09 April 2012

Liveblogging the CASL Strategic Wargaming Roundtable

The crowd is assembling at the Center for Applied Strategic Learning as we're getting ready for tonight's festivities. Mike Markowitz and Joe Saur are tonight's speakers.

Stay tuned for the liveblog...
(Rex, wish you were here!)

Incidentally, you can also check into other presentations/events from CASL at this archive site.

Tim's throwing us off by talking from the back of the room instead of up front.
He's doing it b/c they're streaming the content live, which means he's got to in front of a mic.

Brief infomercial for MMOWGLI, and looking for some interested parties to help with the testing on a new project from that team.

First presentation is starting, with Mike Markowitz talking about Wargaming Irregular Conflict.
Discussing work with TRAC (TRADOC Analysis Center), building a tabletop wargame for dealing with insurgency at the tactical level. TRADOC asked CNA to design a game for battalion/brigade level ops, such that the units being maneuvered were companies/platoons.
Original iteration was a province in Iraq, later adjusted scenario to Helmand Province.

Operational context for a complex multi-player tactical wargame, played on a computer network.
OWA - Operational Wraparound.
Represent operations of 3 BN-level entities to address the question of "what's going on next door?"
Feeding actions of players into Cultural Geography and Nexus Network Learner models
Simulation of C2 in flow of brigade "assets" into a particular BN level
Integrate the model results
--- Cultural Geography model outputs affect infrastructure "income" and intelligence
--- Nexus Network Learner adjudicated KLEs and other interactions

What makes war "irregular"?
state vs stateinsurgency & resistance
fire & maneuverPresence
capture territoryControl population
decisive campaignsprotracted conflict
clear distinction of combatants"unlawful" combatants
combined arms & joint operationslight infantry and policing
attrition-based modelsattitude and behavior models

more after the jump!

4-player tabletop game
Blue: US Forces
Green: Afghanistan
Yellow: DoS, NGOs, PRTs, etc. All those human-focused resource-control organizations that can affect the necessary factors for success
Card-driven mechanics from the "effect deck" that gives outcomes/effects from the encounter based on a probability distribution of outcomes
Excel worksheet covered scheduled/scripted events

(Details on Helmand Province scenario, with a laydown of the geography)
Focused on 4 most heavily populated districts of Helmand
Sub-divided into neighborhoods

District display board looks an awful lot like a Cataan-style resource tracking tool that was adapted to Power Grid and then rendered by the guys from Star Fleet Battles.
Each district had a unique board that tracked what's relevant there (for instance, Kajaki district doesn't have a "legal" tracking tool)
Holding boxes for personalities (KL's) and their roles.

Distinctions between "representation" and "modeling"
Markowitx's contention that modeling needs to be reduced to mathematical definitions, where as representation can be more loosely defined
quote on the slide: "modeling is about physics, representation is about storytelling"
What needed to be represented:
- Time (weekly)
- Space (districts/neighborhoods)
- Forces (company-sized elements)
- Effects (what are the verbs)
---- attrition (red)
---- rest & refit** (blue/green)
---- Civilian casualties
---- Infrastructure (build/destroy)
---- Change "attitudes" of key leaders
** blue doesn't really get eliminated the same way red does. if blue gets their butt kicked, they get pulled out of the line, but aren't' truly removed from the game

Sample counters on screen
Blue maneuver assets had 2 sides, kinetic, non-kinetic (hey, we've seen this idea before!)
Blue combat multipliers: lift, UAV, EOD team, HUMINT, CAS, arty, etc

Three categories red combat units: foreign fighters (jihadists from out of town); full-time fighters; part-time fighters (very local, very fragile)
Red assets: IED, Leader, Civilians
IEDs act as representations of campaign of IEDs within an area, and are persistent within the game
Civilians as decoys/dummies
Leaders as red commanders with multiplier effects (lesson learned on the ground from the Taliban is that Taliban will fight hard rear-guard actions to allow key leaders to escape)
--- Forces on map as question marks, allowing for fog of war - turned up once revealed

Sample effects card on-screen... kind of tough to describe, so see the slides once they're ready
Advantage - can resolve a LOT of combat quickly by drawing cards rather than looking up die rolls

Planning boards to allow the plotting of next turn's actions to allow for basically simultaneous movement on the start of the turn (hmmm... another great idea!)

Showing the sample of how the probabilities of certain actions appeared within a 72-card deck.

