Testing Trainmaster

General discussion about TrainMaster
User avatar
RulerofRails
CEO
Posts: 2061
Joined: Sun Dec 08, 2013 1:26 am

Re: Testing Trainmaster Unread post

Sorry that this one reminds you of a bad time. Condolences for then and now. (**!!sry

To address your previous comment, the only thing that the real count for Iron/Steel should practically prevent is taking Iron or Steel from Warrington station.

If we look at the design of Trainmaster in general, small amounts of cargo are needed at many industries all throughout the map. For example the Machinery needed to make Glass (1920 onwards). So, in practice we could see at least a dozen cities on a TM map where just letting the train run on "Any Cargo" consist will disrupt reliable production.

If you see Gumboots' acronym "HAB", that's short for "Haul Anything But." This is a "wish-list" feature that many of us wish the game could do. So we would say that in Warrington station the train shouldn't pick-up any Iron or Steel. Or at the 20th century town with a Glass Factory it shouldn't pick up any Machinery or Chemicals. But it could pickup anything else (that will make a profit). Unfortunately such a feature does not exist. :cry:

In 1.05 custom consists are for-profit. Without HAAL we could do this:
Custom consist 1.05.jpg
Additionally in 1.05 the industrial model is a lot simpler so there are a lot fewer cases where we need to even consider such a measure.

In 1.06 and TM, the HAAL coding "hi-jacked" the custom consist setup. If we used the above consist, the train will haul that cargo no matter what, without considering profit. In the case of supplying industries, we don't care if the train is running at a loss (as you said this was what HAAL was intended for). But it disrupts the normal profits from distribution of consumer products, like Food, Goods, Furniture, etc.

So how do we manage these things in TM? The following is the philosophy I have developed. It may not fit with all play-styles.

In the 19th century there is a lot express in TM, including Troops. Lean into this. Most connections are worthwhile if you only run trains set to "Any Express."

TM is all about industry. In the 20th century, especially mid- to late, lean into industries. Industry revenues are high.

Think about cities individually. In a 19th century game I have probably "express" connected all the map. That means I connected the cities for the express revenue. I will just use "Any Cargo" trains for this initially. Later when I am starting some industries up, I will add a prefix to the station name for example "=Crewe" to denote that trains stopping there should pickup express only. At the same time I will update the train roster for express only orders at the trains visiting that station. It's important to keep trains organized to find them easily. I just renumber them so that trains on the same route are consecutively in the list. (See this topic on that: https://hawkdawg.com/forums/viewtopic.php?f=5&t=4133) The freight trains are generally specific.

I build small stations for deposit of resources to industries/ports. The main station is for pick-up. The purpose of these is to drop the cargo on the correct economic cell to feed the industry directly.
Industrial depot.jpg
Maybe there is a better way to illustrate. Let me know if this shot is unclear.
I try to buy an industry that I take all the trouble to "feed." This means that freight profits are less important.


How to run distribution of consumer cargoes from these cities?
If resources are plentiful and we don't mind wasting a few we could set some our trains with a custom consist like this:
Custom 2 freight, 5 express.jpg
Or we can specify them exactly and bring the loss-making feature into play. Remember that keeping a price equilibrium means probably a low volume or infrequent visits (see final option).
Custom, Goods, Meat, 5 express.jpg
In a high-volume situation we can use the main page to set the whole train for a particular cargo type. We could also do this with an infrequent setup, the train will take a longer route with intention that it takes about once a year to do a full "trip."
Goods only.jpg
Can you see why TM has a reputation for being "micro-management" heavy? It's true. Personally I found planning around cities as a whole rather than individual trains keeps the brain from frying with complexity.

Something that will also make a big difference (if you can build industries) is pairing similar industries on the same economic cell. As an example: Oil Refinery/Plastic Plant, Or Electronics/Machinery Plant. Building on the same cell as a resource and especially the agricultural communities which have multiple outputs cuts down on haulage needed. For example Paper Mill/Cereal Company pairing.


About Warrington Wire:
My memory is a bit hazy, but I think the GV4=1 event is for the express haulage share your company has versus the AI. It should be (Loads of Express you hauled/Total Express hauled by all companies) in the current year. Maybe there is some error that I didn't catch when testing the scenario. Can you describe in a bit more detail what is happening?

