Military Model IV

[Mark_Everson]

It's time for Military Thread IV!

For reference, the beginning of the immediate predecessor, the Military III discussion, is here.


A brief status report is in order, and seeing as there's no one in charge of the military model design at the moment, I can't stick anyone else with doing it...

The current model:

For an overview of what has gone before you should look at the Military Page at the Clash web site. For a brief idea of what we're doing, look at the introduction there. After that, the next important thing you should check out is the battle and units model worked out by Harli and Krenske. Here is a link to the Scouting Model of their design. Don't follow the links in it, more current links to the other modules (Manuver, Assault, and Units) are found in the Updated Model. Please go down to Krenske's post of February 2nd for brief discussion of the revised model.

Although we have a lot of the bare bones of the model, there is an Enormous amount of specification, and refinement, to be done. If you are a Military Guru and would like to apply for the position, please let me know in this thread, or e-mail me. We are also interested in getting whatever comments people have at any level on the basic models, or things that were set in the other discussions. But the most important part is that we will soon have code to test further parts of the model, which leads to the...

Coding Status:

Laurent D. is busy getting the new military code off the ground. I will leave it to him to update the status at a future date.

[Edit 3/28] Put in link to updated model up top.
[This message has been edited by Mark_Everson (edited March 28, 2001).]
[This message has been edited by Mark_Everson (edited March 28, 2001).]

[Lord God Jinnai]

I like the solution 1, but with some changes. First, a province need not have an economic center if it has a military center such as a fortress or encampment that should solve most of the problems, but for the few remaining ones, the AI or the player can designate a point within the province for these units to be constructed. This may also leed to a small city-like boom because the resources will be centered there. Generally if the AI plaaces it, it should be for strategic area first and then possible city growth as secondary.

In cases where 2 or more such city/fortress combinations occur within 1 province, the player can either focus on having them be built in only 1 place, which default should be the largest fortress or have the option of pooling resources from the neighboring squares of each.

[Mark_Everson]

One thing has come up in my Econ model coding that relates to the military model. It involves govt orders to build military units. What I envision is that the player or AI orders such and such a queue of units to be built and can have either resource constraints (like use local taxes only) or time constraints (extra money needed then comes from the treasury). There will also be effectiveness constraints so you don't build legions in places where it is especially expensive to do so. And of course you can specify what region you want things built in also. So far, so good, its flexible and doesn't require much micromanagement to get what you want.

The problem happens when running the economy at an individual map square level like I'm doing now. If the govt order is "Build Legions everywhere using 35% of tax revenues" it is simple to do just that. But you won't get very satisfactory results, because squares with small economic capabilities will take Forever to build one legion.

Solution #1:

You could just build all the legions in cities, with their presumably much larger economies, and just take cash taxes from the little places. Over time this economic stimulus will grow the city economies even larger and stronger. This works fine where a few squares in a province are predominant economically. It doesn't work when the economic oomph is spread evenly.

Solution #2

I have a crude solution that I'm not quite happy with. I'll present it, and then we can see if anyone else has a better idea. You can have all the squares work on legions, but not require each square to build a whole legion itself. FE every time a square gets to, say, 10% of the size of a legion it marches off to the provincial capital to be unified with other bits into a whole 'standard' legion. Optionally these 'stub' units could just teleport there 'cause they'd be really annoying on the map IMO.

Any suggestions from the croud?

[Richard Bruns]

Possibility:

Define how much time you want to take to build the stuff. Then pool squares into "production units" so that each production unit takes the specified amount of time to build the unit. For example, 20 underdeveloped squares combine into a production unit to make 1 legion in over a period of 2 years, while a city square makes a legion in 2 years all by itself. The unit, when built, is moved to either the border or a nearby city. This has the added bonus of alloting defense according to economic capacity, guarding more valuable regions.

[LDiCesare]

Caravans? Solution 2 reminds me of building wonders in civ: You could have caravans that are built (fast) in one place, and that are building a given kind of work (here armies). Once the caravan is built, you move it to someplace and it helps building. That's solution 2 but more general (not only units could be built that way). You could either show the caravans as units (so you can attack them, they cannot defend) or teleport them with a cost (loss of time, loss of production) or make them "trade routes" (just lines drawn -or not- along the map which could be pirated if we add that one day).

