# V2 Product Review — Fighting Game (Unreal Engine 5) Status: independent product review — revised and expanded in a second meticulous pass on 2026-07-07 (§10, Additional considerations) and a third pass on 2026-07-07 covering monorepo integration, shared functionality, and TODOS/ roadmap alignment (§11) Reviewer: Claude (product-perspective deep review) Date: 2026-07-06 (first pass) · 2026-07-07 (second and third passes) Scope: `V2/V2_features.md` + all `V2/features/*` companion pages, `V2_TODOS.md` completion state, `V2/balance/pricing|partnerships|launch`, `apps/v2/web/`, `VERIFICATION_REMEDIATION_2026-06-12.md` Method: full deep-read of the V2 docs corpus, then critique against the shipped competitive set (Street Fighter 6, Tekken 8, Mortal Kombat 1, WWE 2K25, EA Sports UFC 5, Guilty Gear Strive, 2XKO, Fatal Fury: City of the Wolves) and current industry state of the art. Market facts reflect knowledge through early 2026. --- ## 1. Executive summary V2 is the most ambitious fighting-game design document I have ever reviewed: a single UE5.5 title that promises **seven faithfully-distinct fighting rulesets** (MK, SF, Tekken, WWE, UFC, SC, Def Jam — plus a second-class KOF), a **≥60 fighter launch roster**, **52 modes**, a first-class creator suite, and — as a peer genre in the same box — a **full AAA arcade racing game** with an open world, heists, and a fighting/racing crossover ("Get-Out-and-Fight"). The spec corpus is unusually rigorous and unusually honest: validated C++ data catalogs, real deterministic-replay infrastructure, fail-closed rights gates, and a verification log that caught and fixed its own fabrications. **The product's four genuinely differentiated bets** — per-ruleset feel with no blends, the Defensive Options Library, deterministic server re-simulation anti-cheat, and a consumer-friendly monetization posture (no loot boxes, cosmetics-only pass, PPP pricing) — are each individually stronger than anything the current competitive set ships. **The product's central risk is arithmetic, not vision.** No shipped fighting game attempts a third of this scope; the "complete at launch, no coming-soon panels" promise currently sits on top of an 8-fighter presentation vertical slice, zero cooked builds, zero authored art/VO, and 18 partner-IP fighters with no executed licenses. The docs know this (the honesty discipline is exemplary), but the _product promise_ has not been re-scoped to match the reality, and that gap is where launches die. Verdict: **exceptional design foundation; needs a scope-honest launch strategy, a sharper competitive identity (which ruleset is THE esport?), a matchmaking liquidity plan, and about a dozen delight-level fixes** detailed below. --- ## 2. What V2 is (product identity) - One fighting game whose experience spans MK finishers/Krypt, SF motion-input depth and World Tour, Tekken juggles and side modes, WWE booking/Universe/ MyGM depth, UFC ground-game sim, SC weapon combat, and Def Jam's hip-hop brawling — implemented as 8 ruleset IDs over one shared combat resolver, with crossover rulesets ("Tekken with WWE stipulations") as a headline novelty. - Plus a peer AAA racing product (NFS/Blur/Split-Second lineage): ≥70 vehicles, ≥40 tracks, 25 race modes, open-world Metro City, police heat, heists, demolition derby, and fighter/driver crossover. - Platforms: PC, PS5(+Pro), Xbox Series, Switch 2, Steam Deck verified; cloud streaming; VR/AR surfaces for racing-adjacent content. - $69.99 premium (+$89.99 Deluxe), PPP-scaled across 18 markets, no loot boxes, 60-tier cosmetics-only battle pass, 2 paid DLC fighters/year minimum. - Rollback 1v1/2v2 at 60Hz; client-server for big-body modes; dedicated servers for 100-player surfaces. --- ## 3. What is genuinely strong (keep and market these) 1. **Per-ruleset feel with no blends** is a real answer to a real problem. Every platform fighter and crossover tag game (MvC, Smash clones) homogenizes its guests into one system. Nobody has shipped "Ryu, but under Tekken movement rules, as a deliberate mode." If the feel layer delivers even 70% of the promise, this is the most novel fighting-game product of the decade. 