Open RT2 Wish Lists

[jeffryfisher]

BTW, didn't our conversation go a little off the topic? 

OK, I have one, I want to be able to move a train in the train list, or add a search feature to the train list.

Don't worry about going off-topic in this thread. When I see a good wish (or backing for a wish I "don't understand"), I add it to one of my lists on page 1. The rest of the thread can be free to go anywhere.

 

I'm sure I already listed wishes for filtering and sorting lists of stations and trains, as well as a wish for selecting multiple trains or stations to receive upgrades / engines en mass.

[outofmage]

Oh, I forgot we can edit here. Well, the list is a bit long, I can't remember every point of them :P

[j7n]

I just read through this thread of valuable observations. I generally agree with everything.

Inflation would be good to make runaway profits from interest less likely (but still possible), although it must not cause overflows or loss of precision, which happened in TTD. Prices must be computed from a fixed base year, instead of the previous period. Perhaps inflation could equal interest payment during prosperity; and during boom times when interest is low, the value of cash should actually decrease. I don't think interest should further increase costs compared to revenue, as that already happens with revenue. Unless, the new inflation model would be a replacement for the existing revenue decay.

A few ideas of my own I could add.

Weaknesses:

Currently trains lose too much speed on curves; the momentum just disappears without the train derailing. It is too beneficial to build straight track and even cross mountains in a straight line (Alps without going through the Brenner Pass). I don't quite get how map scale could be incorporated here. An improvement could be to use a similar model as TTDPatch. If a train makes one 45° turn it should lose a little speed. If it makes another turn in the same direction within a certain distance, it loses much more speed.

TTD takes into account the length of the train (instead of a distance threshold), which wouldn't translate well into RT2 where trains are shorter, and have disposable cars. It wouldn't make sense to have a deadhead engine race curved track at its top speed. Perhaps the distance could be made a little dependant on the mass of the train. Trains could be allowed to make a second turn in the opposite direction (_/ˉˉ) with little penalty because such a junction would represent a gradual shift of the track to the left or right, technically limited by the granuality of the map cells.

Undesirable:

Train priorities, when deciding when to yield to an oncoming train, are difficult to understand and manage. Trains that take long to accelerate may be stopped too easily, even when the other train is of the same broad priority class. Upon meeting, both mass and current velocity of two trains should be compared, taking into strong account the priority class.

A rough formula of computing the priority score could be: mass * (velocity + (20 * (priority - priority[other])) - 10 * relative_heading + 5 * grade). A higher priority train could always pass, but I don't play with priority classes myself, so I'm fine either way. It seems to me though that if the another train has picked up very good speed, it shouldn't be stopped by a possibly erroneously set class. Relative_heading is simply a small bias to prevent a train being run over from behind, which doesn't look good. I'd like to hear what the current system is aimed for.

This would prevent two or more trains stopping and jamming forever on an uphill stretch, where a single train happened to stop (due to a breakdown or an express overtake), as they erratically switch priorities. Only one train should remain stopped then. Currently it seems that if a train has yielded for a very long time, it will take over the track eventually, even on foreign track.

The throttle setting should not be taken into account when computing the priority score, only the current velocity attained. This causes a practical problem and a glitch. If the throttle has been adjusted, it is very hard to return it to the exact previous value. And it seems that currently a train with lower throttle will have a much lower effective priority score causing it to yield for no apparent reason.

A broken down train may be made to yield by lowering its throttle. A broken down or crashed train is presumed to be blocking the track permanently until the section is restored. Yet the crash site may be micromanaged and cleared immediately via the throttle setting.

Stations do not regenerate their demand if only a low demand facility exists in their catchment area. Examples of low demand facilities are a Port (for example, in the Antarctica scenario it demands raw materials) and a Cannery for aluminum. Demand starts at 3 or 4, and, once met, never increases above 0. (Produce demand at the same cannery does go up.) High demand facilities eventually have their demand increase to 9 if not supplied. Low demand should go up some too. Otherwise these routes are hard to make profitable (like in Antarctica where there is little other industry).

Nice to have features:

I personally quite enjoyed the more challenging Second Century campaign, yet some things seemed to be follow logic. Improve the post-Geocore economy, which could be switched on with a dedicated flag on maps. Prices of cargo could stay the same to preserve most of the challenge. But towns could once again start demanding milk, produce and coffee. Different, smaller passenger cars could be used. If no new cars can be modelled, the pre-1950s version without the roof windows could fit better.

