Pulling the odds, probabilities and the full engine breakdown for this fixture.
scheduledVeikkausliigaFinland
Turku PS🇫🇮FinlandvsIlves🇫🇮Finland
Here's everything the engine sees for this match — odds, probabilities, value, and risk.
watchlist
Watchlist — historical and context evidence is too thin for publication.
match winnerTPS Turku
Price caseModel versus consensus and executable price
OddsPadi44%
Book fair36%
Fair odds
2.26
Quoted odds
2.70
Probability edge
+8.0%
Expected value
+19.3%
Analysis only — publication blockedEmpirical 95% value floor is unavailable for this runtime.
Watchlist
Empirical value floor unavailableNo active, model-matched calibration profile is allowed to influence this decision. Raw EV remains a point estimate, not statistically verified value.
Unverified
Best executable quoteBetfair / 19 Jul, 13:49 UTC. The consensus probability is calculated separately.
Betfair
Broad market agreement19 bookmakers / 2% widest probability gap.
58% prior influence
The consensus calibrates model probability and supplies the no-vig comparison. The named bookmaker quote remains a separate executable price for expected-value analysis.
Confidencelow confidence
Riskhigh risk
Decision clock01:51 PM
Key risks
empirical 95% value floor is unavailable for this runtime
engine action is avoid
engine actionability is blocked
Slip action lockedOnly a current, published value pick can be added.Discuss match
Community opinion
What the crowd sees
Fan votes and accountable tips sit beside the model, never inside it.
Separate truth lane
01 / fan pulse
Who wins?
0 votes
02 / reasoned tips
Tipster notebook
0 published
Selections lock 30 minutes before kickoff. Original tips cannot be rewritten after publication.
Model boundary: votes and community tips never change OddsPadi probability, edge, confidence, publication status, or model accuracy.
The full bar is 100% of the model's 1X2 distribution. Evidence quality is shown separately so confidence is not confused with probability.
Turku PS
44%
Draw
25%
Ilves
31%
Evidence quality92%
Historical evidenceHistorical calibration range
Not available
No statistical band is shown.
Empirical interval unavailable: no active, model-matched calibration profile is allowed to influence this decision.
Model evidence
How the engine reached this view
stable
This is the deterministic calculation path, not a generated explanation. Positive factors support the selected side; negative factors pull against it.
Probability journeyTPS Turku
123
1
Starting prior36%
2
Model output44%
3
Final posterior44%
Runtime probability trace is blocked: final probability 44% is preserved, but it does not clear the active action gates.
Model and calibration provenance
Historical evidence model
football-poisson-v5
Runtime parity
exact runtime parity
Engine version
decision-engine-v2
Learning profile
shadow-only
Calibration
inactive
Historical sample
4,135
Held-out sample
1,241
Normalized Brier
0.197
Log loss
0.991
Calibration error
0.013
Player-form coverage
65% (1,871 fixtures)
Training player form
65%
Holdout player form
100%
Latest real-data backtest is available for shadow comparison only; live guardrails require an active model-bound calibration promotion.
Live outcome calibration: Probability calibration remains inactive until a real-data backtest receives explicit live-guardrail promotion.
Player-form coverage uses only provider-capable fixtures and requires acceptable or strong chronology-safe history in both training and holdout windows. Normalized Brier averages squared probability error across the market outcomes; lower is better. Calibration error measures the weighted gap between forecast probability and observed frequency.
Decision factor contributionWeighted points
Value edge+23.0 pts
Expected value+16.0 pts
Confidence guardrail+11.0 pts
Data quality+8.0 pts
Case memory+2.0 pts
Outcome variance+1.0 pts
Decision-risk profile53/100 diagnostic index
Data coverage uncertainty80/100
Context uncertainty60/100
Model-market disagreement54/100
Robustness and review uncertainty80/100
Weighted evidence-risk index, not a statistical confidence level.
Data coverage uncertainty. Downgrade or block public trust until data coverage uncertainty is addressed. Confidence penalty -14.7%.
Market analysis
“Value edge” is our probability minus the bookmaker's fair probability (margin removed). Positive edge means the price is better than it should be. Current best edge: +8.0%.
