Interview: James Sterrett, Professional Wargamer

Americanians, did you know that 0.000000000323% of every dollar you pay in tax is spent ensuring a man in Kansas by the name of James Sterrett gets any PC wargame he desires? Recently I crossed Flare Paths with this mysterious freeloader and asked him to justify his lavish government-funded strategy habit. He pointed out he was Deputy Chief, Simulations Division, Digital Leader Development Center, at the Command & General Staff College, Fort Leavenworth and the games in question were used to hone and assess the skills of the US Army's captains, majors, lieutenant-colonels, colonels, and planners. Totally unfazed, I stared at my shoes for a bit then mumbled “Ah. Right. I see.”

It turns out that James' sizeable wargames budget has some sizeable strings attached. In addition to the fun-sounding stuff - organizing the College's game-based exercises, and researching new software (“We are all encouraged to look for new simulations. If they look promising, or if they seem to have something we can learn from them, then the examination may be pretty detailed. For example, we spent a couple of Friday afternoons this fall playing GMT’s boardgame Andean Abyss”.) his employers also expect him and his team to train other trainers, and do a whole heap of mundane administrative stuff. Reluctantly shelving my Pulitzer dreams and “PENTAGON PAYS MAN TO PLAY COMBAT MISSION ALL DAY” headline, I set about finding out a little more about Dr. Sterrett and his intriguing relationship with digital warfare.

RPS: How did you end up doing what you do?

James: I’ve been a gamer since childhood. I was given a copy of SPI’s Strike Force One, followed by a variety of other SPI and Avalon Hill games. Later, in graduate school, my wife and I discovered a group that plays the original Prussian General Staff training game, Kriegsspiel, in London. Kriegsspiel is umpired and double-blind; everything gets processed at the umpire’s map and reported back out to the players.

About a year later, TacOps came out, and a group of us on the mailing list wanted to play it with teams, which was not possible with that early version. I realized I could umpire it in the manner of Kriegsspiel. Many of the players were serving or retired military, and they taught me a great deal about running exercises.

In 2004, while waiting for my dissertation defense, I heard on the TacOps mailing list that CGSC’s Simulation Division used TacOps, and was hiring. I figured that I’d never forgive myself if I didn’t apply, and they chose to take me on board. In fact, of the five Army Civilians in the Simulations Division, three come from the TacOps mailing list. So, in a sense, those three of us owe our jobs to TacOps’ creator, Major Holdridge.

In several senses, I owe my job to Lieutenant von Reisswitz, who created Kriegsspiel. Not only was Kriegsspiel my gateway into training simulations, but it is also the great-grandfather of virtually all military simulations, so without Reiswitz’s Kriegsspiel, my job might not exist!

On a different note, I owe my brother’s life to a computer game. Back in the early 1990s, we had a surgery computer game, Life and Death. The initial stages of the game require the player to diagnose four or five different lower abdominal pains, such as gas, constipation, and appendicitis. As a result, you learn the symptoms of appendicitis, and you learn the lethal result of letting it go untreated.

My brother tends to try to sleep things off instead of calling the doctor… but when he got a sharp, stabbing pain in the lower right quadrant of his abdomen, halfway between the hip bone and the pubic bone, he recognized the symptoms as appendicitis. Instead of trying to sleep it off, he got himself taken to the Emergency Room right away. The surgeon told him later that it was the most-inflamed appendix he had seen in 20 years. If my brother had delayed, he would be dead.

Handled correctly, by both the creator, instructor, and student, games are a phenomenally effective teaching tool. Experience is a great teacher, and well-designed games can deliver experiences that are tailored to drive home learning.

RPS: I believe you use Decisive Action at CGSC. What makes this game such a useful teaching aid?

James: Decisive Action’s design is tightly focused on making decisions about the synchronization and orchestration of forces as a commander and key staff officers of a division or corps – precisely what we need to run an exercise in which the students take on the roles of the key staff officers in a reinforced division.

Decisive Action is so closely tailored to our needs for two reasons. First, it began life at CGSC: then-Lieutenant Colonel Jim Lunsford was a tactics instructor at CGSC in the mid-1990s, and created the first version for us in his own classroom. He worked some kind of deal to retain the rights to the program when he retired (which is quite unusual), which brings us to the second reason: CGSC has periodically contracted Jim Lunsford to modify Decisive Action to suit us better.

