propulsion power of bidirectional trains

[Engimage]

Don't you agree that, in the demonstration video, it is very ugly to have to place two locomotives right next to each other, just to imitate a bidirectional locomotive?

I think it is good to have the ability to control the way your train acts. As pointed in RoRo vs Terminus thread both behaviours have their pros and cons so let the people choose what they want.
However having rediculous trains like 3-8-3 or even 2-4-2 where you have the number of locomotives almost equal to the number of cargo wagons to make a train fast enough is not cool at all.

[Tekky]

I think it is good to have the ability to control the way your train acts. As pointed in RoRo vs Terminus thread both behaviours have their pros and cons so let the people choose what they want.

I agree. However, I think it would be better to add an option "restrict train to forward movement" or "allow train to reverse" instead of requiring a second locomotive.

[Engimage]

I agree. However, I think it would be better to add an option "restrict train to forward movement" or "allow train to reverse" instead of requiring a second locomotive.

Making a 2-way train symmetric is a good thing. It is good for visual identification and its good for station design as asymmetric trains will be able to enter a station with both head and tail and cause some bizarreness in loading/unloading.
The only scenario where this might help is a train consisting of the single locomotive used for PAX but you can always just R the train and do your job. No need for that in automatic scenarios at all.

[ssilk]

May I point (again) to the basic thought of why I think this distinction is good:
If you have a map of about 1000 tiles size, the bidirectional trains is a very good idea: They save you a lot of space, it is much easier to handle them. But if you go above that size, let's say your map is 5000x5000, well, then you should ask you if double-sided trains are still THAT good idea.

It's a simple decision the player needs to make, like the decision between steel- and electric-furnace.

[Engimage]

May I point (again) to the basic thought of why I think this distinction is good:
If you have a map of about 1000 tiles size, the bidirectional trains is a very good idea: They save you a lot of space, it is much easier to handle them. But if you go above that size, let's say your map is 5000x5000, well, then you should ask you if double-sided trains are still THAT good idea.

It's a simple decision the player needs to make, like the decision between steel- and electric-furnace.

I'm not arguing about the concept of bidirectional trains vs single direction. I'm telling about the requirement to have locomotives facing both sides for bidirectional trains. Its not a good thing imo to have a bidirectional train with a locomotive facing one side. Also its not a good idea to have a single direction train with locomotives facing both sides. So I am against the option for disabling/enabling reverse in train interface.

[Machine Medic]

As long as there is at least one locomotive on the leading end of the train, all other locomotives in the train should contribute full power, regardless of orientation.

Modern diesels use a large diesel generator to deliver AC or DC power to a set of traction motors underneath the frame. Reversing is accomplished by reversing the DC polarity or AC phase rotation being delivered to the traction motors - thus maximum power is always available in either direction.

https://youtu.be/bQPOXbQ4B9k?t=8
^example of both reversed and forward locomotives doing the same work.
https://youtu.be/b1syD6Nu3kw?t=103
^pictures of a reverser used to reverse a DC locomotive.
https://youtu.be/Virnj9vW070?t=738
^ride in a locomotive pulling a train in reverse.

[IronCartographer]

First: Factorio's design principle of trains moving automatically only in the direction(s) they have forward-facing locomotives makes sense from a gameplay perspective. It makes train behavior evident with no locomotive-specific UI at all.

However, I agree with the suggestion of locomotives supplying power in reverse. Modern trains work that way, and the player can drive in reverse, albeit slowly. Dead-weight locomotives make no sense.

Making multiple terminus stations feeding the same system (to compensate for the delay between trains vs. RoRo) might seem undesirable, but an interesting thought occurs to me: With two trains and circuit network logic to alternate between them, you could eliminate station buffer chests entirely.

[Optera]

Making multiple terminus stations feeding the same system (to compensate for the delay between trains vs. RoRo) might seem undesirable, but an interesting thought occurs to me: With two trains and circuit network logic to alternate between them, you could eliminate station buffer chests entirely.

The primary function of station buffer chests is not to act as buffer, but to speed up (un-)loading.
Belt > Chest or Chest > Belt is limited by belt speed while Chest > Chest is only limited by inserter rotation speed.

[Engimage]

Making multiple terminus stations feeding the same system (to compensate for the delay between trains vs. RoRo) might seem undesirable, but an interesting thought occurs to me: With two trains and circuit network logic to alternate between them, you could eliminate station buffer chests entirely.

The primary function of station buffer chests is not to act as buffer, but to speed up (un-)loading.
Belt > Chest or Chest > Belt is limited by belt speed while Chest > Chest is only limited by inserter rotation speed.

I would argue about this. I has to be a buffer as the gaps created due to train absense at the station will lead to production disruptions and problems.

