not a great many people at present working on it. Any assistance that you are able to offer, in terms of producing graphics (easier than it might first appear using Blender), researching historical detail, finding bugs, obtaining balancing data, conducting rigerous in-game testing, producing or modifying data (.dat) files for individual vehicles, ways, etc., would be much appreciated if you are so inclined. The whole project is open source (Github pages for the Standard ( and Experimental ( versions are in the preceding links).

Thank you again for your feedback; it is excellent that you enjoyed the game, and my commiserations to your wife for you having discovered it.
Title: Re: comments on playing simutrans
Post by: The Hood on April 13, 2013, 09:44:54 AM
Thanks for the very detailed feedback. I do find the balancing tedious and I much prefer drawing new objects. At some stage an overhaul of the cost and gear balancing is long overdue (last done before the release of 1.0 about 5 years ago!) - if you or anyone else would like to do this you'd be more than welcome - let me know so we can coordinate this.

Edit: on reading through your list as you say a lot you could do yourself. You can download sources from SVN and you will want to find and edit the relevant parts of the dat file (fairly self-explanatory). If you wish to do this, go ahead and post the updates here or PM me and I can then include them. I'm prefectly happy with most of those suggestions.
Title: Re: comments on playing simutrans
Post by: The Hood on April 15, 2013, 07:54:11 PM
•   Class 123 seems to have a large gap between 1st and other 3 cars (middle 2 offset to rear, as overlap 4th). A graphics glitch I assume.

Note for jamespetts/experimental: this can be solved by switching the DMS and DMBS vehicles.
Title: Re: comments on playing simutrans
Post by: jamespetts on April 15, 2013, 09:55:32 PM
I think that this has been fixed in Experimental some time ago by reversing the order in which the images are drawn as follows:

Code: [Select]


Title: Re: comments on playing simutrans
Post by: The Hood on April 16, 2013, 07:45:48 AM
No - as the vehicles have one end as the yellow cab?
Title: Re: comments on playing simutrans
Post by: jamespetts on April 16, 2013, 09:04:18 AM
What do you mean? I don't follow, I am afraid.
Title: Re: comments on playing simutrans
Post by: The Hood on April 17, 2013, 07:23:54 AM
The two vehicles that are misaligned are ones with yellow cab-ends, so merely displaying them back to front surely just places the gangway end at the front of the train and the cab in the middle?
Title: Re: comments on playing simutrans
Post by: jamespetts on April 17, 2013, 08:17:17 AM
I might have changed the graphics when I fixed this - have a look at the latest versions on my Github page.
Title: Re: comments on playing simutrans
Post by: The Hood on July 27, 2013, 07:55:20 PM
LUL '67 100kmh, yet tube tunnels limited to 80kmh (also hits '59s ability to make a profit), this affects all subsequent LUL stock in the same way that city road limits destroy the ability to use the higher speed of later busses to pay for their running costs. Fair enough in cities (50kph), but either tube tunnel limits could be raised, or a second tunnel type available say from the mid/late 60s which offers a higher speed.

I've now added a modern tube tunnel from 1950 which is capable of 100km/h to fix this.
Title: Re: comments on playing simutrans
Post by: jamespetts on July 28, 2013, 11:21:48 AM
I've now added a modern tube tunnel from 1950 which is capable of 100km/h to fix this.

Hmm - I don't think that this is right, is it? The Victoria ( and Jubilee ( lines both have 80km/h speed limits, and those are the newest Underground lines. I'd suggest having the new tunnels have an 80km/h limit, and the old ones having a 56 or 60km/h speed limit.

Also, you might want to have retire dates for the older type of tunnel now.

Edit: Actually, according to places like this (, a 70km/h speed limit might be more appropriate for the older tunnels (45mph, or about 72km/h).
Title: Re: comments on playing simutrans
Post by: kierongreen on July 28, 2013, 12:58:14 PM
Much of the speed limit is down to sharp curves winding between building foundations, utilities and so on.