QuintAce · Product
Last updated 2026-04-30 Q2 Week 5 v0 draft
v0 draft — auto-generated from Q2_Tactical_Execution_vs_Planning.md. Updated weekly before execution meetings.

Q2 2026 — Product Execution Tracker

Weekly execution tracker — maps actual progress against the Q2 plan. Each theme has a target experience and success gate. Features tracked weekly. Themes ordered by experience layer (L1 first → L8 last).

Status
Q2 Week 5 — v0.6.0 QA staged (RankUpLoop / Leak Review / IC / MCP Expansion); v0.7.0 QA scoped (Multi-Game Format + Guided Navigation); MTT Compete step-1 prototype shipped
Quarter
April 1 — June 30, 2026
Last Updated
2026-04-30

Meeting Notes This Week (Apr 27–30)

Done
On track
At risk
Not started
Blocked
P1P2
🔒 = locked, can't slip

Experience Gates Dashboard

Read this first. If a gate is red, that theme needs immediate attention. We don't move to higher layers until lower layers pass.
GateThemeTargetCurrentStatusLayer
All game formats playableArena: Multi-Game4 games in Compete+Practice3 in PROD (NLHE + PLO + Squid); MTT step-1 prototype Apr 30 (Mariana); MTT Analysis Mode v0.1 spec Apr 29 (Svilen); v0.7.0 QA scoped Multi-Game FormatAt riskL1
D7 Retention >20%Coach: Activation>20%No baselineNot startedL1
First loop <24hCoach: Activation>50% complete in 24hNo baselineNot startedL1
Quint visible in onboardingCoach: Conversational>3x usage/sessionNot measuredNot startedL1
Hand Import 5+ platformsCoach: Analysis5 platforms1 (V1 in QA)At riskL2
Session Review surfaces leaksCoach: AnalysisActionable leaks shownIn progressAt riskL2
Personalized drills from real leaksArena: PracticeDrill completion >60%Spec draftedNot startedL2
Casual Practice Journey Ch.1Arena: Practice>40% complete Ch.1Spec + prototype doneNot startedL3
D30 casual improvingCoach: Learning8% → 15%No baselineNot startedL3
Exploit view in Range ExplorerArena: Exploit CoachUsers toggle exploitAPI released, FE in designAt riskL4
Non-default personas selectedArena: AI PersonasUsers pick personasBase 3 shippedAt riskL4
Viral coefficient measuredCoach: Social & ViralBaseline setNot startedNot startedL5
15+ coaches with communitiesCoach: Communities15 active6 engaged (no communities yet)At riskL5
Organic signups >500/moCoach: Web Platform500/month from webBlog nearly liveNot startedL6
Badge unlocks drive returnsArena: GamificationDaily challenges >30%Art ~60%Not startedL7
Subscription infra liveCoach: MonetizationJuneDeferred to P2 — revisit when L0-L5 passNot startedL8
A

AceArena (Poker Engineering)

Q2 Goal

"Every game format playable — compete, practice, exploit, badges. No broken states."
Gate: All formats Apr 30. Exploit + Personas May 31. Gamification Jun 30.

L1 — Live Gameplay & Multi-Game

"I can play any game format — PLO, Squid, MTT, Guandan"
GateAll formats in Compete + Practice. No broken states.
LayerL1
MetricActive Players (3K+)
FeaturePriStatusThis WeekBlocker
Game logic all variants
Backend API for all game formats
P1🔒API delivered
Range Explorer per game
Format-specific strategy grid for all game types
P1🔒~95% dev
PLO Compete
Full PLO cash game in Compete mode
P1🔒v0.5.0 in PROD; DS quality fix shipped; PLO4 pipeline 67%DS update Apr 27 (Aliak): action-scores API quality fix shipped (ADS-886) — "wrong/inaccurate" label values corrected. PROD data backfill complete (ADS-888). ADS-530 epic (QuintAce NRT to PROD) closed. PLO4 (ADS-870 epic): 6 pipeline tasks done; luck score blocker resolved by Alex/AceExperts (ADS-875 In Progress); action-scores pipeline + rank notification (ADS-874) in progress. PLO4 progress: 67%.RL EV accuracy for 4-bet pre-flop in 3-blind half-anti (still open)
PLO Practice
PLO drills and practice hands
P1🔒v0.5.0 PROD shippedMaria working on spot drills first, then league drills (Apr 20 Weekly).
PLO Strategy Grid combo grouping
Group PLO combos by category (169 → ~9 groups)
P1PLO bucket fix landedBreno fixed PLO bucket issue weekend Apr 19-20 (Apr 20 Weekly). Bet sizing API path now via Glue.
Squid Hunt Mode V1
Squid game format with hunt mechanics
P1🔒v0.5.0 in PROD; Squid pipeline 64%DS Apr 27 (Aliak): Squid pipeline (ADS-848) has 5 stages done — Kafka contracts (ADS-844), pre-enrichment (ADS-845), enrichment (ADS-846), statistics (ADS-847), Mongo sync (ADS-848). Mongo schema + DBSync silent data loss fix (ADS-953) in review (Sasha). API connection (ADS-849) in progress (Henry). New tasks scoped: action-scores API (ADS-909), player rank notifications (ADS-910). Squid progress: 64%.
Squid Practice + Explorer
Drills and range exploration for Squid
P1🔒Designs in progressDmitry Sergushkin started multi-game Practice + Explorer; new drill setup panel (classic/hunt mode + custom Squid distribution). Player labels switching from P1/P2 → UTG/HJ/CO/BTN/SB. Same modal pattern for Range Explorer settings (UX<>Product Apr 21).
MTT Compete V1
9-player freezeout SNG tournament mode
P1🔒Step-1 prototype shipped; in v0.7.0 QA scopeMariana shipped Tournament Creation step 1 (Basics) prototype Apr 30 at tools/ai_designer/specs/mtt_tournament_creation/prototype/step_1_basics.html — links forward to step_2_tournament_setup.html. v0.7.0 QA doc (Yannick, Apr 28) lists Multi-Game Format integration as main feature. v0.6.0 QA doc explicitly disables "New games (MTT, SNG, ...)" — MTT gated behind v0.7.0.Data-services PLO4/Squid pipeline completion (DS Q2 tracker Apr 27: PLO4 67%, Squid 64%); ICM data missing from imported tournaments
MTT Practice + Explorer
Tournament-aware drills with ICM evaluation
P1🔒Analysis Mode v0.1 spec landed Apr 29Svilen shipped MTT Analysis Mode spec v0.1 Apr 29 (specs/mtt/mtt_analysis_spec_v0.1.md, 387 lines). Defines: game-format filters (QuintAce + Partner), player-card with tournament overview stats, hand-stats with stage breakdowns, two-level navigation (tournament list → single tournament view), saved hands, current/next prize logic with late-reg awareness. Out of scope V0.1: Strategy Review (disabled), ICM overlay, Session Review.Data contract alignment
Partner Games (Guandan)
Chinese card game via AceSense partnership
P1🔒Olena offered to take overOlena to discuss with Eric re: Guandan ticket (Eric loaded; friend system not ready). Horizontal layout pursued as quick solution; portrait deferred. Luiz to share Guandan links + portrait concepts with Olena/Arturas (UX-Product Apr 24).