RPS: Can you go into details about the kind of scenarios you and your team use?

I can talk, in general terms, about the scenario used in our Command & General Staff Operations Course [a year-long course for majors, intended to take students who have been company-grade officers and teach them to be field-grade staff officers at brigade and above] division exercise.

In the first of three nested scenarios, the students plan the entry of a Joint Task Force (a corps-level force with Army, Navy, Air Force, and Marine assets) into a theater in order to stabilize the region; country A (OPFOR, Opposing Force) is threatening to invade country B (we’ll call it Bakerland) and is fomenting an insurgency in Bakerland as well.

The second scenario is the division exercise, and it begins as the Army division in the JTF begins to approach the area it intends to defend in order to prevent OPFOR from being able to invade Bakerland. When the division is a few hundred kilometers away (the exact distance varies a lot from instructor to instructor), OPFOR decides to invade anyway. Its direction of advance is across the direction of advance of the US force, so the US force has the opportunity to attack rapidly and cut off country A’s forces if they attack quickly and well, but OPFOR’s force has four divisions and significant support, so it’s a fairly difficult task.

The third scenario focuses on one brigade from the division, tasked with stability operations in the area that was the objective in the division scenario. In theory, anything the students blew up in the second scenario is still blown up in the third.

Keep in mind that the scenarios are designed to teach students about the staff process, not to necessarily make perfect sense or prepare for operations in any specific part of the world. Thus, the attack when the US is on the verge of being in position is a bit silly, but it creates the conditions for various dilemmas the instructors want the students to face. For example, because OPFOR’s attack on Bakerland is still in progress, they have not solidified a defense against the US division yet. The major physical barrier to the US division is a river. OPFOR has a division guarding half of the river, on the direct route to the division’s objective, and the sector that has most of the bridges. The students have to make a decision: attack through that division to make use of the shorter route, or bypass and screen the division, accepting a longer route but avoiding combat until they are closer to the objective? What the instructors grade the student’s planning and execution on is not which direction they choose – it is whether the students understand the dilemmas each route poses, and whether they have done a good job of planning for the problems and dealing with them as they arise. If they attack directly, how will they prevent OPFOR from blowing the bridges, and what is their contingency plan if bridges are blown? If they bypass, how will they deal with the longer routes, the small number of bridges, and the need to secure their flanks against the bypassed division?

RPS: When most of us balls-up a digital battle, we reload or walk away and forget about it. What happens when one of your students makes a frightful hash of things?

Keeping in mind that I’m not a tactics instructor – when students screw things up, they can expect to have a discussion about it with their instructor, focused on how to avoid the mistake in the future. Mistakes in the classroom are embarrassing, but we take them seriously so they don’t become battlefield mistakes that cost lives. Students often describe the exercises as interesting, or engaging, or educational. They don’t often describe them as fun.

Also, what students are graded on is making decisions for the right reasons, not on the outcome. Much of CGSC is teaching process, on the assumption that good process will lead to victory. That seems strange until you consider that the disorganization caused by bad staff process dooms an army – critical information doesn’t get to commanders, orders are garbled, supplies don’t get dispatched to where they need to go, and so on. An efficient, effective staff makes victory possible.

RPS: What other sorts of exercises do you run at CGSC?

James: CGSC structures its curriculum according to the “Experiential Learning Model”. In principle, any block of learning begins with a Concrete Experience (CE), which is intended to get the students interested in the topic. If you have a good CE, you won’t be able to prevent students from talking about it; in this Publish and Process (P&P) stage, the instructor wants to let the students discuss, and collect themes to discuss further in the Generalize New Information (GNI) stage. Traditional GNI is a lecture; but there’s no reason it cannot be a guided discussion, and it’s entirely possible for the students to be unaware of the discussion shifting back and forth between P&P and GNI as various topics come up and are thought through. Finally, there is the Apply stage, which is traditionally some kind of test, and may also be the CE of the next round of the cycle.

The division exercise I’ve described is very much an Apply-phase event.

