Latent Spark Calibrated production and market understanding for the game industry.
Patient Cartographer — Capabilities
The technical foundation
for live strategic pathfinding.
Patient Cartographer maps the paths a game and studio can still take, tests them against uncertain creative, production, and market outcomes, and keeps the strategy current as new evidence arrives—from hands-on planning to governed autopilot.
- 01
Map the possibility space
Explore how different creative, production, and commercial choices could unfold before committing the studio to one roadmap.
Pathway simulation
- Explore thousands of plausible tactical sequences with Monte Carlo simulation
- Carry uncertainty forward across budget, team, audience, brand, capability, and custom resources
- Model each tactic with hit, base, and flop outcome distributions
- Set probabilistic goals instead of relying on one expected outcome
- Test whether multiple resource thresholds can be reached together by a target date
- Model discrete production decisions and continuous change between milestones
- Apply deadlines, milestone gates, release windows, and Steam availability constraints
- Stress-test plans against adverse shocks, delays, and resource shortfalls
- Compare competing plans under the same assumptions and uncertainty
- Work backward from a goal to find pathways capable of reaching it
- Visualize pathways as decision trees and strategic flow graphs
- Checkpoint, resume, shard, cancel, and retry long-running simulations
Pathway search & confidence
- Generate, diversify, and refine candidate strategies through multi-stage pathway search
- Search large tactic spaces with genetic algorithms
- Balance exploration and exploitation with Thompson selection
- Measure when simulation results are precise enough to support a decision
- Use Wald precision, R-hat, and effective sample size diagnostics
- Set release gates for convergence and simulation health
- Identify which assumptions most affect the result with Sobol sensitivity analysis
- Stream progress while durable planning runs are in flight
- 02
Compare viable ways forward
See the strongest current route, the credible alternatives, and the tradeoff between confidence now and options later.
Strategic variants
- Group successful pathways into coherent strategic variants with DBSCAN clustering
- Compare families of strategies rather than brittle one-off schedules
- Select a representative pathway for each variant with medoid-style scoring
- Rank viable strategies by success probability, robustness, and maneuverability
- Show how much of the total success probability belongs to each strategic variant
- Fall back to pathway-level recommendations when the evidence does not support stable clusters
- Adjust scoring weights to reflect the studio’s priorities
- Persist the recommended golden path and its supporting evidence for every run
Decision intelligence
- Show where the studio still has multiple viable choices
- Identify low-optionality steps where the plan is becoming difficult to reverse
- Highlight decisions that materially change the probability of success
- Map decision criticality across the planning horizon
- Compare optionality, certainty, and downside across strategic variants
- Identify bottleneck resources before they become delivery crises
- Explain near-miss pathways and what caused them to fail
- Suggest mitigation tactics for fragile or narrowing paths
- Show credible intervals around projected resource trajectories
- Return plain-language explanations for teams, leadership, and agents
- 03
Learn while making the game
Update the planning model as prototypes, launches, experiments, forecasts, and real operating outcomes reveal what is actually working.
Studio tactic model
- Start from a built-in library of common studio tactics
- Register custom tactics for the studio, title, genre, or operating model
- Represent tactic outcomes with ordered hit, base, and flop LogNormal mixtures
- Model the resources each tactic consumes, creates, or puts at risk
- Define preconditions, durations, dependencies, and production constraints
- Analyze dependency graphs, critical paths, and executable tactic waves
- Organize tactics with templates, categories, tags, and metadata
- Validate tactics against the resource dimensions used by each plan
- Export dependency graphs and tactic structures for review
Calibration & drift
- Record real outcomes individually or in batches
- Recalibrate tactic distributions with fast quantile methods or full Bayesian MCMC
- Schedule recurring recalibration as new operating evidence arrives
- Detect when tactic performance has drifted from historical expectations
- Skip, widen, or preserve uncertain tactic estimates according to policy
- Fit specialized tactics from causal evidence where identification supports it
- Promote or roll back tactic versions under governance
- Condition outcomes by segment, title, market, or operating context
- Validate calibration with temporal cross-validation and posterior predictive checks
- Assess whether a tactic model meets configured quality and validation thresholds
- Condition commercial tactics with Anchored Horizon forecasts
- Condition experimental tactics with Subtle Beacon posteriors
- 04
Explore, replan, and commit
Keep exploration separate from commitment, update plans when the evidence changes, and choose how much of the process runs automatically.
