# V4 Product Review — Tactical Action Universe (Unreal Engine 5) Status: independent product review — revised and expanded in a second meticulous pass on 2026-07-07 (§9, Additional considerations) and a third pass on 2026-07-07 covering monorepo integration, shared functionality, and TODOS/ roadmap alignment (§10) Reviewer: Claude (product-perspective deep review) Date: 2026-07-06 (first pass) · 2026-07-07 (second and third passes) Scope: `V4/V4_features.md` + all `V4/features/*` pages, `V4_TODOS.md`, `REMEDIATION_2026-06-12.md`, planning docs (market analysis, risk register, concurrency/monetization forecast, ADR 0011 cell kill criteria), and the modes/esports/liveops/spectator/store/companion/web subtrees. Method: full deep-read, then critique against the shipped competitive set (Rainbow Six Siege X, CoD Black Ops 6/7 + Warzone, Battlefield 6 + REDSEC, Ready or Not, Arc Raiders, Tarkov 1.0, Delta Force; Hitman World of Assassination + Freelancer; MGS Delta; Commandos: Origins, Desperados III, Shadow Gambit; Black Myth: Wukong, Elden Ring Nightreign; StarCraft II, AoE4, Stormgate, Tempest Rising; Contra: Operation Galuga, Cuphead) and industry SOTA. Market facts reflect knowledge through early 2026. --- ## 1. Executive summary V4 is one premium UE5.5 title spanning six ruleset cells — tactical FPS (Raven Shield planning, R6-style 5v5, CoD-style campaign/6v6/100-player BR), Splinter Cell/Hitman stealth, Commandos/Desperados real-time stealth tactics, a Wukong-lineage action RPG (30 bosses), a dual-mode RTS (StarCraft-like asymmetric + AoE-like historical), and a Contra-lineage run-and-gun — bound by one 34-operator roster, one "Shadow War" narrative, shared gadget/AI/ progression systems, and per-cell netcode (dedicated servers, lockstep, rollback). Launch is fixed at 2026-10-01, $69.99, cosmetics-only store, no loot boxes, 250k CCU planning, $424M base-case Year-1 net. **The corpus is the most self-aware of the V-series.** Its own market analysis states the base rate for this product shape is zero ("nobody has survived the attempt") and names the corpses (XDefiant, Stormgate, Concord, Mimimi). It ships a de-scope ladder with kill criteria, a falsifiable bundle-thesis metric (≥20% of open-beta players touch 2+ cells in week 1), a five-year sunset plan preserving offline play, and quantified, release- blocking per-cell "feel gates" — product discipline none of its competitors practice in public. **Three product-level problems dominate.** (1) **The launch window is the worst in gaming history**: Oct 1, 2026 is seven weeks before GTA VI, inside the CoD corridor, one year after Battlefield 6's resurgence — a premium 6-genre newcomer will be starved of attention exactly when it needs oxygen. (2) **The beachhead cell has the hardest competitive slot**: a $70 5v5 tactical shooter versus free-to-access Siege X, F2P Delta Force, CS2 and Valorant is a price-structure mismatch the "most game per dollar" pitch doesn't fix, because PvP populations don't pay $70 to try a ladder. (3) **The readiness optics are inverted**: 1,441 checked tasks and genuinely excellent simulation cores, but zero cooked assets, no reachable mode in a running client, health-check-only backend routers, and RTS combat resolution barely begun — the plan is real; the game is not yet. Verdict: **the best-engineered product plan in the series, aimed at the wrong date with the wrong SKU shape for its riskiest cell.** Re-sequence (campaign-rich premium launch off the fall corridor, F2P competitive slice later), elevate the cross-cell mashups from Year-1 curiosities to the launch identity, defer RTS ahead of Contra, and let the feel-gate discipline — the thing nobody else does — carry the marketing. --- ## 2. What V4 is (product identity) - Six engine cells (marketing says five — see §6.1), each a GameFeature plugin family over shared cores: one pawn, one ability enum, 60-gadget library, 9-takedown library, one AI perception/crowd/schedule brain, dual XP tracks, three currencies, codex, 31 rivalry pairs. - Per-cell feel contracts, quantified and release-blocking (Wukong parry = 6 frames ±1; Hitman 60fps with 120 NPCs; RTS SC2-opening parity ±2%; Contra 60fps couch co-op P95 ≤16.67ms). - Netcode per cell: dedicated client-server