Balance and realism issues with RT3 v1.06 locomotives

[Gumboots]

If I ever manage to win the lottery, I'm gonna see if I can buy the rights to the source code and open source the thing. There's too much hidden voodoo and guesswork.

[Lirio]

If I ever manage to win the lottery, I'm gonna see if I can buy the rights to the source code and open source the thing. There's too much hidden voodoo and guesswork.

Making Railroad Tycoon 4 would be a far better use of the money.

[thietavu]

If I ever manage to win the lottery, I'm gonna see if I can buy the rights to the source code and open source the thing. There's too much hidden voodoo and guesswork.

Making Railroad Tycoon 4 would be a far better use of the money.

Well, those things could be combined? :) I can't really tell how AMAZED I've been these years (!) at RRT3 while developing my silly, insanely big and complex "Great China 4" simulation on RRT3 TrainMaster edition... Over 230 cities, over a million trees, over 600 trains easily running, etc etc have revealed to me the greatness of the core AI engine! As result, my East Asia truly "lives" - everything seems realistic, and it really feels like one is managing a huge railroad empire where things live and change constantly. Industries come and disappear, and seem to make sense. Of course, the graphics could be improved, but even as they are now, I find it kind of thrilling to drive through the endless high plateaus of Tibet on some powerful steam engine on its way to Lhasa, seeing the Himalayan mountains in the distance... Or the jungles of Myanmar. Or the fields of Bangladesh. And so on... I never grow tired of finding new ways to build routes to tricky places, or to develop industry in far-away places. :)

So, I wouldn't change the basic concept much, or at all! Just improvements here and there, optimizations for modern hardware, even larger (!) maps (yes, I'd definitely build the Trans-Siberian Railway scenario! ;) ), ... The danger is, you see, that RRT4 would become something entirely different again. Perhaps not a "simulation" anymore, but a simpler, faster, more streamlined, prettier thing the way Civilization 5 was (in my opinion) partly destroyed compared to Civ 4. Unfortunately for players like me, very detailed, complex, time-consuming simulations are out of fashion these days. Maybe people have lives these days? ;)

[Lirio]

The depth and complexity of Railroad Tycoon's economic engine is, to my mind, the core of the game. i do not think a better, faster, more streamlined, and functional iteration of RT need sacrifice that core. There are a huge number of things that could make gameplay smoother without making it shallow and unchallenging. For example a decent automatic track laying algorithm would be a god-send because it would both make connecting cities easier and quicker, and make the AI companies stop laying brain-dead straight track across all obstacles. The option to lay track by hand would still be there for those who enjoy it! Which includes myself, i should add. So there you would have easier and simpler without losing any actual complexity.

For a real instead of theoretical example: Crusader Kings, a dynasty management game that came out a year after Railroad Tycoon 3 had a sequel released for it just last year. CK2 is better than its predecessor in pretty much every single way, being smoother, easier to play, and far more in-depth and complex. It could be even smoother if Paradox could make an intuitive and usable interface to save its life, but c'est la vie.

Anyhow, this all beyond the scope of this thread. We're not likely to get RT4, nor are we likely to get the source code for RT3. Better we focus on what can be done, and at present that's improving RT3 v1.06.

[thietavu]

I agree. As great as RT3 already is, I can't but wonder what longer consists (maybe 10..12 cars), photorealistic graphics, more than 2 tracks, better control over what the trains haul, and (especially) an improved editor could do... (sigh...)

One question would be the balance of RT, the game being both a business simulation and a railroad-building simulation. I never deal with stocks, etc. in RT3, unless absolutely necessary, since I'm not really interested in business - only in trains, railroads and the ecosystem they make. For me, the magic is in creating good railroads between places of importance - and keeping them alive. That goal needs some business decisions, but they are "necessary evils", not the main thing for me. Then again, for many others, the emphasis is probably just the opposite, trains being the necessary evil to do business and beat the others in that. These "opposite" motives for playing RT are so different that if RT4 emphasized one clearly over the other, some of us would probably leave the game.

