Cross-Market Pricer#

Normalize and compare odds across Polymarket, Kalshi, and traditional sportsbooks.

When to Use#

Use this skill when the user asks about:

• Comparing odds across different platforms (sportsbooks vs prediction markets)
• Normalizing prices to a common format
• Side-by-side pricing for the same event on Polymarket, Kalshi, and a sportsbook
• Whether there's a pricing gap or mispricing between platforms
• Converting between American odds, implied probability, and contract prices

Conversion Reference#

American to probability: negative → |odds|/(|odds|+100), positive → 100/(odds+100) Polymarket to probability: price is already probability Kalshi to probability: contract_price (already 0.00–1.00 from API) Probability to American: >0.5 → -(prob/(1-prob))*100, <0.5 → +((1-prob)/prob)*100

Operations#

1. Fetch Sportsbook Odds as Normalized Probability#

Pull American odds from The Odds API and convert to implied probability inline:

bash

curl -s "https://api.the-odds-api.com/v4/sports/SPORT_KEY/odds?apiKey=$ODDS_API_KEY&regions=us&markets=h2h&oddsFormat=american" \
  | jq '[.[] | {
    event: "\(.away_team) vs \(.home_team)",
    start: .commence_time,
    source: "sportsbooks",
    outcomes: [.bookmakers[0].markets[0].outcomes[] | {
      name: .name,
      american_odds: .price,
      implied_prob: (if .price < 0 then (-((.price)) / (-((.price)) + 100)) else (100 / (.price + 100)) end) | (. * 1000 | round / 1000)
    }]
  }]'

Replace SPORT_KEY with the appropriate key (e.g., basketball_nba, americanfootball_nfl).

2. Fetch Polymarket Prices as Normalized Probability#

Pull market prices from Polymarket's Gamma API:

bash

curl -s "https://gamma-api.polymarket.com/markets?closed=false&limit=10&tag=CATEGORY" \
  | jq '[.[] | {
    event: .question,
    source: "polymarket",
    outcomes: [{
      name: "Yes",
      polymarket_price: (.outcomePrices | fromjson | .[0] | tonumber | (. * 1000 | round / 1000)),
      implied_prob: (.outcomePrices | fromjson | .[0] | tonumber | (. * 1000 | round / 1000))
    }, {
      name: "No",
      polymarket_price: (.outcomePrices | fromjson | .[1] | tonumber | (. * 1000 | round / 1000)),
      implied_prob: (.outcomePrices | fromjson | .[1] | tonumber | (. * 1000 | round / 1000))
    }],
    volume: .volume,
    liquidity: .liquidity
  }]'

Replace CATEGORY with the relevant tag (e.g., "sports", "nba", "politics", "crypto").

3. Fetch Kalshi Contract Prices as Normalized Probability#

Pull contract prices from Kalshi's public market data API:

bash

curl -s "https://api.elections.kalshi.com/trade-api/v2/markets?status=open&limit=10&series_ticker=SERIES_TICKER" \
  | jq '[.markets[] | {
    event: .title,
    source: "kalshi",
    ticker: .ticker,
    outcomes: [{
      name: "Yes",
      kalshi_price: (.yes_ask / 100),
      implied_prob: (.yes_ask / 100)
    }, {
      name: "No",
      kalshi_price: (.no_ask / 100),
      implied_prob: (.no_ask / 100)
    }],
    volume: .volume
  }]'

Replace SERIES_TICKER with the Kalshi series ticker for the event category.

4. Unified Cross-Market Comparison#

When the user wants to compare the SAME event across platforms, fetch from each source and use Python to produce a unified table:

bash

python3 -c "
import json, subprocess

def american_to_prob(odds):
    if odds < 0:
        return round(abs(odds) / (abs(odds) + 100), 4)
    else:
        return round(100 / (odds + 100), 4)

def prob_to_american(prob):
    if prob >= 1: return -99999
    if prob <= 0: return 99999
    if prob > 0.5:
        return round(-(prob / (1 - prob)) * 100)
    elif prob < 0.5:
        return round(((1 - prob) / prob) * 100)
    else:
        return 100

