What makes a game a good training tool?

[Dook]

For discussion from the TRADOC Capabilities Manager for Gaming Blog

What makes a game a good training tool?

This was one of the questions we asked ourselves when we started developing the requirements documentation for gaming. After a few heated discussions, we determined what makes a simulation such as JCATS a good training tool applies to gaming as well. This is especially true considering the initial focus for gaming of small unit collective training and key leader development. It’s the tools such as terrain editors, scenario editors, after action review capability, interoperability, C2 stimulation, etc that make a good simulation.

With this in mind we began to document the requirements in the form of a capability production document (CPD) for the first increment of gaming capability.

There are three Key Performance Parameters, and nine Key System Attributes.

- KPP 1 Semi-Immersive Training Capability - A semi-immersive training capability enables Soldiers and leaders to train on tasks in the COE when live training resources are not readily available or appropriate at home station and while deployed.

- KPP 2 Leader-Centric Training Capability - Leader-centric training capability prepares the leader to make swift and intuitive decisions in uncertain situations in a training setting prior to deployment and when live training resources are not readily available or appropriate at home station.

- KPP 3 Soldier Enabled Training Development Capability - The Soldier enabled training development capability provides Battalion (leaders & staff), company & below units the ability to quickly modify the gaming content/environment to meet their training objectives.

- KSA 1 Environment - Environment provides the underpinnings upon which the complexity and multi-dimensional nature of real-world operations play out, regardless of the operational environment in which actions are to occur. Commanders, leaders, and Soldiers must understand and contend with the COE variables and their impacts in order to (IOT) prepare for both current and future operations.

- KSA 2 Terrain and Atmospheric Effects - Complex and urban terrain, as well as man-made and natural features provide the basis for physical actions and reactions with other models in the simulation. Operational execution of geospatial information by using systems, units, or forces via collection, generation, storage, management, fusion and dissemination, vertically and horizontally, from peer to peer and National to Soldier level and back.

- KSA 3 Audio and Visual Realities - Audio and visual realities create distracting stimulus that impede users’ ability to focus on the mission and key decision points and provide a more realistic training environment.

- KSA 4 Behaviors - Autonomous and semi-automated forces reduce the resource requirements (cost, personnel and time).

- KSA 5 Equipment - The simulation of actual equipment creates realistic training conditions. The equipment editor will greatly enhance the ability for soldiers to create a more realistic training environment in the simulation. Further, it gives the trainer the ability to adjust the level of difficulty of the training.

- KSA 6 Personnel - Human elements (enemy, friendly, irregular, and neutral forces) create realistic training conditions during military operations. The personnel editor will greatly enhance the ability for soldiers to create a more realistic training environment in the simulation. This makes for a more realistic simulation and provides the trainer the ability to change the level of difficulty to enhance unit training.

- KSA 7 Support - This will support Sustainment in the Warfighting Functions that are required for realistic training conditions to track and ensure units have the correct Unit Basic Load (UBL), classes of supplies, maintenance and medical support to conduct military operations.

- KSA 8 Interoperability - Interoperability is required to produce the necessary stimulus to support the KPPs. The unit uses Battle Command Systems to view and gather information necessary to craft their staff plans and unit orders. Populating a unit’s Battle Command System is required to allow a unit to train as they fight.

- KSA 9 Performance Collection and After Action Review - Commanders are required to assess performance and the after action review process facilitates performance evaluation and corrective action.

The gaming CPD has been staffed through CAC, TRADOC, DA and is currently being staffed with the Army Requirements Oversight Council (AROC). Once the CPD passes the AROC, it is the gaming requirements document for the Army.

[Chelco]

Hi,

Does any off-the shelf war game satisfy those parameters or has those attributes?

I don't know if this is correct, but TacOPs, Decisive Point, Point of Attack 2 and the early games by Pat Proctor (ATF) were made with the military customer/student in mind and then released to the civilian market (is that correct?).

The only games I know that were initially targeted to the civilian market, and then after that were modified to cater the military customer/student are Steel Beasts (now Steel Beasts Pro), Close Combat The Road to Baghdad (now Close Combat Marines) and Operation Flashpoint (as the VBS1/2 series).

And a final and most important question: how does CMSF fares against the parameters/attributes of your post?

Cheers,

[Battlefront.com]

The parameters for military training sims, for official and widespread use, are so esoteric that they actually are harmful to commercially targeted games. The only way to go is to make a game for the commercial market and then make modifications to military specs after a deal is signed. Any other way than that is just foolish.

