Long trains, or short ones?

[aridus]

Which do you prefer?

I've seen people say to build for example a train with 9 coal cars on it for coal runs. This does make good money, and initially has a lower running cost.

I prefer however to run two or three short trains per job, with four or five cars each. Especially for coal. It costs more for engines and running cost at first, but in my experience it makes money faster, for a few reasons:

1. short trains load faster, therefore get out of stations faster, allowing other trains to get in and already be loading while the first one is about to drop off.

2. short trains clear signals faster and tie things up for only half the time.

I've had three short trains running a three lane station coal mine making loads of money by year two before, with three trains there is always one loading constantly, one unloading, and one coming in or just arrived where the previous one left. I usually give them conditional orders also to maintain by themselves when their reliability gets below a certain point. The profit they all make together more than pays for the trains IMO.

[Emperor Jake]

For raw material freight trains, I generally start short and soon add more wagons once the mine or farm has picked up its production. I usually end up with 20 to 30 wagon freight trains, and longer if it's extra long distance or super productive industry. I also often have feeders from nearby mines.

For processed freight trains, such as goods, I use longer trains for transport and shorter ones for distribution but it depends on how my network is set u.

For passenger trains the length depends on the traffic (ie population of cities), capacities of the wagons in the trainset, and space I have to build stations in cities. The longest my passenger trains are is generally 16 vehicles (8 tiles)

- Jake

[aridus]

Good stuff, sounds like you really know what you are doing.

I just had a race with a 9 car coal train vs. a 5 car one, running to the same plant but on two different sides of the mine so they could load simultaneously. The 5 car one beat out the 9 car, surprisingly, by over 5,000 a year. It also maintained reliability a lot more (80 something vs 60 something for the 9 car. I guess long waits between services take its toll)

[V453000 :)]

After playing for some (possibly insanely long time) you will discover that money is nothing in openttd. From my point of view, trains long 3 tiles are best because:

1. You need curves only 3 tile long to keep maximum speed at all times.
2. They fit in shorter areas, so making a waiting bay or station for them is smaller.
3. The game is much more fun with many trains.
4. Good acceleration with one or two engines (depends on which engine)
5. Their tile-length is an odd number. If you plant signals every second tile, odd lengths fit better when jammed.

The last thing I am too interested about is capacity, running cost or maximum speed, because sooner or later my network will be flooded with trains anyway, shall they have 128 or 265 kmh max speed, 120 or 300t coal capacity or whatever else. If you create massive networks, you will get to understand why I say it, even though I believe you currently think I talk nonsense.

[Eddi]

there's a strong dependency between length of train and network layout. the longer your trains are, the longer you can handle lots of cargo with very few tracks (e.g. single track with sidings). that is especially when speed is no big issue (e.g. with wagonspeedlimits).

i often use 10, 12 or 15 tile trains for cargo

[mfb]

5. Their tile-length is an odd number. If you plant signals every second tile, odd lengths fit better when jammed.

On the other hand, trains with even TLs can unjam quicker, as the next train can start earlier.

That comes out of an unjamming line: Note that the longer trains with lower acceleration and a larger spacing in the jam (6 instead of 4 tiles per train) have a smaller distance now.

unjamming.png (11.08 KiB) Viewed 3345 times

The size of everything is a very good argument to use a TL of 3.

[Jacko]

Long Trains ARE the best because they will carry more cargo, so... going long distance a 14+ car train will get you more from long distance, and you would need less long trains than short trains to manage that --> save on eingines

[planetmaker]

Long Trains ARE the best because they will carry more cargo, so... going long distance a 14+ car train will get you more from long distance, and you would need less long trains than short trains to manage that --> save on eingines

Aha.

And long trains are worst because
- they load longer, your cargo ages longer --> less revenue every trip
- they load longer, you need fewer trains --> trains arrive less frequent -> station rating becomes worse --> less cargo to transport or low production industries close in the first place --> less revenue (and fewer trains needed)
- they need much longer curves --> travel slower on most networks due to curve length restrictions -> less revenue
- they need much longer curves --> track building needs to be much more extensive and expansive and longer -> less revenue


Long trains are best because
- on average you might need some engines less --> less money spent at one time
- on heavy-duty networks they increase the cargo transported-per-tile ratio as you have less space used by the distance between two trains
- Jacko says so

hm... so...?

[Jacko]

Long Trains ARE the best because they will carry more cargo, so... going long distance a 14+ car train will get you more from long distance, and you would need less long trains than short trains to manage that --> save on eingines

Aha.

