Below is a game‑theoretic plan for Days 2 and 3 at Gettysburg. I model each day as a sequential, incomplete‑information game between the Army of Northern Virginia (Lee) and the Army of the Potomac (Meade), where each side’s best move depends on what the other is likely to do, how quickly they can shift reserves (interior lines), and the payoff trade‑off between ground gained and casualties incurred. I then translate the equilibrium logic into concrete strategy and tactics for each day. [1]
Model primitives (what matters in the game)
- Payoffs: Attacker’s value of a breakthrough V must outweigh expected casualties C; attack is rational only if p·V − (1−p)·C > 0, where p is the probability of success given terrain, coordination, and surprise. Defenders maximize holding key terrain while minimizing risk of catastrophic rupture. [2]
- Information and beliefs: Each side is uncertain about the other’s true strength at each sector and how rapidly reserves can be shifted; artillery preparation and feints act as signals that can be either noisy (cheap talk) or credible (costly to fake). [3]
- Timing/coordination: Simultaneous pressure across multiple sectors can raise p by preventing the defender from massing at the true point of attack; mis‑timed, serial attacks let the defender redeploy and lower p. [4]
- Repeated dynamic: Day 3 updates beliefs from Day 2 outcomes. If Day 2 shows the defender can absorb blows without collapsing, rational attackers pivot away from frontal, high‑loss actions and try to change the game (maneuver, threaten lines of communication, or disengage to entice pursuit). [5]
Day 2 (July 2): Recommended strategy and tactics
Confederate options and game‑theoretic recommendation
- Strategic aim: Achieve a locally overwhelming, time‑coordinated shock at one flank while pinning elsewhere, raising p enough that the expected value of attack turns positive; avoid attritional, sequential assaults that let the Union shift reserves along interior lines. [2]
- Mixed strategy across sectors: Allocate a majority of striking power to a single flank that promises steep payoff if rolled up (high V), but run credible, costly demonstrations on the opposite flank and limited pressure at the center to saturate Union decision‑making. The costliness (artillery ammunition, visible troop movement) turns the demonstrations into believable signals and reduces Union’s ability to tell where the real blow will fall. [3]
- Coordination device: Time‑on‑target artillery and synchronized stepping‑off orders to create near‑simultaneity; if corps‑level delays or terrain friction raise the risk of desynchronization beyond a threshold (thus lowering p), the dominant play is to abort large‑scale attack rather than feed brigades piecemeal. This is a commitment rule that avoids getting stuck in a dominated, high‑casualty subgame. [4]
- Screening and deception: Use cavalry and skirmish lines to blind Union reconnaissance at the true axis while allowing the demonstrations to be seen; this implements a separating equilibrium where the real attack is hidden and the false attack is observable. [1]
- If no flank looks promising ex ante: Choose the “option value” move—maneuver to threaten Union lines of communication overnight (east/southeast approaches), trading immediate attack for a state change that may force the Union to attack on Day 3, flipping roles in a Stackelberg sense. [5]
Union best responses on Day 2
- Interior‑lines doctrine: Keep a sizable central reserve under strict mobility discipline; pre‑commit to rapid lateral shifts along the ridge line so that Confederate simultaneity must be near‑perfect to succeed. This credible commitment deters deep Confederate pushes by lowering their p. [2]
- Threshold defense rule: Establish a hard “no rupture” threshold—if any sector’s line integrity probability drops below θ, immediately reinforce with reserves and guns, even at the cost of thinning quiet sectors. This selects a robust equilibrium against Confederate mixed strategies. [3]
- Counter‑signals: Conduct limited spoiling attacks and conspicuous artillery movements at threatened sectors to inflate Confederate beliefs about Union strength; cheap talk can be effective if it exploits the attacker’s coordination risk. [4]
- Information systems: Push cavalry screens and signal stations to reduce uncertainty about Confederate massing; better information raises the marginal value of reserves and improves best‑response timing. [1]
Bottom line for Day 2
- Confederates should attack only if they can generate near‑simultaneous pressure and true surprise at a single flank while pinning elsewhere; otherwise, the best response is to preserve strength and maneuver to change tomorrow’s game rather than accept sequential, high‑loss assaults. [4]
- Union should bet on mobility and reserves, not on evenly strong lines everywhere; a centrally positioned reserve and rapid artillery shifts make most Confederate attack profiles unprofitable. [2]
Day 3 (July 3): Recommended strategy and tactics
Updated beliefs from Day 2
