• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Lennartos

BL-Logic
11 Badges
May 9, 2005
1.368
6
  • Hearts of Iron II: Armageddon
  • Crusader Kings II
  • Europa Universalis III
  • Divine Wind
  • Hearts of Iron III Collection
  • Europa Universalis III Complete
  • Europa Universalis III Complete
  • Victoria: Revolutions
  • 500k Club
  • Europa Universalis III: Collection
  • Hearts of Iron II: Beta
This thread is for suggestions regarding the TC usage and supply system of the HOI series.


According to the suggestions this first post will be edited, and the armageddon improvement thread will be updated with the "conclusion" (best solution?) reached here.


It seems that we have found a good solution:
(i count all silent lukers as silent accepts of the idea :D )

General discription:
We need logistical nodes, wich we here will call Supply Centers or SC in short.
All provinces will belong to the nearest SC,therby defining that supplycenters distribution area. (like the Area of influence from trade centers in EU2).
Like EU2/3 that area is flexible and will vary depending on the relative SCs free TC capacity.
Resources will not be send to and distributed from the capital alone, but from and to each SC.
To simulate the inportance of the infrastructure, each resource/supply movement costs TC according to distance and infrastructure.
The general TC linked to IC will be removed and replaced by a system of buildable and assignable(to SCs) TCs.

Each SC has its own TC burden. It will distribute and collect to and from all provinces in its assigned area, and freight resources to other SCs. SCs capabilitys can be improved by upgrading the SC level or improving infrastructure.

Supply centers also limit the stockpile allowed.

Picture of germany with 4 SCs:
DynamicAreas.jpg


gameplay and rules:

Max SC size and max infrastructure level buildable is limited by techlevel.
When occupying an SC it will downgrade permanently.

Trade: trade will go through the SC/convoy system just as every other resource. A new diplomatic option will be added: "Trade agreement", where a cost / TC value can be agreed upon. This allows transporting through neutral land.

The transition from steam/oil based will be made posible with a slider(?).
A steam based infrastructure will use energy(coal) as primary TC cost source.
A Oil based infrastructure will have added oil cost, but have a much lower TC usage cost in low infra provinces.

Gameplay Actions:
Strategic Layer(things that the player should do)
Production of SCs and TCs.
Placing of SCs

Planning Layer(optional automatation)
Prioritys of SCs
Should SC build up the supplies/oil storage?(if yes maybe a desired value)
Should supplies and oil be prioritized over resources?
Should this SC be shut down and/or dismantled?
Should i make a shortcut with convoys over water?
Assigning of TCs to SCs*
Specifying which SC-to-SC routes are valid/possible for each commodity*

Execution Layer (things that should be automated)
From where should i take my demand?
Where should i send surplus?
How much should be send from A->B and B->C?
Assigning of TCs to SCs*
Specifying which SC-to-SC routes are valid/possible for each commodity*
Specifying in which direction goods should flow along these routes
Specifying what order of priority should the different routes have

* = Can be both places depending on implementation

current example topics:
How to model the transition of steam / oil based infrastructure.
How to make optimal routing mechinisms?
How do we balancing the game?
 
Last edited:
Upvote 0
Lennartos said:
We can all agree that constant TC usage regardless of distance to capital is a no go. But how does one model it correctly?
Good concept for a thread - I'll try to keep just to logistics, although the various systems tend to interact somewhat (as they should!)

My two penn'orth, FWIW:
Shoud railways be modeled, and if yes... how to do that correctly?
I don't think it's neccessary to split out railways, specifically, but I do think that TCs should work basically like convoys (though by land) and could be regarded as being mostly rail transport.

Should moving resources cost TC? (making IC placement a TC factor?)
Yes, absolutely! This is why the placement of stockpiles is important - getting them close to industry can leave vulnerable points for bombing and/or enemy attack.

Perhaps the best way to treat it would be to use a simple pool of TC for all industry. Requirements are calculated as for 'land convoys' (see below) using the route between the IC and the nearest eligible supply base (i.e. supply base that sends energy, metal and rares). Captured ICs take a proportion of the industry supply TC pool with them (some of which the capturing force gains).

Should there be a oil usage when transporting supplies in low level infrastructure provinces?
Arguably, TCs (being mostly rail transport) should consume Energy rather than oil. Supply from Supply Bases to the actual units I suggest should be handled separately, anyway, using attachments that will consume oil if they involve motor transport.

Should there be supply center, and if yes... how should they be modeled, and how do we prevent micromanegment?
Absolutely yes. Placement of these is a key strategic consideration for many reasons - opint, strategic planning of offensives, local objectives for military operations in a theatre, limitations on long term speed of advance - the list goes on.

