More industry of the same type

[Myrth]

Since I didn't find any topic devoted to this subject, I herebly start one about RT2. The question is pretty simple. Does more industry of the same type affect production or demand by any amound?

 

And to elaborate. Let's say a station got a textile mill. It requires wool or cotton to produce goods. It generates certain demand for those two things. When we deliver either of them, they are converted into goods, in 1:1 scale. Now what happens if we have two or more textile mills in the same station? Production is doubled? Demand is doubled? The second building is just for show? And what if I bring only a single shipment of cotton, while there are two mills in need for it? Which one will get the load? Or both will?

[MaglevForever]

This is what I know.

 

Production is not affected.

 

Demand is affected. When multiple industries are present the maximum level of demand will be higher and demand will recover more quickly. When two identical industries are present demand is not doubled, my estimate would be around 30% increase in recovery time and demand 2 levels higher. Obviously maximum demand level is 9 so you can't get any more. There are differences in demand among industries, less valuable cargo and industries that require multiple inputs may have problems with demand.

 

You asked about which textile mill would receive a single shipment of cotton. I don't know how to tell which industry will receive the cargo, but if you bought both mills you would find that they share the revenue very unequally. One mill will receive the lion's share and the other a dribble. I wait until one is lucrative and then buy that one. If I am washing those mills with wool and cotton I would consider buying the other one also. It might push 8% annual return if I'm really lucky.

[Myrth]

From my own observation, the demand for certain amound of time is not affected in case of gaining 2nd building by city growth. But I don't know if that's till the end of the year or is simply an effect of already covering whole demand that year, so there couldn't be bigger demand at that moment, even after gainging new building. Or something entirely else.

 

So anyone know how it's estimated which building get your shipment?

[outofmage]

Well, actually it's very hard to measure how quickly demands recover in a normal game, because you constantly supply the industry, and it's very hard to remember how much and how frequently you supply them. So if you see a demand level remains low even after another industry appears, maybe you just over supply it too much or you simply hasn't given it enough time to recover. 

 

What you should do to measure how quickly a demand recover is to create a map with one station covering one texitell and another station with 2 texitells and supply them with 4 cotton at the same time and see if two station recover demand at the same speed.

Here is what I got : the one texitell started on 5 demand, two texitell started on 7 demand. After I delivered 4 cotton, one texitell station dropped to 3 demand and two texitell dropped to 6 demand. Then I stopped the train. After 4 months, the two texitell station recovered to 7 demand while one texitell station remained demand 3. After 10 months, the two texitell station demand level raised to 8 and one texitell only just started to recover to demand level 4. 

 

So the conclusion is obvious: more industries lead to more demand.

[jeffryfisher]

What's really odd is that building a second station next to one factory will allow you to sell twice as much with the same impact on demand (i.e. the demand at each station in completely independent of the activities at other stations).

[Gwizz]

If I remember correctly, Pop Top said the industry closest to the station gets served.

The industry furthest away from the station will  hold its' demand high since it is not served. 

I have placed a small second station close to that industry not being served and was benefited by the high $ demand at that industry.

[outofmage]

I agree with Gwizz. I noticed that too, but not sure, if Pop Top said so, I think that's the deal. Also, similar things happens on supply industries. Cargoes from farms closer to the station will be picked up first and leads to better profits for them.

[Myrth]

What's really odd is that building a second station next to one factory will allow you to sell twice as much with the same impact on demand (i.e. the demand at each station in completely independent of the activities at other stations).

 

So you are telling me that I can build two stations next to each other, covering the same buildings and they won't interfere in any way with supply and demand of each other? I never tried that one, being afraid that there could be some problems with determining which station gets what.

 

Does the same rule applie to supply industries? If yes, my whole life was a lie...

[jeffryfisher]

I previously overstated station independence. It only applies to demand for deliveries at each station. I did not mean to imply that pick-ups were also independent. On the contrary, when any train picks up any (initial, not drop-shipped) cargo, that cargo vanishes from the building that generated it, which in turn makes it disappear from all stations serving that building.

 

Therefore, building a station is a demand doubler, but it's not a supply doubler.

[Myrth]

Still sounds like a good way to create additional routes for passangers if there is a sulprus of them.

 

Thanks a lot for help

[loco_motive]

One station to increase demand also works for different industries using the same inputs, although that can require some micromanagement. I find multiple stations are better used for traffic control, especially when all the demands have likely been reduced to zero anyhow.

 

As far as who gets the delivery, playing on Expert, if you deliver all the wool to the one station but build two other stations covering one mill each, you can see that they both produce goods, and even without the stations, they clearly both will become profitable eventually. Do it for a long time and see what statistics evolve. It would also be interesting to see the results for grain to bakery and distillery, with both cases of which one is furthest away. Steel to T&D vs Auto, which requires 2 inputs, would also be interesting. Also one might question whether results would differ if the wool was delivered one at a time or 6 at a time.