I have only recently started to play RT3. Overall I'm pleased to see how it's maintained most of the PopTop spirit in it. Splitting of food, goods and grain into a few distinct types is handy. I would like to have separate farms for industrial grain (corn) and grain for human consumption (wheat or simply grain). Although I admit it would remove some of the challenge of managing the grain flow.

You may stop reading here, as the following is just some vague impressions.

Some great points made there by Minneapolitan. I too have played only offline and still learning how to communicate.

I agree that Track Maintenance could be a reasonable method to control and reduce the frequency of breakdowns/crashes.

Passenger destinations are a great addition to RT3, and they shouldn't have the basic "demand" making a long express train more or less valuable depending on loads delivered locally.

 

Some of the railcars in the game are a bit ridiculous. First off, a load of Alcohol is 50 tons?? That's disproportionately heavier than anything else in the game, and totally random. A boxcar full of kegs and bottles won't be any heavier than a load of IRON ORE.

I think the alcohol car is unfinished. It is of the same weight and size throughout the game, which makes it out of balance early but reasonably in line when all cars are heavy. The heaviest cars (like gravel, and IIRC bauxite and coal) are actually 65 tons. I think Iron, along with Cement, Waste and Uranium, are intentionally small cars, making the player assess if any gain from shipping them faster would outweigh the costs. I don't understand however, how so much steel can be produced with so little iron.

 

I've never liked the way Goods looks. That's one of the most basic items (vague as it is) to be found in a boxcar, not a giant crate strapped to a flat car. And when has anyone here seen a high-walled gondola with new tires poured into it like coal? The Tires car needs to be a boxcar as well. I think the Rubber car is certainly questionable...

The cars that are exposed to the weather and the sun are definitely of concern to me. TTD has even more issues when it ships grain and paper in this way... I would welcome an improvement here, but only if it doesn't make all cars look indistingushable. RT3 ships many cargos in boxcars and also has "realistic" but very similar industry buildings. These RT2 designs are somewhat symbolic and shouldn't be taken literally. This is my biggest gripe with RT3 which looks too realistic (as was the trend with 3D games).

 

Building Towns - I think we'd all like to be able to place a settlement where we'd like one, just like the railroads did in the old days.

Definitely agree, and also building of industries, both at prohibitely high cost to make that unavailable at the start, but to make the endgame when the player is a god more interesting.

 

Flour Mill: I think it's stupid that grain straight from the farm goes to a bakery. Flour milling is a major industry anywhere, for example, in Minneapolis.

I miss the bread chain in RT3. Flour definitely makes sense. Thank you for your insight in its history. The game could be expanded with more types of food, perhaps as a replacement for the raw products that towns stop demanding.

 

Scrap: We could have cities convert finished goods into (fractional) raw materials. When you deliver three autos to a city, the city could produce one load of scrap (or iron). Hauled to a recycler, the load of scrap could become a load of steel. Similarly, two or three loads of paper could turn into one load of pulp.

I recall reading many years ago that cities (8 houses) produced Waste in RT2, yet I have not seen this in my game. Their production should be constant and low in my opinion, depending on the actual number of houses. This product, while rather difficult to collect from a number of cities at once, could be shipped to a new type of recycling industry for processing. But it would have to be a separate product from radioactive waste, which is not good for recycled paper nor metal.

[MaglevForever]

j7n, I couldn't agree more about the problems with the way corners are represented. I like your suggestions to slow down for multiple corners within a certain distance. I have mentioned this before, but grades shouldn't have anything to do with how slow a train should go around a corner.

 

I mainly wanted to post to say thanks for posting the tip on the effect of the throttle on breakdowns. I didn't know that. Put a smile on my face. Makes the throttle more useful now. Why should a breakdown block the track for so long anyway? Wouldn't a real railroad send out a helper engine or two to take it to the shop for repairs? Or at least move it to a siding. This seems realistic that traffic stops for a little while, but only that particular train is affected. Other traffic is free to move soon after. Thank-you.

[outofmage]

j7n, although you are new here, but you seem to have a lot experience in playing RT2. 

I have some comments about your ideas.