They should be a base similar to naval and air bases (and deployment and strategic redeployment of land forces should be limited to between these bases). The size of the base determines how much they can store - excess is lost. Size may also determine how many units can be in the build queue and/or upgraded there at the same time.

Transfer of material (Energy, Metals, Rares, Supplies, Oil, Reinforcements) should be by convoy - either sea convoy (between Supply Bases that are in Port provinces) or land routes (using TCs). The capacity of the convoy would be calculated by:
  • TCs allocated to convoy divided by effective route length.

    Effective route length = sum of province sizes divided by infrastructure, divided by a factor for terrain (basically the same as the 'movement modifier' for an appropriate troop type from the terrain effects charts). Count the provinces at the two ends of the route are in at half value - making a single full value if they are the same province (only applies to supply from Supply Base to Units - see below).
The capacity of the route counts in both directions. Types of material can be turned 'on' or 'off' in each direction (as with convoys now) and the total capacity is divided by the number of eligible materials.

Each supply base should have a 'target stock' for each material of a set percent of the total capacity (around 10% total) plus 20 times the demand for that material last day. For each route, material should flow from the base with the highest percent of target to that with the lowest percent of target at the capacity of the route.

Supply from Supply Base to units (covers supply, oil and reinforcements) should be done by transport elements organic to each unit. All are subject to the 'supply efficiency', which is calculated as a factor (I suggest 4, but this would need balancing and might be 12 divided by the types of supply needed out of supply, oil and reinforcements - so selecting 'no reinforcement' means better supply) divided by the Effective Route Length (see above) to the nearest Supply Centre with available supply. This 'supply efficiency' has much the same use as the ESE currently has and is capped at 100%.

Naval units need organic supply elements unless they are to have a time limit for operations (after which they head for base automatically) and can trace supply to any Supply Base in a Port province, with the route going by sea and infrastructure and terrain modifiers treated as 1.00. For Air units it is the Airbase that must be supplied, with all air bases assumed to have a standard motor supply element for each air unit based there (unless you want to have supply attachments apply to air units, too, maybe??).

How will the TC changes affect gameplay and how to do the balancing?
The system should make planning of offensives more critical and limit long-term offensive speed. It would also provide near-term objectives during an advance, since captured supply bases would come with a proportion of the contents and TCs associated with it, as well as denying the source of supply to the enemy).

The main balancing factor would be the need to have an AI algorithm for Supply Base placement - this should be based on one for all troop positions, plus ones close behind the front prior to an offensive. ICs will also need them, but most of these can be placed from game start - still, a 'need detector' for the AI might be required. These AI algorithms should be available also for players who think logistics are beneath them.

Supplementary points are that TCs should be manufactured and vulnerable to losses from interdiction attacks on supply routes. Losses to convoy ships or TCs should lose a portion of what was being transported. Strategic redeployment (including of attachments for unit upgrade) should be via 'convoys', taking an extra 'channel' on the route.

Taken together, I think this should add considerable depth to logistical aspects of strategy, without forcing too much micromanagement on the unwilling player.
 
Good god what a post.... :D

i will have to split it up and answer as i get the time....

I don't think it's neccessary to split out railways, specifically, but I do think that TCs should work basically like convoys (though by land) and could be regarded as being mostly rail transport.

The idea behind railroads was the implementation of some simple ideas.
You can not strategic redeploy while not in a province with railroads.
Transport of resources/supplies costs oil when transported through a non railway province.
The question is the definition of railroads...
Mybe basic railroad = true if infrastucture >= 40%? ( requirement of building IC)

The "strategic redeployment rule" you solved by only allowing them between supply centers...
That also seems like a reasonable solution, but i find it hard to see all the consequences...it depends very much on the concept/usage of the supply centers.
 
Last edited:
Should moving resources cost TC? (making IC placement a TC factor?)
Yes, absolutely! This is why the placement of stockpiles is important - getting them close to industry can leave vulnerable points for bombing and/or enemy attack.

i think this is a hard one...
Lets divide it into several points:

TC usage should depend on infrastructure.
But instead of terrain type having an influence on TC usage have infrastructure cost and time be much higher depending on terrain.
(i see 100% infra as "there is a tunnel trough all the mountains, and a bridge over all the marshes")
Also this will increase the tracing speed.


On supply depots used for resources:
I like the idea of having a max stockpile depending on supply depot.
We would like TC usage depending on resource tranports.
so having IC in resource rich provinces means there is less transporting going on.

Simple solution:
Use the current 1 supply center system.
Add TC cost for resources and transport resources from neaby provinces first ( go for lower TC usage).
Only transport resources that are used. (supply depot is automatically made containing remaining resources.)

Pros:
No extra micro manegment.
its fast.
it adds a little more TC "feeling". ( infra helps TC usage)
Cons:
It does not check the location of the resources if they are not in the same province.
TC usage wil not be realistic in far away provinces.