- If Day 2 failed to achieve rupture and revealed strong Union lateral mobility, posterior beliefs imply a lower p for any daylight frontal assault across open ground; with unchanged or higher casualty costs C, the attack condition p·V − (1−p)·C > 0 is unlikely to hold. Rational play shifts from “assault to win now” to “change the state to win later.” [5]
Confederate options and game‑theoretic recommendation
- Do not launch a massed, deterministic frontal assault against prepared center positions; with the defender’s artillery and interior lines, this is a dominated strategy given updated beliefs. Instead, choose among three higher‑EV plays: [3]
- Maneuver/turning movement: Before dawn, reposition to threaten the enemy’s supply/communication routes and compel Meade to attack or withdraw; this changes the leader–follower structure and can create a favorable counterattack opportunity. [1]
- Concentrated flank jab with real surprise: Only if reconnaissance shows a temporarily under‑defended flank and you can credibly synchronize fires and infantry; otherwise abstain. This is a contingent strategy with a hard stop‑loss if simultaneity fails. [4]
- Strategic disengagement to a prepared, defensible line: Invite pursuit on ground of your choosing; in repeated‑game terms, you convert a low‑probability knockout attempt into a higher long‑run payoff by preserving combat power and forcing the opponent into the attacker’s role. [2]
- Signaling plan: Use heavy but inaccurate artillery demonstrations to pin while the true objective is maneuver, not breakthrough; by making the “assault” a costly signal without committing infantry, you mislead about intentions while avoiding the bad subgame of an actual frontal charge. [5]
Union best responses on Day 3
- Commitment to defense-in-depth: Publicly signal readiness to absorb attacks at the center and punish crossing of open ground; this lowers Confederate expectations of success and can deter frontal options entirely. [3]
- Anti‑maneuver posture: Protect lines of communication with mobile reserves and cavalry screens; avoid being lured into a hasty pursuit that flips the game to Confederate advantage. The dominant response to a Confederate feint‑and‑shift is patience plus reconnaissance. [1]
- Limited, conditions‑based counterattack: Only after repulsing an enemy assault with clear disorder on their side, launch localized counterpunches to improve terrain; otherwise, maintain the high‑payoff equilibrium of strong defense rather than risking a role reversal. [4]
Why this differs from high‑loss frontal assaults
- After a stalemated Day 2, a large, visible, center‑mass assault is predictable, gives the defender maximum time to concentrate fires, and produces a low p with very high C; game theory flags this as a dominated move once beliefs update. Maneuver, credible feints without commitment, or operational disengagement strictly improve expected value while preserving forces for later stages of the repeated campaign. [2]
Practical “if‑then” rules you could have issued as orders
For the Confederates
- If artillery and infantry cannot strike within a 15–30 minute window across at least two sectors, cancel the assault; commit instead to a turning movement overnight. This avoids the bad subgame of serial, telegraphed attacks. [4]
- Do not commit more than X% of total infantry mass to any action lacking flank surprise or artillery overmatch; preserve strength to change the game tomorrow. [5]
For the Union
- Keep a central reserve ≥ Y brigades on wagons ready to shift within Z minutes to any threatened sector; prioritize artillery counter‑mass at the likely crossing point of open ground. [3]
- Launch counterattacks only on evidence of enemy command/communication breakdown; otherwise accept the draw and conserve strength. [1]
Net assessment
- Equilibrium on Day 2 favors the well‑positioned defender if reserves are mobile and coordination problems plague the attacker; rational attackers either create true simultaneity at a flank or avoid battle‑of‑attrition assaults. [2]
- On Day 3, after observing resilient defense and failed coordination, the attacker’s best response is to refuse a frontal fight, maneuver to change constraints, or disengage to tempt a pursuit on favorable terms; the defender’s best response is to hold, scout, and punish only when the attacker hands them an exposed subgame. [5]
Sources
1
Game Theory 2nd Edition by Michael Maschler (Author), Eilon Solan (Author), Shmuel Zamir (Author)
2
The Art of Strategy: A Game Theorist's Guide to Success in Business and Life Illustrated Edition by Avinash K. Dixit (Author), Barry J. Nalebuff (Author)
3
Game Theory 101: The Complete Textbook Paperback – September 3, 2011 by William Spaniel (Author)
4
Algorithmic Game Theory 1st Edition by Noam Nisan (Editor), Tim Roughgarden (Editor), Eva Tardos (Editor), Vijay V. Vazirani (Editor)
5
Theory of Games and Economic Behavior by John Von Neumann and Oskar Morgenstern | Dec 20, 2021
In addition:
Here are additional, more technical game‑theoretic details for Day 2 and Day 3 at Gettysburg, plus concrete decision rules you could have issued as orders.