Market
Selection
Odds
Bookmaker
Model
Raw implied
No-vig implied
Margin
No-vig edge
EV
Match winner
TPS Turku
2.70
Betfair
44%
37%
36%
+9.7%
+8.0%
+19.3%
Match winner
Draw
3.70
Betfair
25%
27%
26%
+9.7%
-1.4%
-7.3%
Match winner
Ilves Tampere
2.60
Betfair
31%
38%
37%
+9.7%
-6.6%
-20.0%
Match winnerTPS Turku
2.70
Model chance
44%
Fair market
36%
Edge
+8.0%
Expected value
+19.3%
Price details
Bookmaker
Betfair
Raw implied
37%
Bookmaker margin
+9.7%
Match winnerDraw
3.70
Model chance
25%
Fair market
26%
Edge
-1.4%
Expected value
-7.3%
Price details
Bookmaker
Betfair
Raw implied
27%
Bookmaker margin
+9.7%
Match winnerIlves Tampere
2.60
Model chance
31%
Fair market
37%
Edge
-6.6%
Expected value
-20.0%
Price details
Bookmaker
Betfair
Raw implied
38%
Bookmaker margin
+9.7%
Odds movement and freshness
Decision generated7/19/2026, 1:51:49 PM
Price expires7/19/2026, 2:49:36 PM
Verified price tapeMatch Winner movement
306 stored snapshots
TPS Turku2.62 to 2.70drifted 0.08
Draw3.70 to 3.70unchanged
Ilves Tampere2.56 to 2.60drifted 0.04
Decimal odds from stored pre-match provider snapshots. Lower odds mean the market has strengthened that outcome; higher odds mean it has weakened. This is observed movement, not proof that the move is correct.
OddsPadi explanation
OddsPadi estimates Turku PS at 44%, the draw at 25%, and Ilves at 31%. The current prices do not show a clear positive value edge, so the responsible call is to avoid forcing a pick.
The model weighs ratings, recent form, attack and defensive strength.
Bookmaker margin is removed where possible before value edge is calculated.
Confidence stays cautious when the edge is weak or data quality is limited.
Sports outcomes are uncertain. Predictions are model estimates, not guarantees.
Advanced engine audit
Audit-only detail cannot override the canonical public decision above. Candidate markets below show the engine's working, including blocked opportunities, but only the canonical status is publishable.
Decision engine
decision-engine-v2 - rules and model reasoning
avoid
Decision engine is avoiding TPS Turku: model 44%, no-vig implied 36%, edge +8.0%, EV +19.3%, fair odds 2.26.
Research brief
Statusblocked
Decision clockBelief expires at 14:51 UTC; next review is 13:56 UTC.
Turku PS vs Ilves is blocked from a public recommendation.
Decision engine is avoiding TPS Turku: model 44%, no-vig implied 36%, edge +8.0%, EV +19.3%, fair odds 2.26. The research brief says block public recommendation until failed gates, missing data, or invalid memory conditions are repaired.
Model thesisTPS Turku is the leading thesis because the model price beats the no-vig market by +8.0% with +19.3% EV per unit.
Market thesisMarket disagreement exists: model 44% versus no-vig 36%, edge +8.0%, EV +19.3%, quoted odds 2.70.
Risk thesisRisk thesis blocks the recommendation: Monitoring state: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve..
Research checks: Refresh bookmaker odds: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Injury and suspension news: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates. Confirmed lineups: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates. Runtime holdout economics: Do not promote the decision until this abstention gate is cleared by fresh model, market, or provider evidence. Keep as analysis-only watchlist unless confidence improves or risk falls.
Decision notebook
Statusblocked
Assumptions5
Falsifiers5
Next review01:56 PM
Notebook is blocked: 2 falsifier or abstention condition(s) prevent public trust.
assumption - high - openModel probability remains valid
TPS Turku is currently modeled at 44% with +8.0% no-vig edge.
Action: Rerun the sport model after any lineup, injury, weather, odds, or live-state update.
assumption - critical - openMarket price still supports the thesis
Quoted odds 2.70 imply EV +19.3% after margin removal.
Action: Refresh bookmaker odds and recompute raw implied probability, no-vig probability, edge, and EV.
assumption - medium - openMissing context does not overturn the thesis
Open context gaps: Injury and suspension news, Confirmed lineups.
Action: Fetch the highest-priority missing provider signals before keeping the same action.
falsifier - critical - openOdds shorten enough to erase value
Remove the thesis if refreshed edge or EV is no longer positive; current fair odds 2.26.