 

The profitability of an industry, on the other hand, is strictly a function of its own supply and delivery to a demand site. Station demand level and the distance shipped are irrelevant. Depending on various things this can range roughly from 8 1/3% for supplying one item to a 2-demand industry to 25% for a fully developed delivery system for the same industry. Etc.

 

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

 

Here is a demand anomaly that has to do with ports, at least in the TSC campaign scenario #18 Antarctica Rising.

 

Coal, as the first of 4 demanded cargos listed at the ports, gets a small bonus in demand for a station covering one port (4,3,3,3) but a huge bonus in demand from a station covering 2 ports, where the other items don't get any increase in demand at all (9,3,3,3), and it sustains its demand level very well. I tested this by switching iron and coal, but have not tested this in a completely different scenario.

[j7n]

I generally build 2 stations around Steel Mills, an In and an Out, so that trains at the Out can start up without yielding to an empty train that just delivered coal or iron. Most other kinds of industry are usually served by the one train. I do sometimes take advantage of separate demand to drop off some coal at the Out station.

The problem with demand at Ports is a general problem with secondary demand in industries. That demand is low and does not recover. Jeff's Mod fixes most industries, but with the case of ports, I think that demand is recorded in the scenario file. The Antactica scenario is rather difficult because one must earn enough money to hurt or buy out competitors, without profitable industry. The last competitor stubbornly refused to be liquidated, and I purchased him with all his bonds.

[loco_motive]

Did a small test anyone could easily do of deliveries of grain to 2 bakeries approx 2 and 8 units from the station.

 

First delivery; closer got 3, further got 3. Second delivery; 1 & 5. Third delivery; 0 & 6, followed by 3/3, 2/4, 1/5. An apparent 26:10 preference for the further bakery.

 

But wait a minute! Since when does anything in this game work sensibly? In the first test, the further station had the lower y-axis coordinate. I reversed the situation so the closer station had the lower y-axis coordinate and guess what? The situation reversed and deliveries were distributed 6/0, 5/1, 5/1, 6/0, 3/3, 3/3. An apparent 28:8 preference for the closer bakery.

 

I would formulate that, for 2 bakeries, the one with the lowest x coordinate, then the lowest y coordinate in case the x coordinates are the same, gets at least half the grain, the other one gets the rest. [Ed.]

 

Now I guess if grain was delivered one at a time, it would always go to the same bakery, but that's untested.

 

This follows somewhat my theory of successive barrels of water in this game, when the lowest ordered barrel is sated, water starts running into the next barrel, etc. in some manner. Something similar happens when a huge increase in supply occurs where there are several AI stations; the lower numbered company buys a s***load of engines for starters. Also something similar happens in train-filling order when several are waiting in the same station

 

Well, who wants to pursue this any further.

 

BS is where you find it.

[Myrth]

So the game is not using any particulary spectacular not even factor-based way to determinate who gets what, but only use the position on the grind? Are you sure about that? When I tested this tonight, the results were rather random, being more affected with distance from station than the position on the grind. And yet it was still pretty much random.

[loco_motive]

Going one step further into the murk:

 

In distributing 6 grain per load between a distillery and a bakery, the bakery was favored about two to one, based on 48 runs and 288 deliveries.

 

Note that food is ID# 10 compared with ID# 35 for alcohol, probably significant. The distillery got more grain than the bakery only 5 out of 48 times. In this case of mixed industries, neither the position of the industry nor the station nor the economic state appears to be relevant but perhaps just the ID # of the manufactured product.

 

Test 1. Food 7 squares away, distillery 2 squares away, everything shipped, 8 trials: 30F, 18A.
Test 2. Food 2 squares away, distillery 7 squares away, everything shipped, 8 trials: 32F, 16A.
Test 3. Food 2 squares away, distillery 7 squares away, nothing shipped,    8 trials: 28F, 20A.
Test 4. Food 7 squares away, distillery 2 squares away, everything shipped, 8 trials: 37F, 11A.
Test 5. Like 1 except economic state at 4 throughout: 31F, 17A
Test 6. Like 1 except economic state at 0 throughout: 33F, 15A

 

Here is the distribution respectively of 6 5 4 3 2 1 0 food per load: 5 9 19 10 5 0 0. This is approaching some kind of statistical

distribution which they could have easily applied.

 

So in any case these results for these runs are 191 food, 97 alcohol, very close to 2:1, but there's a fair variation shown. For

example, test 2 had a run of 5 3-3 combinations and test 4 food had all 4's 5's and 6's.

 

I wonder if the grain from the 6 fields I planted is picked up in any order.