Design of wargames is in 2 phases:
Proliferation of complexity (bowl of spaghetti)
Ruthless simplification (meat cleaver)
The more iterations through this, the better the game)

Lessons learned:
--- Direct manipulation of physical components
--- Concrete representations of abstractions
Quality of players determines success
--- Experience, maturity, patience, subject matter expertise

What was left out:
The Opium Economy
ROE, JAG, Lawfare
Competing Narratives
"The Strategic Bonehead"

Joe Saur, talking about some DIME/PMESII modeling.
(it's Joe's first trip up here from GT for a roundtable... he's a pretty prominent dude in the serious wargaming world)

How to put in a DIME action and how would that action feed into other models for 2d- and 3d-order effects downstream?
Put in a DIME COA and see what would come out.
Example: if you build a wall between 2 neighborhoods in Baghdad, you can stop the fighting between them, but you might wreck the economy between them, too.

How to visualize the key issues within the area, especially for the non-map-related issues, such as political machinations? Who are the key imams? What are the key family relationships? What are seasonal activities and what does the rhythm of life look like during that time?

Build a COA and include the time-steps through the COA and lay it out on a matrix to allow visualization of the relationships between D-I-M-E and look for where effects should be visible downstream.

BAE and LockMart both competed for opportunity to work on these types of games/models.
Both had a variety of systems dynamics models and/or other applications to track the activities.
(comments on the ways in which the 2 contractors approached the problem... not going to repeat them here, so as not to poke at them)

(a list of senior mentors - pretty impressive resumes, but not going to try to repeat it all here)
Mentors were shuffled through the different teams during the experiment

Requirement was that BAE/LockMart models produce more valid results than the O&A database team.
How to assess "valid" results?
Teams would assess whether or not the results could statistically fit within the bell-curve, and that was the best assessment they could provide.

Validity of models are much harder to approve/assess within social modeling - i.e., DIME/PMESII
Models can describe what factors may have been important, but can't tell you whether or not the COA is actually going to work. Can't describe concrete individual events on the battlefield.

Describing how to visualize shared data points/predictors between the two different models.
What variables were the key to analyze for how that COA ran through the system? Trace those variables through the model to see where they affected the model.

Sandbox/search engine to share data/info in a visual way
List a key leader and then build model around it: what affects that person, in what direction, and at what strength?

Allows intel, ops, pol-mil, etc to build their own models to approach understanding the on-the-ground situation, and then cross-pollinate the models to see what variables they were sharing, in what ways, to determine key variables.

Q&A starting now, with Rex probably asking 20-30 questions from afar
(I may not keep up with too much of this)

Question for Markowitz: "What are the defeat mechanisms? In other words, how do you win?"
(sarcastic comment from Dr Falken Peter Perla, sitting next to me: "don't play the game")
Markowitz: within the time spans in this game, it wasn't likely to happen, b/c the game was much broader than the local area

Question from me jumping in on how the HTT folks would've tied in with either of the projects folks were working on. The Map-HT outputs and HTT reports would've been very useful in either game. Another commenter jumping with a thought about non-geo-specific date being represented "off the map" so as not to tie to to geographical locations and give it a "territorial control" paradigm.

How to map the non-geolocated world? Markowitz: How do you "map" the electromagnetic spectrum on the battlefield when fighting electronic warfare?

Comment about a new dimension of warfare: "bureaucratic dominance" - the ability to push through any sort of activity that would've otherwise been stomped on by a higher level headquarters as being outside of the scope of your mission. If you want to do "something you're not supposed to" how do you get your warrant expanded to include it so you can do it?

Question to Markowitz about what he would've put in if given the option (from his list above).
Markowitz: Better modeling of the economy, especially the local economic area and subsistence levels.

Complex question from online about how Markowitz's project fit into the COINtra/COINdinista intellectual conflict.
Training: acquisition of skills against a particular standard.
Game was not intended as an educational tool.

Wrapping up now with some comments about other CASL initiatives.

By: Brant


Rex Brynen said...

Well, I'm there virtually!

Brant said...

You're probably getting a better sense of it listening to Joe than me trying to type along with him...

Rex Brynen said...

I only asked two! (And only really got an answer to one, but the second was a bit complicated for a virtual question.)

Anonymous said...

When's the next one?

Thanks for the live updates, Brant.


Jon Compton said...

Really sorry I missed this one.

Brant said...

Jon, you would have enjoyed it

Guardian said...

Really interesting stuff. It *almost* makes me miss my old days of being knee-deep in the military-industrial complex :).

Brant said...

Michael - nothing scheduled yet. Usually quarterly, which would put us into July. But Connections is scheduled for July, so I don't know if there's one then.