2. **The Defensive Options Library** (13 canonical defense mechanics as per-ruleset toggles: Parry, Just Defend, Push Block, Burst, Guard Impact, Drive Reversal…) is a genre-literate flex none of the incumbents can match — and it powers the crossover rulesets cheaply. 3. **Deterministic replay + 1% server re-simulation anti-cheat** is beyond current fighting-game SOTA. SF6/T8/MK1 all rely on client heuristics and reports; frame-drift detection riding a deterministic sim is the right architecture, and the team already fixed the cross-process hash bug that would have silently broken it (`V2StableNameHash`). 4. **Monetization posture**: no real-money loot boxes, competitively-relevant items never paywalled (enforced in code, not just policy), configurable spend limits, minor caps, PPP pricing with anti-inversion floors. This is cleaner than MK1 (Krypt/store backlash), Tekken 8 (Tekken Shop/battle-pass backlash was the genre's loudest 2024-25 monetization story), and WWE 2K MyFACTION packs. This is a _marketable_ trust asset — put it in the reveal. 5. **Accessibility depth** (auto-block/parry assist tiers with per-tier ranked flags, Quadstick/adaptive controllers, eye tracking, photosensitivity hard limiter, blind-play audio cues, `certGateBlocksRelease: true`) exceeds SF6 — the current accessibility leader — on paper. If certified, it's a press-release-grade differentiator. 6. **Training suite** (8-slot dummy recorder, Frame Trap/OS detectors, 160 trials, coach lobby, Replay Takeover) matches or exceeds the T8 "My Replay & Tips" + SF6 training gold standard. 7. **The honesty discipline** ("what ships, honestly" sections; fail-closed rights gates; the remediation log). As a working method this is rare and valuable; it makes the plan auditable. --- ## 4. Product gaps ### 4.1 Gaps vs. shipped competitors | Area | Competitor state | V2 state | Gap | | -------------------------- | ----------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | DLC fighter cadence | SF6/T8: ~4/yr; MK1: ~6/yr via Kombat Packs | "2 paid DLC fighters/year minimum" | **Below genre standard.** 2/yr reads as a dying game to the FGC. Commit to 4/yr; the 60-fighter launch roster does not remove the _cadence_ expectation — cadence is a liveness signal, not a content volume signal. | | Player-mimicking AI ghosts | Tekken 8 Super Ghost Battle (learns your style; fight your own/others' ghosts) is the single most-loved T8 innovation | Adaptive AI + AI Director exist, but no player-style ghost is specified | **Missing SOTA feature.** V2 already has deterministic replays + per-match analysis — the ingredients for best-in-class ghosts. See §5.1. | | F2P on-ramp | 2XKO is fully F2P; Fatal Fury/SF6 run free-weekend + cheap Year editions; MultiVersus (RIP) proved and burned the model | $69.99 premium + free weekend + 8-fighter demo | **Structural exposure.** A premium-only 8-ruleset game competes for the same newcomer 2XKO acquires for $0. See §5.2. | | Esports flagship clarity | CPT (SF6, $2M Capcom Cup), TWT (T8), EWC slots — each game IS its format | Crown circuit spans 8 rulesets + racing with no declared flagship format | **Identity gap.** TOs, broadcasters, and pros need to know what "V2 at EVO" means. See §4.3. | | Season-1 balance trust | T8 Season 2 (2025) balance fiasco → mass review-bombing; SF6 conservative patches praised | Monthly balance patch cadence promised; per-ruleset feel CSVs | Monthly is _fast_ for a fighting game — SF/T8 patch quarterly precisely because comp players hate churn. Recommend: quarterly balance for ranked-eligible rulesets, monthly only for casual/crossover queues. | | Kameo/tag runtime | MK1 shipped Kameo as its core; tag is 2XKO's core | Kameo = data only (launch gate); tag mechanics = catalog, no GAS abilities | Known gap, correctly gated — but note both are _headline_ systems in the docs. Don't market them until the runtime exists. | | Kernel anti-cheat | EAC/BattlEye table stakes on PC | Planned, plugin absent | Table-stakes gap on PC at launch; the server re-sim layer mitigates but does not replace it (re-sim catches sim divergence, not input-automation/ESP-adjacent tooling). | ### 4.2 Gaps vs. industry SOTA (beyond direct competitors) - **UGC combo-trial sharing.