Action: Downgrade to avoid when no-vig edge or EV falls to zero or below.
falsifier - high - openAdverse injury, lineup, weather, or live event appears
Action: Apply the provider context update and rerun actionability before showing the pick.
falsifier - high - openBelief expires before refresh
Belief is moderate: TPS Turku at 44% with +8.0% edge, +19.3% EV, uncertainty 31/100, expires in 60 minutes.
Action: Hide or downgrade the recommendation if the belief expires without a fresh decision run.
Operator checklist: Refresh bookmaker odds: Complete this check, then rerun the decision engine before trusting the current posture. Injury and suspension news: Complete this check, then rerun the decision engine before trusting the current posture. Confirmed lineups: Complete this check, then rerun the decision engine before trusting the current posture. Runtime holdout economics: Complete this check, then rerun the decision engine before trusting the current posture. Keep as analysis-only watchlist unless confidence improves or risk falls.: Complete this check, then rerun the decision engine before trusting the current posture.
Agent calibration
Healthstable
Reliability84/100
Calibration actiontrust
Contradictions0
Decision can be shown with normal responsible-use language, while still requiring fresh odds before action.
Data coverage
Statuspartial
Score55/100
Provider-backed6
Mock0
Missing5
Stale0
Data coverage is 55/100 with partial input support and 6 production gap(s).
provider backed - pre match - api-footballFixture for the day
Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
provider backed - current - api-football-standingsLeague standings
Turku PS rank 7, 22 pts; Ilves rank 9, 16 pts.
provider backed - pre match - api-football-recent-fixturesHome/away performance
Home/away strength uses provider recent-form windows when available, alongside team and league strength; data quality 92%.
provider backed - pre match - api-football-recent-fixturesRecent form
Recent form is loaded from the provider feed.
missing - missing - missing-providerInjuries
No provider-backed injury feed is connected.
missing - missing - missing-providerSuspensions
No suspension provider is connected.
missing - missing - missing-providerLineups
Confirmed lineups/starters are not connected.
Required before trust: Injuries: No provider-backed injury feed is connected. Suspensions: No suspension provider is connected. Lineups: Confirmed lineups/starters are not connected. News signals: No news provider is connected. Team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached. Historical training corpus: Latest real-data backtest is available for shadow comparison only; live guardrails require an active model-bound calibration promotion.
Odds intelligence
Statuspositive ev
Markets1
Positive EV1
Actionable1
Avg margin+9.7%
Best actionableTPS Turku
Watch leaderNone
Odds intelligence found 1 actionable value candidate(s) across 1 market(s); best is TPS Turku.
value - medium confidence - medium riskTPS Turku: 2.70
Model 44%; no-vig 36%; edge +8.0%; EV +19.3%; fair 2.26. TPS Turku has positive no-vig edge +8.0% and EV +19.3%. Score 0.140. Shortening tolerance 16%.
avoid - low confidence - high riskIlves Tampere: 2.60
Model 31%; no-vig 37%; edge -6.6%; EV -20.0%; fair 3.25. Ilves Tampere is priced efficiently or short versus the model after margin removal. Score -0.456. Shortening tolerance 0%.
avoid - low confidence - high riskDraw: 3.70
Model 25%; no-vig 26%; edge -1.4%; EV -7.3%; fair 3.99. Draw is priced efficiently or short versus the model after margin removal. Score -0.496. Shortening tolerance 0%.
value found - margin +9.7%Match winner
Match winner has 1 actionable value candidate(s); best is TPS Turku at +19.3% EV.
Avoid notes: Draw: Draw is priced efficiently or short versus the model after margin removal. Ilves Tampere: Ilves Tampere is priced efficiently or short versus the model after margin removal.
Market movement intelligence
Statusresilient
Current odds2.70
Fair odds2.26
Shortening room16%
TPS Turku has a resilient market buffer: current odds 2.70, fair odds 2.26, and 16% shortening tolerance before EV reaches zero.
Current price - action consider2.70
Edge +10.4%; EV +19.3%. Current quoted odds and model probability before any market move.
Odds shorten 3% - action consider2.62
Edge +9.7%; EV +15.7%. Small price move against the model thesis.