We also use a Concrete Experience game called Future Force, which is built to drive one student, or a small group of students, through an exercise lasting an hour or so in which they wrestle with planning the Army’s force structure. They don't necessarily get a lot of planning time, the sim is very easy to use, and the purpose is to get the students thinking about various issues in Force Management, so the AAR is very gentle. Force Management has traditionally been a dry, bureaucracy-laden topic that was routinely blasted in student feedback as the most-hated block of instruction at CGSC. Future Force hasn’t made the topic loved, but it is no longer routinely hated, because the simulation helps introduce students to the bigger picture of why it matters to them.

The least common place to use simulations is in GNI. They can be extremely effective at teaching new information and concepts, but they are time-consuming in doing it (in principle, this is the stage of learning that taught my brother to diagnose appendicitis). This is, however, where some instructors will include more use of games such as Decisive Action and Combat Mission Shock Force; this is also where we have some experiments running with games such as Command Ops and Scourge of War.

RPS: Are there dangers as well as advantages to using computer wargames as training tools?

James: Certainly. Any simulation is a simplification of reality – and that’s why we use them, to avoid the cost in money and lives of conducting more realistic training. After all, we could conduct perfectly realistic training by issuing live ammunition to soldiers and having them fight. This would arguably produce battle-hardened units, but the cost in lives would be frightful. Thus, we need to have unrealism in simulations.

The trick is in controlling the realism and the unrealism in order to ensure that you teach the lessons you intend to teach, without teaching bad habits. For example, VBS2 (which is nearly the same as ARMA II) can be very good at teaching teamwork and battle drills, but it has some limitations. On the lesser end, moving a mouse has nothing to do with aiming a rifle, but this is obvious enough that no corrective action is likely to be required. More powerfully, the fact that you don’t suffer when your avatar dies can encourage soldiers to take greater risks in the game than they ought to. I’ve heard of a variety of creative punishments used to ensure that virtual death carries some real-world price, in order to keep this in check.

The issues in our sims are more subtle, but I know one instructor has ended exercises by assigning to students, as homework, the task of writing letters to the mothers of three soldiers who died in the simulated battle they just fought, in order to drive further reflection on the seriousness of getting their planning right.

RPS: What's your take on the difference between a game and a sim?

James: I tend to use the terms interchangeably. However, the definitions can be fun to argue about, and there’s a definite need to struggle against the common informal definition that “games are bad, simulations are good”.

That usually refines into “simulations are realistic, games are for fun”. This definition doesn’t really get you a useful distinction, especially because some things that are sold as games are also sold for military training; examples include TacOps, Steel Beasts, and Armed Assault II/VBS2.

The quick, but sometimes useful, definition is that there is no difference – it all depends on the purpose for which it is being used. Thus, the same piece of software is a game when used for entertainment, and a simulation when used for training. This definition is useful when somebody is complaining that they are using an unworthy game and should be using a wonderful simulation, and moving them to a much more useful discussion of whether or not the specific tool they are using suits their training needs. If it does, then it’s good. If it doesn’t suit the intended use, then it doesn’t matter if we call it a game, a simulation, or a gold-plated frog leg: no name is going to make it better.

Finally, the official AMSO (Army Modeling & Simulation Office) definitions turn out to work remarkably well. Paraphrasing, AMSO defines a model as a mathematical description, a simulation as a model iterated over time, and a game as a simulation with goals and decisions. By this definition, the troops-on-the-ground exercises run at the National Training Center are games – and the military used to call them “wargames” – and Candyland is not a game. (Candyland is a simulation: Candyland derives from Snakes & Ladders, and Snakes & Ladders derives from an Indian game in which the chutes are vices, the ladders are virtues, and the goal is Buddhist Enlightenment… a simulation of the Buddhist worldview.) However, while these definitions are logical, they don’t do much to handle people complaining.

(I’m intrigued that AMSO’s definition of a game is remarkably similar in its overall result in to Greg Costikyan’s definition in his excellent article, “I Have No Words and I Must Design”.)

RPS: Putting aside the stuff games can't simulate (responsibility for actual flesh-and-blood soldiers, personal danger, Napoleon's haemorrhoids etc.) are there aspects of military command that, in your experience, most games simulate poorly?

James: We usually don’t plan very well when we play games as civilians, not least because good planning is hard work! However, leaving aside the difference in the consequences of our decisions, and the fact that the real world doesn’t allow for setting up the pieces for a second try at the scenario, there are three things gamers usually do not have to deal with.

