How to simulate Sypply routes / TC correctly

[Lennartos]

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 )

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:

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?

 

[Lennartos]

[*]Some routes may be too dangerous to use. The route from Paris to Essen may be within the range of Allied bombers, the route through the Western Mediterranean may be too vulnerable to U-boats, etc.
[*]Some routes may suck up too many TCs. Low Infra routes you may not want to use at all because the TC load they use is too high/you consider it a waste of TCs for the tiny flow they manage.
[*]You want to run down an SC or let it build up because of some future plans you have - maybe you want to build a new IC near it, or maybe you plan to move a force there to board assault vessels prior to a seaborne invasion...
[/list]
Without some basic control over routings, many real-world dilemmas just don't signify.

One that springs to mind is the trade of metal ores from Sweden to Germany. The metals in Sweden arose in the North of the country, so Sweden's preferred source SC was in the North (where the mines were). The ores might be traded from Stockholm, say, but (a) Sweden would want more in return, since they have to use TCs moving the stuff south overland and (b) it would be vulnerable to USSR submarines crossing to Germany.

To get the best deal, Germany can accept the ores in occupied Noway (which is in fact what they did), at Narvik SC. This is a short hop from the Swedish mines, so few TCs are needed for the transfer and Sweden may well be happy to provide them as part of the deal.

So now Germany has metal ores in Narvik and needs to get them to where the ICs are - Germany. There are two options - overland down the rough country of Norway, or by sea. Overland is poor infrastructure country and quite a long way - it would hoover up lots of TCs to be effective or use valuable local TCs for little effect. The sea route, on the other hand, is vulnerable to the Royal Navy and the Royal Air Force - but is close enough to Norway's coast to be defensible by the Kriegsmarine based in fijords and the Luftwaffe based on Norwegian airfields.

This was in actual history a major issue that demanded a good deal of German high command staff time. Without the ability to define 'valid routes', how do I assign TCs to Narvik to collect local resources and service the Sweden trade route, and yet not see those TCs drained off to ship ore by rail down country?

The choice of oil routes from the Caucasus or the Middle East I see as also possibly quite critical. In a German invasion of the USSR, does oil get passed along from one front-line SC to another to feed the tanks before Moscow, or does it get fed back to a behind-the-lines distribution point to be passed forwards, minimising exposure to Soviet raids and bombers? For goods generated in British Empire provinces in southern Africa, how do I prevent them wastefully being moved overland around Africa rather than funnelled to the nearest port for outward shipment?

More generically, how do I stop stuff moving all over the map and arrange for it to be collected together regionally and shipped down one, big, transport route? Say I, as USSR, take Germany and France. How do I stop a 'wide, shallow river' of supplies, oil and resources flowing accross the whole of Europe, and arrange for all French resources to be collected in Lorraine (say) and shipped through Germany along a big, heavily defended supply corridor?

"Say I, as USSR, take Germany and France. How do I stop a 'wide, shallow river' of supplies, oil and resources flowing accross the whole of Europe, and arrange for all French resources to be collected in Lorraine (say) and shipped through Germany along a big, heavily defended supply corridor"

Thoughts:
The game always routes through the best route first. Then the next best and so on. So resources will always go in a straight line or close to it.
Do you really think you can transport all frances and germanys resources through one SC?
When forcing routes to safer places you must also calculate extra TC cost. making even fewer resources movable.

Current solution:
IF the allied are bombing north france(the best route), then infrastructure will fall, and another route is chosen automatically. No big deal. TC will always choose best(and therefore the safest) path.

Your addition:
All in all, what you want to do, is to have a ability to close some routing possibilitys.. thats fine.
The AI however must be able to work without this feature.
For micro manegment enthusiast i have added:
"Specifying which SC-to-SC routes are valid/possible for each commodity"
witch covers that point perfectly.


"For goods generated in British Empire provinces in southern Africa, how do I prevent them wastefully being moved overland around Africa rather than funnelled to the nearest port for outward shipment?"

You make a convoy "shortcut".


Regarding Sweden:
Sweden could not get the metal through the inner country, and it could not move it through the east-sea from august to may.(frozen water). So it had to be transportet through norway to get there.
Historically, that says that you must have a trade agreement with norway to get to the resources in winter... Gameplay wise the winter is not even implemented... so it doesnt realy matter either way.

1) You want swedens metal.
2) Do you care how much from each province you get?.. no
3) Do you care IF it gets there?.. yes!

so in the end, it would be a bad idea to choose: i want 30 metal from province A and 15 from province C, as it doesnt add anything to the game.

 

[Balesir]

"Say I, as USSR, take Germany and France. How do I stop a 'wide, shallow river' of supplies, oil and resources flowing accross the whole of Europe, and arrange for all French resources to be collected in Lorraine (say) and shipped through Germany along a big, heavily defended supply corridor"

Thoughts:
The game always routes through the best route first. Then the next best and so on. So resources will always go in a straight line or close to it.