L2-L3 — Practice Engine

"I practice exactly the spots that cost me money"
GatePersonalized drills from real leaks. Drill completion >60%.
LayerL2-L3
MetricD30 serious retention
FeaturePriStatusThis WeekBlocker
Drill engine for all formats
Unified drill system supporting PLO, Squid, MTT
P1In devNew Practice Mode drill setup panel design (multi-game) starting — Squid Lab/Squid Value renamed for capitalization consistency (UX<>Product Apr 21).
Personalized Drills
Practice specific mistakes (spot drills) or leak patterns (leak drills)
P1In devMaria progressing on spot drills, then league drills (Apr 20 Weekly).
Bet Sizing Simplification
Three-tier bet sizing progression by rank level
P1🔒Spec ready, in designSpec ready and in design — soon ready to push to team (Svilen, Apr 20 Weekly).
Casual Practice Journey
Skill-based journey: 11 chapters, 5-tier mastery, connected path UX
P1Taxonomy review pre-reqThanh: review skills + leaks taxonomy before starting casual practice journey. Svilen to schedule meeting Thanh+Svilen+Miguel this week. Miguel to share taxonomy code with Svilen (Apr 20 Weekly).Taxonomy review

L2-L5 — Social Competition

"I play with friends, in my community, or in ranked leagues — and I share my results"
GateLeagues across 2+ formats. Private tables weekly. Share triggers clicked.
LayerL2-L5
MetricWAU growth, Viral coefficient
FeaturePriStatusThis WeekBlocker
Leagues V1
Multi-game leagues building on V0 (6 levels, 50-player pools, 1-month seasons)
P1Working in devLeague functionality working; reset done quickly for testing (Vasyl, Apr 20 Weekly). Bryan redesigning communities screens after meeting with Svilen (UX<>Product Apr 21).
Compete Mode Redesign
Three entry points: Play Ranked / Community Play / Play with Friends
P1Communities-first; friends scrapped for nowInitial technical design for "communities + friends" not viable. Friends scrapped for now — communities first, friends to follow. Svilen drafting communities landing page spec (Apr 20 Weekly + UX-Product Apr 24). Brian assigned card-drill sharing; Olena gets friend system once unblocked (UX-Product Apr 24).
Matchmaking
Skill-based opponent matching for Compete mode
P1Groundwork via AI personasVasyl: matchmaking not yet in dev — groundwork via AI Liquidity Injection + AI personas with different styles/ranks to match league tiers (Apr 20 Weekly).
AI Liquidity Injection
AI fills empty Compete seats, transparent labeling, LP-safe
P1Foundation for matchmakingConfirmed as foundation for matchmaking strategy (Vasyl, Apr 20 Weekly).
Share triggers in session results
Share buttons on post-session screens (WhatsApp, Discord, X)
P1Not started
Private tables + share link
Invite-only tables shared via link
P1Design scope startingTable Share Dialog starts W16 (Bryan) — Arturas scopes technical constraints first, design follows. Shared visual = image snapshot of inviter + table.Technical scoping
Community tables + AI backstop
Always-on tables for communities with AI filling empty seats
P1Communities first, then friendsDefault "Friends Community" with 4 AI personas; 3-community user limit; weekly rank-points leaderboard. Friends scrapped for now per Svilen (Apr 20 Weekly).

L4 — Exploit Coach

"I know HOW to beat my specific opponents, not just what GTO says"
GateUsers toggle exploit view in Range Explorer. Practice against exploit profiles.
LayerL4
MetricD30 serious retention
FeaturePriStatusThis WeekBlocker
Exploitative ranges in Strategy Grid
Toggle GTO vs exploit view in Range Explorer
P1🔒API released, FE in designCarry from Apr 9. Thanh testing exploiter interface this week (CAI–Product Apr 24).
BEX integration
Connect exploit engine to Quint AI coaching layer (CAI + Gameplay AI)
P1Ian progressingMiguel: Ian made BX changes and brought it down for testing (Apr 20 Weekly). Yannick OOO until next week; Ian's focus this week = Exploit Coach + Multi Game Format + V2/V3 RL/DS integration (#cai_general).
Range Viewer extension
Dropdown to pick opponent profile in Range Explorer
P1Not started
Opponent archetype dialogue
Pre-hand info showing opponent type and suggested adjustments
P1Not started
Practice against exploit opponents
Drill hands against specific exploit profiles
P1Not started

L4 — AI Opponents & Personas

"I play against realistic opponents that feel like real players"
GateUsers select non-default personas. Persona challenges shared virally.
LayerL4
MetricActive Players, Viral coefficient
FeaturePriStatusThis WeekBlocker
6 opponent profiles
Add The Caller, The Folder, The Maniac to existing 3
P1Bot profile work transferred to JonathanDavid: handing off bot work to Jonathan, who will use different bot profiles (aggressive/tight/coach) to run DA pipeline and see how article results change by coaching style (CAI/AceExpert Sync Apr 23). Xiangzhi continuing 5-bot creation with shifted strategies — cleaning PR for aggregation tool concurrency (Apr 23).
AI Persona gameplay (stat → FCR)
Convert stat profiles to actual gameplay decisions
P1Not startedCarry.
Persona selection UI
Browse and pick AI opponents before playing
P1Not startedP2 — carry.
Persona branding
Name, avatar, backstory for each AI persona
P1Pivoting to 12 special charactersVitalii dissatisfied with Roussan's avatar work; team pivoting from current avatar system to 12 special characters (opponents + coaches) — easier to handle by available concept artists. Larger avatar project deferred until ecosystem use (achievements) defined. Roman trialing artist Andrea for full-time Quint art lead role (UX-Product Apr 24).Art lead hire
Operator-specific profiles
AI mimicking real platform tendencies (e.g., "WPT Global Caller")
P1Not startedThanh: 100+ profiles needed (50–75 strategy/rank + influencer + operator imitation models). Apr 20 Weekly.
User-configurable profiles
Custom opponent stats (VPIP, PFR, 3-Bet, AF, etc.)
P1Not started

L7 — Player Identity & Gamification

"I see my progress and earn badges that mean something"
GateBadge unlocks drive return visits. Daily challenges >30%.
LayerL7
MetricD7 retention
FeaturePriStatusThis WeekBlocker
Badge system (10 badges, B/S/G)
10 skill badges with Bronze/Silver/Gold levels
P1Tied to taxonomy reviewThanh uncomfortable with current badge system (tied to taxonomy review). Gamification focus shifting to rank-up loop polish with animations + clear progress visibility instead (Apr 20 Weekly). Arturas: badges visible on homepage but missing from league preview dialogues — needs hookup.Badge art + taxonomy review
Badge triggers
Auto-award badges on gameplay milestones (hands played, rank-ups, streaks)
P1Not started
Basic badge celebration
Animated reward moment when earning a badge
P1Not startedRank animation to use current graphics temporarily until avatar direction locked.
User Profile
Player profile page with stats, rank, and history
P1🔒~80% ready; navigation pattern resolvedProfile = popup with close button (not back button — web-app navigation issue per Svilen). Logout moves from sidebar to profile page. Toggles for settings (preferred over chevrons). Tablet uses popup behavior. Avatar+rank+name combined into single container; edit icon next to avatar. Olena to apply changes (UX<>Product Apr 21). Tablet-specific design also requested.
Daily challenges
Daily gameplay tasks with rewards (e.g., "Win 3 hands with pocket pairs")
P1Not started
Friend profile page
View another player's profile and stats
P1Not started
Friend list on profile
See and manage friends from profile page
P1Not started
Achievement milestones
Long-term progression goals (e.g., "Play 1000 hands", "Reach Gold")
P1Not started
C

AceCoach (Web Application)

Q2 Goal

"Users get coached, follow a plan, invite friends, join communities. Prove the product retains before scaling."
Gate: D7 >20% by Apr 30. Casual retention improving May 31. Monetization deferred — revisit when L0-L5 pass.

