Model vs market
We model football probabilities, then test them against the market.
The model estimates each outcome's probability, compares it with the bookmaker's margin-free probability, and tracks what happens when the two disagree.
- matches priced next 72 h 25 leagues
- 110
- signals across 68 matches
- 121
- newest price each price carries its own time
Research model · Not betting advice
Not betting advice. Estimates from a statistical model, published for reading. A positive edge is a statement about an average over many comparable prices, not about the match in front of you.
Current evidence · Retrospective record · Rule v2 Retrospective evidence The record of what today's rule finds in past prices, each read at its own evaluation moment; nothing in it was published at the time, so it cannot demonstrate an edge. In the glossary
No profitable edge demonstrated
Lost money on this record (interval below zero)
- ROI
- −17.7%
- 95% interval
- −25.3% to −10.2%
- Sample
- 1,697 signals 1,023 matches
Prospective validation
Prospective validation The ledger of signals as they were published live, settled once their matches are played; only it can show an edge, and only once it reaches the promotion floor. In the glossaryIn progress
190 of 400 signals and 84 of 250 matches settled so far.
How this is computed
- Units
- −300.49u
- Rule
- v2
Today's rule, version 2, applied to each historical selection at its last evaluation before kickoff. These were not necessarily signals published live at the time, and a quoted historical price may not have been available to you. This record can hold the verdict down; it can never raise it. Only signals published live, in the forward ledger, can.
The 95% interval resamples matches rather than selections, 2,000 times from seed 20261102, so selections on one match move together.
11 voids are staked and counted. 929 settled rows are left out of every figure: their price was captured after the evaluation it was paired with.
Retrospective: today's rule asked about each selection at its last evaluation before kickoff, not a list published at the time. A quoted historical price may not have been available to you. Computed .
Computed . Explore the evidence
Today's model–market disagreements
Each match's largest price that meets the current signal rule, with the model's probability set against the market's. A signal is not a recommendation.
Disagreement Disagreement Model probability minus market probability, in percentage points: it says the two disagree, not which is right. In the glossary Model-implied edge Model-implied edge Model probability × market odds − 1: what a stake of one would return on average at the offered price if the model is right, not an observed return. In the glossary
-
Süper Lig
Kocaelispor – Fenerbahçe
Fenerbahçe at most 1 goal · Team goals
- Model probability
- 63.6%
- Market probability
- 42.3%
- Disagreement
- +21.3 pp
- Model price
- 1.57
- Market price
- 2.20 Bet365
- Model-implied edge
- +40.0%
Live signals like this one: 2 settled on 2 matches; validation needs 400 on 250.
Details Hide details : Kocaelispor – Fenerbahçe
Why the model expects Fenerbahçe to score at most once
How the model gets this chance
- Projected goals 1.36 – 1.25
- Score matrix Binomial · Dixon–Coles
- This selection 63.6% Derived prediction
Fenerbahçe's projected goals and Kocaelispor's set a score matrix of two moment-matched binomial goal counts with the Dixon–Coles low-score adjustment; the scorelines this selection covers add up to 63.6%.
- Projected goalsDerived prediction
- 1.25
- Fenerbahçe's scoring rate in the score matrix the probability is read from
- VenueModel input
- Away
- The model uses the competition's away scoring rate
Ranks, recent window and sample
- Attack v defenceModel input
- 5th vs 6th
- Fenerbahçe's attack against Kocaelispor's defence, by fitted rating
- Recent xGContext
- 2.15 vs 1.12
- Not a model input: Fenerbahçe's xG per match against Kocaelispor's xG conceded, last 6 matches
These are the model's inputs for this match, not a causal attribution, and not advice.
- Price at which a 5% model-implied edge begins
- 1.65
- Bookmaker margin
- 7.4%
- Raw implied probability (1 / odds)
- 45.5%
- Margin removal
- Multiplicative, over the complete book
- Odds captured
- Price age
- 16 h 6 min
- Time to kickoff
- 2 d 6 h
Similar past signals (retrospective, model-implied edge 30–40%): Insufficient sample (26)
- Versions
- Model v1 · rule v2 · model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Eliteserien
Molde FK – Bodø/Glimt
Bodø/Glimt not to score · Team goals
- Model probability
- 15.5%
- Market probability
- 10.6%
- Disagreement
- +4.9 pp
- Model price
- 6.44
- Market price
- 9.00 Bet365
- Model-implied edge
- +39.7%
Live signals like this one: 3 settled on 3 matches; validation needs 400 on 250.
