{"id":2880,"date":"2026-05-14T10:00:00","date_gmt":"2026-05-14T10:00:00","guid":{"rendered":"https:\/\/oddspapi.io\/blog\/?p=2880"},"modified":"2026-04-12T15:24:53","modified_gmt":"2026-04-12T15:24:53","slug":"consensus-odds-fair-odds-calculator-python","status":"publish","type":"post","link":"https:\/\/oddspapi.io\/blog\/consensus-odds-fair-odds-calculator-python\/","title":{"rendered":"Consensus Odds: How to Calculate Fair Odds from 350+ Bookmakers"},"content":{"rendered":"
What Are the True Odds? 350+ Bookmakers Have the Answer<\/h2>\n
Every bookmaker bakes a margin into their odds. That margin \u2014 the vig<\/strong> \u2014 means no single book shows you the “true” probability of an outcome. But what if you could strip the vig from 100+ bookmakers and average the result?<\/p>\n
That’s consensus odds<\/strong>: the market-implied fair price after removing each bookmaker’s margin and aggregating across every available source. It’s the closest thing to “true odds” that exists \u2014 and the more bookmakers you include, the more stable and accurate the estimate becomes.<\/p>\n
The problem? Most odds APIs give you 40-85 bookmakers. That’s a noisy sample. At 130+ books \u2014 including sharps like Pinnacle, Singbet, and SBOBet \u2014 you get a genuine market consensus that outliers can’t distort. This tutorial shows you how to calculate it in Python with free data from OddsPapi<\/strong>.<\/p>\n
Why Sample Size Matters for Consensus Odds<\/h2>\n
With 130+ bookmakers including sharp books, exchange prices, and regional soft books, the median converges to a stable fair price. Remove 10 books? The result barely changes. That’s the definition of robust.<\/p>\n
Before you can calculate consensus odds, you need to understand vig removal. Every bookmaker’s implied probabilities sum to more than 100% \u2014 the excess is their margin.<\/p>\n
Basic Conversion: Odds to Implied Probability<\/h3>\n
def implied_prob(decimal_odds):\n \"\"\"Convert decimal odds to implied probability.\"\"\"\n return 1 \/ decimal_odds\n\n# Pinnacle on Chelsea vs Man City (1X2):\n# Home: 3.60 Draw: 4.15 Away: 1.98\nprobs = [implied_prob(3.60), implied_prob(4.15), implied_prob(1.98)]\nprint(f\"Implied probs: {[f'{p:.4f}' for p in probs]}\")\nprint(f\"Sum: {sum(probs):.4f}\")\nprint(f\"Overround (vig): {(sum(probs) - 1) * 100:.2f}%\")\n# Implied probs: ['0.2778', '0.2410', '0.5051']\n# Sum: 1.0238\n# Overround (vig): 2.38%<\/code><\/pre>\n
Pinnacle’s implied probabilities sum to 102.38% \u2014 that extra 2.38% is the vig. For a soft book like 888sport (vig ~8%), the distortion is much larger. To find fair odds, we need to remove it.<\/p>\n
Multiplicative Method (The Default)<\/h3>\n
The simplest and most common approach: divide each implied probability by the total. Three lines of code, accurate for 95% of use cases.<\/p>\n
def remove_vig_multiplicative(probs):\n \"\"\"Remove vig by proportional normalization.\"\"\"\n total = sum(probs)\n return [p \/ total for p in probs]\n\nfair = remove_vig_multiplicative(probs)\nprint(f\"Fair probs: {[f'{p:.4f}' for p in fair]}\")\nprint(f\"Fair odds: {[f'{1\/p:.3f}' for p in fair]}\")\n# Fair probs: ['0.2713', '0.2354', '0.4933']\n# Fair odds: ['3.686', '4.249', '2.027']<\/code><\/pre>\n
Pinnacle’s 2.38% vig is now gone. The fair odds are slightly higher than the raw odds \u2014 that’s the margin that was baked in.<\/p>\n
Power Method (Optional Upgrade for Quants)<\/h3>\n
The multiplicative method distributes vig proportionally across all outcomes. The power method (sometimes called the Shin method) adjusts vig non-linearly \u2014 favourites lose less vig, longshots lose more. This is more theoretically sound because bookmakers tend to shade longshot prices more heavily.<\/p>\n