[aober93]

Why remove the variety the game has to offer. The logistics part of the game is where everyting can be improved regardless of how often you change it because of the drawbacks everything has. This makes for a very individual playstyle. Why remove this?

[BenSeidel]

Making multiple terminus stations feeding the same system (to compensate for the delay between trains vs. RoRo) might seem undesirable, but an interesting thought occurs to me: With two trains and circuit network logic to alternate between them, you could eliminate station buffer chests entirely.

The primary function of station buffer chests is not to act as buffer, but to speed up (un-)loading.
Belt > Chest or Chest > Belt is limited by belt speed while Chest > Chest is only limited by inserter rotation speed.

I would argue about this. I has to be a buffer as the gaps created due to train absense at the station will lead to production disruptions and problems.

It's only a buffer if you class the train itself being a buffer, or belt as being a buffer. You can compensate for the disruption to supply by the train leaving/entering quite easily without any chests, the chests are only there to get the train out of the station so it can go and pick up the next load. The speed that a full set of stack inserters can empty a train when putting the items into a chest is crazy fast.

[IronCartographer]

The primary function of station buffer chests is not to act as buffer, but to speed up (un-)loading.
Belt > Chest or Chest > Belt is limited by belt speed while Chest > Chest is only limited by inserter rotation speed.

Speeding up unloading during the times when a train is actually present so that the belts can remain full all the time is exactly what a buffer is/does: Evening out the variations in throughput.

If you unload wagons directly onto the belt, and switch between trains rather than unloading them at the same time, the dual-station cargo wagons can serve this purpose.

[MKNZ]

The OP keeps being derailed.

No power added isn't realistic, it doesn't matter what extent you want to change it, it's not up to OP, or general posters, the point is we all agree no power isn't realistic correct?

The minimum should be 'reverse' power, previous to the conversation of 'equal power directional trains'. We all agree with that right? If not, why not?

[AileTheAlien]

Train engines (running in automatic) should stay single-direction, because it allows the player to make an interesting gameplay choice:
- use less space for tracks, but use more engines?
- use less engines, but use more space for tracks?

This game has aliens, robot arms that use no complicated cameras, portable fusion reactors, and working laser turrets. It's not "How Trains Actually Work In Real Life Simulator 2017".

[IronCartographer]

Train engines (running in automatic) should stay single-direction, because it allows the player to make an interesting gameplay choice:
- use less space for tracks, but use more engines?
- use less engines, but use more space for tracks?

This game has aliens, robot arms that use no complicated cameras, portable fusion reactors, and working laser turrets. It's not "How Trains Actually Work In Real Life Simulator 2017".

In real life, we don't have to use Terminus stations to avoid train pathing through undesired and wasteful loops. The "it's a game" argument goes both ways. People who want to use RoRo stations and loops with one-directional trains are free to do so for aesthetic or personal reasons, but those who use bidirectional trains are simply forced to endure slower acceleration and arbitrarily longer acceleration ramps.

This game has a philosophy that it is not about "exploration" and generating unnecessarily large maps. I think that applies similarly here: Forcing longer separation in rail lines to compensate for the lower acceleration is just annoying, when the dead weight is caused by the need for Terminus to compensate for other limitations.

With 0.15, better fuels will yield better accelerations and therefore throughput for intersections. That is fun! "More, longer track (without even the station layout considerations of loop designs) to get the same effect" is not.

From the moment I started using trains in Factorio, I liked how simple it was to see a train's behavior from the locomotive placement, but the dead weight seemed pointless and broken.

I think a good compromise would be for the "reversing power" to be applied in automatic mode as long as another locomotive was facing forward: There would still be less acceleration than the equivalent number of locomotives facing forward, but it would actually be consistent with locomotive behavior while driven manually. "Bidirectional trains waste fuel!" would continue to be an argument for not using them.

[BenSeidel]

Train engines (running in automatic) should stay single-direction, because it allows the player to make an interesting gameplay choice:

People who want to use RoRo stations and loops with one-directional trains are free to do so for aesthetic or personal reasons, but those who use bidirectional trains are simply forced to endure slower acceleration and arbitrarily longer acceleration ramps.

The only gameplay downside with single headed trains is the slightly larger overall size of the station as you need a loop. The trains have faster accelerations (or two additional cargo wagons per loco), less time with the station being empty, and are less likely to get a "no path" error. The only reason that they aren't used is the path finding destroying UPS, and that issues is sooooo large and important that no matter how good you make single headed trains, double headed trains will always "win".

But, if we exclude the UPS issue and only look at the gameplay mechanics there is too much of a skew towards single headed trains. They are simply too efficient when compared to double headed trains. If you make the trains supply power when going backwards then the only time you will ever really have a single headed train is a 1-X train. Not really an interesting mechanic.