with 300ms-clamped lag comp; 25Hz lockstep with desync recovery for RTS; 8-frame rollback-style prediction for twitch PvP. - Nine day-one platforms (incl. Switch 2 and M-series tablets, cell-stripped at cook); 12 VO + 14 subtitle languages; VR/AR variants; adaptive-trigger, HOTAS, wheel, adaptive-controller, eye-tracking peripheral support. - Live service: 90-day seasons, 100-tier pass, cosmetic-only store, 30-minute data-table hotfix channel, public balance ledger and MMR histograms, 12-event $500k-major esports circuit, Twitch drops, creator codes. --- ## 3. What is genuinely strong 1. **Feel gates as release-blocking automation.** Quantified genre-feel contracts (parry frames, breach timing, build-order parity, frame-time percentiles) enforced by Gauntlet CI is beyond anything Ubisoft/Activision publicly practice. It is also the only credible answer to the "no compromise blends" promise — keep it sacred, and _publish_ the gates; it is a trust asset the FGC-adjacent tactical community will respect. 2. **Balance transparency stack**: public balance ledger, Designer's Tier List, published matchup charts, public ARPG frame data, `/ranked` MMR histograms, and a matchmaking diagnostic that explains your match quality. No live shooter does all of this; several (Siege, CoD) are routinely criticized for exactly its absence. 3. **Product discipline artifacts**: ADR 0011's de-scope ladder with kill criteria; a falsifiable bundle metric; a program risk register that names the fixed date itself as a risk; a Year-5 sunset plan preserving offline modes and export windows — the latter is precisely aligned with the post-Stop-Killing-Games regulatory climate and is a marketable pledge no AAA competitor has made. 4. **The cross-cell mashups are the genuinely new thing**: an RTS commander injecting economy into a battle royale ("Commander Economy BR"), a tactical operator and a Wukong monk in one asymmetric room ("Temple Breach"), Hitman assassinations unlocking CoD campaign branches via world-state leaks. Nobody can copy these without building six games. 5. **Consumer-friendly economics**: no loot boxes anywhere, reskins with identical stats, PPP pricing with a public `/store/pricing` page, content-not-power DLC, archived (non-FOMO-recycled) passes. 6. **The honesty loop again**: the remediation pass corrected fabricated community baselines, un-checked ~40 falsely-green backend tasks, fixed an esports calendar that began before launch, and left the final-polish gate deliberately red. The plan can be audited; that is rare. --- ## 4. Product gaps ### 4.1 The bundle thesis vs. the specialist problem Each cell fights a specialist: Siege X (free, 10 years of operators and map knowledge), Hitman WoA (a decade of sandbox refinement + Freelancer), Wukong (20M+ copies of art-budget spectacle), AoE4/SC2 (decades of ladder culture), Cuphead/Contra nostalgia. The bundle wins only if circulation is real — the docs' own ≥20% metric. Two product moves de-risk it: - **Market the beachhead, bundle the rest.** Sell V4 as _the tactical shooter with four complete campaigns and an RTS included_, not as "six games in one" (which reads as six compromises). CoD itself is the precedent: campaign + multiplayer + zombies is a bundle nobody calls a bundle — because one identity leads. - **Put a mashup in the first hour.** Cross-cell events are Year-1/Year-2 roadmap items; the bundle thesis needs its proof _at first contact_ — e.g., the campaign's second mission hands you the Raven Shield planner over a Commandos-style isometric view before a first-person breach. If circulation is the bet, the onboarding must _be_ circulation. ### 4.2 The launch window (highest-severity, cheapest fix) Oct 1, 2026: seven weeks before GTA VI, mid-CoD-corridor, against BF6's established Year-2 and Arc Raiders' momentum. The corpus prices this as R-06 but treats the date as immovable. Every comparable success in the premium single-player-plus space (Elden Ring, KCD2, Monster Hunter Wilds) won a Q1 window. **Move to Q1–Q2 2027.