Big grain solution:
Each area is a hidden supply depot containing the production and resource usage of the area.
Each day the remainder is calculated...
if there is a defecit it will get resources from nearby areas
(using a simplefied tracing method, not a fine grain province based approach)

Pros:
No micro manegment.
Reasonable fast...
Much more accurate than the simple solution.
Cons:
Still not 100% "realistic" TC usage... no ability to trace TC land convoys directly.

Fine grain Solution:
Each province is a supply center and traces to each other.
(placing a supply center on each province will have the same effect)

Pros:
No micro manegment.
100% "correct" TC usage..
TC traceability
Cons:
Extremely slow... (that is to say... unusable)

supply depot solution:
Is if i understand it correctly a mixture of the fine grain and big grain solution.
Each province trances to the nearest supply center.
Then we trace from supplycenter to supplycenter if there are some defecits.
Questions:
Why build supply depots for resource distribution?
(meaning: what is the bonus on building supply depots?)
How to prevent placing a supply depot on each province?
What happens on conquerred islands without supply depot?

Pros:
more accurate than the Big Grain spolution.
TC traceability
Cons:
requires manegment.
medium to rather slow and speed decreases with the amount of supply centers placed.
 
I don't have much time right now, but just want to try to respond a bit here:

Lennartos said:
But instead of terrain type having an influence on TC usage have infrastructure cost and time be much higher depending on terrain.
(i see 100% infra as "there is a tunnel trough all the mountains, and a bridge over all the marshes")
That's not a bad idea for the 'convoy' routes, at least. Simplifies things somewhat. For the unit supply leg this might switch to 'average of infrastructure and terrain speed modifier', to represent the fact that this stage doesn't need roads or rail (although it clearly benefits from them) and that inevitably not all of the trip can be on the best roads in the province...

supply depot solution:
Is if i understand it correctly a mixture of the fine grain and big grain solution.
Each province trances to the nearest supply center.
Then we trace from supplycenter to supplycenter if there are some defecits.
Questions:
Why build supply depots for resource distribution?
(meaning: what is the bonus on building supply depots?)
Because all resources have to go via a Supply Centre (which is where 'assembly' of land units happens and where supplies are administered). Good selection of sites is important because the TCs to get from two distant Resource Provinces to a Supply Centre is likely more than to get to a nearby 'collection centre' Supply Centre and then 'convoy' to the distant Supply Centre. E.g. resources from Scotland are likely better gathered in Glasgow and either distributed directly to local ICs or 'convoyed' to London than shipped from each separate province to London using 'pool' TCs and then shipped back to Glasgow/Edinburgh ICs using more 'pool' TCs.

How to prevent placing a supply depot on each province?
Except for the slowdown from much number crunching I don't think you need to - besides which Supply Centres would not be free to build. The choice between a few, big, well defended SCs and a distributed network of small ones is a viable strategic choice, IMO.

What happens on conquerred islands without supply depot?
You need to build one PDQ! Level 1 SCs should be pretty quick to build, I think (just off the top of my head), but bigger ones should take exponentially longer while requiring relatively fewer resources (i.e. IC-days for a Lvl.2 SC is less than twice the IC-days for a Lvl.1, but the time required is more than doubled).
 
Thinking further about this, I'm not sure that the use of Infrastructure is quite right. By just relating 'effective route length' to Infra. it means you can just add more TCs to bring the route effectiveness up. This might be OK for the final 'Supply Centre to Unit' leg, but the idea of adding extra trains to a single track route to get more throughput doesn't make much sense.

Maybe there should be a limit on the number of TCs that can be assigned to a route based on the minimum Infra province in the route? This might make it a bit confusing when combined with automatic route selection, since adding a TC might mean a longer but higher Infra route is used - or maybe the shortest route should be used for as many TCs as possible, then the next shortest, and so on? That's probably best, actually, so additional TCs are gradually less effective until eventually they do no good at all as all of the possible routes are full. This would make carefully targetted logistical strikes potentially very useful...
 
That's not a bad idea for the 'convoy' routes, at least. Simplifies things somewhat. For the unit supply leg this might switch to 'average of infrastructure and terrain speed modifier', to represent the fact that this stage doesn't need roads or rail (although it clearly benefits from them) and that inevitably not all of the trip can be on the best roads in the province...

The unit speed is completely seperate from the TC calculation so thats as it would be yes.



On the supply depot solution:
If it was me, i would place supply centers on each province to get a better TC usage...(as TC cost is wasted from and to the nearest supply center)
treating them as little more than a garrison on anti partisan(TC preservation)
This will kill the gamespeed no matter what system you own:

Making a routingtable for 2100 provinces takes: 40ms
Routing from one baspoint through 2100 provinces takes about 50ms.

we now have 230 basepoints (capitals).
But if we assume that one cannot route through allied countrys that still leaves us with 2100 base provinces. But it still gives us some overhead including multiple of borderprovinces.