CM:SF, as it is today, satisfies most of the points listed above. Where it has issues are as follows:

#2 - Leadership training. While CM games of all flavors score very high with junior officers and down, the upper end want something which basically amounts to what we call CoPlay. To simulate a Battalion they want no single officer to handle more than about a company's worth of stuff and to micromanage the combat elements as little as possible. They also want the officers to have to deal with logistics and other non-combat elements that gamers would reject (for the most part). More on this last point in a sec.

The primary problem with this is that the Army likes to train in a stratified way. The want all Captains working on one exercise, or all Majors, or all Lt. Colonels/Colonels. At least for the most part. What they don't really see the value in, at least as far as I can see, is combining training at several levels concurrently. At least not for the classroom stuff. There are sound educational reasons for doing it this way, but unfortunately it is difficult for us to cater to because it would involve vast amounts of AI programming so the lowest level stuff (tanks, squads, etc.) would be largely autonomous. No commercial release could support such programming, so it would have to be done after a military contract was signed.

#4 - Behaviors - largely covered in my previous comments.

#7 - Support - the military is trying to train officers to juggle their responsibilities and to take into account the ramifications one type of decision has on the other areas. Casualty evacuation, supply, special combat capabilities, etc. are all things that the real life commander has to deal with, but 98% of wargamers don't care about. Or at least don't care about once they find out it dilutes their combat experience.

#8 - Interoperability - we don't even have the specs for this sort of thing (I think they are available, though) and don't care about them. It's a military feature only, so if the military wants it in... fine, but they definitely have to pay for it because there's no utility for it in the civilian market.

#9 - Performance review - the simplistic AARs found in commercial wargames don't go nearly far enough from the military's standpoint. However, by and large this stuff is pretty easy for us to support because the data is in there. The only thing we'd have to do is find out how they want it presented (which is likely to take on different forms). The big hurdle for us would be to have the entire game recorded as a single, fully reviewable "movie". We can do it, but it would involve a dedicated "recording" PC and quite a bit of work on our end. One of those features that's pretty straight forward, but involves a lot of grunt work to make it work. Because a dedicated PC would be needed to "record" the whole battle, we don't see this being a commercially viable feature for us to support for some time to come. Highly desirable, absolutely, but impractical from a cost standpoint.

Steve

[salwon]

it would involve a dedicated "recording" PC and quite a bit of work on our end.

This is one of those features that would be so unbelievably awesome, but is completely impractical on a modern PC. Maybe when quad core becomes standard...

[dan/california]

It depends on the specs for the recording PC. If the one that I am going to replace sometime in the next year anyway is adequate to the task it wouldn't cast me a dime. If it needs the same specs as the machine running the game it is a much bigger problem.

[Dook]

Thanks for the reply Steve. Your response is exactly the kind of discussion I was hoping to stimulate.

[Battlefront.com]

The main problem is I/O (input output) and RAM more than the CPU. What has to happen is the recording PC needs to buffer a LOT of data for it to be able to keep up writing it to disk since disk writing is inherently slower. There is no way that a single home PC can keep up with both the needs of generating the data and storing it concurrently.

Remember, the way it works now the system only has to deal with what it's currently working on. Everything that has already happen is purged from memory, everything yet to come isn't touched yet. That's much easier to do. As soon as data has to be remembered, the RAM requirement goes sky high.

We tried using "seeds" back in CMx1 in order to make TCP/IP work more efficiently. This is the technique of saving key data at key points and simply feeding the seeds into the equations instead of literally storing each result as a stand alone result. What we found was that different processors, even within one family (some Pentium chips actually had bugs in them!), handled floating point differently. This caused data divergence when two systems were rounding off numbers differently. There are, apparently, complex work arounds to ensure that the results are consistent, but the last time we checked into them we still felt it wasn't practical. Obviously if we go to try capturing the entire play we're going to have to look into that again since that's the best way to go.

Steve

[Chelco]

Oh, and I forgot this piece of s**t, Full Spectrum Warrior which ended up as garbage for both civilians and the US Army.

(Steve, for goodness sake throw us a sentence or something about the Marines module release please).

[Field Marshal Blücher]

I'm not Steve, but if you pre-ordered you can start downloading . . .

[Chelco]

Roger!

Stuck at the office.

[LongLeftFlank]

Unless you've lost your BFC.com password in which case there's a LONNNNNNNNNNG wait while it gets to your email (been over an hour now and still nada). :-(

[MikeyD]

One thing trainers like is reproducibility. Give 'em exactly this tactical problem let the player get exactly those expected results when they try out X, Y, or Z tactic. Trainers tend to be tailored towards teaching the biggest idiots in the class, which puts a bit of stranglehold on the product. You might say BFC had done too good a job of recreating the fog of war to get a *consistent* reproducible training aid out of it.