Peter Perla discussion at NWC on “next generation wargaming”

This recollection is just that, one of those “stream of consciousnesses” things. I tried to listen and think hard about what was being said, not take copious notes, so if you were there, it is not meant to be a chronological reconstruction of the discussion, simply a “story about the discussion topics” as I remember them (and embellished it, twisted, spindled and mutilated and supplemented with a few of my own asides to…). So if I grossly mischaracterized anything, feel free to correct me. My goal here is to present some of the topics discussed and the flavor of the discussion. There were 3 different groups Peter addressed, the Wargaming department at large, the Halsey Groups, and then a still smaller group. Since the topic thread wove through all 3 groups I’m sticking to the generalities of the discussions, not which group said what.

The following day I then had a 4 hour 1 on 1 with Peter that ended with a mutual “our brains are full”. I can’t describe what an intellectual “high” one walks away with after such a discussion. It will be literally weeks to fully digest all the ideas that flew back and forth that morning! We need to do that more than once every 10 years, Peter!

The discussion looked at the future in 3 general areas – technology, adjudication, and basic design. There were other things, but I will try to staty focused on the “next generation” stuff as the conversations ranged a fair bit off that topic. Much of the discussion also fell back to “function determines form” and how easy it is to forget that.

The technology piece touched on most of “tech d’jour” from touchscreen tables to MMOG’s and virtual worlds to intelligent agents. There was a lot of push back on the notion that it is “tech” that will determine the “next generation of wargaming” but there is a bit of chicken and egg to it. Without some investment in new tech, its hard to determine what doors it opens. So since the discussion started with Tech, I’ll start with some of the bits discussed.
There have been a number of attempts at the “multi-touch game table” from DIY like
To Geek Chic’ high end “furniture grade” offerings :

The “desired endstate” for most is the combination of a large in table horizontal surface as shown above, with the ability to manipulate “your stuff” either by direct screen touch, or for “hidden objects” via an iPad. So the “big screen” shows the info available to everyone and your iPad shows you your private “COP”. For MMOGs, the idea would be to allow for a combination of face to face social and traditional MMOG play by allowing “cells” at various peoples houses, were a portion of a clan can meet and enjoy the game together, while participating in a broader game experience.

Its that idea of cell linking that seems to capture the imagination of some. They see the ability to have individuals or small teams rather than be in a room with powerpoint, providing written orders, that a “distributed gaming system” along these lines would solve a lot of problems. Even to the point of letting the teams interact with an MMOG “world out there” with the sort of dynamic challenges they would face on real irregular warfare battlefields. The “coolness factor” is just oozing out this concept!!

I’ll discuss a few other things before the bucket of ice water comes out for that…

Virtualization came up a bit, but most of the comments were non-committal at best. There is considerable work going on looking at “2nd life” style virtual worlds. The big question is “in order to…?” What can you do in a virtual world you can’t do in a real one that makes sense to do? It did not come up in the discussion but NUWC is working on using virtual worlds so that you can easily “instrument the avatar” in ways you can’t instrument a human. There is also some potential to use virtual environments where you have multiple level security challenges and various national and international groups collaborating, but these are not as applicable to wargaming as other challenges right now.

An area where virtual collaboration came up was the Naval Postgraduate School’s MMOWGLI system. There a few “oh geez” and rolled eyes when it was brought up, many I think because its moniker has as much to with MMOGs and wargames as LCS has to do with Littorals and Combat… As a “virtual yellow sticky drill” it does have some usefulness. It’s not what its name purports it to be, alienating some because of that.

The other tech issue that was discussed was virtual agents, both from the “JFAF” (joint FULLY automated forces) vice “JSAF” (Join SEMI-automated forces). There was also discussion of the use of “agents” that could collect data reducing the stress on the white cell. The discussion weaved around but always came back to the question of did you think that operational level outcomes are “Newtonian” interactions of mechanistic agents, or did operational level outcomes represent emergent behavior of a complex system. The consensus appeared unanimous that it is the latter. This meant that “Newtonian” agents were insufficient and that the outcomes you get from even simple “complex agents” was unpredictable (Conway’s game of life was mentioned as an example of a deterministic, yet unpredictable system) and you did not know what might come of using them.