Planning runs & replanning
- Validate objectives, tactics, and constraints before expensive simulation begins
- Run durable planning workflows with checkpoint and resume
- Follow the full pipeline from simulation through convergence, clustering, scoring, and selection
- Replan from an updated resource state, tactic library, forecast, or evidence set
- Maintain versioned base plans with optimistic concurrency controls
- Record operator decisions, commitments, and milestone gates
- Preserve lineage for audit, reproducibility, and later comparison
- Publish lifecycle notifications as planning runs progress or change state
Sandbox & commitment
- Explore hypothetical amendments without changing the committed plan
- Resimulate staffing, scope, launch, funding, and sequencing alternatives in a sandbox
- Compare multiple sandboxes under the same uncertainty
- Adopt an explored pathway into a new base-plan version only when ready
- Keep exploratory strategies visibly separate from committed decisions
- Require notification, confirmation, or approval before high-impact commitments
- Allow agents to propose amendments under the same governance rules
- Export decision briefs and plan summaries for studio review
Autopilot to copilot
- Choose how much Patient Cartographer runs autonomously
- Turn a plain-language goal into a configured planning run
- Run a closed loop from configuration to simulation, recommendation, and brief
- Pull supporting evidence or propose a Subtle Beacon study during the planning loop
- Recalibrate and replan automatically when policy allows
- Use manual-studio, trusted-automation, or agent-heavy operating presets
- Set notify, confirm, and allow rules for different actions
- Give humans and agents access through the same governed capability contracts
- Run disciplined pathfinding without requiring a planning specialist on every decision
- 05
Coordinate without micromanaging
Give studios and publishers a shared view of material uncertainty, risk, and decision timing without exposing task-level operations.
Studio & portfolio planning
- Plan across multiple titles, initiatives, and strategic horizons
- Compose publisher scenarios from consented studio-level signals
- Link production references across studio feeds without merging tenant workspaces
- Model shared capital, capacity, and portfolio constraints
- Align milestones and decision gates across titles
- Compare portfolio strategies under shared uncertainty
- Refresh linked studio statistics according to explicit policies
- Support studio, publisher, incubator, and portfolio deployments
PC2PC collaboration
- Create and revoke explicit publisher-to-studio sharing relationships
- Share consent-scoped sufficient statistics instead of raw pathways
- Share gate timelines for milestone and commitment alignment
- Share golden-path probability mass without exposing studio internals
- Surface decision entropy where strategic uncertainty remains high
- Provide decide-by signals for material publisher decisions
- Share resimulation sensitivity when assumptions or support levels change
- Share aggregated tactic-category evidence without exposing individual outcomes
- Enforce tenant opt-out and policy gates on every publish and read action
- Apply role-based access control to granting, revoking, publishing, and consuming signals
- Broker privacy-safe cross-studio learning and federated tactic calibration
- Support portfolio-level views of uncertainty, risk, and creative-capital allocation
- 06
Work alone or close the loop
Use Patient Cartographer as a complete strategic pathfinding system, or connect it to Iridae and the studio tools already in place.