L1 — App Experience & Activation

"The app is intuitive from first open — I know where to go and what to do"
GateD7 >20%. First loop completion within 24h. No navigation confusion.
LayerL1
MetricD7 retention, Activation rate
FeaturePriStatusThis WeekBlocker
New Onboarding (multi-plan)
User picks learning plan on first open, learns core loop
P1In QA; Arturas working on game-format presetsArturas working on onboarding changes including game-format presets and rank-up loop improvements (Apr 20 Weekly).
Revised Onboarding Flow
Streamlined first-time user experience
P1🔒In devCarry from Apr 17.
Homepage & Sidebar Redesign
Direct action cards per game format, remove broken progress UI
P1🔒Integration section in devOlena landing page slider finished except missing one image for "We Poker" (Chinese site file). Colors+assets handed to DC for tools/lessons section (Apr 20 Weekly). Olena to coordinate with Marketing on gradient change urgency — current purple too similar to Lucid/Phenom Poker; team exploring new gradient direction (UX<>Product Apr 21).Chinese partner screenshots; gradient direction
Header & Nav Consistency
Standardized header/nav across all pages (desktop + mobile)
P1Redesign in progressCarry.
Enhanced celebrations (4-tier)
Scaled animations: small win → rank-up → badge → milestone
P1Not startedLuiz W16: team proactively identifying gamification moments (streaks, rank-ups, rewards) as features ship.
Re-engagement
D3/D7/D14/D30 win-back sequences via email + push + in-app
P1Spec done
Streak system
Daily streak counter with recovery mechanics
P1Not started
Homepage gamification UX
Activity rings and daily goals on homepage
P1Not started
Landing page
Public marketing landing page
P1Hotfix Apr 14 + updates in progressHotfix Apr 14 to add mailchimp tag for new users (Sebastian #dom_quintace). Coin Case integration screenshots missing (UX-Product Apr 17). Olena working with Mykola to update landing.
Limited Access Beta
Invite-only access via codes
P1LiveConfirmed invite codes already live (Weekly Execution Apr 13).

L1-L2 — Conversational Coaching

"I ask Quint anything about my play and get a solver-backed answer"
GateQuint AI used >3x per session. Instant Coach visible during onboarding. Latency <3s.
LayerL1-L2
MetricD7 retention, Activation rate
FeaturePriStatusThis WeekBlocker
Instant Coach optimization
Tune IC triggers for onboarding "aha moment" on good AND bad plays
P1🔒O9 In QA — IC + HUD + Bot HUD all in v0.6.0CAI Q2 tracker Apr 30: O9 (IC New Flow) — AA-5373/5374 (Mistake & Good Play Adjustments, EV-based) In QA (Gytis); AA-2994 Opponent Profiling/HUD Done; AA-4867 HUD for Bots In QA (added Apr 27). Fix version quintace/0.6.0 applied to all O9 In QA tickets. v0.6.0 QA doc lists IC as main feature.Web App integration
Quint AI Button
2-4 contextual questions per page, deep-linking to drills/hands
P1🔒Guided Nav 94% (17/18) — final FE integrationCAI Q2 tracker Apr 30 (Yannick): O6 Guided Navigation 94% (17/18 Done). Last item: FE integration AA-5123 (Mykola, Committed). Aman closed /cai-guided-navigation slash command against Thanh's review feedback + 62/62 preset-queries QA pass. Web team (Mykola) re-slated Context Chat + predefined-question flows to 0.7 release (V4 cutover, V2 stripped); CAI-side adapters being aligned to QuintacePresetQueries schema. MCP QA doc for Guided Nav added Apr 30 (Aman, cai/shared/docs/guided-nav-mcp-qa.md) — Claude Desktop testing path with mock + real dev player personas.FE integration
Chat UI/UX improvements
Better chat interface, response formatting, conversation flow
P1🔒Embedded Quint UX expansion proposedThanh: explore generic mechanism for embedded Quint — additional box below decision overview / range explorer, OR floating button like Instant Coach. Luiz to design pattern (Apr 20 Weekly).
Additional IC flows
New Instant Coach scenarios beyond current coverage
P1🔒In progress
Exploit Coach in Analysis
Exploit recommendations integrated into Analysis mode views
P1🔒O5 66% (4/6 QA Validated); David started quality reviewCAI Q2 tracker Apr 30: O5 Exploit Coach 66% — 4/6 QA Validated; AA-4707 In Dev (David Robinson started quality review). EV-of-fold bug from Apr 23 (3-blind half-anti, RL post-processing assumes fold EV=0) — no commit-level update this week; carry. Ian: Squid info-format + MTT params + BB-sizing tools API shipped (O2). Ante double-scaling fix + stack-size bug diagnosed with AceExpert (Ian, O3).RL team EV fix (carry)
Quint explains exploit "WHY"
Quint narrates why exploit differs from GTO in plain language
P1Not started
Quint in persona voice
Quint adopts personality style matching selected AI persona
P1Not startedDavid: prompting needs to change to reflect coach type (tighter player / exploiter) instead of assuming default GTO when calling strategy service with profile param (CAI/AceExpert Apr 23).

L2 — Analysis & Insights

"I import my hands and see exactly where I'm losing money"
GateHand Import V2 supports 5+ platforms. Session Review surfaces actionable leaks.
LayerL2
MetricD30 serious retention
FeaturePriStatusThis WeekBlocker
Hand Import V2
Import from PokerStars, GGPoker, ACR, WPT, 888
P1🔒Disabled in v0.6.0 QA — gated for v0.7.0v0.6.0 QA doc (Sebastian, Apr 28) explicitly lists "Updated Hand Import flow (has now a status and report)" under Disabled Features. Feature is built but gated for the next release.DS Kafka topic stability (carry)
Session Review & Leak Reports
AI-generated session review with actionable leak identification
P1🔒Leak Review in v0.6.0 QA scope (Homepage + AnM)v0.6.0 QA doc lists "Leak Review (Homepage + AnM)" as main feature alongside RankUpLoop and IC. Analysis Mode prototype updates shipped Apr 28 (Nick): analysis_hub.html + leak_detail.html + leaks_discovery.html + session_mistakes.html + ui_spec.md (~2,600 LoC).KI infra for DA pre-population (carry)
Session Results Multi-Game
Post-session summary supporting all game formats
P1🔒In progress
Multi-Game Analysis: PLO
PLO hand analysis with format-specific leak detection
P1v0.5.0 PROD shippedPLO bucket fix landed (Breno, weekend Apr 19-20). Strategy grid now logical for PLO (Apr 20 Weekly).RL EV accuracy
Multi-Game Analysis: MTT
Tournament hand analysis with ICM-aware evaluation
P1Spec v0.1 landed Apr 29 (Svilen)Svilen shipped MTT Analysis Mode spec v0.1 (387 lines): tournament-list → single-tournament view, stage-based hand stats, saved-hands per-tournament + cross-tournament, current/next-prize logic with late-reg awareness. Two-column expandable hand-history layout (no sticky right panel issue) carried in. ICM analysis overlay still out-of-scope for V0.1.Data contract; ICM data
Rank-to-Percentile Mapping
"You're Gold = top 25% on PokerStars" (winning players only)
P1Alex feasibility doc requestedAlex Bondarev to conduct experiments + write attention-wrap-up document on ELO ranking feasibility for V2 scoring stakeholders (Poker Analytics Apr 22).David + Aleksandr
Cross-game Dashboard
Unified progress, leaks, and rank view across all formats
P1Not started
Multi-Game Analysis: Squid
Squid hand analysis with format-specific metrics
P2Designs in progressNew drill setup panel (classic/hunt mode + custom Squid distribution) being designed by Dmitry Sergushkin. Player labels switching to UTG/HJ/CO/BTN/SB. Same modal pattern for Range Explorer (UX<>Product Apr 21).Data contract alignment