 

I wonder what happens for steel to a T&D, steel mill and cannery.

 

Anyhow, it seems like its always going to be better to maintain control of resources through the use of separate stations as the

game progresses. Then you don't need to know any of this stuff, but you still may need to continually bull-dose that pesky T&D that

keeps popping up and stealing the iron from your steel mill, or the steel from your auto factory, or build some track onto that location.

 

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

 

In view of Myrth's remarks about my previous post, I think it might be a good demonstration of how careful you have to be in doing this kind of statistical analysis, drawing results from such a slim sample. It was stated as just a hypothesis though, based on those slim results. I do, however, continue to feel confident that the bakery with the lowest x, then lowest y coordinates gets the most grain, just not as much more as the previous results show. For example, I went back and did 16 more runs and got a 54/42 split for the bakery furthest away with the same x but lower y-coordinate. It included two 2/4 splits, unseen in the first tests. This combines with the 26/10 and 28/8 of the first tests to produce 108/60, getting closer to the 2:1 range  observed above.

.

[Myrth]

And I would like to apologise. I've run more careful tests once again and their results more or less matched your pattern about grind relation. I've also make sure to check it with other kinds of industry than bakery. You know, just in case if bakery was affected by grind position and let's say umber mill not. But nah, the grind seems to work for every kind and type of industy in the game. I didn't check ports, mostly because their supply and demand depends from map to map, but let's hope they also use grind position.

 

I always wonder in such situation how does the developing team feels after all those years, where there are still players who devote quite a lot of time to crack how does exactly the game works.

[jeffryfisher]

"Grind"? Do y'all mean "grid", as in {x, y} Cartesian co-ordinates?

 

Allocation of delivered loads should follow demand, not grid position. I suppose I should add a wish to the wish-list thread...

[outofmage]

Jeff, you must have noticed that the profitability difference between the same type of industries within a station's radius. If delivery follows demand, they should get evenly supplied and equally profitable. Same thing happens to farms and mines, closer they are to the station, more lucrative they get. I think this mechanic is reasonable, industries far from a station makes it more expensive and takes longer to transfer their goods between their properties and train stations. So they have lower priority when you supply/carry away goods to/from them. 

[loco_motive]

For deliveries to an industry, as seems pretty clear to me from my testing details above, that distance from station is not a factor. I don't understand the basis for denying this.

 

For things like grain and cattle, I suspect that they are picked up in the order they are produced, as they would also be assigned at that time an initial value for rot (deterioration of sale value over time) and days to deliver (length of time it will survive in the station). So efficient pickup and eventual industry profit if delivered to a demand location would have nothing to do with location within the station's zone. If it can even happen that 2 items can be produced at the same time, there may be tie-breaking rules if only one of them is picked up. But that is going to second order precision, not really necessary.

 

Do the testing to prove it's wrong, but don't just say it's wrong in the face of evidence to the contrary.

[outofmage]

I did some tests today, and I think it depends on RRTII versions. 

Distance from station was a factor in RRTII original version(1.50) and TSC(1.55), but not in Platium Edition(1.56).

(different versions have slightly different mechanic in this issue)

 

I create identical maps from scratch in each version of the game. Build a station covers 2 meat packing plant, one is closer than another to the station. Build another station to cover houses and cattle yards. Type cheat "overtime", use a train to haul only 6 loads of cattle and 6 food back in the first year. Buy the meat packing plant after annual report to see their profits.The result is: in v1.50 and v1.55 (Chinese version), the close to station one makes $44K, the far away one makes $12; in v1.56, however, they both make $38K.

 

They really twisted some small things that they thought not important enough to be listed in the History file. It's quite insteresting that we make this revealed. 

This is a link to download my tests:

https://dl.dropboxusercontent.com/u/77269283/industryTests.zip

 

[Edit]

I am sorry for not reading your posts completely loco, but I do have some spare time currently. I get your whole idea and I will do your "successive barrels of water theory" in more details. This theory is indeed all over the game. For example, it happens in AI buying up the oversold stocks too. When you dump a lot of stocks into the market,  AI will buy them up to the price they thought to be reasonable. It may takes several AI's purchase power to bring up the stock price, and they do this in the order of id. When one computer maxout his leverage, next one fill it in, some only use up their cash, some use up all their purchase power depending on their style until the stock price bounce back to its "true value", left with the last AI not buying any share. This also happens in a 32-AI game, when there are limited cities to be connected, the front-end-id AI will start companies, the middle-end AI invest in large, the last-end just do nothing with their 100K cash.

 

However, we need to confirm the "successive barrels of water" theory in this industry demand scenario by tests. we need to know the capacity of each barrel(the yearly demand for an industry). By controlling the throttle to slow down the train or reduce the loads for each run, we can figure out how much loads each bakery needs to fill up its capacity. And if we under-supply them, the water only falls in higher barrel, nothing into the lower barrel, if we over-supply them, water will fill in all barrels and overflows into somewhere.