- Formal decision model and thresholds
- Attack decision: Commit to a major assault on sector i only if p_i·V_i − (1−p_i)·C_i ≥ 0, where p_i is the success probability given terrain, surprise, and coordination; V_i is the payoff from rupture (e.g., rolling a flank or cutting the enemy line); and C_i is expected casualties plus loss of maneuver options if the attack stalls. If uncertainty about p_i is large, use the robust rule p_i ≥ C_i/(V_i + C_i) with a safety margin ε; if the lower‑bound of your credible p_i interval is below this threshold, abort. This operationalizes “don’t feed the defender sequentially.” [2][4]
- Interior vs. exterior lines: The defender’s interior lines increase dp_i/dt (the defender’s ability to raise their success probability over time) unless the attacker synchronizes pressure across multiple sectors; thus the attacker either creates simultaneity or cedes the advantage to time. This makes “piecemeal assaults” a dominated subgame for the attacker once the defender’s reserves start flowing. [2][4]
- Value of information (VOI): Allocate cavalry/skirmishers to reconnaissance if VOI > opportunity cost, where VOI ≈ E[max_j EV(plan_j | better intel)] − max_j EV(plan_j | prior). If VOI is high on a suspected weak flank, delay to gain intel can strictly dominate a rushed attack. [1][3]
- Day 2 (July 2) – deeper recommendations
Confederate play
- Mixed strategy with credible demonstrations: Concentrate striking power on one flank with the highest V_i and plausible surprise while running costly demonstrations on the opposite flank and light pressure in the center; the costliness (ammunition, visible maneuver) turns the feints into believable signals and forces the Union to hedge, raising your focal p_i at the true axis. If demonstration costs start to cannibalize the main effort’s logistics or timing, pare them back—feints are useful only insofar as they raise p at the main blow. [3][4]
- Synchronization/abort rule: Issue a standing order that if artillery preparation, flank jab, and fixing attacks cannot commence within a tight window (e.g., within one “step” of each other), the corps cancel and re‑set; this prevents sliding into the dominated subgame of serial, self‑disclosing assaults that let the defender reallocate reserves along interior lines. [4]
- Screening and deception: Mask the real axis with cavalry and skirmish screens while allowing the demonstrations to be highly visible; by making the true attack hard to observe and the false attack easy to observe, you create a separating equilibrium that pushes the defender to misallocate initial reserves. [1][3]
- Option value of maneuver: If pre‑assault reconnaissance lowers your p_i below the threshold, bank the option—march overnight to threaten the enemy’s communications and force them to either retreat or become the attacker on Day 3. Preserving combat power to change tomorrow’s game can strictly dominate a low‑EV assault today. [5]
Union play
- Interior‑lines doctrine: Keep a mobile central reserve and pre‑arranged lateral routes so you can mass artillery and infantry faster than the attacker can exploit any local success; your rule is “do not defend all sectors equally—defend the rupture threshold everywhere, then surge where p_attack appears to rise.” This maximizes the marginal defensive value of each reserve brigade. [2]
- Trigger policy: Define observable triggers (e.g., two enemy brigades seen forming beyond cover plus 10+ guns unlimbered) that automatically pull central reserves and counter‑battery assets toward that sector; formal triggers beat ad‑hoc judgment under fog of war and defeat the attacker’s timing game. [3][4]
- Counter‑signals: Use limited spoiling attacks and conspicuous artillery movements to inflate Confederate beliefs about your local strength; judicious “cheap talk” can be effective when it increases the attacker’s coordination risk and delays their main effort. [4]
- Day 3 (July 3) – deeper recommendations
Confederate play
- Post‑update choice set: After Day 2 shows the defender can shift reserves quickly and absorb blows, the posterior p_center for a daylight frontal march across open ground falls while C rises; a massed frontal attack is dominated unless fresh intel reveals a transient weakness. Prefer one of three higher‑EV moves: (a) pre‑dawn flank maneuver to threaten lines of communication, (b) a tightly synchronized flank jab only if genuine surprise and artillery overmatch are attainable, or (c) strategic disengagement to a prepared line to invite pursuit on better ground. [2][5]
- Signaling without commitment: Use artillery demonstrations to fix the defender while you maneuver; demonstrations should remain unaccompanied by massed infantry so you don’t get locked into the bad subgame of continuing a failing assault because of sunk costs. [3][5]