Are we talking about routes 'hopping' from SC to SC, here, or direct routes from source to destination SC? Because I thought we had decided that the second one was impractical from a coding point of view (would take the PC too long to calculate routes, basically)? My view is that SC-to-SC routes will be auto-calculated anyway, but which SCs are allowed to connect with which other SCs defines the broad/rough route to follow and should be specified (whether by player or by program routine).

Assuming this is so, the question becomes 'how is movement driven between SCs?'. The obvious answer is 'according to demand' or, more correctly to allow for stores already available, 'according to the shortfall against what stock is needed to adequately cover demand'. If that is so, then one 'best' route of supply (of any commodity, not just 'supplies') may not be able to satisfy demand - either in the long term or just temporarily. Unless particular routes can be 'shut off' completely, I can't see that there won't be situations where the supply logic will say 'we need more - the best remaining is here on the far side of the Alps...' when, frankly, you would rather just suffer the shortage!

Do you really think you can transport all frances and germanys resources through one SC?

All the commodities that need to be shipped East, yes! Remember that an SC represents a structure in a whole province. For a level 10 SC think in terms of Rotterdam or similar!

When forcing routes to safer places you must also calculate extra TC cost. making even fewer resources movable.

Absolutely - that is the whole point of it! To paraphrase Sun Tsu 'to win a battle without even striking a blow - this is the quintessence of military excellence'. If, by threatening my enemy's supply routes with my bombers, I can make him use inefficient routes and use more TCs without often actually raiding with my bombers (and thus keeping my losses low), I would say I have achieved a major success!

Current solution:
IF the allied are bombing north france(the best route), then infrastructure will fall, and another route is chosen automatically. No big deal. TC will always choose best(and therefore the safest) path.

The problem with this is that it is exploitable by the player. First bomb a bit - destroy some enemy TCs and resources/supplies and lower Infra. Enemy reroutes TCs. Stop bombing until the Infra recovers and the enemy returns to routing his TCs through vulnerable territory - then bomb again, wiping out some more TCs, resources and/or supplies. Rinse and repeat...

The problem is that you have then programmed in the sort of dumb behaviour that the AI so often is accused of (with some justification). A well-constructed AI would close the vulnerable route altogether until the bombers responsible were believed to be elsewhere or on another task.

Your addition:
All in all, what you want to do, is to have a ability to close some routing possibilitys.. thats fine.
The AI however must be able to work without this feature.

Or work with it. I think there are circumstances where the AI should close down routes, too - either because, like the coal in the mountains example we had a while back, they are just not worth the TCs needed to run them or because enemy threats make them undesirable (lowering the 'effective Infra' for the AI's calculation of 'best/viable route', maybe?).

For micro manegment enthusiast i have added:
"Specifying which SC-to-SC routes are valid/possible for each commodity"
witch covers that point perfectly.

It sure does - thanks!

"For goods generated in British Empire provinces in southern Africa, how do I prevent them wastefully being moved overland around Africa rather than funnelled to the nearest port for outward shipment?"

You make a convoy "shortcut".

That would be OK if I want to ship to Alexandria - but if I build a sea convoy to India, it will still try to send stuff overland to Alexandria if Alex is short for a while... Actually, for this sort of thing it might be sufficient to say that there is a 'cutoff' where running a route is just not worth the bother. If I still need more metal in Berlin but the nearest unused surplus is in Siberia it is probably just not worth the effort to ship it in - I built too many ICs in Berlin...

Regarding Sweden:
Sweden could not get the metal through the inner country, and it could not move it through the east-sea from august to may.(frozen water). So it had to be transportet through norway to get there.
Historically, that says that you must have a trade agreement with norway to get to the resources in winter... Gameplay wise the winter is not even implemented... so it doesnt realy matter either way.

1) You want swedens metal.
2) Do you care how much from each province you get?.. no
3) Do you care IF it gets there?.. yes!

so in the end, it would be a bad idea to choose: i want 30 metal from province A and 15 from province C, as it doesnt add anything to the game.

Let me be clearer what I have in mind for trade agreements.

Currently a trade agreement looks like this:

• 'We give x units of commodity a, y units of commodity b, etc.
  and in return
• You give w units of commodity c, z units of commodity d, etc.

What I want to replace that with is this:

• We give x units of commodity a, y units of commodity b, etc.
• At location (SC in) province m
  and in return
• You give w units of commodity c, z units of commodity d, etc.
• At location (SC in) province n

The provinces must be provinces with an SC. They may be in either trade partner's country - 'province n' might be in my country; that would mean that the other party in the trade is responsible for transporting the goods to me at that location using their own convoys, TCs or whatever. If 'province m' is in the other party's country, I would likewise have to set up a route to get the goods to the agreed province.

