War-Gaming Insurgency

[AmericanPride]

In another thread (on military staffs), I mentioned I'm involved with a project for MIOBC. After some consideration, I have decided to switch subjects from military staffs to war-gaming insurgency.

Here's the gist of my argument:


The Army requires a simple, flexible, and objective tactical decision simulator in order to enable officers at all echelons to study and understand the complex environment of counter-insurgency warfare. The Army should create a browser-based, common-access multiplayer COIN simulator.

My focus will be on practical questions: time, personnel, resources, and most importantly IMO, usability. Using a common-program language (such as javascript or PHP), the Army IMO will be able to create an user-friendly, inexpensive, tailorable, and effective COIN simulator for any officer or any rank.

Thoughts?

[wm]

In another thread (on military staffs), I mentioned I'm involved with a project for MIOBC. After some consideration, I have decided to switch subjects from military staffs to war-gaming insurgency.

Here's the gist of my argument:


The Army requires a simple, flexible, and objective tactical decision simulator in order to enable officers at all echelons to study and understand the complex environment of counter-insurgency warfare. The Army should create a browser-based, common-access multiplayer COIN simulator.
My focus will be on practical questions: time, personnel, resources, and most importantly IMO, usability. Using a common-program language (such as javascript or PHP), the Army IMO will be able to create an user-friendly, inexpensive, tailorable, and effective COIN simulator for any officer or any rank.

Thoughts?

I trust the inner quotation above summarizes the two conclusions you plan to draw from this analysis, not your thesis. By the way, without a significant amount of new argument, I do not see how you can infer your second, normative (should create . . .) conclusion from your first, descriptive one. I'm also interested in seeing what kind of premises you intend to marshal to show that the Army needs a "tactical decision simuilator" to train "officers at all echelons." Why do folks working at higher echelons, where they should be doing stuff at the operational and strategic levels, need training in tactical decison making? Ojne last point--no decision is objective IMHO. Each varies depending on the character of the decider and all the aspects of METT-TC knownto and/or understood by the decider at the time of the decision.

Good luck. But you seem to have predetermined your conclusion.

[MikeF]

American Pride,

At the Naval Postgraduate School (NPS), in Monterey, CA, there are two Model and Simulation (M&sS) on-going projects focused on Irregular Warfare that are unclassified. Below is a brief description of each. They may peak your interest.

Furthermore, after you've done so time on the line, you could consider pursuing a master's degree at NPS (through Advanced Civil Schooling (ACS)) with your thesis work in wargaming IW.

Hope this helps.

1. CORE LAB (COMMON OPERATING RESEARCH ENVIRONMENT)
- DEFENSE ANALYSIS DEPARTMENT
-Methodology- Social Networking (think google earth and Anaylist Notebook)

http://www.nps.edu/research/CoreLab/index.html

2. TRAC Monterey and Operation Research Deparment
-Methodology: Wargaming using Values and Beliefs of indigenous population to model and simulate IW environment.

http://www.nps.edu/research/TRAC/

v/r

Mike

[Meh]

Not to sound too harsh, but doesn't Virtual Battle Space (http://virtualbattlespace.vbs2.com/index.php?option=com_frontpage&Itemid=1) already fit the description? If not, then surely the most efficient solution would be to contact the developers, Bohemia Interactive Australia (BIA), in order to work with them on a COIN module?

Fully developed 3D simulators will always beat the living crap out of browser-based apps, both in terms of immersion and flexibility.

Every pogue and his dog is getting funding for simulations of questionable value right now. There's a myriad of different apps floating about within the DoD and abroad. Everyone's trying to reinvent the wheel. NiH, NiH, NiH! Why can't people just settle on something like VBS and then improve that?

Perhaps you could make it your life goal to unify these efforts instead of placing one more small, imperfect fish into the sea?

[AmericanPride]

I trust the inner quotation above summarizes the two conclusions you plan to draw from this analysis, not your thesis. By the way, without a significant amount of new argument, I do not see how you can infer your second, normative (should create . . .) conclusion from your first, descriptive one.

Yes. I also intend to build a demo version of this game over the course of the next several weeks. Anyone with knowledge of a common programming language can do so. And it can also be built to allow anyone to tailor scenarios to their own needs. My intent is not to design a simulator of a particular tactical engagement or theoretical framework, but to build a cheap, user-friendly, common access program that enables anyone to use it.


I'm also interested in seeing what kind of premises you intend to marshal to show that the Army needs a "tactical decision simuilator" to train "officers at all echelons." Why do folks working at higher echelons, where they should be doing stuff at the operational and strategic levels, need training in tactical decison making?

I think the Army could use a better training system for decision-making and war-gaming in a COIN environment. The program will be tailorable for the needs of whatever echelon is using it; the program will only provide the shells of actions, events, demographics, units, etc and how they interact with one another. Users will then be able to input/delete/modify the components of those shells to fit their needs. That 'kit' can then be saved and used elsewhere as necessary. This will allow battalions to model their AOs and combatant commands to model their theater.


Ojne last point--no decision is objective IMHO. Each varies depending on the character of the decider and all the aspects of METT-TC knownto and/or understood by the decider at the time of the decision.