[Mark_Everson]

Hi All: Thanks for the ideas!

I think Richard's idea is clearly better than what I had. I am inclined to just accept his idea for the time being since its farily easy to implement, and yet fits with everything.

I don't really understand your piece LGJ. If you use solution 1 there Must be areas with high economic activity where you build the units. Otherwise it will take forever to build in the middle of nowhere. The resources won't get there by magic... If there are no artisans there, all the raw materials in the world won't do any good, etc. When I get Merchants working, your idea will be possible, but Very expensive, since you need to bring everything, including workers if not there already, to your military outpost. IMO its much better to build the units where the production capacity is, and then ship them off to the fortress or whatever.

The caravans you're proposing Laurent are basically what Merchants are supposed to do. (The only difference is you are tranporting the men as well as the weapons.) But right now I don't have the Merchant code written, so its not a solution .

[Lord God Jinnai]

The one problem you face still is in backwater provinces with no economic activity woth mentioning, not much more than your local village and its not near any major threat so there are no military instalations nearby.

[Krenske]

Hi all,

I am back from the wilds of Africa, Egypt and India, I now have a much better understanding of just what civilisation is. (No it is not the actual buildings but the underlying ideas, education and communication ability. Anyone can take over a civilised infrastructure but give it a generation or 2 and not much is left.)

I have been able to aply some thought to the 40% complete combat model we had in place and have decided to give it some strategic beatings until it reaches 45% or maybe more. It is apparent that the existing version is veryflexible but would also be very difficult to code an AI for.

I am attempting to simplify the model by removing some of the phase repetition (in the scouting and maneuver phases) replacing it with a single step. A victory in the scouting phase will modify the maneuver phase and a victory in the maneuver phase will modify the mass combat\assault phase.

Hopefully the result will be a faster resolution following a systematic approach not requiring a lot of AI intervention in the actual resolution of each battle.


Note I did a quick review during some of my flights of average troop concentrations in battles over the last 5000 years and found that they rarely got larger than 40000. Our current element based system assumes the use of 500 man elements with damage counted in the 100's of men. So almost all historical battles at least up to the age of napoleon would have about 80 elements per side which equates to around 6-10 actual units within each TF. Even modern battles rarely involved more than 40 000 men in any one engagement. Just some information to put things into some scale.

[This message has been edited by Krenske (edited January 23, 2001).]

[Mark_Everson]

LGJ: Well, we'll see how it works out... I don't think its a major issue.

All:

There is some discussion on the building of military units, elements etc. taking place in the Infrastructure Model thread. We are also discussing things like shipyards for building battleships etc. Its just the last few posts in that thread that will be of interest from a military perspective.

[Mark_Everson]

Hi Paul, welcome back!

Your ideas / observations sound good to me. Looking forward to more .

Although what you say about troop densities sounds right, modern battles will be much more complicated, because a single square won't contain the battle. (As you well know.) So there'll be interaction between lots of battles going on in individual squares. That's going to be more of a challenge for the AI than even the complexity of the current single-square battle model! But it sounds like your simplifications to the model won't really cost much in terms of the flavor of the system anyway. You might want to think in the background about how the model handles big modern battles that roll from square to square... is there always a full scouting phase when an existing battle just grows with addition of new troops etc.?

But anyway, for the soonest demos just getting the single-square ancient battles to work right will be enough of a challenge. If you get the chance we should get a few ancient units actually specified so Laurent D. can acid-test the details of the combat model with semi-real numbers!

[LDiCesare]

I have coded the very basic combat system and unit system. If you can provide me with how scouting should affect deploy phase and how deployment should affect fight, along with any ideas to simulate combined arms tactics, I could really begin coding the model. Currently in a TF all units strike the first unit of the next TF, which is quite poor but easily coded.
I can think of making a small grid so that the army would face some direction and use scouting/deploy change that direction.
You could also think of cutting in center, left wing, right wing (additionally), deployment being useful there. Any idea?
Additionally, how should upkeep costs should work? Supposing that armies require only money, I think that the amount should change based on their orders (costs more to move than to stay still, even more to attack) and some economic factors, but which ones?