Details Hide details : Molde FK – Bodø/Glimt
Why the model rates Bodø/Glimt not to score above the market
How the model gets this chance
- Projected goals 1.54 – 1.71
- Score matrix Binomial · Dixon–Coles
- This selection 15.5% Derived prediction
Bodø/Glimt's projected goals and Molde FK's set a score matrix of two moment-matched binomial goal counts with the Dixon–Coles low-score adjustment; the scorelines this selection covers add up to 15.5%.
- Projected goalsDerived prediction
- 1.71
- Bodø/Glimt's scoring rate in the score matrix the probability is read from
- VenueModel input
- Away
- The model uses the competition's away scoring rate
Ranks, recent window and sample
- Attack v defenceModel input
- 2nd vs 5th
- Bodø/Glimt's attack against Molde FK's defence, by fitted rating
- Recent xGContext
- 2.09 vs 1.80
- Not a model input: Bodø/Glimt's xG per match against Molde FK's xG conceded, last 6 matches
These are the model's inputs for this match, not a causal attribution, and not advice.
- Price at which a 5% model-implied edge begins
- 6.76
- Bookmaker margin
- 4.5%
- Raw implied probability (1 / odds)
- 11.1%
- Margin removal
- Multiplicative, over the complete book
- Odds captured
- Price age
- 10 h 6 min
- Time to kickoff
- 32 h 52 min
Similar past signals (retrospective, model-implied edge 30–40%): Insufficient sample (33)
- Versions
- Model v1 · rule v2 · model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Swiss Super League
FC Sion – Servette FC
FC Sion not to score · Team goals
- Model probability
- 21.4%
- Market probability
- 14.6%
- Disagreement
- +6.8 pp
- Model price
- 4.67
- Market price
- 6.50 Bet365
- Model-implied edge
- +39.2%
Live signals like this one: 3 settled on 3 matches; validation needs 400 on 250.
Details Hide details : FC Sion – Servette FC
Why the model rates FC Sion not to score above the market
How the model gets this chance
- Projected goals 1.42 – 1.16
- Score matrix Binomial · Dixon–Coles
- This selection 21.4% Derived prediction
FC Sion's projected goals and Servette FC's set a score matrix of two moment-matched binomial goal counts with the Dixon–Coles low-score adjustment; the scorelines this selection covers add up to 21.4%.
- Projected goalsDerived prediction
- 1.42
- FC Sion's scoring rate in the score matrix the probability is read from
- VenueModel input
- Home
- The model uses the competition's home scoring rate
Ranks, recent window and sample
- Attack v defenceModel input
- 5th vs 1st
- FC Sion's attack against Servette FC's defence, by fitted rating
- Recent xGContext
- 2.34 vs 1.33
- Not a model input: FC Sion's xG per match against Servette FC's xG conceded, last 6 matches
These are the model's inputs for this match, not a causal attribution, and not advice.
- Price at which a 5% model-implied edge begins
- 4.90
- Bookmaker margin
- 5.4%
- Raw implied probability (1 / odds)
- 15.4%
- Margin removal
- Multiplicative, over the complete book
- Odds captured
- Price age
- 13 h 6 min
- Time to kickoff
- 2 d 10 h
Similar past signals (retrospective, model-implied edge 30–40%): Insufficient sample (25)
- Versions
- Model v1 · rule v2 · model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Austrian Bundesliga
Grazer AK 1902 – SK Rapid Wien
SK Rapid Wien not to score · Team goals
- Model probability
- 17.4%
- Market probability
- 11.9%
- Disagreement
- +5.4 pp
- Model price
- 5.76
- Market price
- 8.00 Bet365
- Model-implied edge
- +38.9%
Live signals like this one: 3 settled on 3 matches; validation needs 400 on 250.
Details Hide details : Grazer AK 1902 – SK Rapid Wien
Why the model rates SK Rapid Wien not to score above the market
How the model gets this chance
- Projected goals 1.16 – 1.62
- Score matrix Binomial · Dixon–Coles
- This selection 17.4% Derived prediction
SK Rapid Wien's projected goals and Grazer AK 1902's set a score matrix of two moment-matched binomial goal counts with the Dixon–Coles low-score adjustment; the scorelines this selection covers add up to 17.4%.