I agree. Because the effects of all actions and events will have to be formulized, I think it would be best to allow the user to also modify the impact of those effects based on their own estimates. Before the start of the game, the same administrator (i.e. the S-2) that modified the enemy strength and composition can also apply his estimates of intangiable attributes and relationships.


Hope this helps.

Thanks Mike. I'll look into them.


Not to sound too harsh, but doesn't Virtual Battle Space already fit the description?

Not in my opinion. This is why: 1) It is not server-based. It has specific system requirements to support its engine. A server-based game would only require access to the internet. 2) It has limited tailorability. The idea I envision would allow anyone at any echelon to simulate any COIN scenario (local, provincial, national, etc). 3) AFAIK, it does not model the political, social, economic, and intangible aspects of COIN essential to decision-making. It will also be simple to use, commonly available, tailorable for any user, and objective in its results.


Fully developed 3D simulators will always beat the living crap out of browser-based apps, both in terms of immersion and flexibility.

In immersion, absolutely. In flexibility, I disagree.

[Meh]

Just to ballpark your ideas, what about this?

http://www2.aperianglobal.com/files/gssoldier/

What graphical representations will you use? 3D? Stickmen?

[AmericanPride]

Just to ballpark your ideas, what about this?

http://www2.aperianglobal.com/files/gssoldier/

What graphical representations will you use? 3D? Stickmen?

The Global Smart Soldier is a very interesting product. IMO, however, it seems more focused as a training tool than a war-gaming tool. By that, I mean it seems to support the idea that there is a right and wrong answer to every problem. The idea I have in mind is more sand-box, letting the users define the end-state and the appropriate method to get there.

As for graphical representations, that's a legit question, and one to which I have yet to develop an answer. An idea could be to allow users to upload maps to their kits, scale them, and designate key/important terrain/features on it. This could work for any level of operation. Your question is something I'll be mulling over through the weekend.

[Meh]

I really look forward to seeing what you deliver. My brain just can't figure out how you'd do it.

I don't see how you can overcome the inherent limitations of browser-based programming without a 3D engine. How would you represent the scenario? How much detail? Just maps? Isometric with sprites like the role-playing games of old? Text-based (Zork)? In order to have a sandbox, you need an intuitive interface with no coding and minimal scripting that allows the user to instantly see where she's placed everything. You can't beat 3D.

Why not take a lesson from online games like World of Warcraft? Use client software that feeds "quests" off a central server (America's Army). Or do this:

I presume your target audience is O-4s and below. Most of these guys will have grown up with PCs and consoles. A lot of them are gamers. You could democratise the whole process by creating a game they can relate to - modifying an existing one is easiest - that allows anyone to create scenarios and then post them to a central website where others can download them (e.g. "my time in Shukranville when we received IF while speaking to a local schoolteacher and tended to civ cas while maintaining all-round security which led to Sgt Hickey taking out a guy on a rooftop who was trying to set up with grenades" - L/Cpl Snuffy).

Your key to a successful COIN simulator is not the enemy. It's civilians. Accurately modelling civilians is much harder than creating an AI that will shoot at you. Civilians also require dialogue trees.

Without civilians, just have a combat simulator. There's enough of those around. Combat Mission. VBS. That's not COIN.

I'd actually suggest you get a copy of the 2001 game, Operation Flashpoint, and play around with its editor. It's an excellent commercial warfare simulator (as these things go). Only things missing are dialogue scripting and COIN-related AI coding (particularly civilians). Who knows, it's old enough that it might be possible to embed the engine in a browser-based game.
I hear Neverwinter Nights, a role-playing game, has an excellent scenario editor which allows creation of complex dialogue trees. If you could 'mod' that game to make all the textures and models contemporary, it might turn into a good small-unit simulator.

PS: Far Cry 2 is a good example of a powerful, user-friendly editor that a caveman can use.

[AmericanPride]

Thanks for the suggestions Meh. I have Operation Flashpoint, and it's a great game IMO. No doubt, part of the appeal in gaming (and presentations in general really) is graphics. And that's definitely a hurdle. I do intend to make the game as democratic as possible, allowing users to create scenarios to fit their unit's needs, to store them in a database, and to allow others to access and modify them as necessary. I'm sure it sounds incoherent right now, as I plan on completing the concept sometime this week, and then start coding.

I do not intend to have any AI in the program -- everything will be user-driven. Military units, the enemy, civilians, NGOs, and so on. I'll flesh out a clearer and complete game concept at the end of the weekend. My friends and I have been working on a similar concept to model the Dune universe -- and so I'll basically be reshaping some of those concepts to fit my needs here. Here's an example (http://www.nationmaster.com/encyclopedia/Star-Wars-Combine) of something similar -- but of course be tailored for the military (US Army) audience.

[bongotastic]

I've been wrestling with COIN wargaming for about two years. I designed/umpire a small MBX (http://www.opcon.org/SadrCity/) based on COIN scenario in Sadr City, Iraq.

If you want a 3D environment, what you'll end-up with is a MOUT wargame with assymetric forces and civilians. That's really small unit battles with an exotic ROE. I'd go with Combat MIssion: SF or Flashpoint (and a lot of scripting).