[Mark_Everson]

Hi Laurent:

One comment about armies only requiring money for supply... its true and yet it isn't. At the risk of telling you something you already know, here's a bit from the mil model intro.

quote: The necessity to supply military units with provisions and ammunition plays a very large part in their capabilities. I think we can actually handle supply in Clash without significantly hassling the player. The basic idea is to use dedicated merchants (already need to be coded anyway) to supply the troops. All the player need do is say how much money they're willing to spend to supply x front. The merchant/supply manager does the rest. You would build essentially a special class of merchant that would try to purchase goods from the surrounding area (or home, if transportation is good enough). This supply unit would travel with large armies and provide feedback to the army commander about whether (and at what price) it would be able to supply the army if it went to spot X. Units that were not fully supplied would fight at reduced effectiveness, or have the need to forrage for supplies, reducing movement. Mongol Horsemen could probably forrage fairly effectively, modern armored divisions hardly at all. Using a merchant that goes out and buys supplies for you has the advantage that the player doesn't have to orchestrate the supply itself, but only decide if the price of supply is worth it. A bonus of this supply system is that supply lines can be attacked. This allows proper modeling of a modern envelopment battle, one of the coolest martial activities known to man ;-). In addition the supply system would give a lot of correct flavor in terms of Where large armies could go historically, with little loss in smoothness of gameplay.

So it does only require money for supply, but the money needed won't be the same, even for a given army. It depends on where the army Is. And some places will be prohibitively expensive. Also if there are no paths for the supply line to follow the best merchant quartermaster in the world won't do you any good .

So at least in my thinking supply should be composed of varying amounts of the four basic commodities: food, resources, manufactured goods, and services. Mostly food and manfactured goods in most cases... Under usual circumstances all the player does is sign the check!

[LDiCesare]

I know armies don't need just money, I wrote that in order to simplify things, since I'll probably code supplies of only one thing in order to test it before I switch to lots of stuff.

[LDiCesare]

Currently, fights are coded as follows:
When armies meet, they check whether to flee or not. Attackers may not be able to flee, defenders always succeed (I took the behaviour from Mark's demo4 code).
When they fight, there should be scouting and deployment phases which determine whether they can flee or not, (like you can flee only if you deploy before the opponent).
The actual fight takes place afterwards. Currently the first unit found in a TF is attacked and may be killed in one round. Then I check again for flight and iterate (3 times in my test, which is arbitrary).

Questions:
Should the fight be allowed to end without a victor (both armies still alive, no one fled yet, ready to fight for the next turn)?
To put it another way, how many armies are allowed to die before you break off or stop fighting in order to let reinforcements come in?
Any ideas of how to use deployment? Does right wing, left wing, center front, rearguard and center/HQ make sense?

[Mark_Everson]

Hi Laurent:

Here is my take on your questions. Hopefully Krenske will get his comments in soon too.

Should the fight be allowed to end without a victor (both armies still alive, no one fled yet, ready to fight for the next turn)?

Yes, I think this is essential in a final model.

To put it another way, how many armies are allowed to die before you break off or stop fighting in order to let reinforcements come in?

I think periodically the odds should be re-assessed and the general/AI should make the decision based on available knowledge. I did this in demo 4 and it seemed to work sensibly.

Any ideas of how to use deployment? Does right wing, left wing, center front, rearguard and center/HQ make sense?

That sounds good to me. But we need some more specifics. If deployment is at the unit level then what happens if there are two units do you have just center and right with no left? that doesn't seem correct. So IMO if we use this there needs to be a way to thin units so they do more than one job. I could come up with a proposal, but will wait for Krenske to respond... If you get to a point where this is starting to hold you up, let me know and we can figure something between ourselves to use at least temporarily.

[F_Smith]

Just a random musing . . .

For deployment orders, does 'left wing' or 'right wing', or anything so specifically tactical, really work on this strategic scale?