Odds shorten 5% - action consider2.56
Edge +9.3%; EV +13.3%. Standard pre-action price stress used by the decision engine.
Odds shorten 10% - action consider2.43
Edge +8.0%; EV +7.4%. Aggressive market move against the quoted value.
Odds drift 5% - action consider2.84
Edge +11.5%; EV +25.3%. Market drifts longer; value may improve but could indicate adverse news.
Movement alerts: Remove or downgrade if odds shorten more than 16% from the current quote. A 5% odds shortening does not fully break the thesis, but still requires a refresh.
Belief state
Grademoderate
Believed probability44%
Probability edge+8.0%
Uncertainty31/100
Empirical bandNo data
Expires60 min
Supports7
Opposes0
Empirical interval unavailable: no active, model-matched calibration profile is allowed to influence this decision.
Belief is moderate: TPS Turku at 44% with +8.0% edge, +19.3% EV, uncertainty 31/100, expires in 60 minutes.
supports - medium confidence - model-and-marketTPS Turku model belief
Impact +8.0%. Model 44%, no-vig 36%, EV +19.3%.
supports - high confidence - calibrationCalibration belief
Impact +2.5%. Decision can be shown with normal responsible-use language, while still requiring fresh odds before action.
supports - medium confidence - decision-memoryCase memory belief
Impact +1.5%. Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
supports - medium confidence - modelPoisson expected-goals model
Impact +2.8%. Turku PS expected goals 1.89, Ilves expected goals 1.12; top scoreline 1-1.
Invalidates when: Invalidate if TPS Turku no-vig edge falls to zero or EV turns negative. Invalidate if confirmed lineups, injuries, suspensions, weather, or live events materially oppose the current thesis. Invalidate if bookmaker prices move before the next refresh window. Blocking gate: Runtime holdout economics.
Probability trace
Statusblocked
Market prior36%
Model probability44%
Posterior44%
Posterior edge+8.0%
Posterior EV+19.3%
Runtime probability trace is blocked: final probability 44% is preserved, but it does not clear the active action gates.
model evidence - applied - weight 100%Raw sport-model probability
55% to 55% (0.0%). Probability emitted by the sport model before residual context, promoted calibration, or bookmaker-prior blending.
Trace safeguards: The trace replays actual runtime probability snapshots; it does not apply a second synthetic evidence-fusion pass. Case memory, reliability checks, and abstention rules can gate the action but cannot mutate the published probability. Final probability, edge, and expected value must equal the values used by selection ranking. Fresh odds, lineups, injuries, live events, and stored outcomes can still invalidate the decision. Probability snapshots remain bounded between 0% and 100%.
Decision attribution
Statusblocked
Decisive factorReview loop
Value score100/100
Risk score100/100
Net movement+8.0%
Model-market gap+8.0%
Attribution is blocked: Review loop prevents a trusted public recommendation.
positive - market - impact 71/100TPS Turku value edge
Probability impact +8.0%. TPS Turku has no-vig edge +8.0% and EV +19.3%.
positive - calibration - impact 42/100Calibration health
Decision can be shown with normal responsible-use language, while still requiring fresh odds before action.
positive - price - impact 32/100Market movement
Probability impact +8.0%. TPS Turku has a resilient market buffer: current odds 2.70, fair odds 2.26, and 16% shortening tolerance before EV reaches zero.
negative - risk - impact 88/100Review loop
Review loop blocks the recommendation until failed gates or abstention rules clear.
negative - data - impact 85/100Data coverage
Data coverage is 55/100 with partial input support and 6 production gap(s).
The final action is constrained by a blocker; do not treat the selected side as public value until the negative driver clears.
Decision-risk decomposition
Statushigh risk
Diagnostic index53/100
PrimaryData coverage uncertainty
Penalty-14.7%
Diagnostic uncertainty risk is high at 53/100; primary uncertainty is Data coverage uncertainty.
Weighted evidence-risk index, not a statistical confidence level.
high - data - contribution 16.0Data coverage uncertainty: 80/100
Data coverage is 55/100 with partial input support and 6 production gap(s).
Mitigation: Injuries: No provider-backed injury feed is connected.
medium - context - contribution 8.4Context uncertainty: 60/100
Missing context: Injury and suspension news, Confirmed lineups.