- Projected goalsDerived prediction
- 1.62
- SK Rapid Wien's scoring rate in the score matrix the probability is read from
- VenueModel input
- Away
- The model uses the competition's away scoring rate
Ranks, recent window and sample
- Attack v defenceModel input
- 4th vs 9th
- SK Rapid Wien's attack against Grazer AK 1902's defence, by fitted rating
- Recent xGContext
- 2.68 vs 1.85
- Not a model input: SK Rapid Wien's xG per match against Grazer AK 1902's xG conceded, last 6 matches
These are the model's inputs for this match, not a causal attribution, and not advice.
- Price at which a 5% model-implied edge begins
- 6.05
- Bookmaker margin
- 4.8%
- Raw implied probability (1 / odds)
- 12.5%
- Margin removal
- Multiplicative, over the complete book
- Odds captured
- Price age
- 13 h 6 min
- Time to kickoff
- 2 d 9 h
Similar past signals (retrospective, model-implied edge 30–40%): Insufficient sample (33)
- Versions
- Model v1 · rule v2 · model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Eliteserien
Tromsø IL – Viking FK
Tromsø IL to win · 1X2
- Model probability
- 39.4%
- Market probability
- 26.3%
- Disagreement
- +13.1 pp
- Model price
- 2.54
- Market price
- 3.50 Bet365
- Model-implied edge
- +38.0%
Live signals like this one: 3 settled on 3 matches; validation needs 400 on 250.
Details Hide details : Tromsø IL – Viking FK
Why the model differs
Why the model leans Tromsø IL, but not by much
39.4% Tromsø IL to win
- Attack
- 14th vs 6th
- Tromsø IL attack v Viking FK defence
- Defence
- 2nd vs 3rd
- Defence v their attack
- Recent xG form
- +0.10 vs +0.70
- xGD per match, last 6
- Projected goals
- 1.49 – 1.32
- Home v away
These are the model's inputs for this match, not a causal attribution, and not advice.
- Price at which a 5% model-implied edge begins
- 2.66
- Bookmaker margin
- 8.5%
- Raw implied probability (1 / odds)
- 28.6%
- Margin removal
- Multiplicative, over the complete book
- Odds captured
- Price age
- 19 h 6 min
- Time to kickoff
- 2 d 3 h
Similar past signals (retrospective, model-implied edge 30–40%): Insufficient sample (48)
- Versions
- Model v1 · rule v2 · model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
All prices, next 72 hours
Every priced outcome against the model, largest disagreement first. The table opens on the signals at this moment; All prices shows every evaluated price in the window.
1401–1425 of 1754 prices
- Signal: every check of the rule passed, a model-implied edge of 5% or more
- Stale price, limited data or doesn't beat base rate: never a signal
- Model probability
- Market probability
-
Details: Under 2.5 goals
- Model
- 58.4%
- Market
- 52.6%
- Disagreement
- +5.8 pp
- Model-implied edge
- +5.1%
Model price 1.71 Market price 1.80 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 58.4%
- Market, margin removed
- 52.6%
- Disagreement
- +5.8 pp
- Model-implied edge
- +5.1%
- Model price
- 1.71
- Market price
- 1.80 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.80
- Bookmaker margin
- 5.6%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Draw
- Model
- 26.9%
- Market
- 21.1%
- Disagreement
- +5.8 pp
- Model-implied edge
- +21.1%
Model price 3.72 Market price 4.50 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 26.9%
- Market, margin removed
- 21.1%
- Disagreement
- +5.8 pp
- Model-implied edge
- +21.1%
- Model price
- 3.72
- Market price
- 4.50 Bet365
Details
- Price at which a 5% model-implied edge begins
- 3.90
- Bookmaker margin
- 5.2%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Both teams to score
- Model
- 57.8%
- Market
- 52.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +4.0%
Model price 1.73 Market price 1.80 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 57.8%
- Market, margin removed
- 52.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +4.0%
- Model price
- 1.73
- Market price
- 1.80 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.82
- Bookmaker margin
- 6.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Le Havre to score
- Model
- 74.5%
- Market
- 68.7%
- Disagreement
- +5.8 pp
- Model-implied edge
- +1.7%
Model price 1.34 Market price 1.36 Bet365
Why the model differs
- Model
- 74.5%
- Market, margin removed
- 68.7%
- Disagreement
- +5.8 pp
- Model-implied edge
- +1.7%
- Model price
- 1.34