L3 — Learning Plans & Journeys

"I have a clear path from beginner to expert — Quint tells me what to work on next"
GateUsers follow recommended plan without prompting. Casual Practice Journey: >40% complete first chapter.
LayerL3
MetricD30 casual retention
FeaturePriStatusThis WeekBlocker
Rank-Up Plan
Homepage cards guiding next action in the rank-up loop
P1🔒In v0.6.0 QA scope (main feature)v0.6.0 QA doc (Sebastian, Apr 28) lists RankUpLoop as main feature. Carry from Apr 13 (Sebastian): logic stays in AceCoach for Q2; preferred-game stack edge case still relevant (hands excluded from AnM but counted for RankUp).
Poker Basics Simplified Plan
Beginner track with simple language, no solver theory
P1Spec + 6 prototypesNo new signal Apr 13-19.
Dual-Lingual Journeys
Chinese + English learning paths for Asian market
P1i18n in progressCarry from Apr 9.
Beyond GTO Learning Plan
Expert track: solver theory → exploit strategies → population reads
P1No specMISSING SPEC

L5 — Social & Viral Growth

"I invite friends, challenge them, and we play together"
GateViral coefficient measured. Friend invites generate signups. Share links clicked.
LayerL5
MetricViral coefficient, WAU growth
FeaturePriStatusThis WeekBlocker
Friend system entity
Add/remove friends via share, post-game, search
P1Not started
Shareable Link Infra
/s/{type}/{id} URL scheme with OG meta tags and deep linking
P1Not started
Share-as-invite infra
Every shareable moment = implicit friend add + signup funnel
P1Not started
Add friend dialog + interstitial
Post-game "Add as friend?" prompt + request dialog
P1Not started
No-login content preview
Recipients see shared content without account
P1Not started
Two-sided invite rewards
Both inviter and invitee get rewards on signup
P1Not started
Progressive nudge
Gentle progression from friend connections to community membership
P1Not started
14 reusable components
Shared UI for friend/share/invite across all pages
P1Not started
Friend streaks
Streak counter for playing with the same friend
P1Not started

L5-L8 — Communities & Coach Platform

"I join a coach's community, play at their tables, and get coached via Discord"
Gate15+ coaches with active communities. Community member retention >60% weekly.
LayerL5-L8
MetricCoach effectiveness, WAU growth
FeaturePriStatusThis WeekBlocker
Coach Attribution / Affiliate
Generate codes, track signups, view attribution dashboard
P1🔒UI gap remaining
MCP Foundation
Unified MCP server (3→1), partner auth + scoped access
P1🔒O4 90% — MCP Expansion in v0.6.0 QACAI Q2 tracker Apr 30 (O4): 90% — HandStateSummary shipped across state-modifying MCP tools (Zeyu); DA article tool + v4 test runner shipped; QA 55/55 pass. submit_da_feedback MCP tool submitted (Zeyu, AA-5489); hybrid SKILL.md skill selector opened (Zeyu, AA-5493 In Dev). v0.6.0 QA doc (Sebastian, Apr 28) lists MCP Expansion (CAI) as main feature. Custom-taxonomy LLM agent scoped with Miguel — incremental per-user leak updates every ~20 hands on top of deterministic CSV mapping (Aman, O4). Chris out on holiday Apr 27 – May 6; Chang OOO until May 7 (Valerio covers).
Ecology Reports
Platform integrity dashboard, 5 platforms, fairness metrics
P1Industry report v1 done; multi-game features startingDmitry Sergushkin: completed industry report Apr 20; started multi-game Practice + Explorer feature work (UX<>Product Apr 21).
Community entity + membership
Create/join communities with invite codes, free tier: 50 members
P1Communities-first approach; spec draftingNo new commit signal Apr 27–30. Carry from Apr 24: Brian redesigning communities work after meeting with Svilen, aiming for concrete v1 by end of week. Svilen drafting "communities landing page" spec. Friends scrapped for now — communities only; friends to follow.
Coach profile pages
Public coach pages with bio, community, content (/coach/{slug})
P1Prototypes builtBryan: Community Web Pages started — extends Communities to web surface.
Content Engine
Coach reacts 5-10 min via MCP → auto-publish article + SEO content
P1213-article batch staged; 178 clean / 35 held backCAI Q2 tracker Apr 30 (O1): DA controlled feedback-loop design committed (Chris/Zeyu/Valerio). Evidence Footer (Range-vs-Range / Hand-vs-Range split) + raw-class villain-range donut fix staged for Harold's 213-article batch — 178 clean, 35 held back on LLM prose issues. AA-5489/5490 (L1 feedback loop) In Dev (Valerio); AA-5687–5690 pipeline tickets In Dev; AA-5674 in Code Review (Zeyu). Chris on holiday Apr 27 – May 6.Coach review (Chris OOO)
Community page in-app
In-app community view with leaderboard, tables, members
P1Not started
AI Personas
Persona challenge landing pages + Quint adopts persona voice
P1Not started
Coach Activity feed
Coach's recent actions, curated content, annotations
P2Deferred — paid community feature
Discord integration
Discord server linking with role-based access (Tier 2)
P2Deferred — paid coach feature
Coach tools
Coaches create custom drill sets and label leaks for their students
P2Deferred — paid coach feature
Discord Bot & Async Coaching
Quint AI bot in Discord (/analyze, /quiz, /range)
P2Deferred — paid coach infra
WPT Freeroll/Package
Cross-promotion with WPT ecosystem (freerolls, packages)
P2Deferred — partner monetization

L6 — Web Platform

"I discover QuintAce through free tools, lessons, or a blog article and sign up"
GateOrganic signups from web >500/month. Free tools indexed by Google.
LayerL6
MetricCumulative acquired (5K+), CAC
FeaturePriOwnerStatusThis WeekBlocker
Blog & Content Platform
Blog with 60+ articles, interactive widgets, coach attribution
P1TBDLive on quintace.devBlog live + tools/lessons ready on dev — can be made live this week if approved (Weekly Execution Apr 13). Leak of the Week article published; Thanh+Harold iterating on article voice (Slack Apr 17).
Landing & Homepage web
Public website homepage with "AI coach" positioning
P1TBDUpdates in progressIntegration section done in Figma, handed to Sebastian for implementation with film-strip animation (Luiz W15 Apr 17). Awaiting partner screenshots (Chinese site).
Free Tools & SEO Pages
20 poker tools (/tools/) indexed by Google for organic acquisition
P1DC ✅3 tools + 8 lessons dev-readyEquity Calculator, Hand Analyzer, Range Builder + 8 Duolingo-style lessons finalized + in dev review (Luiz W15 Apr 17). Tools + lessons live on quintace.dev dev env; ready for prod push.
Web Lessons Platform
/lessons/ catalog repurposing 386 widgets, coach-curated paths
P1DC ✅8 lessons dev-readySee Free Tools above. Vitalii working on improved design (Weekly Execution Apr 13).
Coach & Partner Web Presence
SEO-indexed coach profiles (/coach/{slug}), community discovery
P1TBDPrototypes builtBryan Community Web Pages starts W16 (Luiz W16).
Viral Challenge Experience
/challenge/{persona} branded AI challenge landing pages
P1TBDNot started
Outbound & Community Loops
Newsletter sequences, article comments, widget → drill deep links
P1TBDNot started

L8 — Monetization (Deferred)