Mitigation: Fetch or verify Injury and suspension news before raising confidence.
medium - market - contribution 8.1Model-market disagreement: 54/100
Model and no-vig market differ by +8.0%.
Mitigation: Refresh bookmaker odds, compare closing price, and rerun no-vig probability before trusting the edge.
high - robustness - contribution 7.2Robustness and review uncertainty: 80/100
Robustness is 10/100: only 0/6 stress tests survive, so the recommendation is fragile. Review loop blocks the recommendation until failed gates or abstention rules clear.
Mitigation: Bookmaker price shortens: Refresh bookmaker odds and remove the pick if edge or EV is no longer positive.
high - timing - contribution 6.6Timing and freshness uncertainty: 82/100
Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve.
Mitigation: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value.
low - model - contribution 2.5Model uncertainty: 18/100
Model uncertainty is low; model data quality is 92%.
Mitigation: Improve model inputs with provider-backed history, form, team/player availability, and settled calibration.
Decision impact: Downgrade or block public trust until data coverage uncertainty is addressed.
Decision boundary
Statusblocked
Nearest flipContext-shock tolerance: -13.1% past the boundary.
Flip margin-13.1%
Next actionDo not show as public value; clear Context-shock tolerance and rerun.
Decision boundary is blocked; nearest flip is Context-shock tolerance: -13.1% past the boundary.
probability floor - safeProbability floor
Current 44%; threshold 37%; margin +7.1%. Posterior probability is 44% versus break-even 37%; no-vig market floor is 36%.
odds floor - safeOdds floor
Current 2.70; threshold 2.26; margin +0.44 odds. Quoted odds 2.70 must stay at or above posterior fair odds 2.26.
edge floor - safeNo-vig edge floor
Current +8.0%; threshold +3.5%; margin +4.5%. Current edge is +8.0%; conservative baseline is +3.5%.
ev floor - safeExpected-value floor
Current +19.3%; threshold 0.0%; margin +19.3%. Current expected value is +19.3%; EV at or below zero removes value.
score floor - safeDecision-score floor
Current 60/100; threshold 24/100; margin +36 pts. Decision score is 60/100; lean-value consideration starts at 24, while strong value starts at 42 with high confidence.
data quality floor - safeData-quality floor
Current 92/100; threshold 62/100; margin +30 pts. Model data quality is 92% and coverage audit score is 55/100; below 62/100 hard-blocks trust.
uncertainty ceiling - safeUncertainty ceiling
Current 53/100; threshold 66/100; margin +13 pts. Uncertainty score is 53/100; 66/100 or higher is high-risk unless mitigated.
context shock - breachedContext-shock tolerance
Current -13.1%; threshold 0.0%; margin -13.1%. Worst-case stress still leaves minimum edge/EV margin at -13.1%.
price shortening - safePrice-shortening room
Current 16%; threshold 3%; margin +13.2%. The price can shorten about 16% before value disappears; below 3% is execution-sensitive.
Must stay true: TPS Turku posterior probability stays above break-even 37%. Quoted odds stay at or above posterior fair odds 2.26. No-vig edge stays above the conservative baseline +3.5%. Decision score stays at or above 24 and no hard abstention gate triggers. Model data quality stays at or above 62/100. Uncertainty stays below 66/100 and context-shock stress keeps value above zero.
Flip triggers: Context-shock tolerance breached: Worst-case stress still leaves minimum edge/EV margin at -13.1%. Runtime holdout economics: Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive.
AI protocol
Statusblocked
Modedeterministic public audit
Questions answered4/7
Tool gaps2
AI protocol is blocked: 3 failed check(s), 1 triggered abstention gate(s), and current action avoid.
answered - evidence probability-trace-summary, odds-intelligence-summaryIs there a measurable model-vs-market value edge?
TPS Turku has model probability 44%, no-vig probability 36%, edge +8.0%, and EV +19.3%.
Follow-up: Refresh bookmaker odds before public display.
answered - evidence market-movement-summary, decision-boundary-summaryCan the current price survive normal market movement?
TPS Turku has a resilient market buffer: current odds 2.70, fair odds 2.26, and 16% shortening tolerance before EV reaches zero.
Follow-up: Refresh odds, recompute no-vig probability, and downgrade if the latest quote crosses the fair-odds or EV threshold.
blocked - evidence decision-boundary-summaryWhat would flip this action?