- Market price
- 1.36 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.41
- Bookmaker margin
- 6.6%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: OFK Beograd to score
- Model
- 80.8%
- Market
- 75.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +1.0%
Model price 1.24 Market price 1.25 Bet365
Why the model differs
- Model
- 80.8%
- Market, margin removed
- 75.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +1.0%
- Model price
- 1.24
- Market price
- 1.25 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.30
- Bookmaker margin
- 6.7%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Slavia Sofia to score
- Model
- 68.3%
- Market
- 62.5%
- Disagreement
- +5.8 pp
- Model-implied edge
- +2.4%
Model price 1.46 Market price 1.50 Bet365
Why the model differs
- Model
- 68.3%
- Market, margin removed
- 62.5%
- Disagreement
- +5.8 pp
- Model-implied edge
- +2.4%
- Model price
- 1.46
- Market price
- 1.50 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.54
- Bookmaker margin
- 6.7%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Draw
- Model
- 26.8%
- Market
- 21.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +20.8%
Model price 3.73 Market price 4.50 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 26.8%
- Market, margin removed
- 21.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +20.8%
- Model price
- 3.73
- Market price
- 4.50 Bet365
Details
- Price at which a 5% model-implied edge begins
- 3.91
- Bookmaker margin
- 5.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Both teams to score
- Model
- 57.8%
- Market
- 52.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +4.1%
Model price 1.73 Market price 1.80 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 57.8%
- Market, margin removed
- 52.0%
- Disagreement
- +5.8 pp
- Model-implied edge
- +4.1%
- Model price
- 1.73
- Market price
- 1.80 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.82
- Bookmaker margin
- 6.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Dundee United to score
- Model
- 76.8%
- Market
- 70.9%
- Disagreement
- +5.9 pp
- Model-implied edge
- +2.4%
Model price 1.30 Market price 1.33 Bet365
Why the model differs
- Model
- 76.8%
- Market, margin removed
- 70.9%
- Disagreement
- +5.9 pp
- Model-implied edge
- +2.4%
- Model price
- 1.30
- Market price
- 1.33 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.37
- Bookmaker margin
- 5.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Under 2.5 goals
- Model
- 46.0%
- Market
- 40.1%
- Disagreement
- +5.9 pp
- Model-implied edge
- +8.0%
Model price 2.18 Market price 2.35 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 46.0%
- Market, margin removed
- 40.1%
- Disagreement
- +5.9 pp
- Model-implied edge
- +8.0%
- Model price
- 2.18
- Market price
- 2.35 Bet365
Details
- Price at which a 5% model-implied edge begins
- 2.28
- Bookmaker margin
- 6.2%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Śląsk Wrocław to score
- Model
- 74.7%
- Market
- 68.7%
- Disagreement
- +5.9 pp
- Model-implied edge
- +1.8%
Model price 1.34 Market price 1.36 Bet365
Why the model differs
- Model
- 74.7%
- Market, margin removed
- 68.7%
- Disagreement
- +5.9 pp
- Model-implied edge
- +1.8%
- Model price
- 1.34
- Market price
- 1.36 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.41
- Bookmaker margin
- 6.6%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: FC Livyi Bereh Kyiv to win
- Model
- 14.8%
- Market
- 8.9%
- Disagreement
- +5.9 pp
- Model-implied edge
- +48.3%
Model price 6.74 Market price 10.00 Bet365
Why the model differs
- Model
- 14.8%
- Market, margin removed
- 8.9%
- Disagreement
- +5.9 pp
- Model-implied edge
- +48.3%
- Model price
- 6.74
- Market price
- 10.00 Bet365
Details
- Price at which a 5% model-implied edge begins
- 7.08
- Bookmaker margin
- 12.2%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
Full match analysis →: Shakhtar Donetsk – FC Livyi Bereh Kyiv
-
Details: Over 1.5 goals
- Model
- 84.2%
- Market
- 78.3%
- Disagreement
- +5.9 pp
- Model-implied edge
- +1.1%
Model price 1.19 Market price 1.20 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 84.2%
- Market, margin removed
- 78.3%
- Disagreement
- +5.9 pp
- Model-implied edge
- +1.1%
- Model price
- 1.19
- Market price
- 1.20 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.25