Union play
- Commit to defense‑in‑depth at the center and patience everywhere else: Publicly signal a lethal reception for any frontal advance, maintain mobile reserves to counter a turning movement, and avoid being baited into an early pursuit that hands the attacker the choice of battlefield. [3][1]
- Conditional counterpunch: Counterattack only after the enemy’s assault visibly disintegrates (loss of formation, ammunition exhaustion, broken command); otherwise, keep the high‑payoff equilibrium of strong defense rather than risking role reversal. [4]
- Command-and-control as mechanism design
- Principal–agent problem: The army commander’s optimal plan can be degraded by timing frictions and local deviations at corps/division level. Solve with “mechanism design” style orders: (i) simple, binary go/no‑go triggers tied to observable cues; (ii) pre‑authorized aborts if simultaneity fails; (iii) reserve release rules that don’t require upstream micromanagement in the heat of battle. These rules align local incentives with the global equilibrium you want. [1][4]
- Credible commitments: For the attacker, pre‑commit to canceling any major assault that loses surprise or timing; for the defender, pre‑commit to surge reserves even at the risk of thinning quiet sectors. Credible commitments reshape the opponent’s best response before contact. [2][4]
- Reconnaissance, deception, and the value of time
- VOI thresholds: If an additional hour of reconnaissance raises p_i by more than Δp* = C_i/(V_i + C_i) − p_i(current), waiting dominates attacking now; if not, strike while the defender’s beliefs are still dispersed. Time helps the side with interior lines and good signaling; it hurts the side that relies on surprise and simultaneity. [1][3]
- Deception portfolio: Mix cheap signals (false campfires, noisy troop movements) with at least one costly signal (a real artillery deployment you can later reposition) so the defender cannot cleanly separate truth from bluff; the aim is to keep them in a mixed response that dilutes their local strength. [3]
- Worked mini‑calculator (plug your own numbers)
- Suppose a flank rupture would yield V = 8 (operational payoff units) and the casualty/attrition cost if it fails is C = 6. The attack threshold is p ≥ 6/(8+6) = 0.429. If synchronized pressure and surprise lift p to 0.48, attack; if desynchronization drops p to 0.35, abort and maneuver. Add a safety margin ε ≈ 0.05 against optimism bias. This turns staff estimates into a crisp go/no‑go. [2][4]
Actionable checklists you could issue as orders
-
For the Confederates (Day 2–3)
- Only attack if two sectors can fix the enemy while the main blow steps off within one timing window; else cancel and pivot to a turning movement overnight. [4]
- Keep feints costly enough to be believed but not so costly they starve the main effort; if ammunition for the main axis drops below a preset level, halt demonstrations first. [3]
- Never commit more than a capped share of infantry mass to a daylight frontal advance without verified local artillery overmatch and temporary loss of enemy lateral mobility. [2]
-
For the Union (Day 2–3)
- Maintain a central reserve on wheels with pre‑surveyed routes; move on triggers, not on pleas from sectors that merely face noise. [2][4]
- Counter‑battery mass at any obvious assault corridor; rotate fresh brigades into threatened sectors to keep local p_attack depressed below threshold. [3]
- Pursue only on evidence of enemy command breakdown; otherwise, hold ground and let the attacker pay the coordination tax. [1][4]
Net effect
- Day 2 equilibrium favors the defender unless the attacker achieves genuine simultaneity and surprise at a flank; otherwise, the attacker should preserve force and change the game state for Day 3. [2][4]
- Day 3, after negative learning from Day 2, makes a massed frontal assault a dominated option; maneuver, contingent flank probes with hard stop‑loss rules, or disengagement to invite pursuit all yield higher expected value. [3][5]
Sources
1
Game Theory 2nd Edition by Michael Maschler (Author), Eilon Solan (Author), Shmuel Zamir (Author)
2
The Art of Strategy: A Game Theorist's Guide to Success in Business and Life Illustrated Edition by Avinash K. Dixit (Author), Barry J. Nalebuff (Author)
3
Game Theory 101: The Complete Textbook Paperback – September 3, 2011 by William Spaniel (Author)
4
Algorithmic Game Theory 1st Edition by Noam Nisan (Editor), Tim Roughgarden (Editor), Eva Tardos (Editor), Vijay V. Vazirani (Editor)
5
Theory of Games and Economic Behavior by John Von Neumann and Oskar Morgenstern | Dec 20, 2021