def remove_vig_power(probs):\n \"\"\"Remove vig using power method. More accurate for lopsided markets.\"\"\"\n lo, hi = 1.0, 100.0\n for _ in range(200): # binary search for exponent k\n k = (lo + hi) \/ 2\n if sum(p ** k for p in probs) > 1.0:\n lo = k\n else:\n hi = k\n k = (lo + hi) \/ 2\n fair = [p ** k for p in probs]\n s = sum(fair)\n return [p \/ s for p in fair]\n\nfair_power = remove_vig_power(probs)\nprint(f\"Power fair odds: {[f'{1\/p:.3f}' for p in fair_power]}\")\n# Power fair odds: ['3.708', '4.289', '2.011']<\/code><\/pre>\n
For Pinnacle’s tight 2.38% vig, the difference between methods is small (a few cents on the odds). For high-vig soft books (8-15%), the power method diverges more. Use multiplicative as your default \u2014 upgrade to power if you’re building a serious model.<\/p>\n
def get_fixtures(sport_id, days_ahead=7):\n \"\"\"Fetch upcoming fixtures with odds.\"\"\"\n today = datetime.utcnow().strftime(\"%Y-%m-%d\")\n end = (datetime.utcnow() + timedelta(days=days_ahead)).strftime(\"%Y-%m-%d\")\n fixtures = api_get(\"fixtures\", {\n \"sportId\": sport_id, \"from\": today, \"to\": end\n })\n now = datetime.utcnow()\n return [\n f for f in fixtures\n if f.get(\"hasOdds\")\n and datetime.fromisoformat(f[\"startTime\"].replace(\"Z\", \"\")) > now\n ]\n\nfixtures = get_fixtures(sport_id=10) # Soccer\nprint(f\"Fixtures with odds: {len(fixtures)}\")\nfor f in fixtures[:3]:\n print(f\" {f['participant1Name']} vs {f['participant2Name']}\")<\/code><\/pre>\n
Fetch ALL Bookmaker Odds<\/h3>\n
This is the key difference from other tutorials: we fetch every available bookmaker<\/strong> by omitting the bookmakers<\/code> parameter. No filtering, no cherry-picking \u2014 the full market.<\/p>\n
def fetch_all_odds(fixture_id, market_id=\"101\"):\n \"\"\"Fetch odds from ALL bookmakers for a fixture. Returns flat list of dicts.\"\"\"\n data = api_get(\"odds\", {\"fixtureId\": fixture_id})\n bo = data.get(\"bookmakerOdds\", {})\n\n rows = []\n for slug, bdata in bo.items():\n market = bdata.get(\"markets\", {}).get(market_id, {})\n outcomes = market.get(\"outcomes\", {})\n prices = {}\n for oid in OUTCOME_MAP:\n p0 = outcomes.get(oid, {}).get(\"players\", {}).get(\"0\", {})\n if isinstance(p0, dict) and p0.get(\"active\") and p0.get(\"price\"):\n prices[oid] = p0[\"price\"]\n if len(prices) == len(OUTCOME_MAP):\n vig = sum(1 \/ p for p in prices.values()) - 1\n if vig < 0.20: # filter broken feeds (>20% vig)\n rows.append({\"slug\": slug, \"prices\": prices, \"vig\": vig * 100})\n\n return rows\n\n# Example: Chelsea FC vs Manchester City\nodds = fetch_all_odds(\"id1000001761301147\")\nprint(f\"Bookmakers with valid 1X2 prices: {len(odds)}\")\n# Bookmakers with valid 1X2 prices: 115<\/code><\/pre>\n
115 bookmakers with active 1X2 prices on a single Premier League match. The vig ranges from 0.84% (Betfair Exchange) to 18%+ (regional soft books). That spread is exactly what makes consensus calculation powerful \u2014 you’re sampling the full market, not a curated subset.<\/p>\n
Step 3: Four Ways to Calculate Consensus Odds<\/h2>\n
Method A: Simple Median<\/h3>\n
The median is the most robust starting point. It’s resistant to outliers (broken feeds, stale prices) and requires zero configuration. Always aggregate in probability space<\/strong>, not odds space \u2014 decimal odds are non-linear.<\/p>\n