As it is now, trade deals are a reciprocal thing; if one party is short and cannot supply, the whole deal is off for the day. This would remain - in other words, both routes must be set up with enough convoys/TCs and/or enough supply must be available at the agreed SCs, or the trade is not effectual for the day.

As a further thought, provinces 'm' and 'n' might even be allowed to be in a neutral country with whom both trade partners have an access agreement, maybe?

This will mean that the AI will need to be able to value commodities not just in themselves but at a specific location (SC). This should not be too difficult - the AI already has algorithms for assessing general value and will also need the calculations to assess how many convoys/TCs are needed to get commodities from SCs with surplus to those with demand for other purposes. Combine the two with some weighting for TC/convoy value and you have location-specific values.

Now, returning to Sweden for a moment, I can ask for metal in Narvik, in Stockholm or even in Rostock - but given that Sweden would struggle to provide the stuff anywhere but Narvik or the mines, the value they place on it anywhere else (a) will be astronomical and (b) will come in addition with the fact that the trade will run only occasionally, when the goods build up to the required amount. So, I take it in Narvik, and getting it to Germany is my problem... As long as I can stop it feeding Oslo's ICs overland whenever they get a bit hungry, I'm happy!

Incidentally - this opens up the possibility of useful trades in the same commodity (e.g. swap metal for metal). If I have rares in Malaysia and Netherlands have rares in Amsterdam, I might provide rares to Java in return for rares in Rotterdam...

 

[Lennartos]

The problem with this is that it is exploitable by the player. First bomb a bit - destroy some enemy TCs and resources/supplies and lower Infra. Enemy reroutes TCs. Stop bombing until the Infra recovers and the enemy returns to routing his TCs through vulnerable territory - then bomb again, wiping out some more TCs, resources and/or supplies. Rinse and repeat...

The problem is that you have then programmed in the sort of dumb behaviour that the AI so often is accused of (with some justification). A well-constructed AI would close the vulnerable route altogether until the bombers responsible were believed to be elsewhere or on another task.

You just answered your own suggestion.
Even IF you get the AI to cut a connection because its too risky, as soon as the bombing stops+infra is ok again+AI safety value delay, it will reopen it as the area is safe again.
You will always be able to use the method against a AI.

Where TCs even a great target in WW2? i mean planes couldnt hit a bloddy thing even when its as large as a shopping mall.
Of course i dont know the numbers, but i would think that partisans still where the highest danger for TCs.
planes might hit a TC once in a while directly, but that would be more of an accident than calculated.
Hitting stationary targets that can be planned and prepared for(do you know when and where trains are) and calmly aimed agains(like SCs and bridges/infra) are much more likely targets...

 

[Lennartos]

Regarding trade:

There are just sooo much tings that will go wrong...(murphys law)

You want 30 metal fron narvik. ok fine
now norway buys 10 metal from narvik

As narvik only has 30 metal, it sends 10metal from Falun via stockholm to kirkuna SC, just so that norway can transport it down to OSLO again.
-------------------

maybe a simplifcation...

when asking for a trade, you set a trade target.. for example Berlin.
When the trade deal is made resources will be send from SCs with enough surplus.
Not necessarily the same each day... just the one that is closest and has enough surplus for today. You pay convoys for your deal of the bargin, and swedish metal will come with their convoys. ?

 

[Balesir]

You just answered your own suggestion.
Even IF you get the AI to cut a connection because its too risky, as soon as the bombing stops+infra is ok again+AI safety value delay, it will reopen it as the area is safe again.
You will always be able to use the method against a AI.

Why does the break have to be only an arbitrary delay? If the Intel system were a little more 'joined up' (for the AI and for the player, too) then a more 'rational' process for the decision to return to the old route could be used. Like when no bombers are 'seen' within range, or when the same bombers are seen somewhere else.

Where TCs even a great target in WW2? i mean planes couldnt hit a bloddy thing even when its as large as a shopping mall.
Of course i dont know the numbers, but i would think that partisans still where the highest danger for TCs.
planes might hit a TC once in a while directly, but that would be more of an accident than calculated.
Hitting stationary targets that can be planned and prepared for(do you know when and where trains are) and calmly aimed agains(like SCs and bridges/infra) are much more likely targets...

For medium to high altitude bombers, sure - but rocket firing and cannon-using fighter-bombers could and did take out trains, planes, boats and tanks... And bombing of rail yards, terminals and rail bridges, etc. is bound to hit a few trains and rolling stock.

 

[Balesir]

Regarding trade:

There are just sooo much tings that will go wrong...(murphys law)

Yes, indeed - that sounds like a good model of the real world!