** V2 has 160 authored trials; SOTA-adjacent games (KI community labs, Strive mission mode) stop at authored too. Letting players _publish_ trials/situation labs from replays (with the dummy-recorder format as the container) would leapfrog the field at near-zero engine cost — the replay format and trial evaluator already exist. - **Structured crews/clubs.** Co-op crew 3v3/4v4 exists, but there is no persistent club/guild container with progression, club leaderboards, or club wars — standard in every live-service peer (even T8 has Tekken Fight Lounge socialization; Brawlhalla/MultiVersus/2XKO all ship clubs or equivalents). The faction system (unimplemented spec) partially overlaps; see §5.4. - **Companion app is read-only by design.** SOTA companions (EA FC, NBA 2K) let users manage loadouts/marketplaces off-device. Wardrobe (≥500 items), battle-pass claiming, replay-watching, and queue-notify are all safe write/interactive surfaces that would make the app worth installing. - **Watch integration.** No mention of a spectate-from-anywhere / watch-parties surface (SF6 Battle Hub viewing + Twitch Drops exist as separate pieces). In-client "featured live matches" rail (top-MMR matches, replays-as-esports reruns) is cheap given the replay/spectator infra and drives ladder aspiration. ### 4.3 The competitive-identity gap (most important product decision) Eight rulesets × separate MMRs × Kameo/no-Kameo splits × three control schemes × racing MMR × crossover queues is a **matchmaking-liquidity catastrophe** at any realistic CCU. The 100k CCU planning assumption makes it survivable at launch peak; at month 6, niche queues (UFC ranked? SC Kameo-off?) will be unmatchable outside NA/EU prime time, and empty queues read as "dead game." Recommendations: 1. **Declare one flagship competitive format** (recommend the SF-layer 1v1 ruleset — deepest motion-input identity, cleanest esports lineage) as the Crown circuit main event; every other ruleset gets a "specialty title" circuit tier (like EVO side events / WWE 2K's separate esports track). 2. **Ranked queue consolidation**: rotate niche rulesets through scheduled ranked windows (daily 4-hour windows per region, calendar-published — the Killer Instinct/Halo playlist lesson) instead of 24/7 ghost queues. 3. **One cross-ruleset prestige umbrella** ("Champion's Path"): a meta-rank aggregating per-ruleset ranks, so a player's identity is "V2 Grandmaster," not "SF-queue Grandmaster, Tekken-queue Silver." This converts fragmentation into a collection meta — the game's breadth becomes a progression asset. --- ## 5. Ideas that would make the product better ### 5.1 Ghost Dojo (high conviction) Player-style ghosts trained from your deterministic replays: fight your own ghost, your rival's, your coach's, and — with NIL-cleared pros — "pro ghosts" as premium _training_ content (cosmetics-only rule stays intact; ghost packs are training content, not power). Tekken 8 proved demand; V2's determinism + per-match analysis pipeline can do it _better_ (ghosts with explainability: "your ghost jumps too much on wakeup"). Async social loop: leave your ghost in a friend's lobby overnight, get the replay in the morning. This also solves niche-queue emptiness: ghost opponents make dead ranked hours playable without dishonest bots — label them as ghosts, always. ### 5.2 A free tier with a sharp edge "V2: Openweight" — permanently free: 2 rotating rulesets/week, 8-fighter weekly rotation, unranked + Battle Hub + spectator, full training suite with the _free_ fighters. Full cross-buy of progress into the premium SKU. This answers 2XKO structurally while making the $69.99 SKU the "everything, forever" statement. The GameFeature-plugin architecture already supports partial-content SKUs; the store/entitlement stack already models editions. ### 5.3 First-hour "Style Discovery" onboarding Eight rulesets at first boot is a paradox-of-choice failure waiting to happen. Ship a 15-minute guided intake: three short fights under three archetypal feels (motion-input, juggle, grapple), a two-question taste probe, then a recommended ruleset + 3 recommended fighters + auto-configured control scheme. SF6's