def consensus_median(odds_rows):\n \"\"\"Median of implied probabilities across all bookmakers.\"\"\"\n probs_by_outcome = {oid: [] for oid in OUTCOME_MAP}\n\n for row in odds_rows:\n for oid, price in row[\"prices\"].items():\n probs_by_outcome[oid].append(1 \/ price)\n\n raw = {oid: statistics.median(ps) for oid, ps in probs_by_outcome.items()}\n total = sum(raw.values())\n return {oid: p \/ total for oid, p in raw.items()}\n\nfair = consensus_median(odds)\nfor oid, label in OUTCOME_MAP.items():\n print(f\" {label}: {fair[oid]:.4f} (fair odds: {1\/fair[oid]:.3f})\")\n# Home: 0.2680 (fair odds: 3.731)\n# Draw: 0.2440 (fair odds: 4.099)\n# Away: 0.4880 (fair odds: 2.049)<\/code><\/pre>\n
Method B: Sharp-Weighted Mean<\/h3>\n
Not all bookmakers are created equal. Pinnacle, Singbet, and SBOBet are “sharp” \u2014 they accept the highest limits from professional bettors, so their prices reflect the most informed money in the market. Weighting sharps 3x gives them outsized influence without ignoring the soft book signal entirely.<\/p>\n
def consensus_sharp_weighted(odds_rows, sharp_weight=3, soft_weight=1):\n \"\"\"Weighted mean: sharp books count more.\"\"\"\n weighted = {oid: [] for oid in OUTCOME_MAP}\n\n for row in odds_rows:\n w = sharp_weight if row[\"slug\"] in SHARP_BOOKS else soft_weight\n for oid, price in row[\"prices\"].items():\n weighted[oid].extend([1 \/ price] * w)\n\n raw = {oid: statistics.mean(ps) for oid, ps in weighted.items()}\n total = sum(raw.values())\n return {oid: p \/ total for oid, p in raw.items()}\n\nfair = consensus_sharp_weighted(odds)\nfor oid, label in OUTCOME_MAP.items():\n print(f\" {label}: {fair[oid]:.4f} (fair odds: {1\/fair[oid]:.3f})\")\n# Home: 0.2741 (fair odds: 3.649)\n# Draw: 0.2443 (fair odds: 4.093)\n# Away: 0.4816 (fair odds: 2.076)<\/code><\/pre>\n
Notice the sharp-weighted result pulls slightly toward Pinnacle’s prices (3.60 \/ 4.15 \/ 1.98) compared to the raw median. That’s the intended effect \u2014 sharps are better price-setters than soft books.<\/p>\n
The cleanest approach: strip the vig from each bookmaker individually<\/strong>, then take the median of the resulting fair probabilities. This removes the distortion that high-vig soft books inject into the consensus.<\/p>\n
def consensus_vig_removed(odds_rows):\n \"\"\"Remove each book's vig first, then take median of fair probs.\"\"\"\n fair_probs = {oid: [] for oid in OUTCOME_MAP}\n\n for row in odds_rows:\n implied = [1 \/ row[\"prices\"][oid] for oid in OUTCOME_MAP]\n fair = remove_vig_multiplicative(implied)\n for i, oid in enumerate(OUTCOME_MAP):\n fair_probs[oid].append(fair[i])\n\n raw = {oid: statistics.median(ps) for oid, ps in fair_probs.items()}\n total = sum(raw.values())\n return {oid: p \/ total for oid, p in raw.items()}\n\nfair = consensus_vig_removed(odds)\nfor oid, label in OUTCOME_MAP.items():\n print(f\" {label}: {fair[oid]:.4f} (fair odds: {1\/fair[oid]:.3f})\")\n# Home: 0.2677 (fair odds: 3.734)\n# Draw: 0.2448 (fair odds: 4.084)\n# Away: 0.4872 (fair odds: 2.052)<\/code><\/pre>\n
This is the recommended method for serious work. By devigging each book before aggregating, you’re comparing apples to apples \u2014 115 independent estimates of the true probability, all free of margin distortion.<\/p>\n
Method D: Exchange-Anchored<\/h3>\n
Exchanges (Betfair) charge commission on winnings, not vig on odds. Their back price is close to the true price \u2014 especially for liquid markets. Use it as a ground truth benchmark.<\/p>\n