Iridae operating intelligence
- Run stand-alone without requiring any other Iridae system
- Use Latent Spark belief state to condition game, market, and tactic context
- Use Anchored Horizon forecasts to condition commercial and revenue tactics
- Pull Subtle Beacon posteriors into pathway assumptions
- Trigger Subtle Beacon studies from high-uncertainty or high-criticality decisions
- Feed new evidence back into recalibration and replanning
- Federate tactic learning across permitted peers
- Close the loop from belief to forecast to path to experiment to updated belief
APIs, agents & enterprise controls
- Use the planning workspace or invoke the same capabilities headlessly
- Integrate through the capability API, OpenAPI, MCP, and A2A surfaces
- Register agents with scoped, time-bounded delegation
- Receive webhooks and asynchronous task progress
- Use durable checkpoints, resume, cancel, and idempotent operations
- Apply tenant isolation and role-based access control
- Preserve lineage artifacts for audit and reproducibility
- Operate on hosted multi-tenant infrastructure with observability and runbooks
- Apply rate limits, health probes, usage reporting, and entitlement controls
- Support commerce and billing workflows where enabled
Pathway simulation
- Explore thousands of plausible tactical sequences with Monte Carlo simulation
- Carry uncertainty forward across budget, team, audience, brand, capability, and custom resources
- Model each tactic with hit, base, and flop outcome distributions
- Set probabilistic goals instead of relying on one expected outcome
- Test whether multiple resource thresholds can be reached together by a target date
- Model discrete production decisions and continuous change between milestones
- Apply deadlines, milestone gates, release windows, and Steam availability constraints
- Stress-test plans against adverse shocks, delays, and resource shortfalls
- Compare competing plans under the same assumptions and uncertainty
- Work backward from a goal to find pathways capable of reaching it
- Visualize pathways as decision trees and strategic flow graphs
- Checkpoint, resume, shard, cancel, and retry long-running simulations
Pathway search & confidence
- Generate, diversify, and refine candidate strategies through multi-stage pathway search
- Search large tactic spaces with genetic algorithms
- Balance exploration and exploitation with Thompson selection
- Measure when simulation results are precise enough to support a decision
- Use Wald precision, R-hat, and effective sample size diagnostics
- Set release gates for convergence and simulation health
- Identify which assumptions most affect the result with Sobol sensitivity analysis
- Stream progress while durable planning runs are in flight
Strategic variants
- Group successful pathways into coherent strategic variants with DBSCAN clustering
- Compare families of strategies rather than brittle one-off schedules
- Select a representative pathway for each variant with medoid-style scoring
- Rank viable strategies by success probability, robustness, and maneuverability
- Show how much of the total success probability belongs to each strategic variant
- Fall back to pathway-level recommendations when the evidence does not support stable clusters
- Adjust scoring weights to reflect the studio’s priorities
- Persist the recommended golden path and its supporting evidence for every run
Decision intelligence
- Show where the studio still has multiple viable choices
- Identify low-optionality steps where the plan is becoming difficult to reverse
- Highlight decisions that materially change the probability of success
- Map decision criticality across the planning horizon
- Compare optionality, certainty, and downside across strategic variants
- Identify bottleneck resources before they become delivery crises
- Explain near-miss pathways and what caused them to fail
- Suggest mitigation tactics for fragile or narrowing paths
- Show credible intervals around projected resource trajectories
- Return plain-language explanations for teams, leadership, and agents
Studio tactic model
- Start from a built-in library of common studio tactics
- Register custom tactics for the studio, title, genre, or operating model
- Represent tactic outcomes with ordered hit, base, and flop LogNormal mixtures
- Model the resources each tactic consumes, creates, or puts at risk
- Define preconditions, durations, dependencies, and production constraints
- Analyze dependency graphs, critical paths, and executable tactic waves
- Organize tactics with templates, categories, tags, and metadata
- Validate tactics against the resource dimensions used by each plan
- Export dependency graphs and tactic structures for review
Calibration & drift
- Record real outcomes individually or in batches
- Recalibrate tactic distributions with fast quantile methods or full Bayesian MCMC
- Schedule recurring recalibration as new operating evidence arrives
- Detect when tactic performance has drifted from historical expectations
- Skip, widen, or preserve uncertain tactic estimates according to policy
- Fit specialized tactics from causal evidence where identification supports it