World Tour start does the single-ruleset version; V2's whole premise _requires_ the multi-ruleset version. Also: a persistent "ruleset dial" tooltip glossary — the game invents a lot of vocabulary (Kameo vs Assist vs Tag; Blazin' vs Rage Art vs X-Ray) and newcomers should never see two names for one concept without an inline explanation. ### 5.4 Ship the Faction War; merge it with crews MKX's faction war is one of the stickiest light-social loops the genre ever shipped, the spec + balance data already exist, and it's the natural container for the missing clubs feature (§4.2): factions for identity mass-scale, crews (≤50) nested inside factions for social intimacy, weekly faction war scored by crew contributions, World Boss damage counting toward faction totals. One system, three currently-disconnected specs (faction spec, crew co-op, World Boss) — and it resolves the `@v2/aje-faction-governance` naming collision by forcing the rename (see §6, criticisms). ### 5.5 Crossover rulesets as a live-ops jewel, not a settings checkbox "MK with SC ring-outs" buried in custom-match settings will be discovered by 0.1% of players. Make it a weekly featured **Chaos Rules** queue (one curated crossover, leaderboard, cosmetic reward) — the Fortnite-LTM/MK-Towers cadence. This is V2's cheapest recurring novelty engine and its most clip-able content; it should own a home-screen tile. ### 5.6 One-thing-to-fix coaching card The per-match analysis dashboard is comprehensive; comprehensiveness is not delight. After each loss, surface exactly **one** actionable insight ("You got hit by the same frame trap 6 times — practice it now?") with a one-tap jump into a pre-loaded training scenario (the trainers and dummy-recorder can stage it). The dashboard stays for those who want it; the card is what changes player behavior — and it's a genuine differentiator over SF6/T8's passive stats. ### 5.7 Racing×Fighting as event content first "Get-Out-and-Fight" and vehicle-melee are marvelous trailer moments and risky ranked formats. Position the crossover as _scheduled spectacle_ (weekend events, Demolition-Derby-into-Rumble playlists) rather than a persistent queue; measure, then promote what sticks. Blur/Split-Second nostalgia is real (and the arcade-racing market is currently starved — NFS is dormant), but the combined-genre queue population math is unforgiving. ### 5.8 Smaller delights - **24h cosmetic "regret refund"** (unused items) — trust-builder no competitor offers; the idempotent ledger makes it cheap. - **Daily community moments feed**: auto-highlighted top clutch clips (the highlight-reel scorer already ranks moments) with one-tap "download the replay / practice vs this setup." - **Spectator anti-spoiler "watch from round 1"** for featured matches (delay infra exists). - **Karaoke mode duets cross-lobby** (it's a social mode; make it multiplayer social, and clip-exportable with the AI captioner). - **Arcade-cabinet high-score seasons** (the 8 mini-games) folded into the battle pass as zero-pressure XP — WWE-2K-boardwalk energy, retention for non-competitive players. --- ## 6. Criticisms and tweaks (coherence, naming, promise hygiene) 1. **Re-scope the launch promise now.** "Complete at launch, no coming-soon panels" + 60 fighters + 52 modes + a peer racing game is not a plan, it's a press release from a parallel universe. The honest inventory (8-fighter presentation slice, 3 uplugins, no cooked builds) supports a credible phased strategy: launch the fighting core with 3 rulesets _deep_ (SF/MK/ Tekken feel layers + 24-30 fighters fully presented), ship WWE/UFC/SC/DJ rulesets as free "Ruleset Seasons" with their signature modes, and launch racing as its own moment mid-Year-1. The promise then becomes "every ruleset season is complete on arrival" — keepable, and it converts scope into a 2-year drumbeat instead of a launch-day miracle requirement. 2. **Fix the currency naming.** "Krystals" (live-service), "Crowns" (store), "Crown Points" (esports) — three names, two of which collide with each other across paid/earned/prestige categories. Recommend: premium = Krystals; earned = Koin (already in the price-table engine); esports = Crown Points; nothing else may use "Crown." 