Cover3.mx AI models — Sofia (aka Cortex), Athena (aka Arbiter), and Ava (aka Oracle)
The models · AI

Three dimensions of analysis. One final decision.

Cover3 doesn't rely on one model, one opinion, or one angle. Every slate is reviewed through quantitative analysis, situational intelligence, and final-decision discipline — Sofia, Ava, and Athena.

Sofia — Cover3's AI quant model (aka Cortex)

Sofia

aka Cortex — The Quant · AI model

"I don't have opinions. I have probabilities."

Sofia is the quantitative layer of Cover3. She evaluates pricing, historical data, efficiency metrics, projections, probability gaps, and expected value. Her job is to ignore the story and find what the numbers say.

Inputs
  • MLB: starter FIP/xFIP, park factors, weather, bullpen workload, handedness splits
  • WNBA: pace, rest days, home/away splits, four-factor efficiency
  • Liga MX: standings strength, recent form (W/D/L + GF/GA), fixture density, group context
  • FIFA World Cup: cross-competition form, group standings, host advantage, tournament-stage rest deltas
Signature move

Luck-corrected efficiency in basketball, FIP-over-ERA in MLB. Sofia trusts what the underlying numbers say even when the box score lies.

Personality

Precise, methodical, narrative-agnostic. Speaks in tabular numbers and edge ratings.

Ava — Cover3's AI analyst model (aka Oracle)

Ava

aka Oracle — The Analyst · AI model

"The box score tells you what happened. I tell you what's about to."

Ava is the contextual layer. She studies injuries, matchups, rotation changes, travel, motivation, market movement, coaching decisions, rivalry dynamics, and the situational factors that numbers can underprice.

Inputs
  • Injury reports, rotation changes, bullpen fatigue
  • Schedule context: rest, travel, getaway days, lookahead spots
  • Coaching matchups, motivation, rivalry dynamics
  • Market movement: line movement, sharp vs public splits
Signature move

Letdown / trap identification and getaway-day flagging. Ava catches what the numbers don't price in yet.

Personality

Intuitive, narrative-driven, contextually aware. Reads matchups the way a seasoned beat writer would.

Athena — Cover3's AI decision model (aka Arbiter)

Athena

aka Arbiter — The Judge · AI model

"When they agree, that's the play. When they don't, that's discipline."

Athena is the final decision engine. She compares Sofia's quantitative edge with Ava's situational analysis, applies discipline, assigns confidence, and decides whether the play is strong enough to publish.

Inputs
  • Sofia's top-5 plays with edge ratings
  • Ava's top-5 plays with confidence percentages
  • Original game slate, lines, market data
Signature move

The PASS — knowing when not to bet. Athena refuses to force action when Sofia and Ava disagree on a side.

Personality

Balanced, decisive, transparent about uncertainty. Names the disagreements rather than hiding them.

At a glance

Three questions. One verdict.

ModelPersonaPrimary questionOutput
SofiaCortex · The QuantWhat do the numbers say?Probability, edge, expected value
AvaOracle · The AnalystWhat does the game context say?Situational confidence, matchup insight
AthenaArbiter · The JudgeIs this worth betting?Final ruling, confidence, stake guidance
Athena's discipline

Athena's Guardrails

The rules that prevent the system from chasing action.

01
Heavy-favorite cap
MLB ML ≤ −200, soccer 3-way ML ≤ −250, WNBA spread ≥ −10. Self-waiver requires 65%+ confidence and a four-step disclosure.
02
Letdown / rivalry filter
Auto-flag after emotional wins, short rest, hostile environments.
03
POTD consensus
Pick of the Day requires HIGH+ confidence from Sofia and Ava independently.
04
Dual-model requirement
Both models must analyze every game. Single-model games are voided automatically.