However, my observation in the MBX is that the most intersting part of COIN is the intelligence aspect of the battle. Most of my game mechanics derives from a RPG called Twillight 2000. This ruleset does great at soft factors (interrogation, investigation/observation) and to resolve the most creative ideas coming out of the player's imagination. I've got a prototype small unit sim based on this ruleset (loosely).

In a few words, my take on COIN is primarily about intelligence and networks of people. This is probably something that you want to consider if you are aiming to do something beyond a small unit assymetric urban battle game. It can be done fairly easily, but this isn't something all that common in other wargames.

I look forward to hear more about your project.

[AmericanPride]

Bongotastic,

I agree with your points. I aim (hope) to capture those intangible features of COIN that are not present in other systems/programs/sims. I'm almost done with working out the rough edges of the concept, and I'll put that up here later this evening.

[AmericanPride]

Alright, here it goes:

1. Intent. The intent is to build a viable and flexible COIN simulator that successfully captures the political, economic, and social dynamics of the COIN environment in addition to the typical military missions. It is intended for multiple users to play the same session simultaneously. Because there are a large number of theories on the origins and characteristics of insurgencies, the system will have sufficient flexibility to enable the user to modify it to his own needs.

2. System. It will be powered by an online application (I'll be using PHP and MySQL) to enable the creation of an interactive website. This will allow anyone, anywhere to access the program, to create their own sessions, and to modify their scenarios according to their needs. Users will be able to modify everything from the progression of time, the scale of the scenario, objectives, units, actions, and effects. Further, users will be able to upload and modify maps, and identify key terrain and characteristics (similarly to ArcGis, but more user-friendly). All of these modifications can be saved and stored collectively as 'kits', allowing users to share them with one another.

3. Gaming Features. Each user will control a 'faction'. A faction is a military or political organization, such as a combat unit (of any size), a terrorist group, or NGO. Before the start of each session, each faction will be assigned a list of objectives, by which it's credibility will be measured. Factions will have other ratings, such as legitimacy, organization, etc. The user, through the faction, will have control of various units. These units can be as varied as known persons, military vehicles or echelons of any size, or non-conventional equipment. These units will have various ratings such as ethnicity, leadership, speed, etc. Units will conduct actions, such as tactical missions. Each action will have specified effects, such as increasing legitimacy. Each action will also take a specified amount of time (or phases) to accomplish, and will have predetermined indicators. Prior to each session, the user will have the opportunity to create, modify, or delete factions, units, actions, and effects. Users issue orders to their units to conduct actions either in real-time or in phases (determined prior to the session). A resolution calculator will determine combat effects and losses, the successful completion of actions, and random events (if desired). The results of each decision, and the end-state of the scenario as a whole, can be stored and viewed in order to allow analysis of decision points and outcomes. Objectivity will be attained by the use of formulas; however the inputs will be determined beforehand based on intelligence estimates (especially of the intangible values).

The total cost of the program is under 30$ (the cost of purchasing a programming for dummies book).

[bongotastic]

If you throw some javascript and "AJAX", you can break out of the php/mysql server-side mode and truly have an interactive game right out of the browser. One interesting option with browser is the ability to render 3D scenes using a raytracing engine (or an openGL engine) on demand and integrate these to the game interface. Raytraced images can end-up looking even better than interactive graphics.

As I mentioned before, I've got quite a bit of code and a prototype for a combat resolution system. If you intend to go work collaboratively, PM me.

Christian

[BayonetBrant]

Holy Crap... you want *any* type of COIN to be sim-able at *any* level of echelon?

Boy that's a stretch. There's are a *huge* variety of factors that all apply at different levels, and from a pure game-design standpoint, you really, really need to make sure your objectives are clearly defined, the tools available to the players appropriately further those goals, and that the echelons at which the players are making their decisions are appropriate to the echelon at which they should be training (ie, BDE S-4s don't control BN-level LOGPACs).

I'm not saying this can't all be done, but good Lord that's a huge task from a game-design perspective, before you ever start writing a single line of code. No matter how good your code is, if the underlying game-mechanics aren't wired tight, the game will fail as a training tool.

If you want more game design ideas/advice/info - especially about COIN - I recommend reading this article (http://www.escapistmagazine.com/articles/view/issues/issue_167/5248-The-Thinking-Mans-Warfare). Beyond that, feel free to PM me and let's chat some.

[Cavguy]

I just don't think a simulation wargame can account for the human-centric complexity of a COIN/Stability environment. Perhaps with a real-live Red Team, immersed in the thought/thinking context of the host, but predicting second and third order effects of human interaction has remained near-impossible. I can't predict my wife's reactions, so what makes me think I can predict second and third order effects in an alien culture with a computer?

http://blog.wired.com/defense/2009/02/nato-wants-sim.html


Veteran counterinsurgents have long been skeptical of how accurate these models can really be. "Wait a minute, you can’t tell me who’s going to a win a football game. And now you’re going to replicate free will?" retired Lieutenant Colonel John Nagl, who helped write the Army's manual on defusing insurgencies, told Danger Room in 2007. "They are smoking something they shouldn't be," retired Lieutenant General Paul Van Riper quipped to Science Magazine. "Only those who don’t know how the real world works will be suckers for this stuff.”