After all, Sid Meier's "Railroads!" was a warning example... It did many things right. Still, it missed some critical points, like being a credible *simulation*. It was another example of those "light" simulations for people who don't have time or interest in anything deep and complex. And that's exactly why it wasn't interesting enough. Too few locomotives. Far too few, far too small maps. No sense of "real world". Etc.

But now I'm babbling and taking this thread to branch lines, sorry. ;)

One thing about the diesels in RT3: they seem to be pretty miserable in handling steep grades, at least until some monsters become available. This doesn't seem realistic to me. For example: I believe steam engines like the UP Big Boy were abandoned because the diesels offered about the same performance with less costs. Yet, there are very few diesels available in RT3 that can handle really heavy freight over steep grades at all. In a way, it is realistic, since many diesels had much less pulling power than Big Boys etc, but in real life they were used in combinations of 3..6 engines. Just wondering whether it might make sense to change the "twin-diesels" to represent "more than 2 engines -units", giving them the performance and costs of, say, 4 engines? Then a "double GP35" would be comparable, even superior, to a Big Boy.

Silly idea?

[Lirio]

Still steadily working on this mod, doing the 1930s and 40s engines at present. Electrics and diesels are showing up, mostly ignoring them, focusing on steam. Thus far i've made a ton of changes, both small and large. A number of them probably weren't strictly necessary, but i really like to tinker with things. It's possible something in the game balance broke, and i haven't gotten as much playtesting done as i would like. The game still runs though! ^_^;

One thing about the diesels in RT3: they seem to be pretty miserable in handling steep grades, at least until some monsters become available. This doesn't seem realistic to me. For example: I believe steam engines like the UP Big Boy were abandoned because the diesels offered about the same performance with less costs. Yet, there are very few diesels available in RT3 that can handle really heavy freight over steep grades at all. In a way, it is realistic, since many diesels had much less pulling power than Big Boys etc, but in real life they were used in combinations of 3..6 engines. Just wondering whether it might make sense to change the "twin-diesels" to represent "more than 2 engines -units", giving them the performance and costs of, say, 4 engines? Then a "double GP35" would be comparable, even superior, to a Big Boy.

Silly idea?

Not a bad idea actually. It would be a simple matter to rename, say, the GP35 Double to GP35 Multi Unit or GP35 MU, and change the stats to represent multiple engines.

Main issue is figuring out what stats to put in though. As said before though, i'm kind of lost on diesels and their balance points, and for whatever reason it's often harder to get information on them that is convertible into game statistics. The pulling power value for steam engines in RT3 seems to be derived from the real life tractive effort divided by 4000, more or less. Not every value is that, mind, they just kind of cluster around it. The Big Boy and Challenger's pulling power value is their tractive effort divided by 3000, while for the Atlantic and Pacific it's pulling power divided by 6000. Generally high speed engines tend to have their pulling power divided more than low speed engines. This is realistic since most locomotives developed the most tractive effort at speeds below 10-20 mph, dropping rapidly as speed increased. Additionally some values are adjusted up or down for game balance, or to account for the limitations of the game engine.

Anyway, the point is that finding a steam engine's tractive effort is pretty easy. Usually it's just given, and if it's not you can almost always find cylinder bore and stroke, boiler pressure, and driver diameter, which is all you need to calculate it yourself. With Diesels though, no clue. Horsepower is usually the quoted figure, and i don't know how to translate it into actual game stats.

[thietavu]

Hmm... worth thinking, anyway, then, I guess. A GP7 is virtually useless after 1950. A twin GP7 is only "useless x2"". Four GP7s might actually be able to do something useful, like in real life.

One strange notice: there are diesel units from only one US manufacturer in the game (not realistic at all). It doesn't really affect the actual game - it's just strange.

[Lirio]

So, Railroad Tycoon 3 doesn't have any Garrats, which leave the World region deficient in terms of heavy freight engines during the 30s, 40s, and 50s until the QJ comes along. The best solution i can think for this is to simply take one of the big US Mallets and give it World availability as a stand-in for the real life Garrats. Anybody have any suggestions?

[Gumboots]

Sounds as good as any other solution.

[Lirio]

Alright, i think i'll make both the Challenger and the Big Boy available in the World region.