** The forecast's own $184M downside band is mostly launch-visibility risk; three extra quarters also directly attack the two most under-built areas (RTS combat sim, cooked content) and the 900k-line VO mountain. ### 4.3 SKU shape vs. the F2P competitive field Premium buy-to-play protects campaign value but strangles PvP population. Ranked ladders (five of them!) need liquidity a $70 gate denies — and 2026's tactical field is free (Siege X free-access, Delta Force, CS2, Valorant, Warzone, REDSEC). Options, in order of preference: 1. **Free competitive slice** (R6Modern 5v5 + TDM playlist + BR, ranked included) with the premium SKU holding all campaigns, co-op cells, RTS, and cosmetics-status — the Helldivers-adjacent "premium but populated" posture; or 2. Persistent **free weekends per cell** on a published rotation (the GameFeature plugin architecture makes per-cell trial SKUs cheap); or 3. At minimum, a $29.99 "Tactical" edition upgrade path. Without one of these, the five ranked ladders (§4.4) will be ghost towns by month 4 and the esports circuit unjustifiable. ### 4.4 Ladder and mode liquidity Five ranked ladders + casual playlists + BR + co-op queues across six cells, on a 250k-CCU plan, repeats V2's fragmentation problem in milder form. The RTS ladder is the immediate concern (an unfinished combat sim feeding a Glicko-2 ladder with $500k-major ambitions). Consolidate Year-1 ranked to **two ladders** (R6Modern 5v5; RTS Asymmetric 1v1 _if_ the cell survives its gate) and let Spies-vs-Mercs, Historical, and Boss Rush run as seasonal leaderboard events until population proves out. ### 4.5 Missing genre table-stakes (2026 edition) - **No extraction mode.** The hottest shooter format of 2025–26 (Tarkov 1.0, Arc Raiders, Delta Force Hazard Ops) is absent from a game that has all the ingredients (session servers, loot economy, stealth AI, Gulag-style stakes). A "Shadow War Extraction" playlist reusing Warzone's map and the stealth cell's AI would cover V4's largest modal blind spot. - **No horde/zombies-style co-op pillar.** CoD's third pillar exists for retention reasons V4's Spec Ops missions only partly serve; the AI Director + crowd system (400 NPCs) is a horde mode waiting for a playlist. - **Hitman cell lacks a Freelancer analogue.** WoA's Freelancer (roguelike campaign over the sandbox maps) is the genre's SOTA replayability layer; V4 has contracts + persistence + world-state — one roguelike meta-layer short of matching the best thing its inspiration ever shipped. - **RTS vs. 2026 expectations**: automated tournaments (SC2's daily cups via battle.net/ESL), replay-analysis coaching, and a functioning combat sim. The last one first. ### 4.6 Industry SOTA checklist | Capability (2026 bar) | Bar-setter | V4 | | ------------------------------------------ | ------------------------- | ----------------------------------------- | | Release-blocking quantified game-feel CI | (nobody) | ✅✅ beyond SOTA | | Public balance/matchmaking transparency | (nobody does all) | ✅✅ beyond SOTA | | Sunset/offline preservation pledge | regulatory pressure only | ✅✅ beyond SOTA — publicize | | Per-cell netcode honesty (3 models) | genre-standard per genre | ✅ spec'd; determinism tail = R-04 | | Extraction mode | Tarkov/Arc Raiders | ❌ **gap** (§4.5) | | Roguelike sandbox meta (stealth) | Hitman Freelancer | ❌ gap (§4.5) | | F2P competitive on-ramp | Siege X/CS2/Valorant | ❌ **gap** (§4.3) | | Kernel anti-cheat at launch | EAC/BattlEye table stakes | ⚠️ planned; server replay validation real | | Cross-genre shared world-state | (nobody) | ✅✅ novel — under-marketed (§4.1) | | UGC workshop | Steam Workshop norm | ⚠️ R-08 liability; launch curated-only | | 12-language full VO | AAA norm is 8–14 | ⚠️ scope risk R-03 (900k lines) | | Accessibility suite | CoD/Sony first-party bar | ✅ meets/exceeds on paper | | Live AI commentary | experimental everywhere | ✅ beyond SOTA, correctly fail-closed | | Console/PC cert across 9 platforms day-one | rare even for giants | ⚠️ R-02; consider staggering | --- ## 5. Ideas that would make the product better 1. **Shadow War seasonal campaign as the connective tissue**: one persistent faction meta-war (Modern Forces vs Resistance) whose weekly territory state is moved by _every_ cell's matches — your Commandos mission sabotages the supply line that weakens next week's BR loot table; the RTS weekly determines which map variant 5v5 plays. The world-state leak infrastructure exists; making it _legible on one map screen_ turns six games into one war, which is the product's whole thesis. 