3. **"Battle Royal" vs "Battle Royale"** as two different modes distinguished by one letter is user-hostile. Rename the WWE elimination mode to its genre name — "Rumble" — which the docs already use for Royal Rumble. 4. **Resolve KOF's status.** An 8th ruleset ID with no mode tag and "second-class" framing will read as shade to a passionate community — and SNK partner-IP fighters (Terry, Mai, Kyo, Iori) are in the roster plan while their _ruleset_ isn't first-class. Either promote KOF to the Ruleset Season roadmap or fold its mechanics (MAX mode etc.) into the Defensive/ super libraries and stop naming it as a ruleset. 5. **Kameo prose/data mismatch** ("3 assists" vs 12 actions) and the **replay size spec/code discrepancy** (1MB/5min vs 1MiB/min) are documented; fix the prose. Small, but this corpus's credibility is its honesty. 6. **Cut or quarantine the crypto-adjacent fan-token governance** (`@v2/aje-faction-governance`). Linking external token balances to faction votes is a reputational grenade in the current games press climate, it collides by name with the (better) faction-war spec, and its upside is nil for the core audience. If retained for some strategic reason, it must not share a name with the faction system players touch. 7. **Vault inheritance** (cosmetics beneficiary on documented passing) is humane and quietly excellent — but market it never; it's a settings-page discovery, not a bullet point. As designed (estate documentation flow) it's fine; just keep it out of trailers. 8. **Brand partners vs. Extreme gore tier.** Nike/Adidas/Red Bull placements in a game with an MK-grade Extreme gore tier need per-brand context clauses (no logo adjacency to Fatalities/X-Ray). The partnership ethics gate exists; add gore-context as an explicit clause category, or the first Fatality-next-to-a-swoosh clip will do it for you. 9. **AI commentary should default off for ranked/spectated matches.** The FGC's commentator culture is beloved and employment-sensitive; an AI caster _replacing_ humans is a community-relations own-goal, while an AI caster for _your own offline sets and training review_ is delightful. The 800ms KO-latency engineering is impressive — spend it where it wins hearts. 10. **Monthly balance patches** — see §4.1; competitive fighting games have learned quarterly-with-hotfixes is the trust-preserving cadence. --- ## 7. Industry SOTA checklist | Capability (2026 SOTA bar) | Bar-setter | V2 | | -------------------------------------------------------- | ------------------------------ | ---------------------------------------------------- | | Rollback + crossplay + cross-progression | SF6/T8/Strive | ✅ spec'd (1v1/2v2 rollback, unified MMR) | | Modern/simplified controls, ranked-legal with damage tax | SF6 Modern | ✅ Modern 0.80 scalar, Dynamic ranked-locked | | Open-world onboarding mode | SF6 World Tour | ✅ World Tour, 3 cities (content risk high) | | Social hub | SF6 Battle Hub | ✅ 100-player Battle Hub | | Player-style AI ghosts | Tekken 8 | ❌ **gap** (§5.1) | | Replay takeover + inline tips | T8 My Replay & Tips | ✅ Replay Takeover + trainers | | In-client tournaments | SF6 | ✅ Crown Tournaments in Battle Hub | | Broadcast SDK / observer tooling | League-grade tooling | ✅✅ beyond genre SOTA (OBS/NDI/Spout, anti-spoiler) | | Deterministic server-side cheat re-sim | none in genre | ✅✅ beyond SOTA (if kernel AC also ships) | | AI commentary/highlights | experimental industry-wide | ✅✅ beyond SOTA (deployment caveats §6.9) | | Accessibility certification | SF6 sound accessibility | ✅✅ beyond SOTA on paper (`certGateBlocksRelease`) | | Cosmetics-only pass, no loot boxes | 2XKO (F2P), post-backlash norm | ✅ enforced in code | | F2P or free-tier on-ramp | 2XKO | ❌ **gap** (§5.2) | | 4+ DLC fighters/year | SF6/T8/MK1 | ❌ **gap** (2/yr minimum) | | UGC trial/scenario sharing | (nobody ships this well) | ❌ opportunity (§4.2) | | Clubs/crews with progression | genre-wide | ⚠️ partial (co-op crews only; §5.4) | | Kernel anti-cheat on PC | EAC/BattlEye table stakes | ❌ planned, absent | --- ## 8. Risks (product-level) 1. **Scope vs. content reality** — the defining risk; see §6.1. 