So lets say we use all in all 100ms for the routingtable, and 75ms for the routingtime of one supply center per player. (capital)
Thats 175ms
Now each player makes 1 new supply center.. we now have 250ms stall time.

if one is build in each area (for resource distribution), then we have about 420 supply centers. Lets say you have won the game... the whole world is now dark grey, so you own all 420 suply centers.
Then you have the following timetable:
Routing table: 40ms (only one country)
Routing time = 420*50ms = 21seconds.
And thats not even worst case....

so i am inclined to prefer the big grain solution for resource TC usage.
because placing supply centers around each area just to collect resources and send them around is just what the big grain system does...
Worst case:
Routing table: 15ms (only 420 "provinces")
Routing time = 420*2,75ms = 1,155 seconds.
Much less resource usage for almost the same TC usage simulation.
And no extra manegment.. no building of supplybases or anything.

Dont get me wrong...
i also think supplycenters should be in the game...
But i dont think its worth blowing the resource gathering into the most cpu demanding part of the game.
So how about be keep supplycenters for supplies and war.. so no effect on
the resource transportation system.
 
Lennartos said:
Should there be a oil usage when transporting supplies in low level infrastructure provinces?

No. Low infra also mean lack of roads, and thus likely not a good place to transport things by truck. I think that for 90% of the world these transports would have been by horse or riverboats in 1936 (consuming nothing or very very tiny amounts of energy).

But the most burning supply issue I would like to se fixed is that you can sink someones convoys, but not destroy their rail system. Destroying (both bridges and engines) have zero longterm impact. But both bridges and trainengines require large amounts of heavy industry to rebuild and is not something that can be repaired in a few days by locals.

Another important thing if you ask me is convoys that lose both ships and resources they carry, And convoys on both land and sea.
 
Alex_brunius said:
No. Low infra also mean lack of roads, and thus likely not a good place to transport things by truck. I think that for 90% of the world these transports would have been by horse or riverboats in 1936 (consuming nothing or very very tiny amounts of energy).

Point taken... Germany relied on heavily on trains and horses...
But the allied didnt... so then should there be a option to motorize the TC pool?
---- only a thought for further discussion ---
Only horses and limited train transport = 50% base TC.. 0.01 Energy per used TC
Full use of trains = normal base TC.. 0.1 Energy per used TC
Motorized/mechanized trucks in addition to trains = 140% base TC...0.15 energy per TC + 0.1Oil
---- only a thought for further discussion ---


Alex_brunius said:
But the most burning supply issue I would like to se fixed is that you can sink someones convoys, but not destroy their rail system. Destroying (both bridges and engines) have zero longterm impact. But both bridges and trainengines require large amounts of heavy industry to rebuild and is not something that can be repaired in a few days by locals.

Another important thing if you ask me is convoys that lose both ships and resources they carry, And convoys on both land and sea.

The supply systems me and Balesir are talking about incorperates TC usage for transporting resources from and to IC. Bombing infrastructure will have a great influence on availible TC.
had germany any real train deficiency caused by bombing? Was it not much more a infrastructure/TC problem?
 
Lennartos said:
But the allied didnt... so then should there be a option to motorize the TC pool?
Perhaps, or a option to modernize it with diesel useage after enough tech is aquired. If I understand it correctly the first large scale (transport) diesel boats and trains started to emerge during the war.

The big distinction would need too be steam versus diesel because one use energy and the other oil. horse or train is better represented with the generic infrastructure level of the province I think.

Lennartos said:
had germany any real train deficiency caused by bombing? Was it not much more a infrastructure/TC problem?
Im not sure actually, but Ive read alot about allied bombers targeting train stations and railyards so I would be surprised if they didn't.

Some kind of measure of how many trains & trucks are available for land routes would be nice though, since people then could destroy them on purpose (with logistic strike) or build more of them to expand their TC specifically if needed.
 
i have been trying to visualize a marriage between the BigGrain supply center system.
Rules:
Resources/Supplies are only send between the areas supply center.
Only one supply center per area.
Tracing A-->B may not pass area C on the way.
Effect of rule:
When infrastructure changes in an area(bombing?), only that one area has to recalculate the costs only to its direct neighbours.
That way the routing table is always correct.
The routing table has been cut down from 162 possible routings to 32. ( by only having one SC per area)
That means a complete trace build takes 992 steps, instrad of 26081.

BigGrainRouting.jpg
 
Last edited:
Alex_brunius said:
Perhaps, or a option to modernize it with diesel useage after enough tech is aquired. If I understand it correctly the first large scale (transport) diesel boats and trains started to emerge during the war.

