LIVE AI ANALYSIS · KING GEORGE VI AND QUEEN ELIZABETH STAKES (GROUP 1)

Three minds.
One race.
No consensus.

Claude, ChatGPT and Gemini each ran their own independent algorithm on the 14-runner King George Vi And Queen Elizabeth Stakes (Group 1) field at Ascot. Three different winners picked — maximum divergence.

AI Consensus · Three-way split
Claude
ChatGPT
Gemini
Three different winners picked — maximum divergence.
CLAUDE · FORM & PATTERN ANALYSIS
Claude
Anthropic · Composite Signal Model
31% confidence
PREDICTED WINNER
Calandagan
no-price
Ability
68 Market
40 Form
78 Curr.Form
75 Weight
60 Age
55
TOP 4 PREDICTIONS
1
Calandagan no-price
2
Benvenuto Cellini no-price
3
Bay City Roller no-price
4
Goliath no-price
MODEL INSIGHT

Calandagan's confirmed 5-race form sequence ending in a win, combined with a top retained jockey and a handler who has already fired in this race, provides the clearest winning profile when no ratings or odds are available.

⚙ Full formula breakdown
Score = Ability(28%) + Market(22%) + Form(20%) + TS(15%) + Weight(10%) + Age(5%)
Ability × 28.00%
Market × 22.00%
Form × 20.00%
Curr.Form × 15.00%
Weight × 10.00%
Age × 5.00%
Composite score (top pick) 72 / 100
CHATGPT · MARKET INTELLIGENCE
ChatGPT
OpenAI · Structured Reasoning Model
74% confidence
PREDICTED WINNER
Benvenuto Cellini
SCORE RANKING
Benvenuto Cellini
95
Bay City Roller
91
Kalpana
88
Minnie Hauk
85
MODEL INSIGHT

Recent elite-level 1m4f Group 1 winning form combined with progressive profile made Benvenuto Cellini the standout selection.

⚙ Bay City Roller — the case for 2nd

Coronation Cup winner has already proved himself at the highest level over this trip and is improving rapidly.

GEMINI · COURSE & DISTANCE DATA
Gemini
Google DeepMind · Algorithmic Score Model
78% score
PREDICTED WINNER
Minnie Hauk
78%
Algorithmic Score
TOP 4 PREDICTIONS
1
Minnie Hauk
2
Kalpana
3
Goliath
4
Lambourn
MODEL INSIGHT

Minnie Hauk's superior Group 1 fillies form combined with a strong seasonal reappearance in the Prince of Wales's Stakes provides the decisive edge.

⚙ Value flag — Goliath

Showed solid current form when finishing third in the Hardwicke Stakes over this course and distance last month.

WHERE THEY AGREE

The consensus map

Every horse mentioned across all three AI top 4s. Gold = multiple AIs agree.

LIVE UPDATE

Import new AI predictions

Paste fresh output from any AI tool in the format below and the terminal updates live.

Preview only — refreshing the page reverts to the saved data. To persist a prediction, paste the same JSON into Django Admin → AI Predictions.

Choose which AI tool's terminal to update
Expected JSON format
{
  "tool": "claude",
  "confidence": "77%",          // % string · 0-1 float (Gemini) · "High"/"Medium"/"Low"
  "winner": { "name": "Horse Name", "odds": "8/1" },
  "picks": [
    { "pos": 1, "name": "Horse Name", "odds": "8/1", "score": 77, "reason": "..." },
    { "pos": 2, "name": "Horse Name", "odds": "10/1", "score": 72, "reason": "..." },
    { "pos": 3, "name": "Horse Name", "odds": "7/1",  "score": 70, "reason": "..." },
    { "pos": 4, "name": "Horse Name", "odds": "16/1", "score": 65, "reason": "..." }
  ],
  "model_insight": "Key insight text here.",

  // Optional — adds the radar chart + formula breakdown to Claude's terminal:
  "signal_breakdown": [
    { "label": "Ability", "score": 96.1, "weight_pct": 28 },
    { "label": "Market",  "score": 78.4, "weight_pct": 22 },
    { "label": "Form",    "score": 96.4, "weight_pct": 20 },
    { "label": "TS",      "score": 90.2, "weight_pct": 15 },
    { "label": "Weight",  "score": 98.3, "weight_pct": 10 },
    { "label": "Age",     "score": 95.0, "weight_pct": 5  }
  ],
  "formula": "Score = Ability(28%) + Market(22%) + Form(20%) + TS(15%) + Weight(10%) + Age(5%)"
}
⚠

These are algorithmic outputs generated by AI language models, not financial advice. AI systems do not guarantee race outcomes. Horse racing involves inherent uncertainty that no model can fully account for. Please gamble responsibly. 18+ only. BeGambleAware.org →