- Bookmaker margin
- 6.4%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Como – Cagliari Serie A Cagliari 2+ goals Team goals The model-implied edge clears 5% and the price passes every other check of the current rule. A signal is a mechanical filter, not a recommendation, and does not imply historical profitability. How this is decided →
Details: Cagliari 2+ goals
- Model
- 24.9%
- Market
- 18.9%
- Disagreement
- +6.0 pp
- Model-implied edge
- +24.4%
Model price 4.02 Market price 5.00 Bet365
Why the model differs
- Model
- 24.9%
- Market, margin removed
- 18.9%
- Disagreement
- +6.0 pp
- Model-implied edge
- +24.4%
- Model price
- 4.02
- Market price
- 5.00 Bet365
Details
- Price at which a 5% model-implied edge begins
- 4.22
- Bookmaker margin
- 5.7%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The model-implied edge clears 5% and the price passes every other check of the current rule. A signal is a mechanical filter, not a recommendation, and does not imply historical profitability. How this is decided → The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Cagliari to win
- Model
- 15.5%
- Market
- 9.5%
- Disagreement
- +6.0 pp
- Model-implied edge
- +54.7%
Model price 6.47 Market price 10.00 Bet365
Why the model differs
- Model
- 15.5%
- Market, margin removed
- 9.5%
- Disagreement
- +6.0 pp
- Model-implied edge
- +54.7%
- Model price
- 6.47
- Market price
- 10.00 Bet365
Details
- Price at which a 5% model-implied edge begins
- 6.79
- Bookmaker margin
- 5.4%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: ŠK Slovan Bratislava to score
- Model
- 90.4%
- Market
- 84.4%
- Disagreement
- +6.0 pp
- Model-implied edge
- +0.4%
Model price 1.11 Market price 1.11 Bet365
Why the model differs
- Model
- 90.4%
- Market, margin removed
- 84.4%
- Disagreement
- +6.0 pp
- Model-implied edge
- +0.4%
- Model price
- 1.11
- Market price
- 1.11 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.16
- Bookmaker margin
- 6.7%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
Full match analysis →: MFK Ružomberok – ŠK Slovan Bratislava
-
Nice – Paris Saint-Germain Ligue 1 Nice to win 1X2 The model-implied edge clears 5% and the price passes every other check of the current rule. A signal is a mechanical filter, not a recommendation, and does not imply historical profitability. How this is decided →
Details: Nice to win
- Model
- 21.8%
- Market
- 15.8%
- Disagreement
- +6.0 pp
- Model-implied edge
- +30.7%
Model price 4.59 Market price 6.00 Bet365
Why the model differs
- Model
- 21.8%
- Market, margin removed
- 15.8%
- Disagreement
- +6.0 pp
- Model-implied edge
- +30.7%
- Model price
- 4.59
- Market price
- 6.00 Bet365
Details
- Price at which a 5% model-implied edge begins
- 4.82
- Bookmaker margin
- 5.6%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The model-implied edge clears 5% and the price passes every other check of the current rule. A signal is a mechanical filter, not a recommendation, and does not imply historical profitability. How this is decided → The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Over 2.5 goals
- Model
- 61.3%
- Market
- 55.3%
- Disagreement
- +6.0 pp
- Model-implied edge
- +4.2%
Model price 1.63 Market price 1.70 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 61.3%
- Market, margin removed
- 55.3%
- Disagreement
- +6.0 pp
- Model-implied edge
- +4.2%
- Model price
- 1.63
- Market price
- 1.70 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.71
- Bookmaker margin
- 6.4%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Draw
- Model
- 21.8%
- Market
- 15.8%
- Disagreement
- +6.0 pp
- Model-implied edge
- +30.7%
Model price 4.59 Market price 6.00 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 21.8%
- Market, margin removed
- 15.8%
- Disagreement
- +6.0 pp
- Model-implied edge
- +30.7%
- Model price
- 4.59
- Market price
- 6.00 Bet365
Details
- Price at which a 5% model-implied edge begins
- 4.82
- Bookmaker margin
- 5.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Cracovia – Jagiellonia Białystok Ekstraklasa Cracovia to win 1X2 The model-implied edge clears 5% and the price passes every other check of the current rule. A signal is a mechanical filter, not a recommendation, and does not imply historical profitability. How this is decided →