First, we usually have far better knowledge of the situation than is possible for real armies; consider that one of the key pieces of information from ULTRA decrypts was the Axis order of battle in various theaters – simply knowing what units the Axis had was a major intelligence coup, but such information is routinely handed to players. Moreover, the scenario usually tells us what the friendly and enemy win conditions are, while those are often less clear in real life.

Second, in nearly every game, our forces do exactly what we tell them to do, exactly when we tell them to do it. In the real world, subordinate forces need time to conduct their own planning so they can carry out our orders, and they may not go about the task exactly as we envisioned. (The best game I’ve played for experiencing these challenges is Panther’s Command Ops series with the Command Delay set to the maximum value. I’ve also heard good things about
Scourge of War in this regard but have not personally played it.)

Third, gamers are usually planning by themselves, which means they have to explain everything only to themselves and to the game. Military staffs deal with more information than one person can process; even a battalion staff is likely to be several dozen people. Getting this many people to pass information among themselves efficiently, and let alone coming up with a coherent plan that everybody understands, requires practice.

“Plans are worthless. Planning is essential.”, often attributed to General Eisenhower, sums up the Army’s approach to planning. The planning is meant to make the staff analyze the situation, figure out its salient features, figure out what possible plans they and their enemies might use, compare those plans, select the plan most likely to succeed, and, perhaps most critically, to try to forecast the decisions that will need to be made. Gamers tend to make a decent job of all but the last, at least informally. However, figuring out, in advance, what combinations of friendly and enemy actions will require specific responses in order to continue towards success, is fairly rare for gamers. That’s largely because there is much less need to do it; when you need only explain the plan to yourself, your forces do exactly as they are told when they are told, and you have most of the necessary information, it’s easy to change plans on the fly. Figuring out decision points and explaining them to your own staff and to subordinate units is how the Army tries to get a gamer’s level of flexibility built into its plans. It takes a lot more work to figure out what your Decision Points will be, and issue orders to account for them; but once you do, it allows large organizations to react more rapidly to specified events because they already know what the reaction is supposed to be.

The Military Decision-Making Process (MDMP) is a formal planning process: a plan on how to make a plan, so that all the people on the staff know what the dance steps are and thus what they need to be doing and when. It certainly does not guarantee a good plan. However, it greatly improves the probability that a large group of people can come up with a reasonably complete plan, even when they haven’t slept in three days. Much of what CGSOC teaches centers on learning and practicing MDMP, so that our students can be useful staff officers after they graduate.

Even given their advantages, gamers can benefit from learning the planning process as well, because the structured approach can help you think through your own planning more completely and be better prepared for the surprises your opponent throws at you.

RPS: Do civilians feature in any of the wargames you use at CGSC?

James: It’s not quite that simple. First, we draw the distinction not as military and civilians, but as non-combatants and combatants; combatants includes both uniformed military and also the various non-uniformed people who are shooting at you.

Second, civilians are definitely present, both as combatants and as non-combatants, in some of the scenarios, but how they are handled in the simulation varies. UrbanSim, built to be a counterinsurgency simulation, handles both types directly. Decisive Action, built for major combat operations, can handle combatants and non-combatants, but they are events controlled by the instructor – so the instructor uses tools on the host computer to add effects for civilian refugees clogging roads, or insurgents attacking unescorted logistics convoys, or the impacts of students not using appropriate restraint on firepower; but the decision on what happens comes from the instructor, not the simulation.

RPS: Instructors embellishing simulations - jerry-rigging them to better reflect a particular real-life situation - is that usual?

James: Military exercises are generally full of workarounds and “injects”. Workarounds are the things we do in the simulation to make it do things it was never intended to do. For example, a simulation that has no naval vessels needs naval gunfire support, so we put artillery on an island and declare the island to be the ship. Sometimes these work out well, and sometimes they cause further trouble; the ship example is (in)famous because the island couldn’t be sunk by enemy airpower.