2. **Extraction playlist** (§4.5) — reuse, don't build: Warzone map sectors + stealth AI + contracts economy. 3. **Freelancer-style "Syndicate Campaign"** over the 10 Hitman sandboxes (§4.5). 4. **Nemesis-lite rivalries**: the 31 authored rivalry pairs are static lore; add per-player dynamic rivals in PvE (the enforcer who caught you returns promoted, remembers your disguise) — Shadow of Mordor proved the delight; the NPC schedule + world-state systems make it cheap here. 5. **Plan-sharing as UGC-lite**: Raven Shield plans and Desperados Showdown setups exported as shareable challenge codes ("beat my plan") — leaderboard per plan. Safer than the workshop (R-08) and unique to V4's verbs. 6. **Couch identity**: Contra couch co-op + split-screen + RTST hot-seat is quietly the best local-multiplayer package in AAA; brand it (a "Living Room" hub tile) rather than leaving it scattered. It also insulates the Contra cell from the de-scope ladder by giving it a marketing job. 7. **Cell-stripped platform honesty**: Switch 2 dropping Warzone/R6Modern must be on the store page in plain language, or it becomes the review headline. Turn it into a positive: "Tactical Campaign Edition" naming. 8. **Esports: grassroots first.** Flip the circuit — Year 1 = the six $25k Opens plus in-client tournaments; Majors unlock when a ladder sustains a population threshold (publish the threshold). A $500k major for an unproven ladder buys embarrassment, not legitimacy. 9. **AI commentary as accessibility first**: ship the live commentary as an opt-in _blind-accessibility narration layer_ (screen-reader-grade match state) before it's a broadcast feature — same tech, unimpeachable framing, and it de-risks R-14 by proving the pipeline on consenting users. 10. **Publish the feel gates** (§3.1) as a "Genre Integrity Report" per season — turns internal discipline into external trust, and gives press a story that isn't "six genres, really?" --- ## 6. Criticisms and tweaks 1. **Resolve five-vs-six.** Marketing prose says five cells; code, gates, and the remediation say six. Pick six (it's true) and rename the Stealth split explicitly, or the first technical interview goes badly. 2. **Fix the readiness optics internally**: 1,441/3/31 checkboxes with "no mode reachable in a running client" (the `RegisterMode()` boolean bug), health-only routers for 12 of 13 services, in-memory service state, and 0 cooked assets is a plan-complete/game-absent state. The remediation banner says this correctly — the _executive summary_ of every status doc should lead with it, or scheduling decisions will keep being made against the checkbox ratio. Also reconcile the three circulating stand-in asset counts (891 vs 770 vs 504). 3. **De-scope order: RTS before Contra.** ADR 0011 sacrifices Contra first, but Contra is cheap, feel-gate-green-able, and carries the couch-co-op marketing job (§5.6), while RTS has no combat resolution, the largest isolated netcode risk, 3 of 5 factions, 44 maps, two campaigns, and a ladder+esports commitment. If anything fails its gate in 2026, it will be RTS — plan the faction system's two-bloc fallback (already specced) as the _default_, and let RTS join in Season 2 as the "new front" of the Shadow War. A deferred cell arriving as a season is a beat; a deferred cell at launch is a broken promise. 4. **VO scope**: 12 full-VO languages × ~900k lines is Cyberpunk-scale recording for a first title. Cut to 5–6 full-VO at launch (EN, FR, DE, LatAm ES, JA, zh-Hans), subtitle the rest, and say so; R-03 is the quietest schedule-killer in the register. 5. **Roster nomenclature**: 34 "canonical operators" vs 40 R6Modern "operator slots" will confuse everyone (players, press, your own store). Unify the vocabulary: canonical roster = characters; R6Modern slots = loadout archetypes those characters occupy. 6. **Warzone/Gulag/killstreak-name adjacency**: the corpus knows mechanic names are placeholders (R-13), but the _feature docs_ still read as a CoD bill of materials. The rename pass should also be a differentiation pass — ship at least one signature twist per borrowed mode (Commander Economy BR is exactly the right kind) or reviews will score it as a compilation. 