FYI, I felt that the original scenario was a bit boring in the end-game. You setup your Iron/Steel haulage, then you were sitting and waiting to gather the needed points. When fixing the coding errors I also took the liberty to add the express haulage share and industrial profit elements. The hope was that the player wont get bored later in the game.

PS. Here's a patched Meat Packing Plant to go with the files I already shared. The Hides recipe had a mistake: input amount one byte off. It would make Hides even if it wasn't supplied with any Livestock! I thought I had fixed this previously and shared it, but I can't see where now. Here's one I did:
Attachments
Meat Packing Plant.zip
(365 Bytes) Downloaded 112 times
User avatar
Gumboots
CEO
Posts: 4813
Joined: Mon Aug 13, 2012 4:32 am
Location: Australia

Re: Testing Trainmaster - HAAL Unread post

Grandma Ruth wrote: Tue Jun 14, 2022 11:19 amI must admit that HAAL was more to do with industrial production than anything else, haulage targets didn't really figure much.
I gathered that, ^**lylgh but it's fairly crude way of handling industry supply, and as pointed out the supplied cargoes will tend to drift away unless you use other techniques as well. If you do use the other known techniques, you usually don't need HAAL. Then there's the problem it creates with price islands and exploiting them, which led to those being disabled in 1.06, which led to other problems.

All up, I think including HAAL was a mistake that came from people expecting RT3 to be RT2, in terms of how industries handled supplied cargoes, while not understanding the rest of RT3's dynamics. TBH I have wondered how difficult it would be to go through old betas of 1.06, find the place where HAAL was added, and get rid of it.

But there is a built-in downside, don't forget - whenever you ship something at a loss, you're losing money by definition; so you have to compensate for that somehow in any reasonably realistic game.
Unless the loss is massive it should easily be offset by the resulting industry profit, so overall not a loss (which is why you would do it, of course).
Grandma Ruth
CEO
Posts: 1237
Joined: Mon Nov 13, 2006 7:17 am
Location: West Yorkshire, England
Contact:

Re: Testing Trainmaster Unread post

RulerofRails wrote: Tue Jun 14, 2022 4:38 pm If you see Gumboots' acronym "HAB", that's short for "Haul Anything But." This is a "wish-list" feature that many of us wish the game could do. So we would say that in Warrington station the train shouldn't pick-up any Iron or Steel. Or at the 20th century town with a Glass Factory it shouldn't pick up any Machinery or Chemicals. But it could pickup anything else (that will make a profit). Unfortunately such a feature does not exist. :cry:

In 1.05 custom consists are for-profit. Without HAAL we could do this:
The attachment Custom consist 1.05.jpg is no longer available
Additionally in 1.05 the industrial model is a lot simpler so there are a lot fewer cases where we need to even consider such a measure.
Oh I agree, HAB would be a great feature. But you can still do the same profitable custom consist in 1.06 and TM. I have used the consist screen like this:
Custom Consist.jpg
If you pick out only profitable cargoes, the train will make a profit - but it is a time-consuming job to do if you were to do it for every run for every train. A judicious mixture of express, random cargoes and custom consists both profitable and unprofitable is what I find works. I also move the trains about on the list: usually the "any xxx" ones at the top so I can more or less ignore them and concentrate on the ones I need to change frequently.

"Pairing industries" - what a good idea, I never thought of that. When you can build industries, I have always used the "trick" of placing them near the resource. It means you don't make haulage profits, but you do make the industrial ones. On one map (can't remember which or even what version of RT) I had about six textile mills, upgraded, just churning out profits whilst I played with the trains!

I've started a new thread for "Warrington Wire" with a spoiler alert in case anyone else should be visiting the forum.

@Gumboots:
Gumboots wrote: Tue Jun 14, 2022 4:54 pm 1. Then there's the problem it creates with price islands and exploiting them, which led to those being disabled in 1.06, which led to other problems.

2. TBH I have wondered how difficult it would be to go through old betas of 1.06, find the place where HAAL was added, and get rid of it.