You want 30 metal fron narvik. ok fine
now norway buys 10 metal from narvik

As narvik only has 30 metal, it sends 10metal from Falun via stockholm to kirkuna SC, just so that norway can transport it down to OSLO again.
-------------------

If you try to be (too) clever, such things may happen, true. But, mostly, trade will/should be routed along the main, historical arteries - it happened that way for a reason! While, in theory, you could ask to get rares delivered anywhere, in practice most will come from Singapore/Ceylon/Indonesia. That's the easiest, after all.

maybe a simplifcation...

when asking for a trade, you set a trade target.. for example Berlin.
When the trade deal is made resources will be send from SCs with enough surplus.
Not necessarily the same each day... just the one that is closest and has enough surplus for today. You pay convoys for your deal of the bargin, and swedish metal will come with their convoys. ?

This might be a simplification for the player, but for the program it seems to me it will be much more difficult. As a trade proposer I think you have a choice: propose locations where the stuff will be available in abundance in any case (near the ultimate source), or prepare to spend time managing things to make sure the flow continues.

It would make things easier if the 'one-off bulk lot' deals were still possible (as under the 'negotiation' tab, currently). I'm not sure how best to account for the convoy/TC load for these, though. To allow them free passage is obviously a bad idea (everybody will use them instead of trade deals!), but how to do it...?

 

[Lennartos]

When making a one-time deal:
lets just regard it as a value limited trade deal.....

This also gave me a good idea:
Trade agreements should have a optional minimum time or time limit.
No more make one agreement and cancel it the next day.

In effect: If you make a deal you are bound 2 months by that agreement.
Or the agreement is only effective for 2 months.


still, im afraid the trade system may be too complicated.

• The inexperienced user shold not drown in calculations.
• The current trade AI should be able to hande it.

So lets summarize:
Must haves:

• Simulation of bottlenecks in the logistic system.
• show the importance of infrastructure.(as it WAS important)
• get more realistic and historic battle progress.(offensives have to stop to wait for logistics)
• The simulation should not be cumbersome, but intuitive to understand.
• Should be automatable. ( the AI needs to do this too)
• Should not limit current or future functionality.(its a seperate layer)

Nice to have:

• TC attackable.
• TC Cost simulation
• Steam/Oil simulation

I think these are our current primary concerns.

Having SC centers gives us all "must haves".
Having TCs gives ud a more flexible SC system, as TC can move wheras SCs are static.
So far so good.

Now to trade:
Its just a simple I give X, i want Y.
So i would say capital to capital trade is fine. (US->UK;germany->sweden;etc..)
However there could occur some minor exeptions:
....
.... snip....
...
I made a few case.. like jap + axis netherlands...
But after writing here for about an hour i came to the conclusion that its all fine with capital to capital.
If japan wants to have allied netherlands rares, and netherland is out of trade reach (UK blocking).. then no matter what we do, netherland will never get anything from the trade. Netherland may just as well sign control of these islands over to JAP.

its simple its fast its good.

 

[Balesir]

When making a one-time deal:
lets just regard it as a value limited trade deal.....

This also gave me a good idea:
Trade agreements should have a optional minimum time or time limit.
No more make one agreement and cancel it the next day.

In effect: If you make a deal you are bound 2 months by that agreement.
Or the agreement is only effective for 2 months.

Brilliant! Yes - there is no such thing as a 'one-off, lump sum trade deal' but all trades may/must have minimum and maximum durations. That should work fine.

still, im afraid the trade system may be too complicated.

• The inexperienced user shold not drown in calculations.

OK - so a default is offered to exchange some of the selected goods from where your own stock is highest for some of the selected return goods where the partner's stock is highest. The confirmation screen shows an estimate of how many convoys (if not on same continent) or TCs (if on same continent) will be required to service the trade. Inexperienced users (or those who don't want to bother) can just accept this default. A route is set up automatically for the trade upon acceptance.

• The current trade AI should be able to hande it.

All AI trades could follow the pattern given for the default, above. For deciding whether to accept 'variant' deals see my earlier post.

So lets summarize:
Must haves:

• Simulation of bottlenecks in the logistic system.
• Show the importance of infrastructure.(as it WAS important)
• Get more realistic and historic battle progress.(offensives have to stop to wait for logistics)
• The simulation should not be cumbersome, but intuitive to understand.
• Should be automatable. ( the AI needs to do this too)
• Should not limit current or future functionality (it's a seperate layer)

I would add here 'make logistical factors important in picking objectives to capture'. E.g. capturing ports to help bring supply nearer the front lines, take a key province to cut the enemy's best (but not necessarily only) supply route.

I still think SCs covers this marginally and SCs plus TCs and convoys on routes covers this well, though.

Nice to have:

• TC attackable.
• TC Cost simulation
• Steam/Oil simulation

I think these are our current primary concerns.

Having SC centers gives us all "must haves".
Having TCs gives ud a more flexible SC system, as TC can move wheras SCs are static.
So far so good.

Agreed.

Now to trade:
Its just a simple I give X, i want Y.
So i would say capital to capital trade is fine. (US->UK;germany->sweden;etc..)
However there could occur some minor exeptions:
....
.... snip....
...
I made a few case.. like jap + axis netherlands...
But after writing here for about an hour i came to the conclusion that its all fine with capital to capital.
If japan wants to have allied netherlands rares, and netherland is out of trade reach (UK blocking).. then no matter what we do, netherland will never get anything from the trade. Netherland may just as well sign control of these islands over to JAP.

its simple its fast its good.