Injects are used for things the simulation doesn’t cover and workarounds cannot (or should not…) cover them. (In some cases, where no simulation can cover the topic adequately, the injects are the entire exercise!) Each inject should be a specific piece of information that will be given to the students. The full writeup of the inject should include where the information is from, whom it goes to, what the staff is expected to do with it, and what the next inject is depending on what they do – similar to the chaining of paragraphs in a Choose Your Own Adventure. A given chain of injects is called a thread. Each thread should test or teach the staff on one or more themes the exercise is intended to test or teach. When you put all the threads together, they form the Master Scenario Event List, or MSEL (pronounced like measles without the final “s”).

Not all threads will necessarily be used in a given exercise; that is up to the Exercise Controller (EXCON), who makes the final decision on which ones to use, or not, based on how well the staff is doing and what they need to work on.

RPS: Does licensing a game like CMSF mean you can demand tweaks?

James: Not unless we pay for them! If we simply bought a game from a store, then we get the same updates as everybody else. If we bought it via a development contract, then we will specify what features we want added, and we get a lot of bug fixes as well. However, that costs rather more money because we are hiring the developers.

In CMSF’s case, we simply bought sufficient copies from a store to have a license for each computer we expected to use it on, so we have no right to demand anything from Battlefront.

All that being said, developers are usually interested to hear what we might want, and sometimes work those features into subsequent releases if it looks like the feature will both help with their primary market, and might also garner military sales. There’s no contract and no promises involved. With military budgets dropping and the internal hurdles to buying software rising, it’s unfortunately a risky bet.

RPS: Can you think of any wargames currently used exclusively for training, that may have what it takes to succeed in the recreational sphere?

James: Unfortunately, most Army wargames are not much fun to play, especially if they began life as Army sims. Armed Assault II is the sibling of the military VBS2 – but that series began life as the famed commercial hit, Operation Flashpoint. The closest to you can get a sim built initially for the Army is probably Pat Proctor’s BCT Commander, a series originally closely based on the military’s JANUS simulation. Note, however, that BCT has a more user-friendly interface than JANUS.

RPS: Does using wargames for a living mean you're now less inclined to play them in free time?

James: It changes the games I play somewhat – if I’m using a given title at work then I’m much less likely to play it at home. I am still an avid gamer. Strategy/wargames I’ve spent quality time with in the past year or so include XCOM, X-Com 1 & 3, Steel Beasts Pro PE, Civilization V, and Crusader Kings II on PC (does FTL count in this category?), plus Neuroshima Hex and Small General on my Android phone. I’m also a boardgamer; one of the highlights of the past year was getting a group together to play Hube’s Pocket, and we periodically stage a Kriegsspiel game as well.

RPS: Swingeing defence cuts have led to the closure of CGSC, and an unexpected inheritance has left you $10,000,000 richer! What wargame would you make?

James: Several oddball projects come to mind, and I don’t know if they could all be done for $10 million or even £10 million! They also show the influence of years at CGSC, I’m afraid.

First, a game in which the player is the chief of a supporting arm in a staff – chief of artillery, or chief of logistics, for example – and has to support somebody else’s plan. That plan may be a bad one, and the player can try to influence the commander to improve the plan, but the player has to earn influence by making operations succeed through having their branch do a good job. I’ve been battering at how this ought to work for years but never have come up with a really good solution.

Second, a game that was not only decently fun to play but rewarded the players for working together as staff officers and commanders. Team-play wargames are pretty rare as it is, and I don’t know of any in which a player can have an interesting job as a staff officer unless you count the Artemis starship bridge simulator. There’s potentially a lot of fun to be had there, though. Our team has periodically gone to game conventions and run staff exercises for people at the convention – teaching them the staff roles and the Army planning process, helping them through planning, coaching them through fighting the battle as staff officers while we operate the simulation in accordance with their orders, and then conducting an AAR – and typically the people who try it have a great time.

Third, in the daydream department, I’d like to build a version of a table-sized computer display that can run a complex boardgame really effectively. There are three advantages that boardgames have over computer games (leaving aside the numerous advantages computer games have; I am a fan of both). First, boardgames have a decided advantage in letting me see the entire map at once; a screen the size of a table, or a VR system of some sort, would solve that. Their second advantage lies in letting me see all of the simulation’s moving parts and alter them at will – but that can be solved by good documentation and editing tools. Their third advantage is tactile, which perhaps could be solved by VR gloves or a smart surface of some sort.

RPS: Truly fascinating! Thanks for your time.