- Promote or roll back tactic versions under governance
- Condition outcomes by segment, title, market, or operating context
- Validate calibration with temporal cross-validation and posterior predictive checks
- Assess whether a tactic model meets configured quality and validation thresholds
- Condition commercial tactics with Anchored Horizon forecasts
- Condition experimental tactics with Subtle Beacon posteriors
Planning runs & replanning
- Validate objectives, tactics, and constraints before expensive simulation begins
- Run durable planning workflows with checkpoint and resume
- Follow the full pipeline from simulation through convergence, clustering, scoring, and selection
- Replan from an updated resource state, tactic library, forecast, or evidence set
- Maintain versioned base plans with optimistic concurrency controls
- Record operator decisions, commitments, and milestone gates
- Preserve lineage for audit, reproducibility, and later comparison
- Publish lifecycle notifications as planning runs progress or change state
Sandbox & commitment
- Explore hypothetical amendments without changing the committed plan
- Resimulate staffing, scope, launch, funding, and sequencing alternatives in a sandbox
- Compare multiple sandboxes under the same uncertainty
- Adopt an explored pathway into a new base-plan version only when ready
- Keep exploratory strategies visibly separate from committed decisions
- Require notification, confirmation, or approval before high-impact commitments
- Allow agents to propose amendments under the same governance rules
- Export decision briefs and plan summaries for studio review
Autopilot to copilot
- Choose how much Patient Cartographer runs autonomously
- Turn a plain-language goal into a configured planning run
- Run a closed loop from configuration to simulation, recommendation, and brief
- Pull supporting evidence or propose a Subtle Beacon study during the planning loop
- Recalibrate and replan automatically when policy allows
- Use manual-studio, trusted-automation, or agent-heavy operating presets
- Set notify, confirm, and allow rules for different actions
- Give humans and agents access through the same governed capability contracts
- Run disciplined pathfinding without requiring a planning specialist on every decision
Studio & portfolio planning
- Plan across multiple titles, initiatives, and strategic horizons
- Compose publisher scenarios from consented studio-level signals
- Link production references across studio feeds without merging tenant workspaces
- Model shared capital, capacity, and portfolio constraints
- Align milestones and decision gates across titles
- Compare portfolio strategies under shared uncertainty
- Refresh linked studio statistics according to explicit policies
- Support studio, publisher, incubator, and portfolio deployments
PC2PC collaboration
- Create and revoke explicit publisher-to-studio sharing relationships
- Share consent-scoped sufficient statistics instead of raw pathways
- Share gate timelines for milestone and commitment alignment
- Share golden-path probability mass without exposing studio internals
- Surface decision entropy where strategic uncertainty remains high
- Provide decide-by signals for material publisher decisions
- Share resimulation sensitivity when assumptions or support levels change
- Share aggregated tactic-category evidence without exposing individual outcomes
- Enforce tenant opt-out and policy gates on every publish and read action
- Apply role-based access control to granting, revoking, publishing, and consuming signals
- Broker privacy-safe cross-studio learning and federated tactic calibration
- Support portfolio-level views of uncertainty, risk, and creative-capital allocation
Iridae operating intelligence
- Run stand-alone without requiring any other Iridae system
- Use Latent Spark belief state to condition game, market, and tactic context
- Use Anchored Horizon forecasts to condition commercial and revenue tactics
- Pull Subtle Beacon posteriors into pathway assumptions
- Trigger Subtle Beacon studies from high-uncertainty or high-criticality decisions
- Feed new evidence back into recalibration and replanning
- Federate tactic learning across permitted peers
- Close the loop from belief to forecast to path to experiment to updated belief
APIs, agents & enterprise controls
- Use the planning workspace or invoke the same capabilities headlessly
- Integrate through the capability API, OpenAPI, MCP, and A2A surfaces
- Register agents with scoped, time-bounded delegation
- Receive webhooks and asynchronous task progress
- Use durable checkpoints, resume, cancel, and idempotent operations
- Apply tenant isolation and role-based access control
- Preserve lineage artifacts for audit and reproducibility
- Operate on hosted multi-tenant infrastructure with observability and runbooks
- Apply rate limits, health probes, usage reporting, and entitlement controls
- Support commerce and billing workflows where enabled
See Patient Cartographer in action.
Explore how your studio can map viable paths, keep strategy current as evidence changes, and know when to preserve options or commit.