Context-shock tolerance: -13.1% past the boundary.
Follow-up: Do not show as public value; clear Context-shock tolerance and rerun.
needs data - evidence data-coverage-summary, uncertainty-summaryWhich data gaps could change the thesis?
Injuries: No provider-backed injury feed is connected.
Follow-up: Injuries: No provider-backed injury feed is connected.
blocked - evidence attribution-summary, uncertainty-summary, review-loop-summaryDo risk, actionability, or review checks block the decision?
Actionability is 70/100: block a public recommendation until failed gates are fixed. Review loop blocks the recommendation until failed gates or abstention rules clear.
Follow-up: Fresh odds must keep no-vig edge and expected value positive.
answered - evidence historical-discipline-summaryDoes historical backtest discipline allow this decision to be promoted?
No 10-year public historical evidence is attached to this decision run. Run the public historical evidence proof before using history to discipline raw model edges.
Follow-up: Keep historical discipline attached to the audit trail.
Attach public historical evidence or a persisted provider-backed learning profile.
Unlocks: Confirms whether provider-enriched historical results beat market consensus before any raw edge can be promoted.
Reviewer guardrails: Use only supplied evidence IDs and model artifacts. Do not invent injuries, lineups, weather, news, odds, scores, or private facts. Do not expose hidden chain-of-thought; return public audit notes only. Do not upgrade a local avoid or monitor action into a stronger recommendation. Downgrade or abstain when safety gates, boundary breaches, or unsupported material claims remain.
Reasoning graph
Statusblocked
Nodes13
Edges11
Watch nodes4
Reasoning graph is blocked: 6 blocker node(s), 4 watch node(s), and final action avoid.
Provider: api-football-recent-fixtures-form-rating-v1 + missing-player-history. Depends on: fixtures-today.
Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Decision impact: Weak history keeps model strength, form weighting, and learned thresholds conservative.
high priority - ready - standingsLoad league standings
Provider: api-football-standings. Depends on: fixtures-today, historical-results.
Turku PS rank 7, 22 pts; Ilves rank 9, 16 pts.
Decision impact: Missing standings lower data quality and increase context uncertainty.
high priority - ready - recent formCompute recent form and home/away profile
Provider: api-football-recent-fixtures + api-football-recent-fixtures. Depends on: fixtures-today, historical-results.
Recent form and home/away features are ready for this model run.
Decision impact: If form or home/away inputs change, the model probability and value edge are recalculated.
high priority - missing config - injuriesFetch injuries, suspensions, lineups, and news
Provider: Sport-specific context providers. Depends on: fixtures-today.
Injuries: No provider-backed injury feed is connected.
Decision impact: Material team news can downgrade or flip the action before kickoff.
Tool execution audit is blocked: Load team/player historical results, Fetch injuries, suspensions, lineups, and news must run before the current decision can be trusted.
Load team/player historical results is blocked: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Outputs: team-history, player-history, long-form
high priority - executed - api-football-standingsLoad league standings - 1 record(s)
Load league standings executed from api-football-standings with 1 observed record(s).
Outputs: league-standings, motivation-context
high priority - executed - api-football-recent-fixtures + api-football-recent-fixturesCompute recent form and home/away profile - 10 record(s)
Compute recent form and home/away profile executed from api-football-recent-fixtures + api-football-recent-fixtures with 10 observed record(s).
medium priority - waiting - Live score and event providerFetch live score and match events - 0 record(s)
Fetch live score and match events is waiting: Wait until kickoff or in-play mode before live score and event polling becomes active.
Outputs: live-score, match-events, in-play-state
medium priority - executed - open-meteo-forecastFetch weather where relevant - 1 record(s)
Fetch weather where relevant executed from open-meteo-forecast with 1 observed record(s).
Outputs: weather, wind-rain-temperature
Next run: Load team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Control policy
Statusblocked
Visibilityinternal only
Modeblocked
Primary blockervalue-selection
Control policy blocks public display: 8 blocker gate(s), primary blocker Value selection.
Directive: Block public display and collect the required evidence first.
Next best action: Load team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
model - blockValue selection
TPS Turku is the current selected market with edge +8.0% and EV +19.3%.
market - passMarket state
TPS Turku has a resilient market buffer: current odds 2.70, fair odds 2.26, and 16% shortening tolerance before EV reaches zero.
data - watchData coverage
Data coverage is 55/100 with partial input support and 6 production gap(s).