This is a good segue to the discussion of adjudication. There was interesting discussion of some innovative methods of adjudication. The use of “in-stride” adjudication within a line of effort where the white cell provided feedback on player “moves” after a rapid evaluation that seemed to be a kind of “where is the Markov chain of events leading, and what are the sort of events that “naturally follow” from the path the Cell is taking. This obviously works best for strategic and high operational level games where the adjudication is not so much “who shot who” but “how does the rest of the world react”. It also works best where you have several theaters or lines of operation that are sufficiently independent that lack of time synchronization between “Markov chains” doesn’t result in “causal feedback”.

Automation of adjudication was discussed briefly, but lead to a discussion of why you adjudicate? Was it simply to “turn the crank” on the state of the game world? If that is the case then, in the manner of a hobby board game, adjudication was a matter of “dicing for outcomes” say who died and onward you go. That sort of adjudication was easy to get the players to “buy into” if you did not offend their sensibilities (see below) – those that wanted to automate adjudication wanted to automate the ‘hard stuff”. What happened when you performed “influence operations” was the crux of it. Some thought that with a “detailed enough game world” you could either use agents based on the recent “human terrain” research, or you could enlist the players of an MMOG (get ready for that ice water from above) to “be the world”.

This brought up what is one of the “key questions” of “next generation” wargaming. First off is “what happens when I try to influence people to do what I want” even knowable, let alone something that can be gotten from a giant “state machine” populated with software agents? Since the consensus seemed to be that was doubtful, could you populate an MMOG with “the right kind of players” to get the results you want from real people? That was met with hoots and guffaws from those with actual MMOG experience. First off, what would incentivize someone to even play a game where they were supposed to be “regular people going about their daily lives” in the midst of an insurgency? Given the proclivities of most gamers, what stops them from ‘all going Taliban’ on you? Or killing the village elders and trying to take over? Even if you could incentivize then to “play realistically” (one idea was to simply pay them to do so, and they lost money for “going rogue”) how can you be sure a group of them would behave in the same way as a group of “real villagers, in an alien culture, with totally different values, goals and desires” And how would you represent the visceral reality of losing a child? Or your livelihood, or being accused of collaboration when innocent? Despite the “coolness factor” described above, the “MMOG as self-adjucating gamespace” seems to be infinitely harder to get to in practice than it would appear.

So possible feature 1 of next generation wargaming seems to involve “cool new tech” in some way we are not quite sure of. Feature 2 involves getting from force on force adjudication to “influence population to act in a desired way adjudication”. This is the heart of our irregular warfare strategy, and some would argue that this is not just a wargame adjudication issue, but a fundamental philosophical problem with our strategy. Populations learn, adapt and may not even know themselves how they might react when put in a position between emotional and rational responses. Are we predicating our strategy on knowing the unknowable and anticipating the unpredictable?

With that mind-bender we come to one last item in the realm of adjucation that was brought up that started an interesting discussion. If our adjudication is tied to generating decision-making in the players, shouldn’t we be getting decision making information from the White cell about why they decide the way they do? This was met with “the White cell is busy enough” and “they aren’t really players so their decisions are “not the same sort”. If you already have a paucuity of decision-data (its amazing how often the player “let the plan run” and don’t even make any decisions…) does it make sense to throw away what might be way more than 1/3 of the decision making? And from an analysis standpoint, are not the white cell decisions the “most impacful” on the “trajectory of the game”? If we don’t explore what is going on there analytically how do we know that game outcomes are not simply artifacts of adjudication, not player action? The discussion was far ranging but petered out with a “but its just too hard” and “maybe if we had the high tech stuff to make it easier to capture the white cell decision-making without them having to write it down or be interviewed about it?

Well, just when we thought we had recovered from the “influence on population” rather than “force on force” adjudication issue, the “I” word raised its ugly head. How does information affect adjudication? To paraphrase Napoleon “How many battalions has the COP?” Did we really have the handle on force on force adjudication we thought we had? Particularly when it comes to naval battles at long range where Wayne Hughes “Battle of first salvo” is entirely about who has what information when, and realizes that they have it. What does a Global Hawk, or a JSARS, or a shipboard helo tell you” How much is each “worth information wise”. “It depends” was the unison response. It’s the network, not the node, stupid!

