Six models · five daily puzzles · free · no account
Take everything public about an athlete — the box score, the tracking data, the draft slot, the salary, the awards nobody remembers — and squash it down into a single list of numbers. That list is the whole player, as far as the model is concerned.
Then we hide the names and hand you the numbers. The model is wrong all the time — that's the fun part.
Every number below is read from the same JSON the games use. No copy-paste guesses. Six short loops, each one true line that the model actually ships. Watch in order — each builds on the last.
20,719×64-d joint = 12,966 / 5,323 / 2,430
5 dailies live — NBA · NFL · World Cup · Companies · Chimera; tennis ships as model card + playable probe, no daily yet (honest caveat). Then it makes sense: one map, 5 leagues, same engine.
v̂=v/‖v‖₂·cos=v̂·ŵ
live from assets/data/*.json — hoops 12,966 64-d, gridiron 646 projected 2026 (5,323 joint), pitch 2,430 24-d, equities 500 (4,831 company-years), tennis 4,022 probe. Per-sport encoders 64 / 32 / 24 feeding a shared 64 trunk. All vectors L2 unit — cosine is dot. That is why a power forward can sit next to a strong safety: same role geometry.
MTNN · ONNX 549KB · 17 towers hoops
hoops 130 feats → 17 towers → 64-d L2 (10 source files), gridiron 32-d MTNN 13 families 7 files, pitch 24-d MTNN 3 files, equities 64-d 17 towers 7 files (Altman Z, Piotroski F, Beneish M, Sloan, QMJ), tennis 14 files ridge +0.0941 over rank alone, unified 12 files CORAL+contrastive+GRL, scout_cli 6. Each sport owns its towers; joint learns CORAL + contrastive + adversarial GRL to keep role. ONNX ships where measured. If a gate fails, deploy fails.
(seed*1103515245+12345)&0x7fffffff
glibc rand LCG, seed = YYYYMMDD UTC. Example today 20260809→70737614→idx2948/20719. same-link-same-stars ?daily=YYYYMMDD&n=1/3/5. Exposed via window.DAILY_SEED + UNIFIED_CHIMERA_DAILY + console [hub-daily]. deterministic, reproducible, no server dice.
hoops 10 · gridiron 7 · pitch 3 · equities 7 · tennis 14 · unified 12 · scout_cli 6
source_hashes honest list, DM_PROVENANCE shipped. Sport recoverable Δ+0.0593 (G2 acc 0.6851 vs majority 0.6258) — joint still weak sport-blind, we say so. verifyProvenance() auto-runs DOMContentLoaded + 8s idle. If a hash mismatches, UI shows honest FAIL.
PWA v67 offline.html 13k dark #080A0F — shell cached, never big vectors. Week Warrior 7-dot streak localStorage hub-streak hub-best, countdown UTC midnight, toast aria-live, Pack Battle 1/3/5 share haptics vibrate(10) graceful. Copy daily link → clip + confetti #D8452A. Same-link-same-stars means your friend gets your exact board.
All numbers read true from assets/data/hoops.json 12,966 / gridiron.json 646 (5,323 joint) / pitch.json 2,430 / equities.json 500·4,831y / tennis.json 4,022 / unified.json 20,719 / scout_cli.json 8 and assets/models/unified.json 20,719×64-d LCG (seed*1103515245+12345)&0x7fffffff. If a file changes, counts update below.
Same 5 games, filtered by what you actually do. All = no filter. Owner = cap tools / win+valuation. Player = stay on floor / fit. Brand = wins→story / $B. DFS = optimizer / locks. Same engine, 5 sports. Global entry — hoops is canonical, cross-link daily.
One puzzle a day per game. Each runs a real model on real public data — no staged scores, no decorative math. Three sports + one market + one joint cross-sport chimera. One more model — tennis — still ships as a model card with a playable probe and no daily puzzle yet. Unified is now live as the fifth daily: 20,719 player-seasons in 64-d, deterministic dailySeed LCG.
Two real NBA seasons got fused into one impossible player. Name both.
12,966 player-seasons, each one a point in a 64-dimensional space the model learned on its own. Guesses are scored by how close you land — so a wrong answer that feels right usually is.
Who should you actually start this week?
One net predicts a player's fantasy points and his stat line at the same time. Learning both makes it better at each — and the 32-dimensional trunk it builds along the way doubles as the map.
The same trick, played on the World Cup.
Built from StatsBomb's open event data for 2018 and 2022, normalized inside each tournament so a 2018 workhorse isn't judged against 2022 averages. The map is a 24-d MTNN over 2,430 player-rows across 11 tournament and league contexts.