[Steve Blair]

I just don't think a simulation wargame can account for the human-centric complexity of a COIN/Stability environment. Perhaps with a real-live Red Team, immersed in the thought/thinking context of the host, but predicting second and third order effects of human interaction has remained near-impossible. I can't predict my wife's reactions, so what makes me think I can predict second and third order effects in an alien culture with a computer?

http://blog.wired.com/defense/2009/02/nato-wants-sim.html

I'm a much bigger fan of the RPG model for any sort of COIN simulation, which is really where your concern is going, Neil. In order to get a good, 'realistic' simulation you need people in a more or less freeplay environment (with some cultural and other limitations controlled or at least monitored by a white cell).

[Cavguy]

Agreed. Something like a "Knights of the Old Republic" or "Mass Effect" type RPG would give action/consequence feedback - which I think could be valuable.

[MikeF]

To reiterate CavGuy's comments, it's is extremely difficult to model real world behavior in a computer program, video game, or wargaming application. Sec Gates warned (http://www.foreignaffairs.org/20090101faessay88103/robert-m-gates/how-to-reprogram-the-pentagon.html), "No one should ever neglect the psychological, cultural, political, and human dimensions of warfare. War is inevitably tragic, inefficient, and uncertain, and it is important to be skeptical of systems analyses, computer models, game theories, or doctrines that suggest otherwise. We should look askance at idealistic, triumphalist, or ethnocentric notions of future conflict that aspire to transcend the immutable principles and ugly realities of war, that imagine it is possible to cow, shock, or awe an enemy into submission, instead of tracking enemies down hilltop by hilltop, house by house, block by bloody block."

This warning should not suggest that it is not an achievable goal to develop an IW wargaming. It is simply a caution to let you know that it is not easy.

As I posted earlier, there a numerous smart PH-D level operational research and social scientist working on this same topic. They temper there efforts with Sec Gates warning, and they continually ask advice from guys in the field who have been doing this for a while.

From what I've observed, there are two methodologies that they are testing to see if they can wargame IW. Below are a brief description and links to further research.

Take some time to research these theories and see if they help. Again, good luck.

1. Theory of Planned Behavior (http://en.wikipedia.org/wiki/Theory_of_planned_behavior).

Reasoned Action suggests that a person's behavior is determined by his/her intention to perform the behavior and that this intention is, in turn, a function of his/her attitude toward the behavior and his/her subjective norm. The best predictor of behavior is intention. Intention is the cognitive representation of a person's readiness to perform a given behavior, and it is considered to be the immediate antecedent of behavior. This intention is determined by three things: their attitude toward the specific behavior, their subjective norms and their
perceived behavioral control. The theory of planned behavior holds that only specific attitudes toward the behavior in question can be expected to predict that behavior. In addition to measuring attitudes toward the behavior, we also need to measure people's subjective norms -their beliefs about how people they care about will view the behavior in question. To predict someone's intentions, knowing these beliefs can be as important as knowing the person's attitudes. Finally, perceived behavioral control influences intentions. Perceived behavioral control refers to people's perceptions of their ability to perform a given behavior...(Ajzen & Fishbein, 1980)

2. Expected Value-Rational ChoiceTheory (http://en.wikipedia.org/wiki/Rational_actor_model)

Rational choice theory, also known as rational action theory, is a framework for understanding and often formally modeling social and economic behavior. It is the dominant theoretical paradigm in microeconomics. It is also central to modern political science and is used by scholars in other disciplines such as sociology and philosophy.

The 'rationality' described by rational choice theory is different from the colloquial and most philosophical uses of rationality. Although models of rational choice are diverse, all assume individuals choose the best action according to stable preference functions and constraints facing them. Most models have additional assumptions. Proponents of rational choice models do not claim that a model's assumptions are a full description of reality, only that good models can aid reasoning and provide help in formulating falsifiable hypotheses, whether intuitive or not. Successful hypotheses are those that survive empirical tests.

v/r

Mike

[Cavguy]

This warning should not suggest that it is not an achievable goal to develop an IW wargaming. It is simply a caution to let you know that it is not easy.

As I posted earlier, there a numerous smart PH-D level operational research and social scientist working on this same topic. They temper there efforts with Sec Gates warning, and they continually ask advice from guys in the field who have been doing this for a while.

From what I've observed, there are two methodologies that they are testing to see if they can wargame IW. Below are a brief description and links to further research.

Take some time to research these theories and see if they help. Again, good luck.



Both theories are valid, but fall victim to "Black Swans" pretty easily, but that doesn't disqualify their utility.

However, have played a lot of strategy games in the past, especially ones like the "Civilization" series and derivatives therof (space conquest, etc). Almost all feature a negotiation/alliance model for trade and security. However, I have yet to find one that looks like real-world relations for war decisions, even though most are based in rational choice theory. People aren't that mathematical/logical in the real world, and coding of the social/environmental/internal political constraints seems to be the missing factor - but the hardest to quantify.

[Steve Blair]

Both theories are valid, but fall victim to "Black Swans" pretty easily, but that doesn't disqualify their utility.