So how do you adjudicate “information in warfare”. I’ve seen the SPAWAR guys come out of the woodwork at experiments and exercises with their “information relationship diagrams”. A 4 foot wide cylinder of paper that they rooooooooooooooooooooooooooooooooooooooooooooooooooooooolllllll down the length of the wardroom table (and then some) with what looks like an Intel CPU designs on them. “This is how all the information is supposed to move around the battlegroup”. Everyone picks their jaw off the floor and goes “Aaaaah yup…Really?!? I’m supposed to do what with…that…REALLY?!? That of course is just the “how it’s SUPPOSED to work, not the “and all the different ways it will actually work, in the heat of battle.
So we have a third “key attribute” of “next generation wargaming” – how does information and the networks moving it around effect the outcomes of tactical and operational level engagements?

About this time we were on the verge of a “morale check” to see if we would route out of the conference room, but Peter successfully rallied us and returned us to firm ground with the “OK, those are all issues, but lets get back to what it is we are trying to use wargaming to do? What are those issues keeping us from doing that people are paying us to find out through wargaming. (Secret dice roll – they pass! Rout averted!) With the assistance of some Phil Sabin quotes, Peter gets us back to thinking about what things people are asking us to do with wargaming today (warts and all) and is it a problem of needing still more complexity in our games, or is it that we have “lost our way” down the rabbit whole of “realism”, “complexity” and “fidelity” and what we need before we can really address any of those larger things is look at what wargames are asked to do, is wargaming the correct tool to use, and is a return to simplicity and elegance in game design really the “zero-eth order” problem we need to solve to make solving the “first order” problems above meaningful?

The discussion explored the design process and what appeared to be a desire to substitute “game design” with a “world, ready for gaming in” (at whatever level of war or across several levels of war). The latter is potentially of use if what you want is an “experiential” game where players are exposed to and forced to cope within different “warfare environments”. That is not what the vast majority of “war college” wargming is about. Its analytic, not experiential. The objective is not training or education, but acquiring insight. Perhaps about particular kinds of strategic, operational or tactical situations, perhaps about the future roles of technology, or conceptualizing new ways of the things we have interacting, or more mundane “business related” things. Finding insight into these things is not made easier by placing the players into ever more complex worlds, or subjecting their ideas to higher and higher fidelity models. In many cases, one of which I’ve seen recently, the players would never know if you replaced a cadre of pucksters and a multi-million-dollar simulation with an old hobby game CRT and a dice roll.

What is needed is not super-whamadyne sims, but good old-fashioned operations research to identify what aspects of the problem at hand are the most important, and design the game around them. I have fallen into this trap recently with a project that I didn’t feel I had the time to “really dig into” so I made a kinematic spreadsheet model and ran it a gazzilion times to get the feel for the “landscape”. To make a long story short, had I simply taken those spreadsheet results, I would have missed the “key relationships” going on under the hood. I followed my intuition when I got some seemingly contradictory results and found that decision-making elements related to the problem had far more impact than the kinematics. We often make the same mistake by assuming that by using a sufficiently “high-fidelity” sim, we get out of doing that “operations research” stuff that requires time, study, and worst of all, maths that actually will make us break out a linear algebra, advanced statistics or (perish the thought) a calculus-based physics book! Who wants to dig into what is really going on when we can let the simulation do that and just catalog the results?

What the “next generation wargame” is NOT is a replacement for a return to doing old school operations research to understand the context of a game sponsors question. Does the question they are asking really scratch the itch they have to they lack understanding of something, but may not be sure what it is? We owe then more than just turning the crank on a “machine” and telling them “the answer is 42”. This is where Phil Sabin and his “back to basics” approach to manual gaming is perhaps the most important place to begin in our quest for a “next generation” wargame. It looks at wargaming as a tool to be mastered, not technology to be integrated, or a machine to be developed, but a process that can be made more disciplined, rigorous and problem-centric (vice solution-centric).