And long trains are worst because
- they load longer, your cargo ages longer --> less revenue every trip
- they load longer, you need fewer trains --> trains arrive less frequent -> station rating becomes worse --> less cargo to transport or low production industries close in the first place --> less revenue (and fewer trains needed)
- they need much longer curves --> travel slower on most networks due to curve length restrictions -> less revenue
- they need much longer curves --> track building needs to be much more extensive and expansive and longer -> less revenue


Long trains are best because
- on average you might need some engines less --> less money spent at one time
- on heavy-duty networks they increase the cargo transported-per-tile ratio as you have less space used by the distance between two trains
- Jacko says so

hm... so...?

Damn. You shot me down I uh... still think long trains win. Especially on the harder maps when you don't have change to spare on lots of trains

[V453000 :)]

Well, "harder maps" could also be refered to as "totally boring" since you usually buy a train and wait for a year until you can affor another one ... and then you waint another year ... seriously who enjoys playing that?...

[Jacko]

Its called burning hours, and besides, its not THAT bad

[buckethead]

I prefer having long trains, just looks more realistic. And I don't mind if I can make more money using short trains, money pours in like crazy in every OpenTTD game anyway.

[Jacko]

I prefer having long trains, just looks more realistic. And I don't mind if I can make more money using short trains, money pours in like crazy in every OpenTTD game anyway.

One of my Points exactly. Well Said Buckethead

[mrMann]

I always calculate the train length against the initial production of the industry, making sure it's just above the total production. I tend to build 7 tile long stations (so expansion is easy later on), so rarely do my train lengths exceed this.

[aridus]

I prefer having long trains, just looks more realistic. And I don't mind if I can make more money using short trains, money pours in like crazy in every OpenTTD game anyway.

Yeah on large maps where you have room, that is easy. If you want a challenge for efficiency, try a 64x64 map. The "money pours in" theory goes out the window then and you actually have to sell unprofitable trains when your cash starts going steadily negative and you don't even break a million until well into the year 2000, plus towns will hate your guts and make it hard for you to build stations around them and you can't just terraform your way out of it because you don't have money in the first place since all trips are incredibly short.

[andythenorth]

http://www.strategosinc.com/onepieceflow.htm

One piece flow is an ideal model in some contexts. A key benefit is that it delivers value fastest.

There's probably a better explanation somewhere, but I can't find it right now.

It's a production methodology, but has an obvious application to networks, including OpenTTD games. I've seen openttd coop using approximately this tactic in their networks

One piece flow isn't best for every production context, costs can outweigh benefits.

[Qu@rks]

My train length usually depends on the trainset(s) that I use. A lot of times, I use DBSetXL which limits train length for passenger trains because ICE trains can only be 8 tiles (16 cars, 14 + one 2-car engine) long.

Freight trains, however, should be veeeerrryyy long in DBSetXL because many engines (even early electric ones) have hp in excess, thus being able to pull 20 cars or more PER ONE ENGINE (unless you play with a weight multiplier of x4 or higher).

In general, I also prefer long trains. I like how they slowly accelerate and finally reach their maximum speed somewhere in the middle between two stations (and not 10 seconds after they departed one station). This does not apply to regional or S-Bahn trains though, as (in reality) they usually travel short distances at frequent intervals.

I'd say there is no general rule as to build long or short trains. If you aren't restricted by map and/or game settings or a newgrf, try out different approaches and see which one works best in a given context.

[aridus]

Yeah I guess for really loooong trips they can be good. I threw in a 20 car train on an extra productive wood line across the Australia scenario, along with seven other 10 car trains serving the same station. I noticed the efficiency is about even for them most years but the 20 car is slightly more efficient than two of the 10 cars together for every other year.

Together the line moves well over 2,500 tons of wood per trip and makes over a good million credits per trip combined.

[paelleon]

HI, excuse me, but I really don't understand.

I use to build a single train with 6-8 cars (engine excluded) for a single track. This is usually enough to transport more than 70% of cargo production, even if it is a farm (a single train with 4 cars for grain and 4 for cattle). Moreover, most of the time my trains have to wait a long time to get fully loaded! So, to run a single factory smoothly I need to connect it with 3 or more suppliers (farms or steelyards).
Even more problems with Oil fields, they disappears from the map pretty quickly, so that I usually ignore them!
The only cargo that need a really big effort to be evacuated are passengers. For them I use combined stations: buses, trains, airports and, if possible, ports.

How the hell do you manage to raise mines and farms production in order to need more than 8 cars tokeep it empty?

[Brumi]

Sometimes when I find 3 or 4 industries relatively close to each other, I build a train station at one of them, and I use trucks to transfer cargo to the train station from the other industries. This way I can use longer trains efficiently, but long still means 10-14 carriages for me... This strategy is also useful if realistic acceleration is turned off and I'm playing in a mountainous area. Trucks are better in climbing slopes