Except what about the convoys and TCs needed for trade?? I think that 'must haves' from the trade system include that trade can be attacked as any other convoy, requires transport assets and is a way that neutral nations get entangled in the war. The biggest difference between Japan trading from Indonesia and Japan trading with Amsterdam is the distance the convoys have to go! A major 'battle' of WW2 was the 'Arctic Convoys' from the UK to Archangel in Northern USSR. The convoys being attacked were (in HoI terms, at any rate) trade, not internal transport. The by late 1941 the USA was, though not at war with Germany, sinking U-boats on sight due to strikes against US transport ships carrying lease-lend material for Britain...

Maybe this 'simple trading' system could be included as an option - it more-or-less exists in the game now, anyway - that can be selected or not at gamestart? To have it as the only system, though, takes away a major element of strategic relevance, I think.

 

[Lennartos]

"Except what about the convoys and TCs needed for trade?? I think that 'must haves' from the trade system include that trade can be attacked as any other convoy, requires transport assets and is a way that neutral nations get entangled in the war. The biggest difference between Japan trading from Indonesia and Japan trading with Amsterdam is the distance the convoys have to go! A major 'battle' of WW2 was the 'Arctic Convoys' from the UK to Archangel in Northern USSR. The convoys being attacked were (in HoI terms, at any rate) trade, not internal transport. The by late 1941 the USA was, though not at war with Germany, sinking U-boats on sight due to strikes against US transport ships carrying lease-lend material for Britain...

Maybe this 'simple trading' system could be included as an option - it more-or-less exists in the game now, anyway - that can be selected or not at gamestart? To have it as the only system, though, takes away a major element of strategic relevance, I think."

Dont jump to conclusions.
The whole system is based on Convoys and TC. You CANNOT transport without CONV/TC.
Only thing i meant about Capital->Capital trading is that target SCs per default always are the capitals. Selecting a different SC for trade, only gives much grief,without gain.
Japan gets rares from indonesia, and delivers metal to indonesia.. now holland has to send convoys to get these from indonesia(wich they dont have) back to holland.
Holland will not get anything from this deal.. it gets even worse if japan wants oil or money in thetrade too...
now holland has to transport things around the world, just to get a few metal it cannot collect anyway.


So as i see it, trade is either made from capital->capita, or not capped to any SCs at all.(trade will send from where resources are).

 

[Balesir]

OK, I think I see what you're saying, but wouldn't defaulting to the SC with most current stock be better? Plus, remember that money does not need a transport route so its true value and role might be appreciated.

For the Holland case, Japan can ask for rares in Indonesia in return for money in Amsterdam and Holland can buy stuff locally with the money.

 

[Lennartos]

OK, I think I see what you're saying, but wouldn't defaulting to the SC with most current stock be better? Plus, remember that money does not need a transport route so its true value and role might be appreciated.

For the Holland case, Japan can ask for rares in Indonesia in return for money in Amsterdam and Holland can buy stuff locally with the money.

The one with most current stock of WHAT?
You can ask freely on every resource, remember that... also the one with most resources does not necessary have any real surplus.


So what about this:
Primary trade is always capital to capital.
Now the new secondary trade option, that allows the player to buy resources locally with money.(1$/day for 10$Rares/day in SCx)

 

[Balesir]

The one with most current stock of WHAT?
You can ask freely on every resource, remember that... also the one with most resources does not necessary have any real surplus.

For multi-commodity deals I think it would make sense to have one SC defined per commodity - this should not be too burdensome provided that sensible defaults are offered.

So what about this:
Primary trade is always capital to capital.
Now the new secondary trade option, that allows the player to buy resources locally with money.(1$/day for 10$Rares/day in SCx)

If the logistics of resource movement are properly set up I think it will seldom make sense to trade from the capital. It makes much more sense, as a default, to trade from where the resource is generated, surely? Given the concentration of resources near the ICs, maybe stock is not the right measure, but I'm confident there exists some better measure than 'the capital'.

Maybe inflow to the SC from resource provinces minus direct demand? The SC with the greatest resource production minus direct demand gap is the default to trade from?

How about:

• AI nations offer resources each at the SC where they have the biggest production minus demand gap for that resource and
• ask for commodities each at the SC where they have the greatest shortfall against target stock of that commodity?

Both these aspects are simple for the computer to calculate (using last day's figures - or maybe last month's would be more accurate, given WW2 reporting times ). This should give a trade net of reasonable peacetime utility with only default trades (default offered to player = AI standard SCs). Trades where the route(s) involved would be beyond the capability of the offering nation's available surplus convoys or TCs should be excluded from consideration - that should give a reasonable wartime trade network, as well, since blockaded routes would be excluded. Routes that take losses could not be simply excluded, but would need some sort of penalty like that discussed above for specifically land routes. Just thinking 'aloud', now...