Should orders given represent more of a 'strategic' scale than actual location of individual units? More like:

[*]'Refuse battle' [*]'Attack ignoring losses'[*]'Attack until x% losses received'[*]'Defend at all costs'[*]'Fighting withdrawl'[*]'Recon in force'?[/list]

In other words, don't give the Task Force 'deployment orders', but give them 'mission orders' on how to react to an enemy presence?

Just thinking out loud.

[Mark_Everson]

F_Smith:

I think both types of orders have a place in the game. At least left flank will have meaning if the battle model isn't greatly simplified from what it is now. (Though for front-level modern battles such things don't mean as much.)

But I think since Krenske is just working over the battle model that its much more practical to work on the more generic type of orders first. We could go a long way just with the more strategic sorts of orders and a very simple battle model like my old demo 4 one.

What do you think Laurent & Krenske?

[F_Smith]

The main thing I'm thinking about 'tactical' orders is that this will be so dependent on local battlefield terrain and conditions (a scale so small that we won't be including any info on it in the game) that there is almost no way to make that kind of decision in advance.

Agincourt's deployment worked *only* because of the specific terrain and weather. On a flat, open plain with no rain, the tactical choice that Henry made would have been suicide. I just don't see playing this game on that level of detail.

Am I wrong?

[Lord God Jinnai]

I haveto agree with F_Smith here. For most player, deciding whether to hold your cavlry back for a surprise flank won't be happening too much. Also he's right about the terrain. What good will a charge order do uphill against the well defended enemy as apposed to downhill? The player won't know whether the batte is uphill or downhill, but will be able to tell his generals to pull back if there's too much loss (and depending on the general, since ocassionaly the general will disobey because he thinks he can win or doesn't want his men to die needlessly).

Also for what the AI is concerned, when determining how these battles are won or tied, i think we should keep it simple w/o sacrafising too much in strategy because this will really bog down turns later in the game if your generals have all the various type of options your thinkng about.

[Krenske]

At present I am working on having no actual deployment orders as it does not truly fit with the scale and needs to take terrain into account. General execution orders do though have a place. I currently believe that a TF should be given various orders which effectively break down to move, attack, fortify and defend. These will have a intensity qualifier that ranges from light to all out. The order determines the type of engagement and the intensity is used in morale checks to see if the TF has reached its retirement point.

Terrain is taken into account as a modifier to the defence values and as a limiter on opening engagement ranges. The effects of terrain can be reduced/enhanced through scouting and successful maneuvering. The more mobile force, with better scouts and a good general will normally make much better use of terrain.

Please note that with many weapons unit dispersal is minimal and thus they have a very low maximum benefit from terrain. In the days of legions and phalanx's both sides strove to maintain dense formations and as such they have almost no benefit from terrain due to the commanders of each side looking for suitable battle terrain. In fact trying to fight with low dispersal in high terrain values will incurr major penalties to firepower due to disruption of formations.

Tactical orders could be used if actual battle maps were to be used with individually maneuvreing units. That though is not what the initial model is to provide. It should be a way of determining the results of a battle between two forces somewhere within a X Km across square.

[Mark_Everson]

Hi All, good discussion! I think Krenske's approach strikes a reasonable balance. What does everyone else think?

It occurs to me though that maybe we should divert some effort at this point to specing out some units at least approaching the final stats we're going for. The reason is that building of units ties into econ and population (and maybe even ethnic group) areas. So I think it would be beneficial to at least sketch out this area in more detail so we find any inconsistencies quicker. We certainly should be able to build units / TFs in the demos fairly soon.

Because the battle model is more self-contained I'd favor pushing some discussion of building units ahead of it. Since just a raw comparison of military power a la demo 4 can get us fairly far IMO. Opinions, flames, etc.?

[Krenske]

Hi all

I have started a new combat model draft, it is available at http://www.usq.edu.au/users/krenske/game/g/combat.htm .

I will gladly update the pages with any alterations desired and decided upon.

As to the unit based discussion, I am all in favour. Unfortunately I will have be away for 3 days just now but I will be back.