The big distinction would need too be steam versus diesel because one use energy and the other oil. horse or train is better represented with the generic infrastructure level of the province I think.
Could this maybe be represented via a slider? Railway-led to motor-led, so one extreme = high IC/days for TCs, use Energy to run, maybe more expensive Infrastructure build and repair; the other extreme = lower Ic/days for TC and Infrastructure but requires oil to operate?

Techs might act to further modify both ends of the scale - some to reduce the TC and Infra costs and some to lower Energy and/or Oil use by TCs.
 
Balesir said:
Thinking further about this, I'm not sure that the use of Infrastructure is quite right. By just relating 'effective route length' to Infra. it means you can just add more TCs to bring the route effectiveness up. This might be OK for the final 'Supply Centre to Unit' leg, but the idea of adding extra trains to a single track route to get more throughput doesn't make much sense.

Maybe there should be a limit on the number of TCs that can be assigned to a route based on the minimum Infra province in the route? This might make it a bit confusing when combined with automatic route selection, since adding a TC might mean a longer but higher Infra route is used - or maybe the shortest route should be used for as many TCs as possible, then the next shortest, and so on? That's probably best, actually, so additional TCs are gradually less effective until eventually they do no good at all as all of the possible routes are full. This would make carefully targetted logistical strikes potentially very useful...

Hmm... we could make simple models like limiting throughput to supplycenter size and avg. infrastructore in area....
But the real solution would propably be to abandon the global TC value altogether.. and replace it with a regional max throughput (TC). Free from IC condstraints. Upgrading a supply center would simulate assigning more trains to the route.
By clicking on a supplycenter you would get information about current max throughput and current usage by resources and supplies/oil....
That would make supply bases(containg supplies for the offensive) very importand,as would stocking up supplies before the offensive.
(units would have much more supply usage while fighting compared to standing still)
However that would make the game a lot more complicated to learn
so these advanced TC rules should propably be otional... :D
 
Lennartos said:
i have been trying to visualize a marriage between the BigGrain supply center system.
Rules:
Resources/Supplies are only send between the areas supply center.
Only one supply center per area.
Tracing A-->B may not pass area C on the way.
Effect of rule:
When infrastructure changes in an area(bombing?), only that one area has to recalculate the costs only to its direct neighbours.
That way the routing table is always correct.
The routing table has been cut down from 162 possible routings to 32. ( by only having one SC per area)
That means a complete trace build takes 992 steps, instrad of 26081.
This point about 'too many SCs' is a good one. First a few general observations:
  • Even though I propose that trade deals should be set up as discrete convoy routes, be they sea or land, between specified SCs, there is not really one, huge network but several smaller ones (one per nation, in fact). This will simplify things, especially since some of the networks are very simple (e.g. Luxembourg).

    Given the way that the management of them works, you could split all the networks by continent/contiguous land area, plus a sea network, and calculate them and run them separately (land first, then sea - and sea might be only recalculated once a week and run the same for the week following to represent the delays in turnaround time).

    SCs in every province will not be really efficient, in my view, because it will result in much higher levels of stock as material can 'hop' only one province per day. Partisans under the revised system would be represented by randomised attacks (interdicition, logistical strike and base strike, or a combination of all three), so many SCs means many targets...

    Routings etc. actually need to be recalculated only when things change. The routing information calculated at game start should last a while with recalculation being done only for routes including provinces that change and for new routes as ground is captured or new SCs built.
The system of 'one SC per area' would work fine for peacetime, subject to it being 'one SC per area per nation', but once war begins it has two problems:
  • 1) As areas have multiple owners the number of SCs will go up anyway, and

    2) It will have strange and artificial effects on strategy, such as 'if we take these three provinces, that are all in separate areas, first we will be able to set up three SCs to continue the offensive'. If only one SC is allowed per area for all nations it gets even wierder - "we must hold Breslau Supply Centre at all costs to prevent the allies building a Supply Centre in Berlin!"
Routes should definitely not be allowed to go through provinces with an SC in them, though. They must go to the SC and then out the other side. That should cut down the number of viable routes, especially if many SCs are used.

One alternative, trying to limit SC numbers, might be to have increasing numbers of Supply Centres reduce the overall transport efficiency. Given the primitive resources for supply chain planning available at the time this would make some sense and would limit the number of Supply Centres run by any one nation. Technologies and Chiefs of the General Staff would modify the severity of this limitation. Also remember that SCs should take manpower, just as ICs do - and most of the MP investment will be at Level 1, with a level 10 SC having nowhere near ten times the MP of a Level 1 SC. An SC in every province could easily become prohibitive from this angle alone for some nations (and from the IC/days required for others).

For the AI auto-builds of SCs, one per area is a good basis, with a preference order for each area and extras immediately behind the front for major offensives.