92% of TODOS checked ≠ shippable game: zero cooked builds, no art/VO/stages, presentation catalog is an 8-fighter slice. The plan is real; the _game_ is not yet. 2. **Partner-IP fantasy** — 18 guest fighters from Capcom + SNK + WB/NRS + Bandai Namco simultaneously has no commercial precedent outside Nintendo-platform leverage (Smash). The fail-closed gate is correct; the product must be _designed to be complete at 42 originals_, with partners as upside, and marketing must never presell uncleared guests. (The remediation log's catch — hardcoded `bRightsClearedForLaunch = true` — is exactly the failure mode to keep dead.) 3. **Queue fragmentation** — §4.3. This kills more live games than bad balance. 4. **Two-AAA-games-in-one-box install size** — 7 fighting rulesets + racing open world will collide with the console size budgets the docs themselves track. The GameFeature plugin architecture is the answer; surface it as a user-facing modular install manager (CoD Warzone pack model) from day one. 5. **Switch 2 parity for 100-player surfaces** (Battle Hub, BR, Convoy) — degradation strategy exists ("safe degradation"), but the 60Hz rollback promise plus dense hubs on handheld hardware needs early validation; a late Switch-2 cut of headline modes would be a PR wound. 6. **Sensitivity-review pool uncontracted** while shipping locale-aware gore, religious/cultural fighter themes, and 22 locales — the fail-closed ledger currently blocks nothing because nothing is contracted; that line item is cheap insurance against expensive incidents. 7. **Karaoke/DJ music licensing** (≥50 licensed tracks, Def Jam soundtrack feel) — licensing spend and regional music-rights fragmentation are real; the Def Jam _feel_ without Def Jam _clearances_ needs a soundalike strategy decided early, not at cert. --- ## 9. Prioritized recommendations **P0 — decide before any public commitment** 1. Re-scope the launch promise into Ruleset Seasons (§6.1). 2. Declare the flagship competitive format + queue consolidation plan (§4.3). 3. Raise DLC fighter cadence commitment to 4/year (§4.1). 4. Kill or rename the fan-token faction governance (§6.6). 5. Kernel anti-cheat contract signed (EAC or BattlEye) — table stakes. **P1 — build next (highest delight-per-effort)** 6. Ghost Dojo on the deterministic-replay substrate (§5.1). 7. Style Discovery onboarding + unified vocabulary layer (§5.3). 8. Faction War + crews merged social spine (§5.4). 9. Chaos Rules weekly crossover queue (§5.5). 10. One-thing-to-fix coaching card (§5.6). 11. Free tier decision (§5.2) — at minimum, design entitlements so it can be turned on without repackaging. **P2 — polish that compounds** 12. Currency/mode naming cleanup (§6.2, §6.3), KOF resolution (§6.4). 13. Companion app: add safe write surfaces (§4.2). 14. UGC trial sharing (§4.2). 15. Cosmetic regret-refund, community moments feed, karaoke duets (§5.8). 16. Modular install manager UX (§8.4). 17. Brand/gore adjacency clauses (§6.8); AI-commentary deployment policy (§6.9). --- ## 10. Additional considerations (second pass) ### 10.1 The original-roster charisma problem The first pass treated the 60-fighter roster as a scope/licensing issue; the deeper issue is creative. New-IP fighters almost never become icons — the genre's beloved casts took decades and multiple sequels to earn recognition, and even strong recent originals (Strive newcomers, MK1's new faces) struggle against legacy pull. 42 charismatic originals is the hardest creative deliverable in the entire plan, harder than the netcode. The archetype validator guarantees mechanical coverage, not charisma. Recommendation: 24–30 launch originals with _full_ presentation depth (the 8-fighter vertical-slice standard applied to all), a character-reveal marketing cadence (one fighter, one trailer, one week — the genre's proven hype engine), and a "signature three" mandate: every fighter needs one silhouette, one catchphrase, and one clip-able move that identify them in half a second of footage. ### 10.2 Input latency and connection transparency Competitive trust is won on numbers V2 hasn't specified: end-to-end input latency budgets (poll → simulation → photon) per platform, and