A couple of notes though regarding the model as it currently stands. Each strength point of the unit (read as elements) equates to approximately 500 men. The elements would make a suitable unit building block for a square. Just about any square would be able to free up 500 men to form a element.

A legion of 5-6000 men would have about 10-12 elements that can be gathered from a range of squares within a region within a turn. Thus allowing rapid (and realistic) unit construction times. What the elements are equipped with can be the more time consuming thing. Building up a store of 500 suits of armour/assault rifles/warhorses takes time and will most likely be done ahead of time to some degree.

[LDiCesare]

My proposal wasn't so much about tactical orders as about "how can I leverage combined arms tactics?"
It is easier to do it realistically if I know the catapult is behind the phalanx and no cavalry can outflank it.
I agree with Mark that we can wait for such detail.

In devising units, can we start at the archaic and try to think of tech requirements?

[Mark_Everson]

Hi Krenske:

Thanks for putting the new model up, although I could have used to guide to what was old and what was new, since a lot of things looked really familiar . I wrote down some comments to pass on to you, but I left them at work... oops. The one thing I remember from the list is that you have a lot of things like training levels and experience levels that you imply are integers. I really want to stay away from things that change in big jumps. If you just meant that they are reported to the player as "green", "seasoned", "veteran" etc. I got no problem with that. But behind the scenes a unit that is most of the way from seasoned to veteran should fight almost like a veteran, so I think floating point values are the way to go for these things.

Just remembered another thing... in the units writeup it says that elements are 100 men, I take it that 500 is the new definition? That sounds better to me anyway... I look forward to getting you into the unit build discussions.

Laurent:

Yes, I think we should definitely start with ancient units. However, since the technology system really isn't there I'm not sure we gain much by thinking about tech requirements. I think will scratch that I think we will be there in maybe a month or so... The way I would go about development now is for us to pick two or three kinds of units we think would be interesting. (Perhaps the ones we already have graphics for would be best) We could also define some classes of Weapons, and figure out how to build either elements or whole units as defined by Krenske. Not that we should have every value that in his table yet, just a few would probably be sufficient. At that point we could then build units, train them, and also fight very simplified battles just using the units hand-to-hand range and firepower. Having different weapons types could be interesting, especially if you do a "rise of the Hittites" scenario where the player is the only one that has iron weapons. I guess iron weapons would have something like twice the "firepower" of the bronze equivalent.

Anyway, those are my opinions on the direction we should be going. I admit it's partly colored by the fact that I am running the economy model, and would like to start to get an idea of how units weapons and such are built .

[LDiCesare]

I read Krenske's model.
It looks like I'll have a few things to code .

Paul, I currently have TaskForce/Unit subdivision in the code. It seems you also have Unit/Element. I think you unit is a small task force. Currently, TFs are made of Aries which can be either TFs or Units. I'd tend to map my (code) Unit to your Element and the (code)TaskForce to both your Unit and TaskForce. I may be missing something, though. What do you think? Note a side effect of TFs being made of TFs is that you can break down whatever level of detail you'd like to (but I don't think any player would do that).

There are sevral infos about what a unit description should have in the model, but it is very elaborate given the current state of military code.
In order to define our units, we should know what will define them: I write it as a pseudo-xml format because it could be put in text files somewhere and I saw code which looked like paring xml so here it goes:
(I call unit types "archetypes", actual units referring to their archetype - the f's at the end of figures are for 'float', never mind if you do not program)
The phalanx figures ARE SILLY. I just want to show what data we could use (probably I won't code support for much of what I write here within the next weeks but well...)
I put with a ? those that I haven't begun coding at all.

Archetype name = "Phalanx"
Attack value = "1.0f"/
Defense value = "1.0f"/
DistanceAttack value = "0.0f"/
Move value = "2.5f"/
AllowedOrders value ="Attack,Garrison,Fortify"/
!-- note Fortify mayn't be allowed for an elephant unit special abilities like: Spying, Scouting, Skirmishing, can go here --/
RequiredTech name = "BronzeWorking" level = "0.5f"/
!-- power increases when bronzeworking increases beyond that point. --/
?Morale value = "1.0f"/
?Type name = "Man-based"/
?Mobility value = "1.0f"/
?ProductionCost value = "1.0f"/
?Armor value = "0.0f"/
?UpkeepCost value = "1.0f"/
?UseableWeapons list = "clubs,bronze weapons"/
/Archetype

The figures above could be the default for all types of units, we could then precise only those that are specialized when telling of a specific unit.
What info do you think are needed?
Anyone wants to add or remove some?
Paul, which ones do you deem most important if we stick to ancient warfare for now?
Mark, can you tell the econ production figures that are needed for the econ model?