One interesting thing i've learned play-testing the game: acceleration matters. In fact, it matters a lot. The default Pacific is 20 mph faster and pulls about 65% harder than the default Atlantic. The only thing the Atlantic has better is its fast acceleration to the Pacific's average acceleration. i've found that for express trains between cities that are close together the Atlantic does them faster than the Pacific. Though i tend to test with six express cars rather than the maximum eight. After boosting the Atlantic's maximum speed to 85 mph, it turns out that they're Atlantic's just flat out better on runs with very low ruling grades. The Pacific only manages to do the run faster when the line is significantly, or predominantly, hilly. Basically, that extra 10 mph in top speed doesn't matter if the locomotive fails to actually reach it for extended periods.

This is very neat because it offers me a new dimension to balance engines against each other. Though finding the sweet spot often requires actually getting in the game and testing the engines, which can be rather time consuming.

[Lirio]

So, despite ample time and inclination, the last couple of days have largely not been spent doing locomotive research, altering LCO files, and play testing changes. Instead thye were spent diving a lot deeper into the game's files than ever intended. You see, while Bombardier's J1 Hudson is among his earlier works, it's very, very pretty. Bit of a problem with it though: It has no beauty shot and a Pacific's profile picture. It's just something he never got around to addressing, and probably much of the reason why it wasn't included in v1.06.

Well, good news everyone! The time was well spent! ^_^

Below is a picture of the NYC J1 Hudson hauling the Empire State Express to Albany along the New York Central's water level route. You can see that thanks to my efforts it now has both it's own unique profile and a beauty shot. All that remains is to lower its pulling power and it will be good to go!

[arop]

I have waited a long time for someone who could fix it. Another problem is that the loco number and company name is laterally reversed at the other side of the loco. With that fixed it will be among the most succesfull user made locomotives in this game

[Lirio]

Another problem is that the loco number and company name is laterally reversed at the other side of the loco. With that fixed it will be among the most succesfull user made locomotives in this game

Unfortunately, to the best of my knowledge nobody knows how to fix that. It's strange the textures look right on the default locomotives but not on the fan-made ones. Clearly we've been doing something wrong, but no idea what. It's honestly not a major issue, much easier to deal with it than to deal with the wrong profile image. Not being able to tell Pacifics and Hudsons apart on your train list, when you're likely to have both in service at the same time, is rather inconvenient.


Anyway, for bonus points! i've also altered the profile on the Berkshire. Apologies to hoborailcat, but i simply could not stand the sight of the original profile. It looks like a Mikado that rammed into a tank car full of black paint. Now it actually looks like a Berkshire. Admittedly not quite the one hoborailcat made, but a Berkshire nevertheless, and very distinct from the Mikado.

[Gumboots]

I have waited a long time for someone who could fix it. Another problem is that the loco number and company name is laterally reversed at the other side of the loco. With that fixed it will be among the most succesfull user made locomotives in this game

That should be easy enough to fix. It's just a text layer in the tga (or dds, depending on what you're using). There's no obvious reason why the text could not be done as part of a psd before exporting as tga. If it works on the default locos, there's no reason why it shouldn't work on user-made locos.

[Gumboots]

Anyway, for bonus points! i've also altered the profile on the Berkshire. Apologies to hoborailcat, but i simply could not stand the sight of the original profile. It looks like a Mikado that rammed into a tank car full of black paint. Now it actually looks like a Berkshire. Admittedly not quite the one hoborailcat made, but a Berkshire nevertheless, and very distinct from the Mikado.

Nice. It did look terrible before. That was one of the things I was going to fix on the Berkshire, before I lost interest due to not being able to see most of the skins anyway (infamous low res gfx debacle). However, I did get as far as re-alpha-ing several of the higher grade skins for both loco and tender, so they don't glow in the dark. If you want copies of those, let me know.

[Lirio]

That should be easy enough to fix. It's just a text layer in the tga (or dds, depending on what you're using). There's no obvious reason why the text could not be done as part of a psd before exporting as tga. If it works on the default locos, there's no reason why it shouldn't work on user-made locos.