Public companies as vectors — SEC EDGAR XBRL + market tape.
4,831 company-years — 500 tickers across 2015-2024 — as 64-d vectors from SEC EDGAR XBRL. 17 towers: Altman Z distress, Piotroski F quality, Beneish M, Sloan accruals, QMJ, plus market microstructure. One net predicts sector, next-year profile, distress and payout at once. Guess the ticker from its vector.
Daily Chimera — 20,719×64-d joint, dailySeed LCG deterministic, provenance 7/7 honest drift Procrustes glass-box
Three encoders (64-d hoops / 32-d gridiron / 24-d pitch) folded into 64-d with CORAL + contrastive + adversarial GRL. Guess the archetype from a cross-sport neighbour — sport-blind probe, deterministic daily seed — year*10000+month*100+day LCG. Vector Unified is the model, Daily Chimera is today's puzzle — 20,719×64-d native 12,966 / 5,323 / 2,430, archetypes 12 (6 live + 6 deferred), silhouette 0.683, cross-sport NN same-arch 0.9828.
One card per model. Real measured findings, the model’s own limitations quoted from its artifacts rather than paraphrased, and Split Decision — a game that shows you two real entities and asks which one the model rates higher. The answer key is the model’s own numbers, so you are probing the model, not guessing at a future it never predicted.
The joint 64-d model ships from unified_stage2_best.pt (best_epoch 58, 60ep, enc_lr 3e-5, GRL λ 0.10). It was gated on four tests. Every gate was measured against a null that could have failed it. Below are the shipped numbers and what happens when you drop one alignment loss — house rule from SPEC.md §5: drop each loss and measure Δ, each must earn its keep.
scout-cli powers this
Training now runs through scout-cli v0.8 — scout vector train --game hoops, scout vector eval --game unified --gates G1-G4, scout unified ablation --ablate contrastive|coral|adversarial. Router picks 5 tiers: deterministic cheap (no LLM, reads mtnn_meta.json), LLM medium (3-5 sub-swarm via CommsBus), or deep_research heavy 9K (13-swarm checkpointed). Checkpoint Manager writes bundles/ultra/runs/<runId>/checkpoint.json with fields nodeId/agentId/attempt/latency/tokens/status/errorClass so a 60-epoch unified job can pause days and resume. Verification econ budget3 threshold8.0 early-exit delta<0.3 catches mask-as-index pos_drop 0.0 via shuffled null 0.5493.
6
Model cards · 5 with a daily game
21,819
Hoops + Equities + Tennis rows
64/32/24/64/64
Dims · NBA / NFL / WC / EQ / joint
Free
No account · no ads
A player isn't one kind of thing. He's a body, a shot chart, a contract, a draft night, a postseason. Most models pick one of those and throw the rest away.
Vector Hoops runs eighteen separate towers — one each for volume, playmaking, rebounding, defense, efficiency, shot mix, biometrics, tracking, form, market value, roster context, career arc, strength of schedule, team, draft pedigree, playoffs, and honors — then fuses them into a single 48-number embedding. Equities runs 17 towers — Altman Z distress, Piotroski F quality, Beneish M earnings manipulation, Sloan accrual quality, QMJ quality, investment, payout, growth, operational efficiency, market microstructure, and more — into a 64-d company vector.
That embedding gets graded on many jobs at once: cluster the archetypes, name the position/sector, rebuild the box score / financial profile, guess the salary, predict who rises in the playoffs, predict distress, payout, next-year stats. A vector that can do all of that simultaneously has nowhere left to hide a lie. That's the MTNN — a multi-tower, multi-task net.
Vector Gridiron runs the same idea with a shared trunk and multiple heads. Vector Pitch now ships a true MTNN (24-d) that beats the old PCA baseline on role recovery; Equities ships a 64-d MTNN (17 towers) trained with CQS_v2 = 0.4*recall_no_wiki + 0.25*purity + 0.2*next_R2 + 0.15*sector.
Where this is going
The per-sport games still live in their own spaces (64-d hoops, 32-d gridiron, 24-d pitch, 64-d equities). On top of that, a 64-d joint embedding now folds 20,719 player-seasons across the three sports into one shared role geometry — so you can ask what a power forward and a strong safety have in common and get a real neighbour.
Shipped with caveats: sport identity is still partly recoverable from the joint vector (Δ +0.0593 over majority), but the joint daily puzzle is now live — 20,719 × 64-d chimera, dailySeed LCG deterministic, probe daily on the unified model card. Equities is its own market puzzle — guess the ticker from the vector. The five games are the product you play.
Every number on every game is recomputable from public source data — an accuracy harness gates every deploy.
Back to the roster →