However, have played a lot of strategy games in the past, especially ones like the "Civilization" series and derivatives therof (space conquest, etc). Almost all feature a negotiation/alliance model for trade and security. However, I have yet to find one that looks like real-world relations for war decisions, even though most are based in rational choice theory. People aren't that mathematical/logical in the real world, and coding of the social/environmental/internal political constraints seems to be the missing factor - but the hardest to quantify.

And that's one reason I tend to go back to the freeplay RPG style. It's about the only way you can really capture the rather unique elements that go into some human decision-making. Strategy games can capture this to a degree, but some of the better versions I've seen have been "house rules" built for things like Risk and the like. A robust RPG framework with computer aid (for resolution of some tasks like development and local infrastructure issues) still strikes me as the best way to go, although it would also be the most intensive in terms of manpower (training, white cell, and so on).

[bongotastic]

A whole lot can be modeled nowadays. The technology to do it isn't very hard to work with, but it is still mostly used in research labs and proprietary software outfits. There are ways to encode what we know about an issue using knowledge models (or ontologies) and perform reasoning on these.

Furthermore, the ability that software can learn and take decisions in blurry situation is today easy to implement, and pretty robust. The days of neural networks are long gone now, thank God. I know a bit about this because I do research in the field and keep on wondering on how this stuff would apply to modeling 4GW (which unfortunately has nothing to do with what the research money is there for).

Take knowledge models and robust learning algorithm and a lot can then be done. As I mentioned before, COIN isn't going to be all that much about the battles, but about the context of these battles. COIN is definitively a setup for an RPG, but software can do now much more than what most game/sim have to offer.

[MikeF]

Both theories are valid, but fall victim to "Black Swans" pretty easily, but that doesn't disqualify their utility.

However, have played a lot of strategy games in the past, especially ones like the "Civilization" series and derivatives therof (space conquest, etc). Almost all feature a negotiation/alliance model for trade and security. However, I have yet to find one that looks like real-world relations for war decisions, even though most are based in rational choice theory. People aren't that mathematical/logical in the real world, and coding of the social/environmental/internal political constraints seems to be the missing factor - but the hardest to quantify.

True dat. Neil's experiences and COIN efforts in Anbar Province would not translate to success in Diyala. Conversely, my approach to Diyala would not have translated to Anbar. Each province has a unique "conflict eco-system" (Kilcullen) or social fabric.

So, I'm a little skeptical on the ability for the M&S community to develop an application that FORECAST behavior.

However, these programs y'all are working on may help to develop training tools. Back in the day, we used to head to CCTT and UCOFT to practice tank gunnery and maneuver in a 3D environment. While it was not as realistic as the OIF Thunder Runs, it served a purpose of helping to training. I found some utility in taking my platoon there.

Simarly, I am optimistic that someone will develop a useful wargaming or RPG to serve as a training tool for the next level of company grade officers and young soldiers/NCOs to practice PRIOR to experiencing the blunt force trauma of on the job training.

v/r

Mike

[BayonetBrant]

A whole lot can be modeled nowadays.


I guess you need to decide up front if this is just a training tool to help (eventual) practitioners learn when/how to make decisions, or whether or not this is a modelling/rehearsal tool for testing the decisions before deploying them. One requires much greater accuracy than the other.


If you're just trying to help teach the guys who/when/how to make decisions, then you don't need a perfect model of the reactions of people from downrange to learn that the JAG does need to be involved in some targeting decisions and that the Chaplain can be useful for more than just supporting your own guys.


If, however, you want a robust predictive model that says "if we do this action, with these assets, in this province, within this time frame, we should see this specific result within the population" then you're smoking dope - that's just too hard to accurately predict, and when it goes wrong, someone will inevitably scream "but the sime told us this was a good idea!"

[Cavguy]

True dat. Neil's experiences and COIN efforts in Anbar Province would not translate to success in Diyala. Conversely, my approach to Diyala would not have translated to Anbar. Each province has a unique "conflict eco-system" (Kilcullen) or social fabric.

So, I'm a little skeptical on the ability for the M&S community to develop an application that FORECAST behavior.

However, these programs y'all are working on may help to develop training tools. Back in the day, we used to head to CCTT and UCOFT to practice tank gunnery and maneuver in a 3D environment. While it was not as realistic as the OIF Thunder Runs, it served a purpose of helping to training. I found some utility in taking my platoon there.

Simarly, I am optimistic that someone will develop a useful wargaming or RPG to serve as a training tool for the next level of company grade officers and young soldiers/NCOs to practice PRIOR to experiencing the blunt force trauma of on the job training.

v/r

Mike

Mike,

Good points. What I would like to see is a series of "choose your adventure" type game/scenarios where your actions affect the environment - i.e. how you act during a raid, negotiation, etc.

At the risk of making the actions formulatic, it can reinforce best principles, especially if logical consequenses ensue for actions. That's why I have enjoyed BioWare's series of RPG's the past few years - your choices significantly influence the route/outcome of the game, and the actions of those around you.

[bongotastic]

If, however, you want a robust predictive model that says "if we do this action, with these assets, in this province, within this time frame, we should see this specific result within the population" then you're smoking dope - that's just too hard to accurately predict, and when it goes wrong, someone will inevitably scream "but the sime told us this was a good idea!"