Details: Cracovia to win
- Model
- 44.4%
- Market
- 38.4%
- Disagreement
- +6.0 pp
- Model-implied edge
- +5.6%
Model price 2.25 Market price 2.38 Bet365
Why the model differs
- Model
- 44.4%
- Market, margin removed
- 38.4%
- Disagreement
- +6.0 pp
- Model-implied edge
- +5.6%
- Model price
- 2.25
- Market price
- 2.38 Bet365
Details
- Price at which a 5% model-implied edge begins
- 2.36
- Bookmaker margin
- 9.6%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The model-implied edge clears 5% and the price passes every other check of the current rule. A signal is a mechanical filter, not a recommendation, and does not imply historical profitability. How this is decided → The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Both teams to score
- Model
- 64.9%
- Market
- 58.9%
- Disagreement
- +6.0 pp
- Model-implied edge
- +2.0%
Model price 1.54 Market price 1.57 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 64.9%
- Market, margin removed
- 58.9%
- Disagreement
- +6.0 pp
- Model-implied edge
- +2.0%
- Model price
- 1.54
- Market price
- 1.57 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.62
- Bookmaker margin
- 8.1%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Over 2.5 goals
- Model
- 68.0%
- Market
- 62.0%
- Disagreement
- +6.1 pp
- Model-implied edge
- +4.3%
Model price 1.47 Market price 1.53 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 68.0%
- Market, margin removed
- 62.0%
- Disagreement
- +6.1 pp
- Model-implied edge
- +4.3%
- Model price
- 1.47
- Market price
- 1.53 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.54
- Bookmaker margin
- 5.2%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
Details: Coventry City to score
- Model
- 82.7%
- Market
- 76.6%
- Disagreement
- +6.1 pp
- Model-implied edge
- +1.0%
Model price 1.21 Market price 1.22 Bet365
Why the model differs
- Model
- 82.7%
- Market, margin removed
- 76.6%
- Disagreement
- +6.1 pp
- Model-implied edge
- +1.0%
- Model price
- 1.21
- Market price
- 1.22 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.27
- Bookmaker margin
- 6.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
-
FC Dinamo București – FC Petrolul Ploieşti SuperLiga (Romania) FC Dinamo București to score Team goals
Details: FC Dinamo București to score
- Model
- 94.2%
- Market
- 88.1%
- Disagreement
- +6.1 pp
- Model-implied edge
- +2.0%
Model price 1.06 Market price 1.08 Bet365
Why the model differs
- Model
- 94.2%
- Market, margin removed
- 88.1%
- Disagreement
- +6.1 pp
- Model-implied edge
- +2.0%
- Model price
- 1.06
- Market price
- 1.08 Bet365
Details
- Price at which a 5% model-implied edge begins
- 1.11
- Bookmaker margin
- 4.8%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model forecasts this outcome better than its historical base-rate benchmark (replayed over a season in 7 competitions, measured on 17 August 2026). This does not mean it beats bookmaker prices. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
Full match analysis →: FC Dinamo București – FC Petrolul Ploieşti
-
Details: Draw
- Model
- 24.4%
- Market
- 18.3%
- Disagreement
- +6.1 pp
- Model-implied edge
- +22.0%
Model price 4.10 Market price 5.00 Bet365
The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →Why the model differs
- Model
- 24.4%
- Market, margin removed
- 18.3%
- Disagreement
- +6.1 pp
- Model-implied edge
- +22.0%
- Model price
- 4.10
- Market price
- 5.00 Bet365
Details
- Price at which a 5% model-implied edge begins
- 4.30
- Bookmaker margin
- 9.5%
Price
- Price captured
- Bet365
Audit
- Rule
- v2
- Checks
- The price is within the allowance the odds schedule gives it at this distance from kickoff: 6 hours inside the last 24 hours, 48 hours before that. How this is decided → Both sides have a sample of matches behind their rating and the match carries no model warning; this is about how much the probability rests on, not how sure the model is. How this is decided → The model does not predict this market better than a historical base-rate forecast (replayed over a season in 7 competitions, measured on 17 August 2026), so an edge here is never recommended. How this is decided →
- Model
- v1 · methodology model1-rule2-hierarchy0-freshness0-markets0-calibration1-evaluation0
Graded results for everything this board ranked: Results