World Cup RNG — Decision Log v2, complete, June 8 to July 10 2026

2026-07-15  ·  world-cupdecisionloggame-design

Coverage: full #world-cup export June 8 → July 10, huddle canvases, #velocity, design sessions, analytics CSVs. Compression ~16K words → this log (~5%). Format: date · decision · rationale · status. Entries D1-D31 and DL1 from v1 carry forward unchanged unless noted; new entries are E-numbered.

Era 1 — Origin and v1 (Jun 8-13)

# Date Decision Rationale / evidence Status
E1 Jun 8 Project kickoff: football RNG/tycoon targeting the World Cup window, Arda joins as game designer Tournament Jun 11 - Jul 19 as the launch wave Done
E2 Jun 8 Cards ARE the models, no 3D footballers at launch Cuts animation cost (shooting, dribbling), makes the Jun 11 target plausible (Yel + Mike) Shipped, reversed Jul 1 (E31)
E3 Jun 8 Wall vs formation debate → hybrid: formation rows, each row slightly elevated Arda: pure wall kills strategy (positions, performance score become meaningless). Yel: pure formation kills visibility. Elevation gets both Shipped
E4 Jun 8 Real player names + likenesses, risk accepted stefan raised DMCA. Yel: comp game at 20K CCU, 160M visits, 9 months using names + photos, "I'm assuming the risk." Fallback floated: Roblox-adapted figures (Mike) Standing. Risk scales with success, fallback still unwritten
E5 Jun 9 Layered card architecture: back → portrait → front → runtime info. 3D card model with depth (2 materials, swap portrait textures) Same player at any tier, live boosts change numbers, parallax feel Shipped
E6 Jun 9 Kill level-ups, adopt dupe-XP (duplicates convert to XP automatically) No dead pulls (Yel proposed, team agreed) Shipped
E7 Jun 10 Tier ladder expanded to 11 (Common → GOAT incl. Mythical, All-Star, Icon, HoF, Generational) Arda: thin the low tiers, stretch the chase Shipped
E8 Jun 10-11 Pack rip-open reveal modeled on Plants vs Brainrots short, meshes over bones Simpler than rigging (Danyil, Martin). "Getting the opening sequence right is the most crucial" (Yel) Shipped
E9 Jun 13 v1 published. First funnel: 19% drop at step 2. Target stated: 85% conversion First data day Superseded next day

Era 2 — The Pivot (Jun 14-16), the log's strongest entry

# Date Decision Rationale / evidence Status
E10 Jun 14 v1 killed after one day of data. "Too much friction with a lot of depth but not a great core loop" (Yel). Simplify and copy Build a Soccer Squad Funnel evidence + comp proof. Kill-to-new-place in under 24h Executed
E11 Jun 14 Core loop v2 locked: roll dice → choose 1 of 3 → fill 11 → team OVR → OVR converts to luck → reset → repeat Arda: 1-of-3 = sense of choice + position strategy (Danyil questioned, Arda's rationale held) Live, still the loop
E12 Jun 14 Cash currency deleted entirely. Upgrade tree explicitly rejected: "we don't need it" (Arda) Radical simplification. Note the genealogy: cash reborn Jun 24 (E22), tree added Jun 24, tree superseded by fans economy Jul 9-10 (D22) Chain complete
E13 Jun 15-16 Per-player roll weights, not tier percentages (Ronaldo 1/3,000, luck scales weights). Slime RNG pattern Rarity reads per-chase-target (Arda) Live
E14 Jun 16 Bench/collection removed for launch, base tightened, server size pushed 6 → 8-10 "Nothing much going on, more players is better" (Arda) Live
E15 Jun 16 Rebirth debated: Arda pro (cash-cost rebirth, tree persists, permanent multipliers), Yel "maybe no need" Parked deliberately, revisit when leaderboard players feel stuck Parked

Era 3 — Relaunch and the funnel war (Jun 17-23)

D1-D10 from v1 stand (clean place relaunch, 100K ads bypass, Sports genre, no like-gating, autoroll flip-flop, same-country first rolls, nation bonuses, custom events, plot-cap bug 46→50%+, feedback button). New:

# Date Decision Rationale / evidence Status
E16 Jun 21-23 Cheated rolls at positions 4, 8, 11. Roll-6 drop investigated. Camera cinematic limited to first 5 rolls. Roll speed raised, then judged too fast Iteration war on the sub-5-minute drop (27% before min 5) Ongoing
E17 Jun 22 "Roll till you fill your Team 7/11" progress counter Show the finish line, not endless rolling (Arda + Danyil) Shipped
E18 Jun 23 Hit Home Recommendations, ~145 CCU. Yel offers to "pump gas," Danyil declines: "we have a good tempo" Team chose iteration pace over an all-in sprint Noted
E19 Jun 23 Ads kept running alongside recos Mixed cohorts acknowledged as funnel noise (Danyil) Standing

Era 4 — The depth era (Jun 24-30)