[Mark_Everson]

Hi Laurent:

On building units... as far as I know in the OO design F. Smith had thought we should have units (perhaps elements in this case) assembled from Weapons and manpower with Training. Let's leave the manpower aside for now, since we will just arbitrarily grab a chunk of men from the populace until the population model gets further along. The cost to build the unit is simply then the cost for the weapons plus the cost for minimal training. I've already talked about that a little bit in the infrastructure thread, so I will just quote it here. Note that when I give the cost for unit I am lumping together the cost of weapons (pretty much solely manufactured goods abbreviated by P for production) and training which is almost purely services (S). I refer to the cost below in triples of the fundamental goods (F, P, S). If we want to have a system for specifying arbitrary units from pieces then we probably need to add a doctrine like "shock" for guys that charge at you and try to hack you to bits .

quote: Originally posted by Mark_Everson on 01-22-2001 07:30 PM I am now coding building of military units. The approach I'm taking is that when there are orders to build a military unit a temporary infra class is created to build that unit. The cost of that infra class will be the raw materials (food, manufactured goods, and services) amounts needed to complete the unit. So ancient warriors with clubs might cost (0, 3, 2) in the same order as above. The services need is mostly initial training. Additionally the people would also be needed. A legion might cost (0, 70, 50), and a modern armored division (0, 90,000, 120,000). When the cost is reached the temp infra class disappears, and out pops a shiny new unit.

So in the case of warriors, their clubs in light armor might cost 3P, and their minimal training 2S. For the legion it's a lot more of both. I hope this gives you what you need!

[F_Smith]

This may already be a part of the system, I might have missed it. But I thought I'd ask . . .

Will an 'element' be 'supported' by the province that creates it, every turn?

Like the game Diplomacy, since the time frame we're dealing with doesn't really fit with raising a specific group of soldiers (you'd have to re-raise the legion every turn), can an element be tied to the province that produces it, and that province (or mapsquare, I don't really know the specifics) provide replacements?

[Mark_Everson]

Krenske:

Here's the thing that I forgot that troubled me in the unit/element model. The description ascribes defence to dispersion and armor. For close-combat during much of the ancient period, dispersion is of no benefit. I think mixing dispersion with shields/armor is a bit weird. Did I misunderstand, can you comment?

F_Smith:

"Will an 'element' be 'supported' by the province that creates it, every turn?"

Generally not. Read the paragraph on supply in the intro to the mil model on the web page. (you need to scroll down a bit on the mil page to see it)

"Like the game Diplomacy, since the time frame we're dealing with doesn't really fit with raising a specific group of soldiers (you'd have to re-raise the legion every turn), can an element be tied to the province that produces it, and that province (or mapsquare, I don't really know the specifics) provide replacements?"

When the soldiers are at/near home I'd say yes. But if they're a long ways away they will need to re-man with locals. But an army will be able to get as far as men can get in the real world. At least that's the current plan!

[Lord God Jinnai]

I just have a question here, nothing too important right now, but i wanted to see what everyone thought.

If 2 TF units enter the same square from the same civ, but differnt provinces in a very loosely connected civ, either conferderation or loose feudal system, and the player controls in theory both of them, will they fight each other if they haven't a reason not to (ie both heading for a battle against a common enemy) and the leaders of each province hate each other. Remember, i said that this was a loose alignment, not a imperial or federation type gov where such would be more strictly controled from up above.

[Mark_Everson]

I'm not sure how far we can get into inter-provincial rivalries in a feudal or other loosely-controlled system. I think it'll just have to wait 'till we get much further along.