"I upgrade because I hit my limit and the product is worth paying for"
GatePayment live when L0-L5 gates pass. First paid subscriptions. Free→paid conversion tracked.
LayerL8
MetricMRR ($60K+ target at M6)
FeaturePriStatusThis WeekBlocker
Payment & Subscription Infra
Stripe integration, checkout, subscription mgmt, trial system
P2Parked — PoC haltedPoC halted, do not merge (Apr 6). Revisit from architecture standpoint when business requires it.Owner: Arturas
Freemium Gating UX
What users see at limits, upgrade CTAs, trial experience
P2No specDepends on Payment
Community subscription billing
Paid community subs with 90/10 coach/platform split
P2Not startedDepends on Payment
Subscription & Partner Economics
Free/Plus/Pro/Elite pricing with spot-based gating
P2Strategy doc onlyDepends on Payment
S

Shared Infrastructure

ItemDescriptionOwnerStatusThis Week / NotesBlocks
DS NRT Engine → PROD
Near-real-time scoring (24 features) for multi-game analysis
AliakADS-530 NRT epic CLOSED; PROD backfill DONEDS Q2 tracker update Apr 27 (Aliak): ADS-530 epic (QuintAce NRT to PROD) closed. ADS-888 PROD data backfill complete. ADS-886 action-scores quality fix shipped ("wrong/inaccurate" labels corrected). ADS-872 (Strategy API game_mode_code re-alignment) resolved. QAce deployed to PROD (ADS-826). WPT Gold batch to PROD (ADS-704) 13/15 done — promote data pipelines (ADS-716) + API v2 (ADS-912) to Gold PROD remain. O2 enrichment perf optimization (ADS-891) done: 6 code changes (asyncio CPU offload, semaphore scope, in-flight queue right-sizing, connection pool, retry/timeout, zstd compression).Gold Staging API deploy (ADS-936)
Single Endpoint Convergence
Unified scoring/strategy API endpoint across clients
BrenoGlue migration ongoingNo new commit-level signal Apr 27–30. Carry: Strategy Grid → Glue API ongoing; Glue defaulting common params to reduce drift. PLO bucket fix landed weekend Apr 19–20.Consistent scoring across Quintace + clients
Analytics & Measurement
D7/D30 retention tracking, funnel analysis, go/no-go gates
SebastianCarryNo new signal Apr 27–30.L0 gate — all metrics depend on this
Platform Infrastructure
Push notifications, social auth, Kafka events, native app prep
Sebastianv0.6.0 QA scoped Apr 28; v0.7.0 QA scoped Apr 28Sebastian shipped v0.6.0 QA doc Apr 28 (releases/v0.6.0/QA_v0.6.0.md): main features RankUpLoop, Leak Review (Homepage + AnM), Instant Coach, MCP Expansion (CAI). Minor: User profile changes, Web Platform Lessons (FE only), Badges. Disabled: Updated Hand Import flow, MTT/SNG. Aiming staging Wednesday → PROD Thursday. Yannick added v0.7.0 QA doc Apr 28 (releases/v0.7.0/QA_v0.7.0.md): main features Multi Game Format integration (Webapp + CAI) + Guided Navigation (WebApp + CAI).
Design System & Brand
Badge art, visual consistency, design system sync (web + app)
Vitalii + OlenaCarryNo new commit-level signal Apr 27–30 on avatar pivot or gradient change. Carry from Apr 24: 12 special characters approach; Roman trialing Andrea as art-lead. Olena testing new gradient direction.Art-lead hire blocks badges + persona art
Share & Content Tech Stack
OG image generation, SEO infra, content pipeline
Coach teamCarryNo new signal Apr 27–30.
Anti-cheat — GeoComply Phase 3
CatBoost classifier with profile + account features
Ionas PakalnishkisEvaluation phaseApr 30: GeoComply features model doc shipped (anti-cheat/projects/cash-pool-bot-features/geocomply_features_model.md, 173 lines) — Phase 3 enriches baseline with core_view.user_profile_summary + pkw_db_coin.user_main joins; 15 new engineered features (username/nickname/avatar/account-type signals). Production save still commented out. Reference for QuintAce ecology integrity work.
X

Cross-Team Interface Contracts

FeatureInterfaceStatus
Communities
community_id + member_listAceArena owns — starting this month (Apr 6/9)
Compete Redesign
Community entity APILinked to community entity — AceArena
Social & Viral
Share payload APINot defined
Exploit Coach
Strategy Grid componentExisting
Gamification
Badge event triggersNot defined
Multi-Game
Strategy Grid + Decision Overview + canonical data formatv0.7.0 QA scope (Apr 28); Multi-Game Format integration (Webapp + CAI) is the named main feature for the next release
AI Personas
Persona config APINot defined
Payment
GET /user/{id}/tierParked — PoC halted (Apr 6)
W

