“Battle of the video games” [December/January TSJ] questioned the outcome of the “Game After Ambush!” (GAA) competition, which was awarded to the Laser Shot/Bohemia Interactive/Calytrix team for fielding “Virtual Battlespace 2” (VBS2) as the Army’s centerpiece game-for-training effort. The article centered on VBS2, a contract effort selected by the Army Program Executive Office for Simulation, Training & Instrumentation, and “America’s Army” (AA), an Army secretariat effort, and their supposed “duel for U.S. Army training business.” Having served in a leadership role for both of these Army game efforts, I wanted to add a few thoughts on the Army’s clear decision to adopt VBS2. More importantly, the question “What is next?” deserves, but does not have, such a clear answer.

Contractor vs. government

The article posed the question of which would win in a competition between commercial and government solutions for a large-scale serious training game. Having experience on both sides of these efforts, I am convinced that only the commercial sector can perform efficiently. In fact, this is already the case, with both AA and VBS2 being built on commercial game engines. Despite appearances that AA is an in-house government effort, its team — like that of VBS2 — has been staffed by commercial contractors. Game engines are not end products in themselves, but are complex products built to handle the virtual physics computations demanded of games. Engines are made by sophisticated game engine shops and are adaptable to a wide variety of applications. Building and maintaining the game engine is a significant and specialized effort, generally exceeding two to three years of a full team’s effort for the initial product. So, until the Army develops its own game engine from scratch, only commercial contractors will have experience and long-term motivation to program at the source level. Contractors win.

A more relevant issue is whether a contractor or the government could be more efficient in delivering the Army’s training needs. Here, too, I believe the contractor solution dominates. In the case of the GAA contract, the Army required a training game immediately upon award. The Program Executive Office for Simulation, Training & Instrumentation (PEO STRI) contract reflected the needs of the combat developer (Training and Doctrine Command), with delivery, new equipment training and train-the-trainer courses to begin within 60 days. The Laser Shot/Bohemia/Calytrix team had the people and resources in place on Day One, and it began a 54-station delivery over the first seven months of the contract. In contrast, a similar government team did not exist; it would have needed significant time to ramp up and would undoubtedly have pushed the Army’s training timetables significantly to the right.

Also, as economists would note and common sense would dictate, the only effective method of incentivizing performance is economic. To satisfy the Army’s requirement for delivery of a robust training device quickly, PEO STRI structured the GAA contract upon concrete deliverables, with financial payouts after success. An alternative model of a government agency delivery could not be based on economic incentives, but rather on administrative marks on the wall. As I saw many times at AA, the chain of command could create deadlines and self-direct changes for a wide variety of reasons. There was no change process to abide by and no real penalties (economic or other). So all delivery dates were considered malleable. This model worked fine for a free-of-charge public outreach and recruiting game. However, such “squishy” schedules are unacceptable for real-world training requirements.

Overemphasis on code

Much of the TSJ article steered the reader to consider source code as a major issue. One AA leader said “America’s Army” is “obviously going to be inexpensive, because we own the source code.” This statement is certainly open to debate. First, ownership of source code does not imply that the development team can do anything meaningful with it, let alone create a game that would be inexpensive. In video game development teams, modelers, artists and level designers build games on top of a game engine using common, higher level tools and scripting, not by source-level programming. This explains why many games are released with common engines, such as AA, “Unreal Tournament,” “ChaosUT,” “Tactical Ops” and VBS2/”Armed Assault”/”ArmA 2.” Source code changes and additions at AA were always accomplished by the originating commercial engine team, not in-house. So in the end analysis, there was no great cost advantage transferred to the Army by owning the source code of the game engine.

At AA, the source code was owned in order for the government to ensure a safe repository in case the commercial contractor (Epic Games) ever suffered some disaster and to ensure code compliance. For security reasons, this may be good practice. Although the GAA contract has not required source code to date, the Laser Shot/Bohemia/Calytrix team could easily arrange the delivery of source code to the Army if ever desired. (Of course, this would raise the question as to how the Army would maintain the source code.) Certainly, even without VBS2 source code, the Army has not suffered any cost or security injuries.

The supposed duel

The article’s implied duel between competing titans does not exist, nor should it exist. The Army’s virtual training needs are many and varied, as originated and vetted by the combat requirement developers. Given the current state of serious game technology, it is not logical to advocate a single game strategy. I agree with my GAA partner Pete Morrison, CEO of Bohemia Interactive Simulations, that “no one size fits all” in virtual training. Instead, there will continue to be room for many one-off specific solutions — with the caveat that all solutions be capable of integration with other live-virtual-constructive-gaming (LVCG) efforts. The Army, led by PEO STRI, is pursuing this multisolution strategy.