Good discussion! :)
 
Even though I propose that trade deals should be set up as discrete convoy routes, be they sea or land, between specified SCs, there is not really one, huge network but several smaller ones (one per nation, in fact)

You are right..
we have about 230 smaller nets of SC. but please look at the calculation again... even if the nets are smaller ... the global amount of nodes is the same.
Therefore the time to calculate the net is the same(overhead excluded), whether we have one big or several smaller nets.
The problem is that for each SC we have a new Start position from wich our calculation needs to be redone. So for each doubling of SC the time used on routing will double.

The problem of having free acess to SC placement largly due to the lack of borders. if we have one per area, we have a predefined net of areas, giving us small boundarys to do the search... when lokking at sachsen SC,located in leipzig we dont have to test ALL german provinces for SC and calculate the best routes for them.. we only need to test the adjancent areas to get a big picture. fo so each SC we will have a permanent 10-30 provinces to be checked each time a change occurs while with the floating SC, we have a fluid and ever expanding (if you are doing well) number of provinces to check.
if you absulutely want to have the free system at least allow a limited number of hops. ( only check 5 provinces in each direction).
It still will require MUCH more calculation time, but with hop limiting it will be possible to implement.

The system of 'one SC per area' would work fine for peacetime, subject to it being 'one SC per area per nation', but once war begins it has two problems:

1) As areas have multiple owners the number of SCs will go up anyway, and

2) It will have strange and artificial effects on strategy, such as 'if we take these three provinces, that are all in separate areas, first we will be able to set up three SCs to continue the offensive'. If only one SC is allowed per area for all nations it gets even wierder - "we must hold Breslau Supply Centre at all costs to prevent the allies building a Supply Centre in Berlin!"

1) There can only be one FRIENDLY SC in one area... a border is a non-crossable divider. so when capturing the supply depot absorb the resources to the nearest SC, and destroy it.
2) armys will not be bound directly to a SC... but recieve supplies from the nearest SC.... however.. you will NOT recieve resources from the newly taken land if you dont build a SC. (but that could be done, i think);

BigGrainRouting2.jpg
 
Last edited:
Here is the difference between One per area and 5 hops...
and remember this is pnly little germany... once we get to 42.... bah!

One per area:
OnePerArea.jpg


Same amount of SC... 5 hop rule:
5Hop.jpg
 
Last edited:
I think i have it:
Implement an area of influence around each supply center.
(similar to the aoi of trade centers in EU2)

The routing rules are the same:
Only trace to our direct neigbours.
Also every conquered province automaticaly is traced to the nearest SC, giving an ESE value for the units in the province. ( no extra step requirement as with the "fixed area" based approach).

There you have it!
free supply center placement and still limited resource demand.
I still have to make a complete calculation of the resource demand, but it looks good!

a picture of the implementation:
DynamicAreas.jpg



Also im beginning to be more and more in favour of a more regional approach for the TC routing.
Its quite simple:
TC over water = convoys
TC over land = trains
so you can build trains to increase your global TC amount...
The TC limit will be two fold:
1) the size of the supply center will limit the amount of trains sustainable.
2) the infrastructure will determine the amount of resources movable per TC. ( as before a TC multiplicator).
---just a premature example, not real values---
You could haul 10 resources with 1 train in a 100% infra province, or
1 resource with 1 train in a 10% infra environment...
---just a premature example, not real values---
So to improve the situation you could improve the SC to allow it to sustain more trains...
but that costs both for the SC and for the train...
Or you could increase the infrastructure to lower the TC requirement...
however that is a long term investment,as it takes a long time to lay tracks in difficult terrain.
 
Lennartos said:
I think i have it:
Implement an area of influence around each supply center.
(similar to the aoi of trade centers in EU2)
For resource and IC connection this will be the case - and unlike CoTs it won't be affected by 'prestige' considerations, just borders, so it will be relatively static. If you start at the SCs and calculate route lengths to adjacent provinces, you will come to some provinces that have two or more 'ranges'. Simply pick the lowest and link it to the corresponding SC. Carry on calculating ranges to unconnected provinces in the same way - you have only to take into account the single, lowest value for calculated provinces as you roll it out. Once calculated, the ranges remain valid until something changes (new SC, border changes) when you recalc out from the new/changed SC or for the province(s) that change hands. Neat.

I get a general impression that you assume all SCs will have a valid route to all others. I don't see this as so at all. My proposal is to have 'land convoy routes' like the current sea convoy routes - only the ones set up are used. A player doing this manually could in theory connect all possible routes, but I can't believe that would be anything but ineffective. The AI routine would connect only 'nearest neighbours' in each direction.