def consensus_exchange_anchored(odds_rows, exchange_slug=\"betfair-ex\"):\n \"\"\"Use exchange back price as fair odds (minimal vig).\"\"\"\n for row in odds_rows:\n if row[\"slug\"] == exchange_slug:\n implied = [1 \/ row[\"prices\"][oid] for oid in OUTCOME_MAP]\n fair = remove_vig_multiplicative(implied)\n return {oid: fair[i] for i, oid in enumerate(OUTCOME_MAP)}\n return None\n\nfair = consensus_exchange_anchored(odds)\nif fair:\n for oid, label in OUTCOME_MAP.items():\n print(f\" {label}: {fair[oid]:.4f} (fair odds: {1\/fair[oid]:.3f})\")\n# Home: 0.2654 (fair odds: 3.768)\n# Draw: 0.2370 (fair odds: 4.220)\n# Away: 0.4976 (fair odds: 2.010)<\/code><\/pre>\n
Caveat:<\/strong> Betfair liquidity varies wildly by market. For Premier League 1X2, it’s deep \u2014 the exchange price is reliable. For lower-league soccer or niche sports, the vig-removed consensus from 130+ books is often more accurate than a single thinly-traded exchange price.<\/p>\n
Step 4: Practical Applications<\/h2>\n
Find Value Bets<\/h3>\n
Compare any bookmaker’s odds to the consensus fair price. If the bookmaker is offering more than the true price, that’s a value bet.<\/p>\n
Rank bookmakers by who’s offering the best price on each outcome.<\/p>\n
def line_shopping_report(odds_rows):\n \"\"\"Find best and worst prices per outcome across all books.\"\"\"\n for oid, label in OUTCOME_MAP.items():\n prices = [(r[\"slug\"], r[\"prices\"][oid]) for r in odds_rows]\n prices.sort(key=lambda x: -x[1])\n best = prices[0]\n worst = prices[-1]\n spread = ((best[1] \/ worst[1]) - 1) * 100\n print(f\" {label}:\")\n print(f\" Best: {best[0]:20s} @ {best[1]:.3f}\")\n print(f\" Worst: {worst[0]:20s} @ {worst[1]:.3f}\")\n print(f\" Spread: {spread:.1f}%\")\n\nline_shopping_report(odds)<\/code><\/pre>\n
Market Efficiency Score<\/h3>\n
How much do bookmakers disagree? High standard deviation = high disagreement = more opportunity.<\/p>\n
def market_efficiency(odds_rows):\n \"\"\"Standard deviation of implied probs across books.\"\"\"\n for oid, label in OUTCOME_MAP.items():\n probs = [1 \/ r[\"prices\"][oid] for r in odds_rows]\n sd = statistics.stdev(probs) * 100\n print(f\" {label}: std dev = {sd:.2f}pp across {len(probs)} books\")\n\nmarket_efficiency(odds)\n# Home: std dev = 2.74pp across 115 books\n# Draw: std dev = 0.96pp across 115 books\n# Away: std dev = 2.09pp across 115 books<\/code><\/pre>\n
The Draw market shows the tightest spread (0.96pp) \u2014 bookmakers broadly agree on the draw probability. Home has the widest disagreement (2.74pp), meaning there’s more opportunity to find mispriced Home odds.<\/p>\n
Complete Consensus Odds Scanner<\/h2>\n
Here’s the full script combining everything above into a single runnable scanner.<\/p>\n
import requests, time, statistics\nfrom datetime import datetime, timedelta\n\nAPI_KEY = \"YOUR_API_KEY\"\nBASE_URL = \"https:\/\/api.oddspapi.io\/v4\"\nOUTCOME_MAP = {\"101\": \"Home\", \"102\": \"Draw\", \"103\": \"Away\"}\nSHARP_BOOKS = {\"pinnacle\", \"singbet\", \"sbobet\"}\n\ndef api_get(endpoint, params=None):\n if params is None: params = {}\n params[\"apiKey\"] = API_KEY\n r = requests.get(f\"{BASE_URL}\/{endpoint}\", params=params)\n r.raise_for_status()\n return r.json()\n\ndef implied_prob(odds): return 1 \/ odds\n\ndef remove_vig(probs):\n total = sum(probs)\n return [p \/ total for p in probs]\n\ndef fetch_all_odds(fixture_id, market_id=\"101\"):\n data = api_get(\"odds\", {\"fixtureId\": fixture_id})\n rows = []\n for slug, bdata in data.get(\"bookmakerOdds\", {}).items():\n market = bdata.get(\"markets\", {}).get(market_id, {})\n outcomes = market.get(\"outcomes\", {})\n prices = {}\n for oid in OUTCOME_MAP:\n p0 = outcomes.get(oid, {}).get(\"players\", {}).get(\"0\", {})\n if isinstance(p0, dict) and p0.get(\"active\") and p0.get(\"price\"):\n prices[oid] = p0[\"price\"]\n