100% with you. You can't do that. For training though, what's needed is to bend the randomized reaction to account for a number of key factors to reward good decision making. In the end, this can't be anything else than rolling a die.

[Rex Brynen]

And that's one reason I tend to go back to the freeplay RPG style. It's about the only way you can really capture the rather unique elements that go into some human decision-making. Strategy games can capture this to a degree, but some of the better versions I've seen have been "house rules" built for things like Risk and the like. A robust RPG framework with computer aid (for resolution of some tasks like development and local infrastructure issues) still strikes me as the best way to go, although it would also be the most intensive in terms of manpower (training, white cell, and so on).

I absolutely agree—there is so much that unexpectedly crops up in a COIN or peace operation setting that a hard-coded computer simulator just couldn't handle, but which an experienced White Cell can easily deal with.

A case in point, from my own classroom simulation:


One year, the UNICEF player put together an integrated child/maternal health initiative for our fictional war-torn country. She was a very bright international development student, had done her homework, and frankly did an excellent job. Because donor funds were limited, she decided to target its initial application to those areas with the highest infant mortality rates.

The program was to be conducted in conjunction with local NGOs, and those local NGOs also offered a family planning component. Because of the nature of our simulated civil war, infant mortality rates were highest in the southern areas (where the war was largely being fought). Those areas were predominately inhabited by Zaharians, a secessionist ethnic minority.

The ethnic insurgents in those areas, who were in sensitive peace negotiations with the government through UN channels at the time, immediately condemned the UNICEF program as a "UN sponsored eugenics program intended to lower the Zaharian birth-rate." They complained bitterly that most of the areas being targeted had a Zaharian majority. A few insurgent units even went so far as to detain UNICEF and UNDP workers ("to protect them from the righteous wrath of the people," of course).

This was all a cynical ploy to increase pressure on the UN SRSG in the negotiations—right down to the organization of noisy demonstration by diaspora supporters outside UN headquarters in NY ("Peace, yes! Eugenics, no!") It worked wonders, as the SRSG started to press the government for concessions to mollify the angry Zaharians, and pressed other UN agencies to offer increased humanitarian and development assistance in the south.

The UN folks were released a week or so later. The SRSG read the riot act to the UN aid agencies, and insisted on a new structure for UN coordination that would increase political oversight. The UN agencies grumbled. The peace negotiations continued.

That particular intersection of ethnic demographic politics, peace negotiations, development assistance, and internal UN dynamics has only happened that one year out of the ten or so that I've run the sim. It was hugely instructive for the students, illustrating my constant lectures on 'all aid is political" in a way my lectures never could. More to the point here, it seems to me doubtful that an AI-based computer simulation would have been able to anticipate, capture, and moderate it as effectively.

The problem with an RPG or frei kriegspiel approach is that it is very dependent on experienced moderators—its not just something you can ship off to folks for local training.

[Steve Blair]

The problem with an RPG or frei kriegspiel approach is that it is very dependent on experienced moderators—its not just something you can ship off to folks for local training.

Quite so, but that's an issue I'd much rather deal with than an incomplete or broken computer model. However, it might just be possible to have a central white cell and conduct remote training via one of the online gaming models out there (text based might work best, as then folks would focus on the scenario and not the pretty pictures).

[AmericanPride]

I'm not saying this can't all be done, but good Lord that's a huge task from a game-design perspective, before you ever start writing a single line of code. No matter how good your code is, if the underlying game-mechanics aren't wired tight, the game will fail as a training tool.

Good points. My intent is to enable 'kits', which will enable users to define scenarios (factions, units, actions, and effects) according to their needs. There will be base ratings for each category, and how they interact with one another will remain the same, but the relative value of each will differ according to the estimates of the users (established prior to starting a session). So while a faction will always have units that conduct actions that have effects, the types, scopes, and capabilities of those units, actions, and effects will differ. The number and types (determined by the actions available) of units will determine the extent of a faction's capability. So if a faction is an armor company, its units (and their available actions) will reflect that, and the user will be restrained to those choices. Similarly, if a faction is a terrorist organization, its units (and their available actions) will also reflect that, and the user will be constrained to a different combination of choices. Large or small, regular or irregular. Ultimately both are governed by the same operating principles. I think it will help clarify what choices are available, and which are most likely, in particular conditions.


I just don't think a simulation wargame can account for the human-centric complexity of a COIN/Stability environment. Perhaps with a real-live Red Team, immersed in the thought/thinking context of the host, but predicting second and third order effects of human interaction has remained near-impossible.

That is why I do not intend to program an AI. A permanent Red Team is a good idea, but part of my sell for this program is that it will be light on resources. If used as a war-gaming tool, I imagine the S2 section running the opfor and providing the estimates for the impact of effects in the particular AO.

I would label this simulation as a "strategic RPG". Users will roleplay different factions, but will be making decisions toward the accomplishment of particular aims.

[Rex Brynen]

Quite so, but that's an issue I'd much rather deal with than an incomplete or broken computer model. However, it might just be possible to have a central white cell and conduct remote training via one of the online gaming models out there (text based might work best, as then folks would focus on the scenario and not the pretty pictures).

You can also try to combine a primarily rules-based system (either computer-based or otherwise) with a moderator, to develop a system that is both nuanced but can possibly be provided to others as a training kit.