# Date Decision Rationale / evidence Status
E20 Jun 24 Upgrade tree + cash + dice currencies added (cash per pick, golden dice ladder, mutations roadmap) Reversal of E12: players needed a sink and mid-session goals. D25's first-match churn later exposes the limits Superseded by fans economy
E21 Jun 24 Slow reveal VFX only for first-time cards, fast for repeats "What's the anticipation when you know the rarity from the card back... punishing players with higher luck" (Yel), Danyil's fix Shipped, law candidate
E22 Jun 24 Mutations can push OVR above 99, base cards capped at 99 Chase excitement above the ceiling Shipped
E23 Jun 24-25 Moment cards born: Yel compiles top 30 WC moments → 25 tiered cards, "MOMENT" modification, biting-caption voice ("not goal. GOLAZO") The live wedge in card form Shipped, reworked as skins Jul 8 (D21)
E24 Jun 25 Elliptical stadium layout proposed (Danyil): plots inside a stadium ring, center as social space Origin of the city/arena design Adopted direction
E25 Jun 26 Live match fetching ships: results every 5 min, scoreboards, goal boosts + toasts across servers The differentiator goes real (delay accepted, "players never verify") Live
E26 Jun 26 Network optimization: client calls cut from ~150 kb/s to 5-6 kb/s Real perf find (Danyil) Done
E27 Jun 27-30 Match simulation born: commentary sim on PLAY (Yel, FM inspiration), 11v11 HTML (Arda), length fought from 5 min → 1 min → ~40s, cheated first matches added "The crescendo after building the team." Roguelike impact-card direction named (RAIN card makes opponent cards WET) Live
E28 Jun 30 3D player models to replace cards on plots, rarity read via auras Martin's city references + customization 3-theme direction (classic, modern, historic) In progress

Era 5 — Live, 3D, and scale (Jul 1-5)

# Date Decision Rationale / evidence Status
E29 Jul 1 Hair pipeline: 9-style taxonomy from 211 icons, creator-hub assets only, marketplace avoided for IP safety (Danyil) Quiet good call, keep as convention Standing
E30 Jul 1 Duel sync via shared random seed, both clients simulate the same match Arda's elegant answer to turn-based sync complexity Adopted
E31 Jul 2 Cash rework idea born (Yel): tie cash to matches via stadium visitors, tickets, sponsors. "Manager hat and chairman's hat" Origin of the fans economy (lands as D22 Jul 9-10). Sell Lemons assigned as study reference Became the economy
E32 Jul 2 Two-resource cap stated: luck and cash only, "a third dilutes importance" (Yel) F5 law applied in the wild Standing
E33 Jul 3 Betting billboard idea: stake luck on real match odds (Yel) ⚠ Compliance flag: simulated betting on real-world sports outcomes, kids audience. Not built. Needs a deliberate NO in this log before it resurfaces Flagged, undecided
E34 Jul 3-5 Home recos 80K → 150K impressions, 260 CCU, alex added to export the channel "to document process" The brain's origin moment inside the channel Noted

Era 6 — Economy rework and monetization week (Jul 6-10)

D17-D31 and DL1 from v1 stand (parallel duels, console, 2x cash pass, BASS-modeled dev products, moments-as-skins, fans economy, plot L1-3 + tree direction, PvE simplification, first-match churn 8-9%, icon rework, metric goals, 3-week roadmap, data-first, weekly expectations). New from the huddle canvas:

# Date Decision Rationale / evidence Status
E35 Jul 9 huddle Post-tournament plan decided (corrects v1's "loudest silence" flag): nostalgia best-moments week after the Jul 19 final, legend cards (Pelé, Maradona, Messi seasons) + WC meme content post-27th The cliff has a plan. v1's flag was a source-class miss: huddle canvases ≠ channel text Planned
E36 Jul 9 huddle Danyil leaves Jul 17 (back ~Saturday, exam Tue, gone Fri onward). Pre-planned updates for 14/16/21/24, Giorgi handoff before departure Lead dev absent across the final and cliff week, coverage explicit Plan active
E37 Jul 9 huddle Monetization mandate: daily small SKUs (5-10 robux), conversion target 0.47% green, spend/payer → 100 robux, avoid expensive SKUs initially, map the full option inventory before daily adds ⚠ Interacts with the founding "real money off random rolls" principle: OP paid rolls are paid randomness. Publish odds or route through earned currency, decide deliberately Active, flagged
E38 Jul 9 huddle "Never release features without clear success metrics, tracking before launch not after" (Yel). PVP funnel (adoption, retention, card usage) required before expanding social features Law candidate + the pre-registration ritual stated by the founder Adopt as ritual

Genealogies worth keeping whole

  1. Economy: cash exists (v1) → deleted Jun 14 → reborn as pick-cash + tree Jun 24 → superseded by match-only fans economy Jul 9-10 (Fanbase × Capacity × Ticket Price).
  2. The loop: conveyor-pack collection (v1) → killed on data Jun 14 → roll-pick-place (BASS shell) → + simulation Jun 27 → + duels Jul 6 → + club progression Jul 9.
  3. Moments: compiled cards Jun 24 → live-data boosts Jun 26 → skins on default cards Jul 8.
  4. Cards on plots: cards-as-models Jun 8 → 3D player models with auras Jul 1.

Law candidates extracted

L-a: Slow reveals are a reward for novelty, not a tax on luck (E21). L-b: One kill gate executed fast beats ten funnel patches (E10). L-c: Two resources max, a third dilutes (E32). L-d: Tracking before launch, not after (E38). DL1: monetization should help players beat their best XI.

Open compliance items

E33 betting billboard (recommend: deliberate no). E37 paid random rolls vs founding principle (recommend: published odds minimum, earned-currency preferred). E4 likeness fallback doc (recommend: write the brainrot-conversion runbook before the first C&D, not after).

Analytics corroboration (source CSVs, Jun 12 - Jul 9)

D1 retention ~5-7% during the ad-heavy first week → 9-10% → 10.5-12.9% after the funnel/relaunch work. Sessions climbed ~5x Jul 2-9. Average session time settled at 15-17 min. Two anomalies noted (per Yel): the Jun 15-17 D1 dip is discounted as tiny-n, and Jul 7-8 D1 should be read as lagged rather than real.