 

[Lennartos]

Making trades should be fast and painless..
Using 5 minutes for research about where there are surplusses, is absolutely within the micromanegment area.

I would suggest the "dont know dont care" approach, where resources are send where there are suplusses THAT DAY!(or week, for calculations sake)... tomorrow if necessary, use another place.
No SCs are specefied.
Its just treated as ones own resources and will be used to fill holes where needed.)and goto nearest high priority/buildup SC if surplus.

EDIT:
That also has the added improvement that convoy source /destination can be changed due to enemy interference.(danger factor)

 

[Balesir]

Making trades should be fast and painless..
Using 5 minutes for research about where there are surplusses, is absolutely within the micromanegment area.

Agreed - that's why there should be a default representing the current point of overproduction. Those wishing to avoid all micromanagement just accept the default every time - it should not be too bad, most of the time.

I would suggest the "dont know dont care" approach, where resources are send where there are suplusses THAT DAY!(or week, for calculations sake)... tomorrow if necessary, use another place.
No SCs are specefied.
Its just treated as ones own resources and will be used to fill holes where needed.)and goto nearest high priority/buildup SC if surplus.

OK, I'm getting lost in the options, now, but I can think of three ways for this to work:

• 'Convoys' are set up and turned off by the AI in an improved version of the current sea convoy system. This could be set to be optionally player controlled. The problem with using it for trade is that the route finding algorithm gets very much complicated if it has to be done each day/week/month for trade as well as domestic SCs.
• The total demand for movement of commodities is assessed and a percentage is moved according to the ratio of available transport assets to assets needed to move everything. This might be quite simply codable, but it doesn't allow for routes to be targetted specifically and the player is afforded no real control at all over priorities or vetoed transfers.
• Every possible route between SCs is treated as an effective 'convoy'. Transport assets are assigned/deassigned to these on a daily basis. This has a very high calculation load - especially when there is a global trade network - but is quite similar to the first option, above. It would allow the player to have the option of 'fixing' particular routes with particular levels of assets, though, while leaving the rest to automatic control (as opposed to an 'all or nothing' manual control scenario).

Of these, I think the first is the most practical while maintaining the design goals, but I still see a problem if trade convoys must be reassessed by the AI each day. You haven't convinced me yet that trade routes should not be fixed (as this was often specified in the treaty anyway - the value of bulk commodities depend in part on where they are). Give a player the option of control in a variable destination system and I can see exploits fixing a price for metal to the USSR and then delivering it to Vladivostock, just to screw with their railway system!

Finally, the thought of trying to plan supply for an offensive while my transport assets are all over the place (literally!) keeping up with shifting trade routes under AI control - shudder!

Bear in mind that I'm suggesting only that trades have a delivery point fixed - the source could be determined by the AI as per usual. The delivery point just gains a direct demand and, if it is owned by another nation, may not be set to 'fill', 'empty' or any of the other SC settings available for your own SCs.

EDIT:
That also has the added improvement that convoy source /destination can be changed due to enemy interference.(danger factor)

Under these conditions what happens is one side defaults on the contract - force majeur (in its most literal sense!). You renegotiate where the goods are delivered to.

A simplifying step might be to make one side of every deal have to be money. Each separate deal could then be limited to one (fixed) destination. I don't think this is strictly needed, though.

 

[Lennartos]

"Every possible route between SCs is treated as an effective 'convoy'. Transport assets are assigned/deassigned to these on a daily basis. This has a very high calculation load - especially when there is a global trade network - but is quite similar to the first option, above. It would allow the player to have the option of 'fixing' particular routes with particular levels of assets, though, while leaving the rest to automatic control (as opposed to an 'all or nothing' manual control scenario)."

A mixture of the first and the last sounds about right in my head.(convoys are open/close entitys.. land is open to all adancent SCs, unless manually closed)
The Transport layer in the game should always try be "optimal" especially without user intervention, Or else the player again has a too great advantage over AI.
You as the player dont really care where things come from, as long as things arrive where you want it, and its done the most optimal way. (you only care if its done wrong )
Also some sort of tracability is required as you need to get an overview what has to be improved to keep the system flowing.
Also adding the option of opening/closing routes to prevent losses from Air attacks is a good option.(that also is optional automatic)

If the whole transport system is transparent, then adding trade will function as a extension of ones SC network. The seller determines where it is delivered from, the reciever determines where it is useful.

How much can be user controlled can of course be discussed...
If one wanted, one could take control of the whole transport system...
but then we would propably be better off, playing transport tycoon instead...

Once the routing table has been completed, the remaing resource tracing is a pretty light load, as we only have ~400 SCs in the world that can be adressed.
Thought and planning has however to be used to make good alghorithms to archieve a "optimal" tracing system.