The only solution I can see as being viable (that is still fun) is to drop the weight of the locomotives to be the same as that of a wagon. Then it's a case of a 1:1 replacement of loco vs cargo. Single headed trains will still have the benefits above, but double headed trains of the same overall length will have comparable acceleration, and have smaller stations. I feel that this is a difficult tradeoff: station size vs throughput.

[IronCartographer]

But, if we exclude the UPS issue and only look at the gameplay mechanics there is too much of a skew towards single headed trains. They are simply too efficient when compared to double headed trains. If you make the trains supply power when going backwards then the only time you will ever really have a single headed train is a 1-X train. Not really an interesting mechanic.

Reversing power is not as high as the forward power. If we consider the possibility of a locomotive burning fuel at the same rate but converting it less efficiently in reverse:

One-directional trains would still be faster, and convert fuel at the maximum efficiency.
Bidirectional trains would be nearly as fast, but use slightly more fuel with slightly less efficiency/acceleration.

It simply narrows the gap, rather than eliminating it.

[Machine Medic]

I still don't see why you think we should be forcing people to decide between two evils:

If people want to make beautiful, neat, efficient terminal-style stations, then why penalize them?

Bi-directional rail systems with many trains are far more complicated, but can cut signal waiting times in half. If anything, higher throughput per unit track should be used as a strong motivation to make more difficult and advanced bi-directional rail networks and trains.

We shouldn't be holding people back with arbitrary rules like this. Factorio is about facilitating creative innovation, not "I wish I could do this...".

<.>

[jonatkins]

I wrote this as a new post before I found this thread - posting here to preserve my point of view.

After reading some of these comments I also like the idea of allowing locomotives facing any direction to provide full power, but otherwise work as things do now (i.e. one locomotive must be facing the direction of travel). Compatable with all existing setups, and no lugging dead weight around.

========================================================

Currently, locomotives in trains have a driving direction. In manual mode they can be reversed, but when running automatically, they can only go forwards.

This leads to making a choice between eiter using single-direction trains, or pulling dead-weight locomotives around on bi-directional trains.

In reality, real locomotives can run just as well in either direction. However, trains wouldn't run (other than slow shunting manovures) without a locomitive at the forward end.

How about changing Factorio locomotives to work like this?

• Locomotives provide power running in either direction
• Trains need a locomotive at the front to run in automatiuc mode. Locomotive direction is uninportant

Some examples, to help exploain this:

factorio-train-example.jpg (268.17 KiB) Viewed 6269 times

In automatic mode
The top track: both trains can run with the same power - the left one is bi-directional, the right one is single direction only
The middle track: Plain locomotives, just one or a pair, can run in either direction.
The bottom track: The locomotive can't see where it's going, so cannot run in automatic mode in either direction.

Manual mode: rather than locomotives in reverse running but in reduced speed, trains without a locomotive at the front (in the direction of travel) run at reduced speed.

---------------------------------------------------

Additionally, rather than completely banning trains from running without a locomotive at it's front, there could be an option in the train timetable to specity a timetable entry as a shunt manovure. Trains cold then run in reverse, without a locomotive at the front, but at a vastly reduced speed. This would be handy for a mainly one-direction train network where the occasional station needs a train to reverse out of a station, shunting into a separate reversing siding.

[leoch]

Additionally, rather than completely banning trains from running without a locomotive at it's front, there could be an option in the train timetable to specity a timetable entry as a shunt manovure. Trains cold then run in reverse, without a locomotive at the front, but at a vastly reduced speed. This would be handy for a mainly one-direction train network where the occasional station needs a train to reverse out of a station, shunting into a separate reversing siding.

This is about the only new contribution in your post, and is what I was about to write (even slightly better).

Making path finding work as expected is, as I understand it, the only reason locomotives don't go backwards by default. Your suggestion would improve the game without breaking many people's games, with the exception of trains currently running without a locomotive at the front.

The other issue I've had with bidirectional trains is that they can end up turning around unexpectedly. This is fine when the cargo is uniform/symmetric but not for all trains. Being able to specify that the train must in some cases enter a station a certain way around would therefore be useful.

In summary, please:

• let engines apply power backwards in automatic mode (possibly full power, possibly less as when manually driven)
• by default only allow a train to find a path to a station when it has an engine at the front pointing in that direction (or possibly anywhere in the train); if there is an engine at both ends this allows pathing in both directions
• allow items in the schedule to be made a "shunting manoeuvre", meaning do pathing in both directions (possibly restrict speed when there isn't an engine at the front)
• allow items in the schedule to have an enforced direction forwards/backwards, even when pathing in both directions is otherwise allowed (this requires identifying the "front" of a bidirectional train)