Weekly Progress Log

Week of Apr 28 (Q2 Week 5) — v0.6.0 QA staged, MTT step-1 prototype, DS NRT epic closed, CAI roster shift
  • Source caveat: Slack MCP disconnected this run; meeting notes not pulled from Drive. Signals below sourced from product/engineering repo commits Apr 27–30 only.
  • MTT Compete step-1 prototype shipped Apr 30 — Mariana committed tools/ai_designer/specs/mtt_tournament_creation/prototype/step_1_basics.html (490 LoC). First concrete MTT artefact in repo. Forward-links to a planned step_2_tournament_setup.html (next prototype step).
  • MTT Analysis Mode spec v0.1 landed Apr 29 — Svilen committed specs/mtt/mtt_analysis_spec_v0.1.md (387 lines). Two-level navigation (tournament list → single tournament view), stage-based hand stats, saved hands per-tournament + cross-tournament, current/next-prize logic with late-reg awareness. Strategy Review, ICM overlay, and Session Review explicitly out-of-scope for V0.1.
  • v0.6.0 QA doc shipped Apr 28 (Sebastian, execution/releases/v0.6.0/QA_v0.6.0.md). Main features: RankUpLoop, Leak Review (Homepage + AnM), Instant Coach, MCP Expansion (CAI). Minor: User profile changes, Web Platform Lessons (FE only), Badges. Disabled: Updated Hand Import flow (status + report), MTT/SNG. Known issue: mobile issues after global zoom removed. Timeline: staging Wed, PROD Thu.
  • v0.7.0 QA doc scoped Apr 28 (Yannick, execution/releases/v0.7.0/QA_v0.7.0.md). Main features: Multi Game Format integration (Webapp + CAI) + Guided Navigation (WebApp + CAI). Hand Import gated for v0.7.0; MTT/SNG gated for v0.7.0.
  • DS NRT to PROD epic CLOSED (Aliak, Apr 27 — data-services/execution/q2/Q2_Execution_Tracker.md). ADS-530 done. ADS-888 PROD backfill complete. ADS-886 action-scores quality fix shipped. ADS-872 Strategy API game_mode_code realigned. QAce deployed PROD (ADS-826). WPT Gold batch (ADS-704) 13/15: API v2 (ADS-912) and pipelines (ADS-716) remain to promote.
  • DS enrichment perf optimization done (ADS-891). Six concrete code changes: asyncio event-loop CPU offload, semaphore scope fix, in-flight queue right-sizing, connection pool, retry/timeout config, zstd compression level. Strategy API baseline + gap report delivered. Pod sizing documented.
  • PLO4 pipeline 67%, Squid pipeline 64% (DS Q2 tracker Apr 27). PLO4: 6 pipeline tasks done, luck-score blocker resolved by AceExperts (Alex), action-scores pipeline (ADS-874) and rank notifications in progress. Squid: 5 pipeline stages done (Kafka contracts, pre-enrichment, enrichment, statistics, Mongo sync); Mongo schema + DBSync silent-data-loss fix (ADS-953) in review with Sasha; API connection (ADS-849) in progress (Henry).
  • CAI Q2 tracker (Yannick auto-syncs Apr 27–30) — multiple status flips: O3 Production Hardening: 94% (31/33) — 7 Done items moved from Remaining → Completed (AA-5141/5149/5164/5193/5194/2314 + others). AA-5164 Structured Health Checks Done. Remaining: AA-2992 (Stress Test, In Dev), AA-5145 (Chaos Suite, In Dev). O6 Guided Navigation: 94% (17/18) — final item is FE integration AA-5123 (Mykola, Committed). Web team re-slated Context Chat + predefined-question flows to 0.7 release (V4 cutover, V2 stripped); CAI-side adapters aligning to QuintacePresetQueries schema. O9 IC New Flow — AA-4867 HUD for Bots In QA (Apr 27). All O9 In-QA tickets (AA-5373, AA-5374, AA-2994, AA-4867) tagged with quintace/0.6.0. O2 Multi-Game: 6 agent MTT/Squid items Done in Apr 29 sync — chat (AA-5476), decision_analysis (AA-5478), drills (AA-5479), guidance (AA-5480), preset_queries (AA-5484), quintace (AA-5485). All Amandeep/Ian. O1 DA Quality: 26% (10/39). AA-5687–5690 pipeline tickets In Dev. AA-5489/5490 L1 feedback loop In Dev (Valerio). AA-5674 Code Review (Zeyu). AA-5493 hybrid SKILL.md selector In Dev. O4 MCP: 90%. HandStateSummary shipped across state-modifying tools (Zeyu); QA 55/55 pass. Custom-taxonomy LLM agent scoped with Miguel (incremental per-user leak updates every ~20 hands on top of deterministic CSV mapping). O5 Exploit Coach: 66% (4/6 QA Validated). AA-4707 In Dev (David Robinson started quality review).
  • CAI roster change Apr 30 (Yannick): Stefan Juang departed Q2 2026 — removed from TEAMS.md + CLAUDE.md. Oleksiy Dolgykh added as Engineering sub-team hire — onboarding to DA evaluation substrate + regression gate (O1: AA-5688, AA-5690). CAI titles corrected: Yannick = Principal AI Engineer / Manager / R&D / Architect; Chang Cheng = Lead AI R&D; Miguel = Lead AI; Valerio = Tech Lead AI.
  • CAI Datadog production dashboard linked in Q2 tracker (Yannick, Apr 29) — https://app.datadoghq.com/dashboard/s9p-gg7-33s/. O3 production hardening documented in architecture.md §19D with explicit Slack-channel refs.
  • Article pipeline batch staged (Chris, O1) — Evidence Footer (Range-vs-Range / Hand-vs-Range split) + raw-class villain-range donut fix on Harold's 213-article batch: 178 clean, 35 held back on LLM prose issues. Chris OOO Apr 27 – May 6.
  • Squid info-format + MTT params + BB-sizing tools API shipped (Ian, O2). Ante double-scaling fix + stack-size bug diagnosed with AceExpert (Ian, O3).
  • MCP Guided-Nav QA instructions (Aman, Apr 30) — cai/shared/docs/guided-nav-mcp-qa.md (83 lines). Claude Desktop testing path with mock + real dev-player personas, page-scoped scenarios, custom-persona flow.
  • Analysis Mode prototype updates (Nick, Apr 28) — pushed pending changes: analysis_hub.html (625 LoC), leak_detail.html (405 LoC), leaks_discovery.html (563 LoC), session_mistakes.html (392 LoC) + ui_spec.md (619 LoC). ~2,600 LoC new prototype + spec content. Also moved AI Design feedback into tools/ai_design_feedback/ (README + SHIP_PLAN + prompt + scorecard template, 5 files / 335 LoC).
  • Q2 Product Execution Tracker HTML + planning login gate shipped Apr 30 (Thanh, execution/html/product.html 1,123 lines + execution/html/planning-index.html 364 lines).
  • Gameplay AI weekly (Nimit, Apr 30) — autopot10/15/16/18 batch results posted; autopot15: gae_h=16 confirmed median, lr_half logged as cautionary tale; autopot15 added blocker_mean + escape-pipes formatting. gMLP retry-lane validation result RETRACTED for recipe contamination — flagged by Nimit himself, important for downstream consumers. Throughput response posted to Scott's feedback.
  • Anti-cheat / GeoComply Phase 3 model doc (Ionas Pakalnishkis, Apr 30) — anti-cheat/projects/cash-pool-bot-features/geocomply_features_model.md (173 lines). CatBoost classifier with core_view.user_profile_summary + pkw_db_coin.user_main joins; 15 new features (username/nickname/avatar/account-type signals). Evaluation phase — production save commented out.
  • Avatar pivot, Communities redesign, Strategy Service EV-of-fold bug, Brian's communities work, Casual Practice taxonomy review — no commit-level signal this week. Carrying from Week of Apr 21 entry. Slack MCP would resolve some of these on next sync.
Week of Apr 21 (Q2 Week 4) — v0.5.0 PROD shipped, CAI follow-on, communities pivot
  • v0.5.0 PROD shipped Apr 20 — release went out 8 AM UTC. Action-quality inconsistencies surfaced post-release: duplicated logic between lookup endpoint and Data Services. Aliaksandr ran PROD backfill same day to reprocess pro data through streaming engine. Sasha investigating; Vasyl rolled back V1/V2 work last week, so current inconsistencies aren't from that move.
  • PLO + Squid app does NOT use DS scoring (Vasyl, Apr 20 Weekly). DS only covers Hold'em. PLO bucket fix landed over weekend (Breno) — Strategy Grid for PLO now logical.
  • Strategy Service EV-of-fold bug (David Robinson, CAI/AceExpert Apr 23): RL model returns very high positive EV for a fold in 3-blind half-anti game; post-processing assumed fold EV = 0. Affects V1 + V2. Ian removed affected V2 migration component until RL team fixes; David pushing back on RL team. Same model also flagged with inappropriately high EVs for any raise in pre-flop 4-bet spot.
  • CAI v0.5.0 shipped to PROD Apr 24 (Yannick #dom_quintace). Mohamed: session-review.quintace.ai (dive-deeper) NOT yet deployed to PROD — main reason held. Gytis: DA not returning for Lacking/Close actions instantly on PROD (raised Valerio + Yannick).