7. **Wukong cell's specialty modes** (UFC cage boxing, one-hit duels) are tonal strays inside a mythic-ARPG cell; they'd live more coherently as Battle-Hub cabinet-adjacent events than as ARPG modes. 8. **Anti-grind daily XP cap (5000/day)** punishes the marathon cohort the bundle needs most (they're the cross-cell circulators); replace with diminishing returns rather than a wall. 9. **Companion app** ships loadout planner + contracts — good; add the Raven Shield/Showdown plan editor (§5.5's authoring surface) — the tablet is the natural planning table, and it's the one companion feature no competitor could copy. --- ## 7. Risks 1. **The date** (R-06) — see §4.2. Every other risk compounds under it. 2. **Bundle dilution** (R-10) — the docs' own zero-base-rate finding; §4.1's mitigations (beachhead identity + first-hour mashup + falsifiable beta metric) are the counter-programme. 3. **RTS cell state** — combat sim absent, lockstep digest doesn't hash unit state (desync detection gap), no composed match actor; carrying esports promises. Defer or descope early (§6.3). 4. **PvP population under premium pricing** (§4.3) — the beachhead dies quietly if ladders are empty in month 3. 5. **Nine-platform day-one cert** (R-02) with cell-stripped SKUs — stagger: PC/PS5/XSX day one; Switch 2/tablets as a fast-follow with the honest edition naming (§5.7). 6. **Workshop UGC** (R-08) — arbitrary-code liability; launch curated-only, revisit post-launch with the plan-sharing system (§5.5) absorbing demand. 7. **Live AI commentary about real players** (R-14) — correctly fail-closed; keep it esports-and-consenting-users-only until the DPIA-backed track record exists (§5.9). 8. **Shared-roster hotfix contamination** (R-16) — a 30-minute PvP nerf silently re-tuning Hitman/RTST co-op; require the hotfix channel to run per-cell feel-gate smoke suites before propagation (the gates exist; wire them into the 30-minute path). --- ## 8. Prioritized recommendations **P0 — decide now** 1. Move launch out of the fall-2026 corridor (§4.2). 2. Choose the F2P competitive posture (§4.3). 3. Swap RTS ahead of Contra on the de-scope ladder; make two-bloc factions the default plan (§6.3). 4. Resolve five-vs-six naming and readiness-optics reporting (§6.1, §6.2). 5. Cut full-VO languages to 5–6 (§6.4). 6. Wire feel-gate smoke suites into the 30-minute hotfix path (§7.8). **P1 — the identity wave** 7. First-hour cross-cell mashup + Shadow War seasonal map (§4.1, §5.1). 8. Extraction playlist (§5.2); Syndicate roguelike campaign (§5.3). 9. Grassroots-first esports restructure with population-gated Majors (§5.8). 10. "Living Room" couch-play branding (§5.6); platform-edition honesty (§5.7). **P2 — compounding** 11. Nemesis-lite dynamic rivals (§5.4). 12. Plan-sharing challenge codes + companion plan editor (§5.5, §6.9). 13. AI commentary as accessibility narration first (§5.9). 14. Publish Genre Integrity Reports (§5.10). 15. Replace daily XP cap with diminishing returns (§6.8). --- ## 9. Additional considerations (second pass) ### 9.1 Weapon naming and the arms-licensing question The industry abandoned licensed real-weapon names over a decade ago (no royalties or brand association with arms manufacturers); V4's region variants already rename weapons for Australia, which implies real-adjacent naming elsewhere. Adopt fictional/parody weapon naming wholesale — it removes a licensing cost, a brand-safety exposure, and a per-region variant dimension in one decision, and the CoD-lineage audience stopped caring years ago. ### 9.2 Geopolitical content review "Modern Forces vs Resistance" plus real-world-adjacent venues is a region-ban vector (several competitor titles are banned across MENA/Asia markets over faction framing). The AoE-civ sensitivity gate — V4's own build-failing review machinery — should be extended to factions, maps, and campaign framing, with per-region marketing review. The PRC-compliant variant exists; the _prevention_ layer (don't author the incident) is cheaper than the variant layer. ### 9.3 Community servers and what the sunset pledge implies The Year-5 sunset plan promises offline preservation; for a dedicated-server PvP game that implicitly promises **server