in-match connection UI (SF6-style rollback/delay indicators, opponent connection grade). The pre-match handshake grades connections; expose the same telemetry mid-match and publish per-platform latency figures at launch — the FGC measures these with high-speed cameras within 48 hours of any release, so the numbers will exist either way; better they be yours. ### 10.3 Four-player rollback is beyond shipped SOTA Rollback for 2v2 tag means resimulating with four players' inputs inside the 8-frame budget — beyond anything shipped at 60Hz (2XKO's 2v2 uses a different assist model; classic tag rollback games run lighter sims). Treat it as a named R&D risk with a designed fallback (hybrid delay-rollback for 2v2, pure rollback for 1v1) rather than a listed feature, or the netcode promise — the product's most credible pillar — takes its first credibility hit in the mode least able to absorb it. ### 10.4 Cinematic story economics An NRS-grade 12-chapter branching story is its own mocap/VFX production (NRS story modes are estimated in the tens of millions). The Ruleset-Season re-scope (§6.1) should sequence story _chapters_ with the same discipline as rulesets — a 4-chapter launch arc that completes, then season-arc continuations — rather than promising the full branching Light/Dark/Path-Three structure day one. ### 10.5 Hardware legality and platform friction Publish the tournament-legal controller policy early (leverless/SOCD rules exist in code; the _policy document_ is what TOs need), and decide Steam Deck ranked eligibility before marketing "Steam Deck verified" — kernel anti-cheat on Proton requires explicit enablement and historically excludes Linux from ranked queues in competitor titles. An honest "Deck: casual + offline everything, ranked on supported platforms" is fine; discovering it at launch is not. ### 10.6 Voice, AI, and union realities AI commentary, procedural DLC dialogue, and lip-synced caster faces intersect directly with the interactive-media agreements' AI provisions (digital-replica consent, compensation for trained voices). The persona validation pipeline models consent; make it contractual with named talent before any AI-voice feature ships, and expect union questions at every press cycle — have the policy one-pager ready. ### 10.7 Esports operations behind the Crown circuit Beyond prize pools: a published competitive ruleset charter, player conduct code, prize escrow and payment SLAs (the FGC has been burned repeatedly), visa support for international qualifiers, and TO support kits (the offline LAN build + bracket tooling exist — package them). Governance failures, not game quality, are what kill new esports circuits in year one. ### 10.8 Cross-portfolio notes V2 and V4 share roster/GAS/feel-gate architecture — consolidate a combat platform team rather than parallel implementations. The V6 Ori incarnation seam (`v2-fighter` destination adapter already specced on V6's side) is a costless delight: your Egbe companion as a casual-mode guest fighter with its own personality-derived style. And V2's deterministic replay + server re-sim stack is the portfolio's anti-cheat crown jewel — export it as shared infrastructure to V4/V5 rather than letting each reinvent it. ## 11. Monorepo integration & shared functionality (third pass) Audited against DOMAINS/, the 185-phase TODOS/ roadmap, `V_SERIES.md`, `V_SERIES_PLATFORM_CONSOLIDATION.md`, and package-dependency evidence. ### 11.1 The Shakti bridge is the pattern — and V2 is its only proof `libs/shakti/fighting-ruleset-bridge` ("deterministic Shakti biomechanics → fighting-game frame-data bridge," with launch profiles) is consumed by exactly one product in the entire repo: `apps/v2/shakti-ruleset-bridge` — the `EV2ShaktiRulesetId` enum's origin. This is the intended architecture working: a wellness-domain movement library feeding a game's ruleset layer through a typed bridge. V2 should both celebrate it (the "real biomechanics under the rulesets" story is marketable) and extend the pattern — Aja (Phase 16, complete) for animation retargeting/mocap cleanup, Euterpe (Phase 38 + the Phase 132/133 sovereign audio runtime and DAW work) for the karaoke/rhythm