We can make BOGSATs more focused, we can run dozens of small-group games rather than a single “cast of thousands” game, when we do run a “cast of thousands” game we can have built the “game” up from relatively simple vignettes that were prototyped with manual games. We can automate that manual game development process and distribute it electronically. On the technology front, rather than any more simulation software, or model federations, or “fully-automated puckster” tools, I want what I is a mixture of “Outlook rules and Google earth” for wargamers. A tool that lets you define a “sand box” and populate it with entities on the globe, but then have a “rules wizard” like you have in outlook where you can select the game rules that control the interaction of the entities. It would also facilitate the organization of the “operations research” data associated with the project so that the players can access it if they have a question about why certain rules were chosen or why units had certain characteristics. These would be sort of like “apps” that could be configured to help the players understand the connections between the data, behaviors and interactions abstracted in the game.

Back in the “old days” when the Naval War College ran games that constitutes the majority of the curriculum, as important (perhaps more so) than the results of games, was the discussion about the rules. It was in arguing the rules that players shared their perceptions, biases and assumptions about “how things should work”. Building a consensus that helped shape the doctrine we used in WWII. Some of it was wrong, represented groupthink, and missed key adversary capabilities and exaggerated some of our own. What it succeed in however was forge a cadre of officers who shared a common way of looking at the problem, understood that they might not have it all right, because they had been forced to change their minds along the way, and had not necessarily found “the answer” put understood how “the answer” could be unmade, remade, and made irrelevant.

If there is a “next generation of wargaming” to be had out there, it must start with a return to what makes gaming a useful tool, get back to the roots in operations research and employing abstraction to focus the player on what is important, rather than just aiming a fire hose of complexity, at him. Then it can proceed to look at the issues of “how do we incorporate new tech?” How do we get from “force on force” adjudication to “influence on population” adjudication, and in both cases how do we represent the role of information dynamics?

And that only is the tip of the iceberg of what was one of the most thought provoking days I’ve had in a long, long time. A big thanks to Dr. Perla for facilitating a wonderful exchange!

About Paul Vebber

"If you read about something, you have learned about it. If you can teach something, you have mastered it. Designing a useful game about something however, requires developing a deep understanding of how it relates to other things."
This entry was posted in Game Implementation. Bookmark the permalink.

3 Responses to Peter Perla discussion at NWC on “next generation wargaming”

  1. Bill Haggart says:

    Paul: Thanks for sharing. It is a peek at what is on the horizon, and not something I know much about. . As I have mentioned before, one thing facing educators is the fact that information in general is doubling every five years, and that process is speeding up. Our task has changed from teaching information to teaching students how to sift information and find the best for their needs. However, the general reaction to this explosion of information and attendent complexity has been to try and know it all. Textbooks get thicker [30% thicker than twenty years ago, both k-12 and college], longer school days, excelerated classes etc. etc. Curriculum has expanded so teachers are expected to teach 30% more information in a school year than just ten years ago. At some point, it is pouring a quart into a pint jar.

    It strikes me that is similar to the Military’s attempt to embrace all the expanding complexity leading to the ‘rabbit hole’ you mentioned, with analytic simulations being the tool to do that. Let the computer program master the complexity and spit out the answer: 42. The improvements in technology further stoke the hopes that the complexity monster can be tamed. And of course, the continuing worry with such analytic tools is “did we include the right information?’

    That’s the real issue. It’s not how much information can be captured, but rather weeding out the noise of complexity and finding what is the important information to answer the specific question. It’s your Einstein quote: “Everything that can be counted does not necessarily count….” And again, asking what simulations are best used for in that counting, because “not everything that counts can be counted”… the two points you and Peter made. I think that is the challenge facing our civilization, not just the military.

    Best Regards,
    Bill Haggart

  2. I think xkcd has it about right with Crowd Sourcing or MMOGs …

  3. Ian Mitchell says:

    A great article – Applying wargaming approaches to the business world is my current interest. The major differences between civilian society-economy and military operations seem to be:
    Absence of explicit command relationships
    Existence of multi-role participants – the person on the production line of a supplier in the week is also a customer in the shops on Saturday so part of demand.
    Representing group behaviour is common to both. especially the effects of fear and the desire to win. Morale Checks and Victory Points seem just as relevant tp business.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s