3.
But there is a built-in downside, don't forget - whenever you ship something at a loss, you're losing money by definition; so you have to compensate for that somehow in any reasonably realistic game.
Unless the loss is massive it should easily be offset by the resulting industry profit, so overall not a loss (which is why you would do it, of course).
1. I don't understand what people mean by "switching on" or "off" the price islands. I have got two version of TM running now, one with and one without, but what's the virtue in keeping them at all? If they skew the game, don't we just need to incorporate the patch that corrects them?

2. Yes, I should think that as one of us made the HAAL, one of you could "un-make" it! Unfortunately I'm not much use to you in that. I was part of the original 1.06 development team, but only really as a representative of the "end-user", the player.

3. Yes, of course!!!!!!! !facepalm!
User avatar
Gumboots
CEO
Posts: 4813
Joined: Mon Aug 13, 2012 4:32 am
Location: Australia

Re: Testing Trainmaster Unread post

Grandma Ruth wrote: Wed Jun 15, 2022 3:38 am1. I don't understand what people mean by "switching on" or "off" the price islands. I have got two version of TM running now, one with and one without, but what's the virtue in keeping them at all? If they skew the game, don't we just need to incorporate the patch that corrects them?
Price islands don't "skew the game" as such, in the sense that they don't make it worse, or more easily exploitable. One virtue in keeping them is that they greatly reduce the AI spamming of trains. They also mean that human players have to give things a bit more thought.

The catch with having price islands + HAAL is that you can use that combination to massively exploit price differentials. Hauling from a low value area to a significantly higher value area will automatically drag the higher value area down to a very low price for the relevant cargo, after which you can obviously make massive profits by stack/dump tactics. This was recognised early on, so the 1.06 code was edited to get rid of price islands (ie: turn them off). This means AI in 1.06 is more stupid, and re-hauling is easier for human players too.

So it comes down to choosing between a better economic algorithm overall (prices islands "on") but with the opportunity to exploit it via HAAL, or a more simplistic/dumber economic algorithm (ie: price islands "off") that is less exploitable in some ways but more exploitable in others.

2. Yes, I should think that as one of us made the HAAL, one of you could "un-make" it! Unfortunately I'm not much use to you in that. I was part of the original 1.06 development team, but only really as a representative of the "end-user", the player.
It would be an interesting exercise, but I think tackling it would require access to a good array of 1.06 betas and as many of Milo's notes as possible. Unfortunately, I don't think either of those things are available. If Milo ever comes back and shows a renewed interest in the game, it would probably be possible then.
Grandma Ruth
CEO
Posts: 1237
Joined: Mon Nov 13, 2006 7:17 am
Location: West Yorkshire, England
Contact:

Re: Testing Trainmaster Unread post

Last we heard of Milo he had a new girlfriend who kept him away from RT3 !!!!!!!! Probably is the father of a flourishing family by now.
Thanks for the explanation, it takes a great deal of skill and understanding to explain things to non-experts in terms they can understand, without being patronising, and you have explained it admirably.
User avatar
RulerofRails
CEO
Posts: 2061
Joined: Sun Dec 08, 2013 1:26 am

Re: Testing Trainmaster Unread post

Question here: does anybody know if there is a good reason that some cargoes like Bauxite appear way before their first use appears?

From what I am seeing, looks like the cty file has the final* control on when a cargo shows up. Bauxite is set to 1800.
*Another requirement: building that produces it (could be a seeded port/warehouse) must be active.

To me Bauxite in 1800 looks like a clear mistake. The Bauxite Strip Mine itself isn't available till 1886. (I know there are inconsistencies with start of Aluminum chains and the PDF doesn't help.) First industrial production of Aluminum was actually 1856, but 1886 is a good fit because of better process developed then.

But this isn't exactly the question. Chemicals appears in 1800 also. But there is nothing to produce it till the Oil Refinery in 1856. However, some Chemical additions to chains, for example in the Paper Mill exist from 1800 as well.