As I understand it, the World Bank's Carana (http://paxsims.wordpress.com/2009/01/27/carana/) simulation (which explores budget and development prioritization in a fictitious post-conflict country) works this way: part of it looks a bit like a customizable card game (you have limited allocated resources, and an broad array of semi-fixed policy choices), but the presence of a human moderator allows the participants to think outside the box, develop alternative approaches, represent otherwise unrepresented actors and issues, etc. One could easily adapt the approach to focus on, say, the operations of a PRT instead of cabinet-level decisions. The key part of the process is the debate, coalition-building, shared analysis, coordination, etc. among players, in which a traditional pen-and-paper RPG excels.

Maybe, if time ever allows, we could put together something, at least as proof-of-concept.

[Rex Brynen]

This might be of interest, from the Modeling and Simulation Builder for Everyone (mōsbē) software package (http://www.mosbe.com/future/pmesii_models.shtml):


BreakAway has developed tools and editors to enable the scripting of human emotion and behaviors, group and alliance affiliations, and demographic characteristics to the population of a simulated environment. This technology enables a user to assess the overall political, social, economic, and military situation in detail; select and prioritize objectives; plan appropriate strategy and tactics to achieve those objectives; and respond appropriately and effectively adversarial actions to achieve desired end conditions.

The "Force More Powerful" simulation that they showcase on the website has similarities with the mac game Republic: the Revolution (http://www.apple.com/games/articles/2004/02/republic/) (which I happened to pick up for $5 in a bargain bin recently). I sure hope its more stable, runs faster, and has a better interface, though...

EDIT:

There seems to be a whole range of materials on A Force More Powerful, which you'll find here (http://www.aforcemorepowerful.org/game/index.php) and here (http://www.afmpgame.com/).

[selil]

I can do some pretty heavy AI programming with an out of the box ready engine, but from my research the last few week I have found that may be a bad. For the war game I'm considering removing technology is a good idea. I want to figure out the thinking strategies and techniques. So removing crutches or covers is important. In one of the books I recently read (I'm on book 7 of war gaming), the author said chess can be played by a computer, but all you know is the computer knows the rules. Kind of made me think.

I'm slightly stoned on game theory right now...

[bongotastic]

Mike,

Good points. What I would like to see is a series of "choose your adventure" type game/scenarios where your actions affect the environment - i.e. how you act during a raid, negotiation, etc.

I'm paytesting a MATRIX-based game with the French Military. It is a great experience and would be amazing if played with people who were on the ground at some point. The quick gist is that each player gets 5 points. Each point can be used to make an argument, the stronger the argument, the more likely it will pass and influence the game. People can counter-argue, etc. The discussion on the strength of each argument is usually very good. I find very interesting to have to think like an Al-Quaeda operative, or a local Sheik.

It's light on rules, but quite in-depth as far as nailing the issues with COIN.

[BayonetBrant]

For training purposes, the problem with having human arbiters in an exercise is that the lessons that should be learned by the losing team will inevitably be dismissed as arbiter-bias... I've seen that waaaaay too often.

[Steve Blair]

For training purposes, the problem with having human arbiters in an exercise is that the lessons that should be learned by the losing team will inevitably be dismissed as arbiter-bias... I've seen that waaaaay too often.

And you can get the same thing with computerized games....except that it will be the fault of the program or the other team hacked in and cheated. Some people just won't want to learn anything from an exercise.

[MikeF]

For training purposes, the problem with having human arbiters in an exercise is that the lessons that should be learned by the losing team will inevitably be dismissed as arbiter-bias... I've seen that waaaaay too often.

Similarly, while information operations are important when nested within a hollistic COIN strategy, the Great Communicator/Debator approach WILL not unilaterally persuade an insurgent, sheik, or shop owner.

Words are meaningless if not backed with action. One of the reasons an insurgency exists in the first place is the fact that a group of people are so frustrated with their perception of the government that they desire to change it outside the confides of the political process.

v/r

Mike

[AmericanPride]

All,

Thanks for your input. Here are some definitions (that I will use in my brief) to clarify the concept:

A faction is a political or military organization that commands units and is at the direction of a single user.

A unit is a military, political, social, or economic figure or element that conducts actions and is commanded by a faction.

An action is a military, economic, political, or social task conducted by a unit and that has effects.

An effect is a military, political, economic, or social consequence of an action. An effect will be measured in how it alters any of the values for factions and units (ie. increasing/decreasing legitimacy, etc).

As you can imagine, these four components will form the building-block of the sim. I'm busy developing the descriptive values for each (i.e. nationality, ethnicity, religion, cohesion, legitimacy, credibility, loyalty, mobility, survivability, etc). The values can either be numerical or text. As new factions, actions, units, and effects are created, users can mix and match any of them to represent conventional military units, unconventional fighting forces, individuals, or any other kind of organization of any size; military, political, or other.

[defense linguistics]

I'm paytesting a MATRIX-based game with the French Military. It is a great experience and would be amazing if played with people who were on the ground at some point. The quick gist is that each player gets 5 points. Each point can be used to make an argument, the stronger the argument, the more likely it will pass and influence the game. People can counter-argue, etc. The discussion on the strength of each argument is usually very good. I find very interesting to have to think like an Al-Quaeda operative, or a local Sheik.