Multisolution does not imply no dominant solution. By 2009, the Army rightfully needed to create significant commonality in virtual training and designate one primary virtual platform as the standard for which many — but not all — training requirements could be based. This became the full and open competition — not a duel — for an enterprise solution, known as the “Game After Ambush!” In the end, the winner needed to satisfy the Battle Command Training Centers and combat unit needs for virtual training — without waiting. The Army needed a game that was ready to deploy, something literally off the shelf, something that had legs for future development and was versatile enough to be used in each of the LVCG domains. The Laser Shot/Bohemia/Calytrix solution was totally ready, with ready-at-hand interfaces to support user created and edited scenarios, thousands of real-world models, a large usable terrain of hundreds of kilometers and the critically important ability to interoperate in a distributed environment using high-level architecture/distributed interactive simulation (HLA/DIS) standards. AA and the other competing games could not provide such critical features, without significant and lengthy development. In fact, during the technical demonstration phase, VBS2 met 100 percent of the requirements, while I would estimate that AA fell well short — owing mainly to its original design and funding as a public game, not a training tool.

Integration

VBS2 has proven easy to migrate to many applications in the LVCG domains. In its first year, the Laser Shot/Bohemia team has introduced convoy, improvised-explosive-device defeat, call for fire, marksmanship, gunnery and other trainers. The inclusion of the Calytrix “LVC Game” (an HLA/DIS interface within GAA) meant that VBS2 was easy to integrate with other distributed simulations such as Joint Semi-Automated Forces, the Joint Conflict and Tactical Simulation, OneSAF, and SIMPLE (Simulation to C4I Interchange Module for Plans, Logistics and Exercises). And, the ease and speed of introducing real-world terrain data is a key feature that makes VBS2 a serious training and mission-rehearsal tool. Although these features are possible to attain in AA’s engine with skilled effort, the AA game platform could not exhibit such ease of use.

Cost issues

Government development of a serious training game of the breadth and complexity of VBS2 will cost more, period. VBS2 was designed specifically for military use. Its adoption across many military forces at the enterprise level — U.S. Army, U.S. Marine Corps, U.K. Ministry of Defense, Australian Defense Forces, New Zealand Defense Forces — has created a class of privileged stakeholders who share the cost of developing new features, models and terrains, if nonclassified. Cost sharing among the members of this large user community causes the long-run cost of ownership to fall, much like any industrial economy of scale. The military user group also creates a situation of positive network externalities, much like the classic telephone network example, where member benefits increase as total adoption is increased. If AA had been chosen instead of VBS2, these economic effects would be felt only within the Army budget, not across a pool of external users. The total cost of ownership would be borne fully by the U.S. taxpayer instead of being spread across the budgets of many armies.

As a prime contractor for two Army and Coast Guard programs of record, I can see that contractors are best able to account for and control costs. The effect of standardized Defense Contract Audit Agency rules of accounting for overhead, general and administrative expenses, and limits on profit, help to ensure that all costs are properly allocated. Most importantly, full and open competition drives down the total cost. In-house government efforts such as AA need not adhere to cost allocation guidelines nor competition, so the true costs are never really known and inevitably float upward.

What’s next, dual or duel?

PEO STRI’s acquisition strategy for GAA included competition from both government and commercial games. The office made the right choice in VBS2, based on technological and economic rationales, and did not eliminate competitors because they did not have “enough water towers or cars with orange bumpers” as one AA proponent claimed in the TSJ article.

Although VBS2 is now the Army’s single standard for serious gaming, AA could indeed continue as an excellent strategic communication tool for reaching young Americans. AA’s success is well-known and well-deserved. My years serving that program rank as some of the best in my life. But there is a yet another new wrinkle that begs to be considered.

Wisely, in the GAA contract PEO STRI negotiated the inclusion of a version of VBS2 known as VBS2-Lite. Its intended use is to support the nonclassified, small-unit training needs of the Army with an Internet deployable training package for use on personal computers. However, the inclusion of VBS2-Lite seems much more visionary. Did PEO STRI also foresee other uses, such as a new Army game in the public communication realm? Did it see that with VBS2’s real — not simply perceived — “serious” factor, VBS2-Lite could be a major hit with the Army’s target recruiting market? Did PEO STRI contemplate a very economically efficient means to broaden the commonality of Army game usage from pre-recruitment messaging through individual and collective training, as well as mission rehearsal? Whether intended or not, VBS2-Lite can serve these purposes.

So, in addition to its Army-wide training game, PEO STRI owns an option for a VBS2 recruitment game. The existence of this option spurred a question that, when asked recently, I was unable to answer: “Is there an overall strategy for the Army’s two massive game efforts?”

One thing is certain. PEO STRI definitively concluded the “game duel” issue. Remaining on the Army staff’s plate is the “dual game” issue.

Christopher Chambers is president of Laser Shot of Stafford, Texas. He is the former deputy director of the Army Game Project, which produced the “America’s Army” series of video games, and their large supporting network and marketing infrastructure. A retired Army infantry officer, Chambers served in numerous command and staff positions, and as an assistant professor at the U.S. Military Academy at West Point.