  • Strategy Grid → Glue API migration ongoing. Glue introducing default-request defaults to reduce drift across services calling slightly different requests. Single endpoint convergence tracking continues.
  • Communities-first; Friends scrapped for now (Svilen, Apr 20 Weekly + UX-Product Apr 24). Initial technical design for "communities + friends" combined was not viable. Brian redesigning communities; v1 by end of week. Svilen drafting communities landing-page spec (lightweight web discovery for influencers). Olena gets friend system once unblocked.
  • MTT prioritized over Squid for next dev (UX<>Product Apr 21). Eric starting MTT Apr 22. Tournament lobby mobile views: Mariana shipped 3 options — Quintes-style preferred over GG-style and WPT-style. Two-column expandable hand-history layout proposed by Svilen for tournament view (resolves sticky-right-panel issue on Gold). Tournament-specific stats deferred — overall stats stay constant when inside tournament for V1.
  • Profile page navigation pattern locked (UX<>Product Apr 21): popup with close button (not back button — web-app navigation issue). Logout moves from sidebar → profile page. Toggles for settings preferred. Avatar+rank+name combined into single container; edit icon next to avatar. Tablet uses popup behavior. Olena to apply changes.
  • Avatar pivot (UX-Product Apr 24): team moving from full avatar system → 12 special characters (opponents + coaches), easier to handle by available concept artists. Larger avatar project deferred until ecosystem use defined. Roman trialing artist Andrea as full-time Quint art-lead candidate. Vitalii dissatisfied with Roussan's recent avatar work.
  • Gradient change exploration (UX<>Product Apr 21): Olena testing new gradient direction with AI-style stroke animation. Current purple too similar to Lucid Poker / Phenom Poker. Marketing pushing change; mostly affects homepage + marketing materials — in-app swap considered low-effort.
  • Embedded Quint expansion (Apr 20 Weekly): Thanh wants embedded Quint beyond hand analysis — into Range Explorer, especially PLO. Luiz designing generic UI pattern (extra box below decision overview, or floating button like Instant Coach). Connect to right endpoints; simplify prompt management so domain experts (coaches) can write/manage prompts.
  • AI Personas roadmap (Apr 20 Weekly): Thanh sized 100+ profiles needed (50–75 strategy/rank + influencer + operator imitation models). Jonathan owns opponent profiles + integration with Quint mechanisms (David delegated). Xiangzhi cleaning PR for aggregation tool concurrency reduction; took over bot profile work — simulating different RL bot profiles vs each other for mistake scoring.
  • Async DA blog pipeline on staging (Chang Cheng, CAI/AceExpert Apr 23): new tools article2, article_status, article_list — async generation of HTML+SVG artifacts in ~1 min via cloud desktop on staging. Chang OOO until May 7 (Valerio covers).
  • MCP stability concern (Thanh #cai_general): "Dev MCP is very unstable, hangs all the time." Asked Ian + ZEYU for low-hanging fruit. Ian: Exploit Coach needs 3 blinds — possibly cause of stalls; clearer error messages may help. ZEYU shipped structured state feedback on Dev — every MCP tool that mutates a hand returns state snapshot (street/board/pot/stack/positions/cards) so agent doesn't re-parse markdown.
  • PLO equity calculator design (Dmitrii Regent, #cai_general): Monte Carlo over compressed representatives. Sent Breno compression-target request (target=50 for PLO4, 500 for PLO5). Current compression yields up to 28% equity error. PLO RL solutions don't match any known ground truth — David instructed team to assume RL is ground truth for now and feed feedback to RL team.
  • Feedback loop architecture (CAI–Product Apr 24): multi-tier verification pipeline — programmatic (solver) / LLM judge / human expert per layer. Lower layers (action freq) treat unverifiable feedback as issues; presentation layers can auto-optimize. Initial release: feedback selectable by user (not LLM-classified) into MCP interface for coach testing. Valerio targeting end-to-end MVP next week.
  • KI team priority confirmed: full replacement of contextual flow wiring (Thanh, Apr 20 Weekly). Mykola driving with KI team to fix event inconsistencies. Build infra to call/pre-populate DA for all hands lacking decisions so analysis shows up properly inside leak review.
  • Bet Sizing Simplification spec ready, in design (Svilen, Apr 20 Weekly) — soon ready to push to team.
  • Casual Practice Journey blocked on taxonomy review: Thanh wants skills + leaks taxonomy reviewed first. Svilen scheduling Thanh + Svilen + Miguel meeting this week. Miguel sharing taxonomy code with Svilen.
  • 3-blind STR position bug (#dataservices-general): FE treats UTG as normal player; no seat sent for straddle position. Arturas asked DS to map UTG→STR; Aliaksandr suggested FE handle as client adjustment. AA-5567 created for FE adjustment.
  • Hand-import status Kafka integration (#dataservices-general Apr 23): Yifan integrating AnM with dataservices.hand-processing-status Kafka topic. Deployed to dev; debugging missing status for one test hand with Sasha.
  • Brad/Thanh straddle article (Apr 20): Brad confirmed "classic UTG straddle" = 3 blinds. Article hook to emphasize format's trending nature on WPT Gold/Global + Asia (We Poker). Thanh to add hand examples + strategy grids; Brad to recruit students for fact-check.
  • Squid Primer w/ Petrangelo + Dvoress + Andress (Apr 20): experts pushed back on data-heavy approach — content felt mechanical, hyper-focused on minor stats. Pivot to high-level coaching concepts (river bluff catching, c-bet adjustments). VPIP data flagged broken (too low for ante size — needs investigation; likely model query issue). Petrangelo + Dvoress to provide TOC / brain-dump for content regeneration.
  • CAI follow-on issues (Gytis #dom_quintace): MCP UI button not working post-release; Mohamed checking. STR(UTG) action requests failing — created AA-5535 (DS collaboration needed). Translation gaps in component library raised AA-5532, AA-5534 (CPal working on localization for component library).
Week of Apr 14 (Q2 Week 3) — v0.5.0 Release Readiness
  • v0.5.0 PROD release scheduled Mon Apr 20 8 AM UTC (Sebastian #dom_quintace Apr 17). DS, CAI, AceCoach, AceArena all releasing. Maintenance page 30min before; QA regression testing after.
  • Orthodox Easter delay (Apr 13 Weekly Execution): most developers out Apr 13; release candidate review delayed. Team focused on quality/consistency rather than new features.
  • Quality labels broken across all drills (Gytis DM Apr 19) — BIGGEST release risk. Friday staging change to fix CM-history-vs-AnM-suggested-actions quality-name mismatch appears to have inverted the mapping: Close actions saving as Wrong and vice versa. Team will attempt fix Mon.
  • PLO4 issues for release: EVs don't match between suggested actions (EV buckets) and grid — raised 1-2 weeks ago. Pot calculation wrong in PLO summarizer (Thanh QA-requests Apr 17). Range Viewer v2 percentages showing uniform 16.67% / incorrect quads % (Thanh Apr 18).
  • Guandan excluded from next PROD release (Thanh QA-requests Apr 17).
  • Single endpoint convergence (Weekly Execution Apr 13): Breno documenting pros/cons. Glue migration this week; full convergence on single default B2C request in 2-3 weeks. V2 protocol ready pending model training.
  • V2 scoring investigation (Weekly Execution Apr 13): David Robinson found front-end data incorrect — DS likely converting EV V2 into labels incorrectly. Thanh requested DS pick ONE clean V2-correct interface as testing source of truth. Ian integrating EV-based scoring library into lookup tool (CAI Sync Apr 16).
  • Instant Coach v1 validated E2E (Haoyuan #cai_general Apr 14): 6 flows confirmed in dev — onboarding, continuation, session goal, rank progress, in-game Great Play / Review Last Action, play-style change. Amandeep + Haoyuan closing IC for Wed product sign-off.
  • IC feedback from Thanh: too much info on mobile — feels like analysis mode. Prioritize hands by EV. Cadence ~1 good decision shown per 3 hands (Weekly Execution Apr 13). Arturas has prototype.