if len(prices) == len(OUTCOME_MAP):\n vig = sum(1\/p for p in prices.values()) - 1\n if vig < 0.20:\n rows.append({\"slug\": slug, \"prices\": prices, \"vig\": vig * 100})\n return rows\n\ndef consensus_vig_removed(rows):\n fair_probs = {oid: [] for oid in OUTCOME_MAP}\n for row in rows:\n implied = [1 \/ row[\"prices\"][oid] for oid in OUTCOME_MAP]\n fair = remove_vig(implied)\n for i, oid in enumerate(OUTCOME_MAP):\n fair_probs[oid].append(fair[i])\n raw = {oid: statistics.median(ps) for oid, ps in fair_probs.items()}\n total = sum(raw.values())\n return {oid: p \/ total for oid, p in raw.items()}\n\ndef find_value(rows, consensus, min_edge=2.0):\n value = []\n for row in rows:\n for oid, label in OUTCOME_MAP.items():\n fair_price = 1 \/ consensus[oid]\n edge = (row[\"prices\"][oid] \/ fair_price - 1) * 100\n if edge >= min_edge:\n value.append({\n \"book\": row[\"slug\"], \"outcome\": label,\n \"price\": row[\"prices\"][oid],\n \"fair\": round(fair_price, 3),\n \"edge\": round(edge, 1),\n })\n return sorted(value, key=lambda x: -x[\"edge\"])\n\n# --- Run the scanner ---\ntoday = datetime.utcnow().strftime(\"%Y-%m-%d\")\nend = (datetime.utcnow() + timedelta(days=3)).strftime(\"%Y-%m-%d\")\nfixtures = api_get(\"fixtures\", {\"sportId\": 10, \"from\": today, \"to\": end})\nupcoming = [f for f in fixtures if f.get(\"hasOdds\")\n and datetime.fromisoformat(f[\"startTime\"].replace(\"Z\",\"\")) > datetime.utcnow()]\n\nprint(\"=\" * 62)\nprint(\" CONSENSUS ODDS SCANNER\")\nprint(f\" {len(upcoming)} fixtures | Method: vig-removed median\")\nprint(\"=\" * 62)\n\nfor f in upcoming[:5]:\n match = f\"{f['participant1Name']} vs {f['participant2Name']}\"\n print(f\"\\n {match}\")\n print(f\" {'-' * len(match)}\")\n\n odds = fetch_all_odds(f[\"fixtureId\"])\n if len(odds) < 10:\n print(f\" Skipped \u2014 only {len(odds)} books\")\n continue\n\n consensus = consensus_vig_removed(odds)\n print(f\" Books: {len(odds)} | Consensus fair odds:\")\n for oid, label in OUTCOME_MAP.items():\n print(f\" {label}: {1\/consensus[oid]:.3f} (prob: {consensus[oid]*100:.1f}%)\")\n\n value = find_value(odds, consensus, min_edge=2.0)\n if value:\n print(f\" Value bets ({len(value)}):\")\n for v in value[:5]:\n print(f\" {v['book']:18s} {v['outcome']:5s} \"\n f\"@ {v['price']:.3f} vs {v['fair']:.3f} (+{v['edge']}%)\")\n\n time.sleep(2) # respect rate limits<\/code><\/pre>\n
Why 350+ Bookmakers Changes the Game<\/h2>\n
The maths is simple: more independent price sources = more stable consensus. Here's what changes as you add bookmakers:<\/p>\n
OddsPapi returns 130+ bookmakers with active odds on a single Premier League match. That includes 3 sharp books, Betfair Exchange, and 80+ regional soft books that no other API covers. The consensus from that sample isn't just \"better\" \u2014 it's a fundamentally different signal.<\/p>\n
Frequently Asked Questions<\/h2>\n
What's the difference between consensus odds and true odds?<\/h3>\n
\"True odds\" are the actual probability of an outcome \u2014 unknowable in advance. Consensus odds are the best available estimate, calculated by aggregating many bookmakers' prices after removing their margins. With 130+ books, the consensus converges close to the true probability.<\/p>\n
Which consensus method should I use?<\/h3>\n
Start with the vig-removed median<\/strong> (Method C). It combines the outlier resistance of the median with the accuracy of vig removal. If you want to lean on sharp book intelligence, use sharp-weighted mean (Method B). For liquid markets where Betfair is deep, exchange-anchored (Method D) is a useful cross-check.<\/p>\n