 

I do this in 1.56 version to reduce complexity, after all, the other versions are out-dated. I also use grain-food industry as loco did. One station covers grain silo, the other covers 2 bakery.

 

Here is what I got from my tests: 

I set 50% throttle for 3 loads of grain and food back for 20 years

closer get:29

further get:50

 

20% throttle for 2 loads of food for 20 years

closer get:21

further get:34

 

10% throttle for 1 load of food for 20 years.

closer get: 11

further get: 31

All the tests were running on "nowreck" cheat, and on stable economy status. 

 

I don't want to continue. This already proves that the "successive barrels of water" theory doesn't fit in, no matter how few of loads you deliver each year, even less than 1 load per year, there is always a few loads slip to another one and the final delivery ratio is close to 1:2. I can't explain this, maybe it's a mix of demand level of each industry and the location of them lead to this 1:2 ratio. Nevertheless, it does prove that one of the 2 same kind of industry has higher priority than the other.

 

Besides this, I believe it's not coordinates that decides which industry has higher priority, it's by the order of somehow un-shown and un-known industry id. Trains have id, companies have id, players have id, I think industries have id too, although we never see them in games or editor. 

In loco's theory, lower x and y co-ordinate have higher priority in getting served. I think it's because he put down the industry in the editor by the natually habit of putting them from left to right and from top to bottom. By accident, this makes industry located in lower x and y co-ordinate in the front of the id list.

To confirm this, I build two bakery in same places but in different order in the editor. I discovered that the bakery being built later (no matter where it is) has higher priority in getting served.  This is the test I did for you to have a look, you can look at the industry profits to see how each one gets served.

https://dl.dropboxusercontent.com/u/77269283/fd-testz80-1.GM2

It doesn't make sense why it's not the earlier one being built has higher priority, well, at least we have an answer, and it makes this study at least not completey useless. Now we know if a same kind of industry building pops out in a city we already serves one of this kind of industry very well, we need to buy this new industry, because it has higher priority and will become more profitable than the original one.

 

Sorry, it's quite long. To sum up:

1. When deliver to a station covering 2 same kind industry, one of them always get a lot more loads than the other. 

2. I think each industry building has an unique id just like other elements of this game.

3. Industry id has some effect on deliver priority, not the distance, not the grid position.

4. Industry that appears later has higher priority.

[jeffryfisher]

...Unless the game recycles an ID vacated by a deleted building. Then a new building could acquire a low index.

[Myrth]

Run few tests in my spare time during night shifts, just to confirm or disown theory about recycling ID (or whatever else decides which industry is more "fresh"). Using "viagra" cheat and town spawning only bakeries, I was monitoring which bakery get the grain. It's always the new one that pops out after growth of the town, so the IDs are not recycled. I repeat - the IDs are not recycled. Buldozing only leave an empty ID, but new buildings get new numbers. Or at least they are not recycled in 1.56 Platinum

 

Thanks to outofmage I now understands why in my original test of loco's theory I get different results with grid position. In my second run of tests I've putted the industry in order described by outofmage, so the results covered what loco stated. Now we at least know why it showed that grid position is important.

 

And yeah, I meant grid, Jeff. I simply misspelled it ^^

[Gwizz]

One extra point.  The AI RR always seems to pickup and haul the most expensive cargo first.  The AI has a little advantage in that regard.

 

With my play style of having a number of Railroads under my control at the same time,  I have to assign cargo with that in mind. 

I have found it easier to just serve one type of cargo at each station.  This means I normally use small stations for cargo pickup.

[Bossatchal2]

On 10/28/2013 at 11:02 PM, jeffryfisher said:

What's really odd is that building a second station next to one factory will allow you to sell twice as much with the same impact on demand (i.e. the demand at each station in completely independent of the activities at other stations).

Something funky is happening when I am building my second station. Not sure if this is map specific but playing America after the Flood on expert. I build a station at Atlanta. Then I build a second station at Atlanta and the demand for pax drops from 8 to 6 at the main station with no delivery of passengers. The second station only has a demand of 5.

I reload and only build one station at Atlanta and the demand stays at 8.

 

A similar thing happens when I build stations at Patterson. I feel like the stations might be splitting the demand from the houses based on proximity?

 

Moreover, on this map it feels like something strange happens in the year 2031 (profit=740K). I never can make more profit than 2030 (1000K).

 

Does anyone know what is happening here? 

Edited October 10 by Bossatchal2
minor change

[jeffryfisher]

I can't recall seeing that before. I've made a note to check on it (might be weeks before I play though). Let us know if the effect is repeatable, especially on other maps.