Inflation and curving in this game is commonly abused to be unrealistic, I agree with you too.

 

The priority score is very new and creative, although it seems to be a little complex.  I don't know what will happen if it's implemented.

 

Why is by turning throttle can clean a crash scene? I don't get it. 

 

About that demand get stuck in 0 levels, this bug only happens in v1.50, you can try v1.56 (Platinum Edition), it's been fixed.

[j7n]

When a breakdown occurs, the track where the broken train is on is blocked for passage by other trains. But turning down the throttle of the broken train, it is possible to instantly unblock the section, and have the broken train "remove itself" (because it's now on a low priority class). This seems like a cheat.

I'm certain that I played the v1.56 version when the demand problem occurred. I started a new campaign, and didn't load an old save. I just verified this by trying to supply a Cannery on my map.

By searching more, it seems that this problem is already known, and that some modding is required to fix it. As documented by Jeffryfisher, the problem occurs with secondary demand, rather than low like I believed. I don't quite understand the data format yet.

Edit: It is fixed in the Historical US Mod (rt2_plat.exe.j10x2). Same savegames now do generate demand for Tires (auto plant) and Aluminum (cannery). I couldn't check ports in Antarctica yet, because my save is from the very end of the game.

[jeffryfisher]

Curves depend on scale. At metro level, trains should slow significantly for turns (+2% grade). At regional level, trains should slow slightly for turns (+1% grade). At continental level (~6 miles per cell), trains should be unaffected by turns. In fact, 90 degree turns and maybe U-turns should be possible at TGV-speeds.

 

To decide which train of equal priority should stop, I'd like to see a simple momentum (mass x velocity) comparison. The higher momentum should prevail.

 

Don't know if I mentioned this: It would be nice if every station boosted (in proportion to its cost) the odds of houses springing up within its radius. A big station placed between some farms could spawn an unnamed town.

 

About the flatlined demands: Someone else around here analyzed those quite well. Basically, any two-input factory except steel would lose its appetite for input #2 after one delivery. They were bugs in the data tables within the EXE. My mods fix them, but no official patch addressed the issue.

 

Welcome to the gang, and thanks for sparking more discussion!

[outofmage]

j7n, I looked at your link to jeff's modding thread, and I notice the demand level problem still exists. 

I also notice that by turning the throttle down can make a breakdown train passable, I have to say I didn't know this before. Always learn new things!

It seems like Jeff needs to edit this thread again.

[outofmage]

I used to thought the list has covered everything, but it seems not.

Today, I looked through the list carefully to find anything I know, but not in the list. 

1. The way point game crash.

If a new engine was told to start from a way point, the game will crash. Because an engine can only spawn on station front track, but not in the middle of nowhere.

If you bulldoze a piece of track where a way point of a train's next destination is, the game will crash too.

 

2. Unrealistic de-acceleration on uphill grade and acceleration on down hill grade. 

When a train climbing a hill, the speed will slow down. The de-acceleration rate is not determined by the grade, but by the engine's acceleration, which means the faster the engine can speed up, the faster it will slow down. 

For example, when facing a huge grade like this:

If we use Class 55 Deltic, it will reach the bottom at speed of 47mph, and only down to 31mph at the top.

Because Deltic's acceleration is rated "Below Average".

If we use Class 232, the bottom speed is 50mph, top speed is 22mph. Although Class 232 has a better climbing capability than Class 55 Deltic, it drops more speed because its acceleration is rated "Very Fast".

If we use more fast boosting engines like Shinkansen Bullet which gets acceleration rated as "virtually instant". It will face a virtually instant stop hitting this hill.

 

This mechanic is very unrealistic, it makes good acceleration engines have disadvantaged hitting a sudden grade. But it makes a nice-to-have strategy to deal with huge mountains:

[j7n]

To decide which train of equal priority should stop, I'd like to see a simple momentum (mass x velocity) comparison. The higher momentum should prevail.

I would like to see grade, either positive or negative, to be considered also. A train that is already climbing a hill at speed should be allowed to do so.

For example, consider these two trains met on a single track. Their momentums currently are 4800 and 7700. Therefore priority should be given to the food train travelling downhill. If both trains stop, the food train has no trouble accelerating downhill, but the passenger train can only reach 12 mph.