Also im beginning to be more and more in favour of a more regional approach for the TC routing.
Its quite simple:
TC over water = convoys
TC over land = trains
so you can build trains to increase your global TC amount...
I think you have to split the pools for TC and convoys - they are not really interchangeable! ;)

The TC limit will be two fold:
1) the size of the supply center will limit the amount of trains sustainable.
2) the infrastructure will determine the amount of resources movable per TC. ( as before a TC multiplicator).
I commented on (2) above - I don't think it makes good sense any more. Throwing more trains onto a single, rickety track will not increase the amount of material sent down the route!

Rather, the TCs effectiveness should be worked out as follows:
  • The 'route factor' = 2 + distance divided by a 'standard province' (100 km??)

    The capacity of a route is the TCs employed divided by the 'route factor'

    The maximum TCs employable on a route is (the minimum Infra on the route divided by 10, minus 3 (?*)), all times the 'route factor'
    {*: hence Infra 40 is required to have a route at all - not sure if this is required...}

    The capacity of an SC to handle unloading is equal to its size. The number of TCs employed on a route, divided by the route factor for the route, is the capacity of the SCs at both ends of the route that is used
So, to move lots of material, you need big SCs, at least at multi-route hubs, high Infra provinces and plenty of TCs.

---just a premature example, not real values---
You could haul 10 resources with 1 train in a 100% infra province, or
1 resource with 1 train in a 10% infra environment...
See above - 1 TC would shift x resources over a length 1 route, regardless of Infra, but an Infra 40 route could take 1 TC max but an Infra 100 route could take 7 TC, hauling 7x material in each direction.

You need to be able to assign ineffective TC to a route, I think. The 100 Infra route bombed to 62 Infra would handle 3 TCs - the remaining 4 assigned there would, if they had not been blown up ( :D ), do nothing useful as crisis teams struggled to clear the lines and get things sorted.
 
Balesir said:
My proposal is to have 'land convoy routes' like the current sea convoy routes - only the ones set up are used.

Well, lets see... in HOI we have one central SC, and then only convoys to offshore lands. Almost all convoys are routed from the capital.
Sometimes there is an exeption where the covoy creation AI tries to be "smart"..
For example getting supplies to pearl harbor from hawaii...
Resulting in 400convoys shipping supplies to pearl harbour from hawaii...
And then 400 convoys shipping it back(to even things out).
So the few exeptions from the "get everything from capital" is enough to completely throw the automic creation into disorder.
As england/USA ih've had that a few times.( round robin supply routes)
Now what would happen if the whole worlds transports where multidirectional and should be automated by that ai?.... nothing good :D

So you would have to make the convoys manual... that would work nicely...
If the supply creation was totally reworked to be a interactive part of the map mode. The way its now its pretty much unusable, for big style logistics.
Also while making it manually you will have to make good care not to make a round robin connection... (as the AI will propably do)
maybe you should only be able to define one "Source" SC...

Balesir said:
A player doing this manually could in theory connect all possible routes, but I can't believe that would be anything but ineffective.

Agreed!
See above...

Balesir said:
I get a general impression that you assume all SCs will have a valid route to all others. I don't see this as so at all.
Balesir said:
The AI routine would connect only 'nearest neighbours' in each direction.

Here was the great problem...
how does the game know where its neigbours are?...
if you want to find out you have to trace though all province to find the closests...
Now thats where the AOI comes in...
if we know there are only 4 AOI that borders us, we only have to look in these 4 AOI to find the SC we can connect to.
Without AOI in one form or another (i proposed the areas already implemented in the game... but dynamic AOI is better solution :D )
There is no way to determine if you just passed 1 or more SC on the way...
So you could have a connection from leningrad to berlin directly if the SC on the way where in low infra provinces( it would route right besides them).
With a dynamic AOI we know our neigbours, and know we dont route besides them.

Lennartos said:
Also im beginning to be more and more in favour of a more regional approach:
TC over water = convoys
TC over land = trains
so you can build trains to increase your global TC amount...
Balesir said:
I think you have to split the pools for TC and convoys - they are not really interchangeable! ;)

I think you misunderstood:
The idea was just to remove the TC dependency on IC.
Instead you have convoys and trains for water TC and land TC respectilvely.

Balesir said:
I commented on (2) above - I don't think it makes good sense any more. Throwing more trains onto a single, rickety track will not increase the amount of material sent down the route!

hmm...
having 2 trains pulling the same wagons means more speed and more cars pulled per trip. So it will increase the amount of resources transportable per day.
of course the SC will have to be expanded,to be able to service more trains and be able to distribute these resources afterwards.