binaries or hosting tools at EOL** — write that commitment down now (it constrains anti-cheat and service architecture decisions today). Consider community servers for casual/custom modes _earlier_ than EOL: it extends longevity, absorbs the modding-adjacent demand the workshop risk (R-08) defers, and it is the single strongest good-faith signal to the post-Stop-Killing- Games audience the pledge already courts. ### 9.4 Bot and backfill policy PvE squad AI exists; PvP backfill policy doesn't. Commit publicly: bots never in ranked, always labeled in casual, and never counted in engagement metrics. Competitors' silent-bot scandals (and the trust cost when discovered) make this a cheap pre-commitment with real value — and it matters doubly for V4 because the Ghost-Dojo-style labeled-AI stance recommended for V2 should be portfolio policy. ### 9.5 Steam Deck / Proton anti-cheat posture Same decision as V2 (§10.5 there): EAC-on-Proton requires explicit enablement, and kernel-level enforcement differs on Linux. Decide and publish Deck ranked eligibility before "Linux day-one" marketing; an honest tiering (everything offline/co-op, ranked on anti-cheat-supported platforms) survives scrutiny. ### 9.6 VO scale meets union AI provisions The 900k-line, 12-language VO mountain (R-03) intersects the interactive- media agreements' AI clauses: consented, labeled synthetic VO for _secondary_ languages could cut the mountain in half — but only as a negotiated, disclosed policy with named voice talent paid for trained models. Decide this as policy (with talent partners) rather than discovering it as a schedule emergency in cert quarter. ### 9.7 Demo strategy The Raven Shield planning mission is the demo: it is the most differentiated slice of the beachhead cell, it demos in ten minutes, and it photographs unlike anything in the CoD corridor. Ship it at a Steam Next Fest two quarters before launch; ship the Contra cell as a free couch-play arcade demo alongside (it doubles as the "Living Room" marketing beat from §5.6). Wishlist velocity from a great demo is the only marketing that competes inside a GTA VI window if the date cannot move. ### 9.8 Replay-driven creator tooling Deterministic replays + the broadcast SDK make a **replay API** cheap: cinematic camera exports, kill-feed data for YouTubers, clip deep-links. The content-creator program pays for itself when creators can make better V4 content than the marketing team — this is the lowest-cost acquisition multiplier available to a premium title. ### 9.9 Cross-portfolio notes V4 shares roster/GAS/feel-gate architecture with V2 — one combat-platform team, not two implementations (same note in V2 §10.8). V6 Ori operators as co-op guests is specced from V6's side (`v4-operator` adapter) and costs V4 almost nothing. And V4's genuinely excellent discipline artifacts — the de-scope ladder, kill criteria, feel-gate CI — should be promoted to _portfolio standards_: V5 lacks a ladder entirely (its review's P0 #1), and V2's re-scope needs exactly this machinery. V4's best export may be its process. ## 10. Monorepo integration & shared functionality (third pass) Audited against DOMAINS/, the 185-phase TODOS/ roadmap, `V_SERIES.md`, `V_SERIES_PLATFORM_CONSOLIDATION.md`, and package/crate evidence. ### 10.1 V4 is the natural GPS hot-path seed The consolidation doc's language ruling — Rust for hot paths, TS for control plane — describes V4's service tier exactly: of the five parallel online-services stacks, V4's is the real Rust one (`apps/v4/online-services` with an actual `matchmaking.rs`, telemetry-ingest, shared crates). Reality check: the doc says "25 specced, 13 scaffolded"; the directory holds **3 real crates** — less sprawl than claimed, which is _good_ news for consolidation: V4 has the least to throw away and the most reusable core. Recommended posture: V4's matchmaking/telemetry crates become the GPS reference implementation (M1's `libs/contracts/gps/` schemas extracted from them), and V4's backend roadmap stops at "consume GPS" rather than building toward the 26-service manifest. This converts §6.2's health-router embarrassment into the portfolio's consolidation head start. ### 10.2 One combat spine with V2 (Phase 84 has no code; V2 does) Root