Or these two trains, where the passenger train (3200) has caught up with a food train (3500). One of these trains will have to stop. If both stop, the food train can get moving easily, but the passenger train is stuck travelling below 1 mph.

That is an interesting observation, outofmage. I did not know about it, and only attributed the poorer grade climbing of modern trains to car weight. I can confirm this with a GP18 (Above Average acc.) and a Class 232 (Very Fast acc.), where both have comparable mountain climbing capability. GP18 took the grade slowing down to 10 mph, but the Class 232 got stuck travelling at 2 mph well before the apex.

Now we have another explanation why the E111 with Average acceleration is as good as it is.

[SD45]

Interesting observation outofimage.

 

I suppose this is why the Shay works so well. It has pretty poor acceleration, but it can climb grades like the real thing.

[MaglevForever]

outofmage, this is a serious flaw in the game. I always used the slower acceleration engines on grades. Now I have an explanation as to why this works. This explains why anything modern and diesel is second-rate compared to the E111. Great to learn something new to implement strategy for. Wish it could be easily fixed, it would open up the engines a bit instead of the current GG1-E111 dominance.

[jeffryfisher]

j7n, I looked at your link to jeff's modding thread, and I notice the demand level problem still exists.

Which demand level is stuck at zero?

 

It's possible that the mod thread went out of date (that my best mod might be in the US History thread that I was updating more often)

[jeffryfisher]

I would like to see grade, either positive or negative, to be considered also. A train that is already climbing a hill at speed should be allowed to do so.

For example, consider these two trains met on a single track. Their momenta currently are 4800 and 7700. Therefore priority would be given to the food train traveling downhill. If both trains stop, the food train has no trouble accelerating downhill, but the passenger train can only reach 12 mph.

Then perhaps the game could benefit from company-wide prioritization tie-breaker preferences.

 

I've added a few things (in blue) to the page-1 lists to reflect the recent discussion.

[SD45]

Apparently, the acceleration setting does not have a major effect on some locomotives.

 

The AMD-103 and Dash 9 do really good on grades. Even the severe one I set up. The AMD-103 has "very fast" acceleration. Dash 9 "above average".

 

However, the Maglev acts as normal, like it strikes a brick wall.

 

Edit-Did another test. The AMD-103 drops to a realistic mountain railroad speed when it hits a 11.2 percent grade. About 38 mph

 

The TBX-1 just stops dead at 1 MPH

 

The SD45 and Dash 9 did about the same. Stayed at about 40 MPH.

 

Shay did what a Shay does.

 

I may need to look at some other locomotives.

 

All tests were done with two passenger cars.

[j7n]

"momenta" ... thank you for the correction.

Having options to manage train priorities would be even better.

I did just race a few fast trains to see how they did on a practically flat land. And the Thalys slowed down abruptly on 2 squares of 3% grade "under" a bridge. I hadn't noticed this before because I didn't use this type of train.

One more thing: The game should handle more common errors without terminating, if at all possible. It crashes with a Windows error message too often, for example if there's a missing graphic. Currently, if the chosen GM2 saved state file cannot be opened for writing, RT2 will exit without an option to save the game with another name, or to restore write access to the intended filename. This could occur quite often these days with file system permissions. (I have set my finished saves to read-only.)

[outofmage]

Which demand level is stuck at zero?

 

It's possible that the mod thread went out of date (that my best mod might be in the US History thread that I was updating more often)

It's not what I meant. I meant compare to v1.50, in v1.56 this problem still exists, your mod should be fine.

[outofmage]

3.Debt limit merging bug

When merging a company, along with the assets, debt were also merged into us. But a company has 20 bonds limitation. Some of the bonds will disappear in the list after merging. 

For example, if a company has 8,000K debt currently:

Target company has 4,000K debt.

 

There would be 12,000K debt after merging. But the list only contains 20 bonds, missing the last 4 bonds. Although the interest will be calculated through average interest.

If we repay all the 20 bonds. There will be 2000K bonds disappeared. The red figure turns into black. This won't affect the balance sheet, the debt is still there. But we don't need to pay interest on it.

If new bonds is issued, these debt will need interest payment again based on new bonds' average interest rate.

 

I think PopTop has realized this problem and they featured average interest rate to solve this problem, but not completely.

I'd love to see the debt limit is eliminated, we should be free to finance our company as long as we have the credit.