 

[Balesir]

A mixture of the first and the last sounds about right in my head.(convoys are open/close entitys.. land is open to all adancent SCs, unless manually closed)

If routes are all present but may be closed, I think the only difference between that and the 'set up routes as needed' system is one of screen cosmetics, from the player POV - do you want to see all routes on screen, or just the ones you are using? So, for me, either would work we just need to use whichever is easiest to code well.

The Transport layer in the game should always try be "optimal" especially without user intervention, Or else the player again has a too great advantage over AI.

While I agree with this generally, there is one aspect that I think it misses.

If I choose to run/not run a route it might be because it has proven vulnerable because the enemy have attacked it. This is a simple factual state of affairs that the AI could use without problem. However, I might choose to avoid a route because I anticipate that it will be vulnerable to enemy attack - and I might choose to use a route that is under attack because I don't believe that the enemy can maintain the attack for long.

Example: if I see an enemy bomber unit within range of one of my routes, I might decide to reroute because the bombers might attack - or I might not. The AI will need to have some consistent approach to this - a human will inevitably have inconsistent approaches open as well. Whether this is an advantage or not might be a moot point!

You as the player dont really care where things come from, as long as things arrive where you want it, and its done the most optimal way. (you only care if its done wrong )
Also some sort of tracability is required as you need to get an overview what has to be improved to keep the system flowing.
Also adding the option of opening/closing routes to prevent losses from Air attacks is a good option.(that also is optional automatic)

Absolutely - a good map view of convoy and land supply routes is essential. I'm not sure how it could be best done - maybe coloured lines through the provinces traversed? A different colour for each commodity, with a filter to turn commodities off in the view so you can just look at Supplies or just Energy or whatever.

If the whole transport system is transparent, then adding trade will function as a extension of ones SC network. The seller determines where it is delivered from, the reciever determines where it is useful.

Right - but there will still have to be an agreed handover point. This is a real part of any sale/trade agreement - it's fundamental. If I am Portugal, say, rares handed over to me in Borneo are just not the same deal as rares delivered to my own territory. There is a 'Ship To' address with every deal except those that are on 'pick up' terms!

How much can be user controlled can of course be discussed...
If one wanted, one could take control of the whole transport system...
but then we would propably be better off, playing transport tycoon instead...

Heh - actually, I quite liked that game, but I don't want to try playing both it and HoI simultaneously, which is what this would be like!

My 'optimum setup' would be a system that generally runs just fine on its own, but in which I might want to take a detailed look at key sectors and maybe tweak things according to my own priorities and perceptions. As I said above, it's about where judgement and planning are needed, rather than just reacting to events, that I want some discretion.

Once the routing table has been completed, the remaing resource tracing is a pretty light load, as we only have ~400 SCs in the world that can be adressed.

Is this assuming one SC per region? What about contested regions? Does having an SC in a region block anyone else having one there? That doesn't seem to make sense to me...

If one SC is allowed per land province then the maximum is around 1,860. If a maximum of one per region per nation is allowed, the maximum is probably around 1,200 - but, in all cases except 'one per region', I would expect the actual number to be much lower - maybe 500-600 or so.

Thought and planning has however to be used to make good alghorithms to archieve a "optimal" tracing system.

Agreed. The algorithm once the routes are set up (assuming not too many circular routes, etc.) is fairly easy and defined - but the algorithm for setting up routes needs to be worked out. I'll give it some thought - all others are welcome to do so also!

 

[Lennartos]

In short, we are about to reach a agreement...

1) The system should always usa a optimal route, without user intervention. (with some sort of danger detection)
2) the user should have access to open/close a route/SC and prioritize to manipulate the transport layer to do his bidding.
3) A map that shows TC flow/congestion and closed routes.
4) Trade: If we have a system that claculates where resources are most needed there is no need to to control the destination point. ( think one week later.. the excess WILL be transported to the calculated place anyway).
If you need the resources in a SC province xx, the Convoys will transport it there, unless the route is unsafe, then they will try another route or try to transport it via land.

This is also tied closely to the "Off capital prouction".
SItuation: England... you build 2 divisions in india and need 10 extra metal. Now you trade 30 metal from france.
Now here is the deal: You can buy metal from any SC of france as long as there is a surplus. So you would automatically buy 10 metal from asian france(that is transported to india) and 20 from mainland france (to england).

a SC can handle more than 4 provinces...
12-16 provinces per SC sounds more normal... now double up for contested regions, and count a extra 100 for islands... and you will be about 400...

 

[Balesir]

In short, we are about to reach a agreement...

Shock, horror!

1) The system should always usa a optimal route, without user intervention. (with some sort of danger detection)
2) the user should have access to open/close a route/SC and prioritize to manipulate the transport layer to do his bidding.
3) A map that shows TC flow/congestion and closed routes.

Agreed in all points!