This is all a little bit confusing. I'm just putting this here in case someone knows anything. Did someone (Ned?) want early availability for some cargoes at a warehouse/port? Can there be any other sensible reason for cargo to show in the goods list but be impossible to produce with standard buildings? Maybe someone can enlighten me. :-)
User avatar
Gumboots
CEO
Posts: 4813
Joined: Mon Aug 13, 2012 4:32 am
Location: Australia

Re: Testing Trainmaster Unread post

1800 is the default date for cargoes, and having them set to that causes no harm. So, my guess is he just set them to 1800 because he couldn't be bothered thinking up another date at the time, or editing extra files, then just let production be controlled by the date the relevant building appeared.
User avatar
RulerofRails
CEO
Posts: 2061
Joined: Sun Dec 08, 2013 1:26 am

Re: Testing Trainmaster Unread post

Maybe it causes "no harm", but it's also not the normal behavior. We see Bauxite and Chemicals in the UI from 1830, and in the case of Bauxite a solid red (aka useless) price map for them for 50ish years.

TM is admittedly not "very concerned" with a clean-look UI in other cases. I made the edits to enforce a minimum consumption level to alleviate low demands, and at the same time fixed some other obvious issues with the industries etc.. I'm inclined to match the Bauxite and Chemicals with year produced and eventually put all the files in a "pack." But I thought I would ask because in the case of Chemicals, there is an early use for them, just no supply. Hence why I asked for some feedback on if anybody had or thought about using a port/warehouse for early supply. I would assume that nobody had, but you never know and Grandma Ruth was around in the early days of TM.
User avatar
Gumboots
CEO
Posts: 4813
Joined: Mon Aug 13, 2012 4:32 am
Location: Australia

Re: Testing Trainmaster Unread post

Fair enough about cleaning up the UI. Matching cargo start years to production building start years makes sense in that regard.
User avatar
sbaros
Conductor
Posts: 256
Joined: Sun Nov 15, 2020 1:59 pm
Location: Inside the 9th car

Re: Testing Trainmaster Unread post

RulerofRails wrote: Tue Apr 26, 2022 3:00 am Small demands of say 0.03 or 0.05 per year as one of the ingredients on a multiple-input industrial recipe.
Such microscopic demands are too weak to draw or even keep cargo via price demand.
Such microscopic percentages might be preferably represented as "demand only", not included in a product's recipe.
Again, these microscopic quantities might be better limited to exclusively serving purposes of specialised scenarios (for instance timely delivery of components for nuclear weapons) and not included in generic production models. As I have written elsewhere, almost each scenario requires its own, customised Tycoon installation with the intended economy models, rolling stock, architectural styles etc.
Apart from such special cases, I consider it sufficient to cover a minimum of two thirds of a product's final composition by order of significance, including packaging (by weight as well as by volume, since we are interested about carloads after all). The required fuel should be added, especially if it exceeds 1 carload per 3 carloads of finished product.
We must also consider that the obligatory grouping of every possible transportable object into just 52 inclusive categories is not so favorable towards excessively detailed breakdowns. And perhaps rightly so, since a single person is supposed to manage a -hopefuly vast- economic empire.
However, this particular discussion is still premature. There are serious fundamental errors in the production recipes of the official Tycoon distributions which were carried over to the later user-developed versions (including Trainmaster). It is impressive how such grotesque mistakes were made on easily accessible encyclopaedic knowledge. At the bottom line, it counts more to provide a realistic economic model than to faithfully reproduce a locomotive's valve gear. I intend to document this more analytically with figures when spare time allows me.
It will require some considerable effort from our part to correct this situation. Until then, I consider every minute spent on running any Tycoon scenario a waste, since all current versions are rubbish in terms of industry simulation.
If you have no Marxists in the leadership of your trade union, you have no trade union.
Abolish NATO and the (Na)zionist state !
User avatar
Gumboots
CEO
Posts: 4813
Joined: Mon Aug 13, 2012 4:32 am
Location: Australia

Re: Testing Trainmaster Unread post

sbaros wrote: Wed Jun 21, 2023 1:04 pmAt the bottom line, it counts more to provide a realistic economic model than to faithfully reproduce a locomotive's valve gear.
A completely realistic economic model is not feasible within the constraints of RT3. You will always need to compromise to some degree. The real goal is a realistically playable economic model, that allows good scope for a range of scenarios.

And (IMO) figuring out how to fake valve gear on RT3 choofs is fun. :mrgreen: But again, it requires compromise. The game's limitations mean it's impossible to create truly accurate valve gear.
Post Reply