In all honesty, i have no idea what you're talking about. ^_^;

Nice. It did look terrible before. That was one of the things I was going to fix on the Berkshire, before I lost interest due to not being able to see most of the skins anyway (infamous low res gfx debacle). However, I did get as far as re-alpha-ing several of the higher grade skins for both loco and tender, so they don't glow in the dark. If you want copies of those, let me know.

Yeah, sure! If you did the work might as well use it!

[Gumboots]

What I meant by the first bit is that if it's possible to get text flipped correctly on default locos, it should be possible on user-made locos. From a gfx/image point of view, it's trivial (text layers in a psd blah blah blah). If it's not working in the game, that would mean that the game is calling the same portion of the image for the right and left sides, which is a coding problem that should be fixable.

I'll dig out the alpha channels for the Berkshire and do the rest of the skins. The ones I haven't done yet are the lower res ones, so they have far fewer pixels to worry about than the ones I've aready done (which were a fn horrible job). Shouldn't take too much time, so I'll fix them up.

One thing I did notice about the Berkshire was that one set of drivers seemed to be slightly out of whack. You can see it in the first screenshot attached to this post. The third driver is out of line vertically, compared to the other three. Have you noticed this at all?

[Lirio]

What I meant by the first bit is that if it's possible to get text flipped correctly on default locos, it should be possible on user-made locos. From a gfx/image point of view, it's trivial (text layers in a psd blah blah blah). If it's not working in the game, that would mean that the game is calling the same portion of the image for the right and left sides, which is a coding problem that should be fixable.

It should be possible yeah. It's not just the default locos that have it right, a few user made ones also have the text showing up without issue. The Berkshire for example has "Lima Locomotive Works" legible on both sides of the tender. However it seems like even the people who got it right don't know what they did to make it happen.

I'll dig out the alpha channels for the Berkshire and do the rest of the skins. The ones I haven't done yet are the lower res ones, so they have far fewer pixels to worry about than the ones I've aready done (which were a fn horrible job). Shouldn't take too much time, so I'll fix them up.

Pretty sure only the A skin is necessary. The included User Skinning Tools will generate all the smaller images without causing alpha issues. Though if you want to do them all, sure go ahead.

One thing I did notice about the Berkshire was that one set of drivers seemed to be slightly out of whack. You can see it in the first screenshot attached to this post. The third driver is out of line vertically, compared to the other three. Have you noticed this at all?

Just checked it, the third driver is aligned with all the others, it just has a black stain on it that the others do not. It's curious, but not unsightly, just looks like an oil or grease stain, nothing that a real locomotive wouldn't have. Most likely that black stain on the higher res textures is being rendered as a misalignment on the low res texture. There appears to be no issue with the 3DP files.

[Lirio]

Found this in the thread you linked to Gumboots:

We went over this already in one of the other threads. The trick will be to set up a proxy dll file to fool the game into thinking it has an appropriate amount of ram to play with. That means learning Windows-specific coding, whch I will only ever use for this one trick and thereafter have absolutely no use for. Motivation has been somewhat lacking, and other things have been catching my interest.

Nope, you don't have to do any of that. Just follow the instructions here. It should take like 5 minutes total, including the rebooting.

[Gumboots]

One thing I have noticed is that ALL the user-created locos have alpha problems, except for the ones that were manually edited to not have alpha problems. That means WP&P's N&W locos (he really put the time into those) and maybe a couple of others. The rest have more night glow than the main drag of Las Vegas.

The skinning tool is pretty useless if you want a first rate job, and AFAIK will only reskin existing default locos, not create new ones. I wouldn't bother using that tool for this job.

Found this in the thread you linked to Gumboots:

We went over this already in one of the other threads. The trick will be to set up a proxy dll file to fool the game into thinking it has an appropriate amount of ram to play with. That means learning Windows-specific coding, whch I will only ever use for this one trick and thereafter have absolutely no use for. Motivation has been somewhat lacking, and other things have been catching my interest.

Nope, you don't have to do any of that. Just follow the instructions here. It should take like 5 minutes total, including the rebooting.

Ha. In your dreams, mate. Doesn't work on my box. Tried that trick ages ago.