  • IC "much better" during onboarding — non-boring feedback per hand (UX<>Product Apr 14). Decision: don't change "only appears after sitting at table" — just change text. Arturas to adjust.
  • Exploit Instant Coach v2 shipped (Ian #cai_general Apr 17): filters minor % changes, "GTO already optimal" when all players GTO, clearer per-player profile messaging, few-shot examples.
  • Leak fixes on dev (Yannick #cai_general Apr 14): E2E tested, asking Arturas+Thanh to re-review.
  • Communities in review (Luiz W15 Apr 17): Bryan posted 5 screens — Leagues, Community Tab, Discover, Create, Invite. Awaiting Svilen + Luiz feedback. Default Friends Community (4 AI personas), 3-community user limit, weekly rank-points leaderboard (UX<>Product Apr 14).
  • Hand Replayer integrated + live on QuintAce (Weekly Execution Apr 13). Two-mode redesign (interactive Replayer + static shareable Hand Summary) targeting W16 (Luiz W16 Apr 17).
  • W15 done (Luiz Apr 17): Web Tools + 8 Lessons dev-ready (Equity Calculator, Hand Analyzer, Range Builder), Homepage Integration section with film-strip animation, Solution Confidence format-maturity badge locked (Option B.1 top-right), Design System token consolidation → Raf, AI Prototyping grounding file (design_system.md on real Figma tokens).
  • W16 starting Mon Apr 20: Hand Replayer final delivery, Tournament Lobby MTT (table format), My Profile redesign (Olena 80% done + Mariana on settings), Bet Sizing Simplification 3-tier, Table Share Dialog (Arturas scopes tech first), Community Web Pages (Bryan), Hand Bookmark label, Design2Code Track 1 live test Mon (Nick + Mykola + Raf).
  • Tournament Lobby revamp (UX<>Product Apr 14, UX-Product Apr 17): Mariana broke Compete into 4 parts (listings, lobby model, gameplay, create journey). Option 1 chosen for listings (less dev effort). Badges (MT, SNG) to differentiate types. Guaranteed price pool: higher of guarantee or calculated buy-in total.
  • My Profile Page (UX<>Product + UX-Product): Olena ~80% on profile + global settings. Prioritized over Bet Sizing. Layout options: collapsible settings vs always-visible; left-side nav with content on right (Vitalii preference). Mariana stepping in on settings structure.
  • Compete Mode game selector (Dmitry Sergushkin, #quintace_product_design): updated designs for PM-94 (CM Table Creation), PM-91 (CM Lobby), PM-92 (AnM Game Format).
  • Text reduction initiative (UX<>Product Apr 14): Arturas to polish UI text — remove redundant leak review text, rank loop step descriptions (users see them below); keep ranking system explanation (only source).
  • WPTG + Gold WPG (Weekly Execution Apr 13): being QA'd, targeting PROD soon. First release of completely new feature; Global release stress-tests new near-real-time Kafka ingestion pipeline.
  • Guandan decision (UX-Product Apr 17): portrait mode not urgent (no apps to benchmark); Luiz + Arturas agree keep landscape for now. Eric to review old Olena mobile explorations + estimate W16. Thanh excluding Guandan from next PROD release.
  • Chinese Guandan design files + Coin Case integration screenshots still missing (UX-Product Apr 17).
  • Poker Analytics (Apr 15): Alex Bondarev transferred 2B action dataset to Data Bricks but 200 RPS limit → 10+ days for full inference; plan: segment by high-hand-count players. Need Rust equity calculator for PLO pre-flop (Python too slow). 5-bot profile continuation with Xiangzhi.
  • MCP v4 (#cai_general Apr 17): Zeyu all test cases passed. gh-tools 50-hand test: 76% pass; failures = backend instability, not framework. MCP tool for DA blog pipeline sent for PR review.
  • Nick migration 72% cost reduction (Jeffrey Wang Apr 17): overall compute cost $105K/mo → $29K/mo after new cluster migration (not directly Product, but removes infra risk).
  • Landing hotfix (Sebastian Apr 14): mailchimp tag for new users added to v0.4.0 PROD.
  • Eric back from holiday; W16: Reddit App redesign + Poco App both start Monday (UX-Product Apr 17). Delayed partly by DC changing their mind; Roman's team may provide moodboard support.
Week of Apr 7 (Q2 Week 2)
  • Monthly Planning (Apr 9): Communities + entities confirmed as P1 (biggest Q2 features). Exploit Coach = P1.5. Matchmaking + AI personas = P2. Badge system = P1.5/P2.
  • Key ownership changes (Apr 6): Casual Practice Journey moved to AceArena. Community features consolidated under AceArena. Gamification (streaks/daily goals/rings) owned by AceCoach.
  • Monetization PoC halted (Apr 6): Do not merge existing code. Revisit from architecture standpoint when business requires.
  • Canonical data format decision (Apr 6): QuintAce must align with Data Foundation's canonical format for PLO/Squid/MTT. Aliaksandr meeting Yongtao/Jose this week.
  • Instant Coach bugs (Apr 6): Hand reading errors (wrong cards/actions) and incorrect leak categorization. Yannick investigating via Kafka events.
  • Rank-Up Loop bugs (Apr 6): Missing hand counts (15 shown out of 25 played). Incorrect leak bucket assignment.
  • PLO + Squid staging release (Apr 6): Prepared on staging with real-time API from DS. PROD push targeting Wednesday, contingent on QA.
  • Landing page + invite codes (Apr 6): Deploying to production.
  • UX-Product Sync (Apr 10): Homepage messaging adjustments for WPT. Olena assigned My Profile Page + Bet Sizing. MTT design split (4 design + 3 dev). Lessons pages: Bryan working, synced with Svilen. Analysis mode "table bookmarking" and "analyzed vs non-analyzed" completed.
  • i18n decision (Apr 6): One unified mechanism across all teams required. Broadened to multi-language (CN/ES/RU/EN) but not priority this month.
  • Hand Import V2 (Apr 6): Downgraded to V1.1 — no native tracker formats needed, release current version and iterate.
  • Share infrastructure architecture (Apr 6): New service for short codes, preview images, landing pages. Starting viral features this month.
  • Slack (Apr 10): Gytis flagged AA-5323 (filter regression) and AA-5317 (iframe cache issue) to Shashank. Yevhienii (Apr 8) requested all Strategy Grid tasks be coordinated with him.
Week of Mar 24 (Pre-Q2 — Final Q1 Week)
  • v0.4.0 deployed to PROD (Sebastian, Mar 25): Range Explorer, Hand Import, Invite Codes, Revised Onboarding, AnM Vue
  • GTO Solver 500 error on PROD (Gytis, Mar 25) — needs investigation
  • Rank-up loop connected and functional (AceCoach demo Mar 24)
  • Simplified homepage in progress, deploying to dev this week
  • PLO4 display added to analysis mode but PLO data still blocked
  • Leak review integrated with real CAI API data
  • Duplicate hand import prevention shipped
  • Gold emergency hotfix: fullscreen layout broke Analysis Mode (back button missing), fixed within 1 hour
  • Nick: Leagues prototype (5 screens) + Ecology Report prototype shipped via text-to-figma v4.2
  • UX W13: Mariana takes AA-4712, Personalized Drills added to backlog
  • Intermediary screen spec created (Nick)
  • DS NRT engine shipped: 14 functional + 10 non-functional features. Deploying to staging Mar 27. Global + Gold on staging.
  • gto-glue Rust rewrite live: 60 pods → 3 pods, 75% CPU savings, 92% memory savings (Gameplay AI)
  • CAI E3 (Rank-Up Loop) at 100% — Q1 milestone
  • CAI E4 (Instant Coach) backend done — FE integration + Kafka deps still blocked
  • QuintAce NRT staging was blocked 3 weeks on env availability — deploying today
  • League Progress spec v0 created (298 lines, Nick/Svilen) — OQ#1 pending
  • Brad Wilson AMA live + article published
  • 100 opponent profiles in PR #132 (Scott reviewing)
  • Anti-cheat: login-device research found 4 unused high-signal features (carrier 92% ban rate, Build ID 53%)
Week of Mar 22 (Pre-Q2 Setup)
  • Q2 planning complete: Team Charters, Q2 Plan, Q2 Tracker created
  • Q1 files archived to planning/q1/
  • 14 themes defined (Arena 6, Coach 8) with target experiences and success gates
  • Cross-team interface contracts documented
  • Self-sufficiency notes for all 8 cross-team features
  • Experience Gates Dashboard established as primary health check

v0 · last updated 2026-04-30 · auto-generated from Q2_Tactical_Execution_vs_Planning.md