mode's audio spine, and Calliope for AI commentary (V4 already cites the Calliope stack by name; V2's commentary docs describe the same capability without naming the shared lib — name it, share it). ### 11.2 Two combat plans exist; only one has code Root roadmap Phase 84 ("Combat Systems SOTA," multi-genre frame-data engine, SF6/Tekken8/UFC-parity, building on `@maya/games` and `@aphrodite/game-fighting`) is **0/840 tasks** — while V2's own `UV2CombatResolver` (~24k lines, 407 specs) is the most real combat code in the monorepo. Two owners for one capability is how portfolios bleed: either V2's resolver is _declared_ the Phase-84 deliverable (and V4 consumes it — see V4's review §10), or Phase 84 should be closed as superseded. The dossier evidence says V2's resolver wins on merit; make it official before anyone staffs the 840-task ghost. ### 11.3 Service sprawl vs. the GPS plan `V_SERIES_PLATFORM_CONSOLIDATION.md` counts V2 at 50 service libs; the directory now holds **91** — the sprawl grew past its own consolidation doc while that doc sat unstarted (`libs/contracts/gps/` does not exist). The duplication census makes V2 a principal offender and a principal donor: one of ~7 independent matchmaking/Glicko implementations, one of 5 rollback owners, one of ~6 battle-pass implementations, one of 3 anti-cheat stacks. Recommended posture: V2's **season-pass-service** (fully implemented, idempotent, server-authoritative) should be donated as the GPS seasons/leaderboards seed; V2's deterministic-replay re-sim should become the GPS anti-cheat signal bus's premium input; and no new `apps/v2/*` package should be accepted where a GPS domain is specced (the consolidation doc's own exit criterion, applied now). ### 11.4 The orphaned TS game runtime is not V2's answer `libs/aphrodite/` hosts a 27-package generic game runtime including `game-fighting` — with **zero** external consumers. V2's UE C++ path is correct for a rollback fighter; the portfolio decision (adopt-or-delete the aphrodite runtime) merely needs V2's explicit non-dependence recorded, so nobody "consolidates" a shipping-critical fighter onto an unproven TS engine in the name of reuse. ### 11.5 Roadmap assets V2 under-uses - **Phase 82 (voice-as-gameplay-input, 0/145)** — V2's accessibility suite is the natural first consumer (voice-triggered macros for motor accessibility, already ranked-flagged by the assist-tier machinery). - **Phase 137 (platform/live-service sovereignty: console SDKs, matchmaking, anti-cheat, LiveOps)** — overlaps V2's own launch stack; V2 should be named a co-owner rather than discovering parallel work at cert time. - **Themis Music Shield (Phase 74)** — the karaoke mode's licensing exposure (§8.7 first pass) has a completed in-house originality/IP screening layer available; wire it into the track-intake pipeline. - **Concordia (Phase 179)** — V2's anti-cheat appeals already cite a Concordia-style substrate; the real, ~complete Concordia service should back it rather than a V2-local implementation. ### 11.6 Cross-product notes with code evidence The V6 Aye Bridge's `v2-fighter` destination adapter exists on V6's side with no V2-side package yet — cheap goodwill when the roster/GAS stack stabilizes (§10.8 second pass stands). And the V_SERIES honest-status table records V2 at "1 .uasset, no playable content" — consistent with this review's §8.1; the checkbox-vs-cooked-content distinction is now enforced portfolio-wide by `V_SERIES.md` convention, which V2's status reporting should cite rather than restate. ## 12. Closing note The V2 corpus reads like a team that knows exactly how good fighting games work and has encoded that knowledge with rare rigor — the Defensive Options Library alone demonstrates deeper genre literacy than most shipped games. The review's hard message is that V2's biggest competitor is not Street Fighter 6; it is V2's own launch promise. Convert the impossible day-one manifest into a two-year cadence of complete-on-arrival ruleset seasons, pick the esports flagship, keep the monetization halo intact, and this is a generational product. Ship it as currently promised, and it joins the long list of games that were announcements.