It's light on rules, but quite in-depth as far as nailing the issues with COIN.

Hey, there Christian! Reading this thread with interest.

I developed this game with Chris Engle after originally trying to come up with a way of implementing TacOps in my classroom. After several failures I've come to the idea that quantifying COIN simulation is a huge, expensive job--especially if you don't know exactly what you want the simulation to teach. Chris and I wanted to engage my trainees with the conceptual framework of COIN and to provide an opportunity to read important documents and immediately employ the concepts therein. I don't care about training particular battle drills, but I do care about learning how to shape narrative.

I recently gave a short presentation to most of the brigade commanders in the Armée de Terre, the notes of which can be found here (https://docs.google.com/fileview?id=F.9528c166-eae6-4864-b0bc-29677be68cdc&hl=en).

I would welcome anyone interested in watching how this exercise works to register as a guest here (http://defenselinguistics.org/moodle/course/view.php?id=7). If you have any trouble logging in as a guest, please PM me and I'll set you up.

Soon, I will be modifying the game for (an American) friend of mine whose next billet will be as a bn XO.

[BayonetBrant]

Just make sure that the 'actions' you allow to the units/factions are appropriate for their anticipated resources, and that they have some way to tangibly affect the other players and/or the victory conditions. If it doesn't affect the VC or other players, then think long and hard about whether or not to include it, b/c at that point it's purely cosmetic.

To do that, you've got to get your VCs straightened out first. That might seem wierd, but especially for a COIN game, you absolutely have to define 'success' for each faction before you get anywhere close to the tools for them to meet that criteria.

[AmericanPride]

Just make sure that the 'actions' you allow to the units/factions are appropriate for their anticipated resources, and that they have some way to tangibly affect the other players and/or the victory conditions. If it doesn't affect the VC or other players, then think long and hard about whether or not to include it, b/c at that point it's purely cosmetic.

To do that, you've got to get your VCs straightened out first. That might seem wierd, but especially for a COIN game, you absolutely have to define 'success' for each faction before you get anywhere close to the tools for them to meet that criteria.

Good points. I intend to let the user define the objectives (or VC) for the factions at the initation of each scenario. The objectives will be expressed either in the completion of X number of actions or the achievement of Y value in a particular or combination of ratings (a full list of which I have yet to develop). The ratings will capture the intangible concepts implicit in war and politics -- legitimacy, credibility, cohesion, morale, etc. As for 'resources' -- that is an excellent point, and something over which I will have to think. Right now, I have a table of twenty or so ratings for units, while the cost of initiating and completing an action will have a degrading effect on any one or combination of those ratings.

[Jason Port]

Pride,

This is really impressive and I am coming really late to the party here. Brant and I work together and have white boarded out this concept with our development team a few times. Conceptually, all things discussed here are extremely important to factor in. However, there are a couple of concepts which may be being overlooked.

1. EBO already has a good framework for examining the impacts of actions by a faction (to use your terms) in the DIME/PMESII concepts. You may want to consider this.

2. Decide if the tool is going to be a trainer of process or a predictor of behavior. Process is easier to model and easier to teach. Behavior lacks the rules and will be significantly harder.

3. Use of an arbiter is critical. As you know systems follow rules, and these rules remain a constant in game play, until your engine/framework are powerful enough to "learn". A key aspect of human behavior is cavguy's idea that we cannot predict our wives, let alone a bomb maker. Balance the rules with the arbiter. Conversely allow the rules a flexibility through constrained random behavior to keep the player from "learning to play the game well"

4. Don't only use your S-2 as the arbiter. From my experience in working with the civil military folks, they will often bring a non-military solution to issues, which are often right. Not disparaging the MI team, as they have certain grown into civil-military guys in their own right. However, I would suggest that the IO or CMO team could help OC the play

5. Consider a non-server based, LAN instantiation of the system. This type of training is invaluable, but Marines on ship, or guys at NTC might not have the internet access your solution requires.

6. Pick an Echelon - If you are really serious about this as a prototype consider narrowing the echelon. The surest way to fail is to try to please all and ultimately please none. If you can do a proof of concept using the Battalion for example, selling other echelons to stakeholders is easier.

However, with all that being said, I submit the obvious, which has been in the subtext of others posts here. If you are starting out with a development environment and a good idea, you might be starting a little behind. There are several engines which are a good start (I have seen MOSBE and it is sexy, but it too has some limitations). You could spend months just examining the engines and finding the right one, but it will also save you years in new development. Beyond that, I would suggest that once you have an engine, this becomes more of a systems integration and software customization effort than wholesale new development (which saves a lot of time and money, in the long run)

Last, there are ideas which I haven't posted as they are propriety or more sensitive. You can PM me and we can chat offline on the concepts. However, I would suggest that there are several major agencies solving for this right now, and I don't believe anyone has a real comprehensive solution, and ultimately, many of them will wind up shelved when users determine that their value only takes them part way. Again, I am not trying to kill any enthusiasm - I believe in this idea, especially the more we rely on the E-6 squad leader to be an ambassador of good will. Let me know how we can help drive this ahead.