what is a matrix game?

Transcription

1 WG 30 Assessing the Analy-cal U-lity of Matrix Games Rex Brynen Department of Poli/cal Science McGill University what is a matrix game? matrix games are very low- overhead, free- form games that can be quickly designed and implemented originally developed by Chris Engle in the early 1990s matrix games are built around an evolving narrajve, rather than being closely Jed to maps, pre- idenjfied capabilijes, etc. three basic steps in the play of a matrix game: 1. A player/team puts forward: an acjon the consequence(s) of that acjon, if successful reasons why the acjon would be successful 2. Other players/teams weigh in with further reasons why the acjon would be successful or unsuccessful 3. The arguments are evaluated by an adjudicator, the outcome is determined (by adjudicajon or weighted probabilijes), and effects are applied play and the narrajve then conjnues 1

2 why study matrix games? games are easily documented and easy to repeat. several potenjal applicajons: development and rapid assessment of scenarios an alternajve analyjcal method educajon this presentajon discusses some of the early findings on an ongoing project by McGill University and Defence Research and Development Canada to examine the strengths, weaknesses, and techniques associated with matrix gaming. ISIS Crisis game playtests at ConnecJons, UK Defence Academy, McGill University, DRDC and elsewhere parjcipants include military personnel, operajons research analysts, intelligence analysts, subject maber experts, university students the ISIS Crisis the ISIS Crisis developed by Tom Mouat, with SME input by Rex Brynen six players/teams: Iraqi government, Sunni opposijon, ISIS, Kurds, Iran, US one addijonal SME to represent other actors, acjons, consequences map and counters are indicajve only resolujon by weighted probabilijes posijve arguments +1 negajve arguments - 1 die roll to determine outcome players assigned special rules/condijons, for example: Iraq: negajves due to corrupjon/disorganizajon, negajves in Sunni areas ISIS: extra turn with major success US: negajves on escalajon of ground force component Iran: acjons potenjally strengthen Sunni opponents Kurds: negajves outside Kurdish areas 2

3 facilitators should be facilitators, not umpires need to avoid imposing facilitator view of world on game play facilitator will need to adjudicate whether acjons are too ambijous, and break them into muljple acjons: We will drive across the desert, a>ack Najaf by surprise, and seize the city stochasjc outcome determinajon methods raise risk of excessive number of low probability/high impact events or unrealisjc game acjons on the other hand, the game should not be limited to most likely outcome if it is to effecjvely harness the imaginajons, invenjveness, and adaptability of players added white cell/sme as addijonal non- specific actor, able to nudge game back towards more realisjc course, highlight underrepresented issues, represent 3

4 if teams dominated by parjcipants with lible subject maber experjse, the game is at risk of less realisjc outcomes, gamer mode 4-7 teams opjmum for real- Jme game play may be less well suited to one- on- one engagements ideal team size 1-3 for rapid real- Jme play white cell/sme can play unrepresented actors real- Jme distributed, virtual play possible non- RT play possible (by , etc) provides teams with much greater Jme to consider moves loses some of the interpersonal, social, and psychological characterisjcs of the game player percepjons and hence game play may be heavily influenced by informajon provided and physical presentajon map and military unit counters: players tend to take a kinejc approach map showing refugees, DIME/PMESEI factors may result in a less kinejc game lack of explicit rules and capability constraints, coupled with free- form nature of game, may accentuate idiosyncrajc factors narrajve character of game may facilitate player immersion and engagement 4

5 player idiosyncrasies, social engineering, and game output In one game, lead player assignments were heavily socially engineered: Iraqi government and Sunni opposi/on players had preexis/ng friendly rivalry; ISIS and Iranian actors were calm and calcula/ng; US and Kurds were cau/ous. In the game, tensions (and good- natured trash- talking) between the Iraqi Sunni and Shiite bedeviled poli/cal nego/a/ons. intel analyst percep-on game strength, because interacjons reflected the non- material idenjty and ideological drivers of behaviour OR analyst percep-on game weakness, because acjons were not always well considered, dominant personalijes drove behaviour, too much noise game format makes it very easy to idenjfy: 1. courses of acjons considered 2. alternajve and follow- on COAs 3. factors shaping success/failure 4. key informajon and analyjcal needs ( we don t know this ) 5. crijcal juncture and branches 6. more- and less- likely acjons and outcomes through iterajve play 5

6 next steps explore slow- play games over extended period explore distributed online play using roll20, virtual RPG tabletop fuller and more systemajc instrumentajon of games modify game rules and procedures to encourage players to more fully document alternate and follow- on COAs examples of observed themes key need for ISIS to expand, generate sense of momentum military stalemate in Iraq leads to upjck in border and terrorism abacks Iraqi successes in Takrit lead to counteraback in Ramadi crijcal juncture: does Baghdad priorijze the immediate bablefield situajon (via Shiite milijas and sectarian mobilizajon) or sectarian reconciliajon (intended to eventually pull Sunnis away from ISIS)? Kurdish reluctance to advance beyond Kurdish areas, intent to leverage crisis to build own capabilijes many parallel US and Iranian concerns, but ability to act undercut by strategic rivalry Syrian civil war/sanctuary gives ISIS strategic depth foreign fighters offset coalijon abrijon 6

7 addijonal resources PAXsims hbp:// Tom Mouat, Matrix Games hbp:// Chris Engle, Hamster Press Games hbp://hamsterpress.net John Curry and Tim Price, Matrix Games for Modern Wargaming (2014) hbp:// Antony Zegers, Matrix Game Methodology: Support to V2010 Olympic Marine Security Planners (DRDC, 2011) Paul Vebber, AnalyJcal NarraJve Gaming: The Long View, MORS Wargaming COP, 12 November