[outofmage]

4. Improvement of way finding algorithm of two same distance route

If we have two routes with same length to go to the next destination. A train would choose the diagonal track, but I want them to choose the right hand track. 

 

Let me explain it first.

Here is some tracks look like this. The two routes have same length.

In this situation, trains will always choose to turn left.

But in a different layout, trains will always choose to turn right, like this:

 

An explanation is that when facing an intersection, if both routes lead to the same destination and with same length, trains prefer diagonal track than straight track(since an intersection always have one diagonal cell of track and one straight cell of track).

 

 I did a lot of tests and this rule seems to be right.

 

 

This rule is particularly useful when building double oversea bridge. As long as we follow the rules, trains will automatically run on the right side bridge, without setting way points which is difficult to do it right on a big map. 

 

To be continued...

[outofmage]

But the problem is, there are some situations that fits the rule, like these:

but in some conditions, it doesn't fit.

I have to twist the track too much to fit the rule.

 

This becomes annoying. I have to decide whether wasting some fuel and a little bit of revenue to save the effort to make way points.

I hate to make decisions like this.

 

So the algorithm should be able to tell which way is turning right (or going straight), and the right turning has higher priority not the diagonal track has higher priority.

[Gwizz]

Distance on each leg needs to be the same for the right hand rule to always work. Otherwise the trains will almost always take the shortest leg. 

[Gwizz]

Thinking back, the Climax with its two speed gear box was faster than the shay.  The Shay was the most common loco in the woods since it was easer to repair.   Over all I would have preferred a Climax over the Shay.

 

 

I had an interesting read about the reliability and effiency of Steam locos when the first diesels became avaiblable.   The read stated that in the1940s steam locos were reaching their peak in perforamce.  With new technology being developed, they were likely to have continued to beating diesels in both areas.

To top that, diesels spent a lot of time with early rebuilds.  They did have the advanage of a quick startup and it also took less time to train engineers.

The few steamers that came to the USA from China over the last few year are now used in the coal areas of the USA.  They burn coal but without the smoke of the past.  The technology that was put aside is still catching up with steam by improving performace.

 

Lately I watched a video on the end of steam and the coming of the diesel.  I don't remember exactly, but it seems that because the F3s did wear out so fast, the F9s and Alcos was the  replacement for those earlier E-units.  With the vesitility of design, different gearing and changable power units, there are still F3s running today on short line Dinner trains, etc.

 

I believe the use of diesels was more of a political move to reduce polution and increase oil production.  The ease of operation was a minor plus. 

[outofmage]

Distance on each leg needs to be the same for the right hand rule to always work. Otherwise the trains will almost always take the shortest leg. 

The rule is  "when facing an intersection, if both routes lead to the same destination and with same length, trains prefer diagonal track than straight track"

[Gwizz]

With full size trains using single track, the right hand rule mean trains take the track to the right to avoid head-on crashes on sidings.   In RT2 I've never experienced a head-on encounter because one train turned left.

I simply count cells to chart a route.

 

Then again the track in a diagonal cell is longer.  I'll have to watch to see if some trains do turn left and are possibly matched by another train coming in the opposite direction also turning left.  

 

 

Interesting.

[outofmage]

Gwizz, if you use double oversea bridge, you will see head-on encounter happens because a train try to run on the left side bridge. This happens because some trains have been set way points to run on right side, sometimes you just forget to do it. To keep things ordered, you will have to put more carefulness and effort every time you add/change routes (to check if they pass a left side prefer double oversea bridge). 

 

If you just leave it and let the trains always run on the left side bridge, trains won't encounter head-on blocking. But intersection blocking will happen when they return to the main line from the left side. 

 

This is an example of intersection blocking when trains try to go on the left side bridges.

 

To measure which way is shorter is very hard. Because diagonal track and straight track has different length in route distance calculation. It makes counting cells a little complex because you need to keep in mind how many diagonal cells and straight cells there are while counting. This way is practical when the track overall is not long. To measure longer routes, I use parallelogram's property, the opposite sides have same length, with some imagination lines and the help of coordinates, it's easier to tell which way is shorter.

 

I'll have to watch to see if some trains do turn left and are possibly matched by another train coming in the opposite direction also turning left.   

 

I don't know what were you trying to accomplish, can you explain it more detailed?