Phase 84 ("Combat Systems SOTA," 0/840, building on `@maya/games` + `@aphrodite/game-fighting`) plans multi-genre combat V2 has already substantially built (`UV2CombatResolver`, ~24k lines, value-tested). V4's melee/CQC and the Wukong cell's parry/posture systems should consume a shared combat core derived from V2's resolver rather than triggering Phase 84's greenfield — the V2 review (§11.2 there) recommends declaring V2's resolver the Phase-84 deliverable; V4 is the second tenant that makes it a _platform_ rather than a game's internals. The same applies to the roster/GAS/feel-gate machinery (§9.9 second pass): one combat platform team, three products (V2, V4, and V5's melee). ### 10.3 Voice-as-gameplay-input (Phase 82) belongs to V4 Phase 82 (0/145 — voice commands as first-class gameplay input) has no better first customer in the portfolio than Raven Shield's squad-verb grammar: "Red team, stack on door, breach on my go" is literally the phase's use case, and V4's planning-verb vocabulary (MoveTo, StackOnDoor, BreachFlash, WaitFor, MoveOnGo) is already a typed intent grammar the recognizer can target. V6's Vac intent parser (draft-confirm, forbidden- lines) is the proven in-house pattern to borrow. This is a differentiating accessibility _and_ immersion feature that the roadmap already intends to fund — claim it. ### 10.4 The unbuilt shared game substrate is V4's schedule risk too The roadmap phases V4's promises implicitly rest on are 0%: Phase 81 (NPC cognitive core / on-device SLM runtime — the schedule/suspicion AI's growth path), Phase 83 (emergent agency), Phase 176 (world models), and Phase 137 (platform/live-service sovereignty: console SDKs, rollback netcode, anti-cheat, LiveOps — precisely V4's launch stack). Meanwhile the completed domain phases V4 _does_ cite are real: Calliope (Phase 66) for the commentary stack — V4 is the only sibling that names it; keep that discipline. The integration ask: V4 should be listed as co-owner of Phases 82/84/137 in the DOMAINS ownership table so the portfolio roadmap and V4's launch plan stop being separate universes — `V_SERIES.md` records the tracks as deliberately separate, which is exactly how two teams end up building the same anti-cheat. ### 10.5 Roadmap assets V4 under-uses - **Aja (Phase 16)** — mocap retarget/cleanup for the 30k-clip animation mountain; the video-to-motion service could turn reference footage into first-pass animation at scale. - **Hathor (Phase 7)** — quest/dialogue/procedural-narrative substrate for campaign side content; V4's narrative docs don't cite it. - **Themis (Phases 64/75)** — the UGC workshop's IP screening (originality shield) is built; wire it before the curated workshop opens (pairs with R-08 mitigation). - **Kuanyin (Phase 32)** — moderation/restorative-justice layer for the player-conduct system; currently V4 specs its own conduct ladder. - **Aphrodite `game-*` runtime** — as with V2: record explicit non-dependence (UE C++ is the right call) so consolidation pressure never points a shipping shooter at an unproven TS engine. ### 10.6 Cross-product notes with code evidence V6's `v4-operator` destination adapter exists on V6's side only — costless goodwill once the operator framework stabilizes. And note the portfolio pattern the dossiers confirmed: **no lateral V→V package edges exist anywhere except V7→V6/Maya**, so every "V4 reuses V5's X" or "V2 shares V4's Y" claim in any doc is aspiration until a workspace dependency exists — the V_SERIES maintenance rule should be enforced on V4's DEPENDENCIES doc accordingly. ## 11. Closing note V4 is the rare mega-scope plan that argues with itself honestly: its market analysis prices its own thesis at a zero base rate, then builds kill criteria, feel gates, and a falsifiable beta metric anyway. That discipline deserves a launch strategy worthy of it. The product that survives contact with 2026's market is not "six games for $70 in October" — it is a tactical universe with one unmistakable identity, campaigns as the premium spine, a free door into its ladders, its mashups in the first hour, and its RTS arriving as a season-two war front. The engineering culture here can build that product; the calendar entry currently pointed at October 1st cannot.