Balesir said:
Rather, the TCs effectiveness should be worked out as follows:
  • The 'route factor' = 2 + distance divided by a 'standard province' (100 km??)
    The capacity of a route is the TCs employed divided by the 'route factor'

    The maximum TCs employable on a route is (the minimum Infra on the route divided by 10, minus 3 (?*)), all times the 'route factor'
    {*: hence Infra 40 is required to have a route at all - not sure if this is required...}


  • a) The route factor will also have to incorporate the infrastructure...
    steam trains are VERY sensitive to slopes. They are constant pull, not constand hp machines.Thats one of the big reasons why diesel-electric trains are so much used today :) .
    b) As much as i liked this idea(max TC calculation) initially, it just doesnt make much sense when i though it over. As soon as you have one rail in each direction you pretty much dont have any limitation on TC usage... the time to load/unload on each time will be the limiting factor. (as it is today)
    The infrastructure does however limit the speed of the trains, and therefore decreases the trains efficiency. (we even have problem with slow trains in denmark right now because of low infrastructure :) )

    Balesir said:
    The capacity of an SC to handle unloading is equal to its size. The number of TCs employed on a route, divided by the route factor for the route, is the capacity of the SCs at both ends of the route that is used
Balesir said:
So, to move lots of material, you need big SCs, at least at multi-route hubs, high Infra provinces and plenty of TCs.

Agreed.
However as i stated above in my opinion increasing the infra should decrease the TC usage.


Balesir said:
You need to be able to assign ineffective TC to a route, I think. The 100 Infra route bombed to 62 Infra would handle 3 TCs - the remaining 4 assigned there would, if they had not been blown up ( :D ), do nothing useful as crisis teams struggled to clear the lines and get things sorted.

That should be simulated, yes.... Locking the trains into SC routes somehow.
 
Lennartos said:
So you would have to make the convoys manual... that would work nicely...
If the supply creation was totally reworked to be a interactive part of the map mode. The way its now its pretty much unusable, for big style logistics.
Well, you can't make AI nations' convoys manual, so some sort of AI is needed.

Also while making it manually you will have to make good care not to make a round robin connection... (as the AI will propably do) - maybe you should only be able to define one "Source" SC...
I think the key here, as with real life supply chains, is to calculate the correct target stock and have a single route taking flow in both directions. Basically, you call one direction 'positive' and the other 'negative' and calculate the flow (for each material) once for each two-way route. That prevents 'round robin' routes nicely.

Here was the great problem...
how does the game know where its neigbours are?...
if you want to find out you have to trace though all province to find the closests...
Now thats where the AOI comes in...
if we know there are only 4 AOI that borders us, we only have to look in these 4 AOI to find the SC we can connect to.
Without AOI in one form or another (i proposed the areas already implemented in the game... but dynamic AOI is better solution :D )
There is no way to determine if you just passed 1 or more SC on the way...
So you could have a connection from leningrad to berlin directly if the SC on the way where in low infra provinces( it would route right besides them).
With a dynamic AOI we know our neigbours, and know we dont route besides them.
*Click* Now I see what you mean. Yes, the route length algorithm I described feeds into that as it also identifies a 'nearest SC' for each province - these give the 'AoI' for each SC (no limit on SCs is needed) and you allow routes only to neighbouring AoIs when the AI sets them up. Brilliant! That should work!

I think you misunderstood:
The idea was just to remove the TC dependency on IC.
Instead you have convoys and trains for water TC and land TC respectilvely.
:rofl: Sorry! No, I didn't misunderstand, I was just joking! Hence the smiley ;)

I generally use 'TC' as a shorthand to mean 'manuafactured transport units', not necessarily the 'TC = 1.5 * IC' type of TC we have currently.

hmm...
having 2 trains pulling the same wagons means more speed and more cars pulled per trip. So it will increase the amount of resources transportable per day.
of course the SC will have to be expanded,to be able to service more trains and be able to distribute these resources afterwards.
OK, but I was thinking that big volume routes normally go along multiple tracks when the scale of the model is whole provinces. Low Infra = many single line sections, especially through bridges and tunnels etc. in this era.

Maybe it's a bit of both?

a) The route factor will also have to incorporate the infrastructure...
steam trains are VERY sensitive to slopes. They are constant pull, not constand hp machines.Thats one of the big reasons why diesel-electric trains are so much used today :) .
b) As much as i liked this idea(max TC calculation) initially, it just doesnt make much sense when i though it over. As soon as you have one rail in each direction you pretty much dont have any limitation on TC usage... the time to load/unload on each time will be the limiting factor. (as it is today)
The infrastructure does however limit the speed of the trains, and therefore decreases the trains efficiency. (we even have problem with slow trains in denmark right now because of low infrastructure :) )
Slopes is a valid issue - and for motor transport the road grid quality -> speed. Maybe base distance is divided by 50+Infra rather than 100 to get 'route factor'?

That should be simulated, yes.... Locking the trains into SC routes somehow.
Well, if the maximum nubmer of effective TCs is unlimited there is no real point, but if it is limited (and I still think it should be) then yes, transferring TCs should take time - like redeployment.