# Input: user provides these values (agent fills from context)
platforms = {
    'PLATFORM_1_NAME': {'yes_prob': PLATFORM_1_YES_PROB, 'no_prob': PLATFORM_1_NO_PROB},
    'PLATFORM_2_NAME': {'yes_prob': PLATFORM_2_YES_PROB, 'no_prob': PLATFORM_2_NO_PROB},
}

print('=== Cross-Market Comparison ===')
print(f\"{'Platform':<18} {'Yes Prob':>10} {'Yes Amer':>10} {'No Prob':>10} {'No Amer':>10} {'Total':>8}\")
print('-' * 70)
for name, data in platforms.items():
    y = data['yes_prob']
    n = data['no_prob']
    total = round(y + n, 4)
    y_am = prob_to_american(y)
    n_am = prob_to_american(n)
    y_am_s = f'+{y_am}' if y_am > 0 else str(y_am)
    n_am_s = f'+{n_am}' if n_am > 0 else str(n_am)
    print(f'{name:<18} {y:>10.1%} {y_am_s:>10} {n:>10.1%} {n_am_s:>10} {total:>8.1%}')

# Flag mispricings
probs = [d['yes_prob'] for d in platforms.values()]
gap = max(probs) - min(probs)
if gap > 0.03:
    print(f'\n⚠ MISPRICING: {gap:.1%} gap across platforms on Yes side')
else:
    print(f'\nPricing aligned within {gap:.1%} — no actionable gap')
"

The agent must fill in PLATFORM_1_NAME, PLATFORM_1_YES_PROB, etc. from the results of Operations 1–3. If an event only exists on two platforms, remove the third entry.

5. Quick Single-Value Conversion#

For ad-hoc conversions when the user gives a single odds value:

bash

python3 -c "
odds_input = 'INPUT_VALUE'
# Detect format and normalize
if odds_input.startswith('+') or odds_input.startswith('-'):
    odds = int(odds_input)
    prob = abs(odds)/(abs(odds)+100) if odds < 0 else 100/(odds+100)
    fmt = 'American'
elif '.' in odds_input and float(odds_input) <= 1:
    prob = float(odds_input)
    fmt = 'Probability'
elif odds_input.startswith('\$'):
    prob = float(odds_input.replace('\$',''))
    fmt = 'Kalshi contract'
else:
    prob = float(odds_input)
    fmt = 'Decimal probability' if float(odds_input) <= 1 else 'Unknown'

# Convert to all formats
if prob > 0.5:
    american = round(-(prob/(1-prob))*100)
elif prob < 0.5:
    american = round(((1-prob)/prob)*100)
else:
    american = 100

am_str = f'+{american}' if american > 0 else str(american)
print(f'Input: {odds_input} ({fmt})')
print(f'Implied probability: {prob:.1%}')
print(f'American odds: {am_str}')
print(f'Kalshi contract: \${prob:.2f}')
print(f'Polymarket share: {prob:.4f}')
"

Replace INPUT_VALUE with the user's value (e.g., "-150", "0.62", "$0.58").

Output Rules#

1. Always show implied probability as the primary format (3 decimal places, e.g., 0.620)
2. Include American odds equivalent alongside probability for sportsbook-familiar users
3. When comparing platforms, display as a table with columns: Platform, Yes Prob, Yes American, No Prob, No American, Total
4. The "Total" column (Yes + No probabilities) reveals vig — anything over 100% is overround
5. Flag gaps greater than 3% between platforms as potential mispricings
6. Always note which sportsbook was used as the sportsbook source (since books differ)
7. Report The Odds API quota after any sportsbook fetch

Error Handling#

• If ODDS_API_KEY is not set, tell the user to get a free key at https://the-odds-api.com/
• If Polymarket returns empty results for a category, try broader search terms or list available categories
• If Kalshi returns empty results, the event may not have an active contract — suggest checking kalshi.com directly
• If an event exists on one platform but not another, report what's available and note the gap
• If rate limited on any API, report the limitation and suggest which platforms can still be queried

About#

Built by AgentBets — full tutorial at agentbets.ai/guides/openclaw-cross-market-pricer-skill/.

Part of the OpenClaw Skills series for the Agent Betting Stack.