4) Trade: If we have a system that claculates where resources are most needed there is no need to to control the destination point. ( think one week later.. the excess WILL be transported to the calculated place anyway).
If you need the resources in a SC province xx, the Convoys will transport it there, unless the route is unsafe, then they will try another route or try to transport it via land.

The key point, though, is whose convoys/TCs will transport it there? That's what I'm saying - there always has to be an agreed handover point. "I will transport the stuff to <specify SC> - you take it from there". Once that has been agreed it works just like you say - the agreed SC becomes just another node in the transport net of both parties to the agreement - it now just has a direct demand for one party and an available supply for the other.

This is also tied closely to the "Off capital prouction".
SItuation: England... you build 2 divisions in india and need 10 extra metal. Now you trade 30 metal from france.
Now here is the deal: You can buy metal from any SC of france as long as there is a surplus. So you would automatically buy 10 metal from asian france(that is transported to india) and 20 from mainland france (to england).

Indeed - but what I don't want is a situation where France sees available metal in France fall (due to building more ICs or another trade, for example) while a cancelled trade in SE Asia means there is more there - and my British convoys immediately and automatically start shipping it from SE Asia to Portsmouth! My agreement with France should be 10 metal in a specified SE Asia SC and 20 metal in a specified French national SC. How France get the stuff there is their problem - how I get it from there to where I need it is my problem. Both problems should normally be handled by the automagic logistical engine

a SC can handle more than 4 provinces...
12-16 provinces per SC sounds more normal... now double up for contested regions, and count a extra 100 for islands... and you will be about 400...

Ah - so it's an 'educated guesstimate'? Fair enough - I'll buy that. A mega-building player might mess up the maths, but that's arguably their own fault!

 

[Lennartos]

"The key point, though, is whose convoys/TCs will transport it there? That's what I'm saying - there always has to be an agreed handover point. "I will transport the stuff to <specify SC> - you take it from there".

but what I don't want is a situation where France sees available metal in France fall (due to building more ICs or another trade, for example) while a cancelled trade in SE Asia means there is more there - and my British convoys immediately and automatically start shipping it from SE Asia to Portsmouth! "

You are right, that is a "possible" problem.

"My agreement with France should be 10 metal in a specified SE Asia SC and 20 metal in a specified French national SC. How France get the stuff there is their problem - how I get it from there to where I need it is my problem. Both problems should normally be handled by the automagic logistical engine "

But it just gets too complicated here.
I want 5 metal from lorraine to london
4 metal from paris to portsmouth
12 Energy from alsace to some SC in india.
Thats just way too much in the micro manegment area. No person in the whole world can manage a global empire with wars all over, and the sime time manege trade and transport issues withing each SC...IN REAL TIME!

So either we abandon the trade screen as we know it and replace it with a new fancy global screen where you can buy resources from Logistic nets. (mainland france, germany, asia france, china, japan, japan holdings in china... all nets seperated by borders and water).


Or we just let trade prices be determined by transport cost and desirability.
so as england you will automatically trade with asia as long as you have enough convoys to spare(as france will sell these cheaper(it doesnt have to transport it anymore)... once it gets rough with free convoys, it will buy the same wares from mainland france ( paying more ) to save convoys.

 

[Balesir]

I want 5 metal from lorraine to london
4 metal from paris to portsmouth
12 Energy from alsace to some SC in india.
Thats just way too much in the micro manegment area. No person in the whole world can manage a global empire with wars all over, and the sime time manege trade and transport issues withing each SC...IN REAL TIME!

?? Why would you want to set up something this complex? Just offer for 9 metal in Normandy and 12 Energy in Marseilles. If France accepts, they get demands in those SCs and it's their job to get the stuff there (through their normal transport nets). How it gets from these ports to your SCs with demand is handled by your transport net - you don't even need to specify where to get it to - the logistics routines will take it to wherever you have need of it. But the transfer points are fixed - Normandy and Marseilles.

As an aside, I don't see why you should not be able to, but I'm baffled as to why I would want to pick up goods from Alsace, Lorraine and Paris. I would select border SCs in just about all cases, I would think.

So either we abandon the trade screen as we know it and replace it with a new fancy global screen where you can buy resources from Logistic nets. (mainland france, germany, asia france, china, japan, japan holdings in china... all nets seperated by borders and water).

The trade screen will have to be modified, obviously, but only to add the 'transfer SC' to each commodity in the trade. I don't really see that as too much of a burden, sorry.

Or we just let trade prices be determined by transport cost and desirability.
so as england you will automatically trade with asia as long as you have enough convoys to spare(as france will sell these cheaper(it doesnt have to transport it anymore)... once it gets rough with free convoys, it will buy the same wares from mainland france ( paying more ) to save convoys.

This sort of 'free global market' is really a very modern conception. Governments in the 1930s and 40s really did get together and agree this sort of trade stuff - so I think it's reasonable to have it in the game, too.