Required: Injuries: No provider-backed injury feed is connected.
tools - blockTool execution
Tool execution audit is blocked: Load team/player historical results, Fetch injuries, suspensions, lineups, and news must run before the current decision can be trusted. Tool orchestration needs tools: 7/12 task(s) ready, 2 high-priority config gap(s), next task Load team/player historical results, readiness 72/100.
Required: Load team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
risk - blockDecision boundary
Decision boundary is blocked; nearest flip is Context-shock tolerance: -13.1% past the boundary.
Required: Do not show as public value; clear Context-shock tolerance and rerun.
risk - passHistorical discipline
No 10-year public historical evidence is attached to this decision run. Run the public historical evidence proof before using history to discipline raw model edges.
ai-review - blockAI review
AI protocol is blocked: 3 failed check(s), 1 triggered abstention gate(s), and current action avoid.
Required: Injuries: No provider-backed injury feed is connected.
operator - blockReasoning graph
Reasoning graph is blocked: 6 blocker node(s), 4 watch node(s), and final action avoid.
Forbidden: publish as value candidate; show as actionable; upgrade by AI review; invent missing data
Monitoring plan
Statusblocked
Prioritycritical
Cadence5 min
Next review01:56 PM
Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve.
critical priority - odds - due 01:54 PMRefresh bookmaker odds
Current edge is +8.0% and EV is +19.3%; a price move can erase value.
Action: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value.
critical priority - team news - due 01:54 PMInjury and suspension news
This missing signal can move the probability estimate or downgrade confidence before kickoff.
Action: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
critical priority - lineups - due 01:54 PMConfirmed lineups
This missing signal can move the probability estimate or downgrade confidence before kickoff.
Action: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
critical priority - calibration - due 01:54 PMRuntime holdout economics
Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive.
Action: Do not promote the decision until this abstention gate is cleared by fresh model, market, or provider evidence.
Stop when: Stop if TPS Turku no-vig edge falls to zero or expected value turns negative. Invalidate if TPS Turku no-vig edge falls to zero or EV turns negative. Invalidate if confirmed lineups, injuries, suspensions, weather, or live events materially oppose the current thesis. Invalidate if bookmaker prices move before the next refresh window. Blocking gate: Runtime holdout economics. Stop while abstention gate remains active: Runtime holdout economics.
Escalate when: If two high-priority monitoring tasks remain unresolved at the next review, downgrade the action to monitor or avoid. If bookmaker movement removes the value edge, remove the recommendation and rerun the committee. If confirmed team news, weather, surface, or live-event data opposes the thesis, rerun the model before showing the pick. If committee disagreement remains after the next data refresh, keep the lower-risk action. Do not raise confidence from historical learning until the real-data profile is active.
Actionability audit
Statusblocked
Postureavoid recommendation
Score70/100
Failed gates1
Actionability is 70/100: block a public recommendation until failed gates are fixed.
pass - weight 18%Value edge and EV: 96/100
TPS Turku has edge +8.0% and EV +19.3%.
warn - weight 14%Confidence and risk: 57/100
Confidence is medium and risk is medium.
Required: Keep as analysis-only watchlist unless confidence improves or risk falls.
pass - weight 12%Data quality: 92/100
Data quality is 92%.
warn - weight 10%Context coverage: 68/100
Missing signals: Injury and suspension news, Confirmed lineups.
Required: Fetch the highest-priority missing context signals and rerun the decision.
warn - weight 12%Belief freshness: 62/100
Belief is moderate: TPS Turku at 44% with +8.0% edge, +19.3% EV, uncertainty 31/100, expires in 60 minutes.
Required: Refresh odds, context, and model state before trusting this belief.
Blockers: Monitoring state: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve. Runtime holdout economics: Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive.
Before action: Keep as analysis-only watchlist unless confidence improves or risk falls. Fetch the highest-priority missing context signals and rerun the decision. Refresh odds, context, and model state before trusting this belief. Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Do not raise confidence from training until real historical data is active. Refresh bookmaker odds: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Injury and suspension news: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates. Confirmed lineups: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
Treat the output as statistical analysis, not certainty. Do not use this audit as staking, bankroll, or financial advice. Refresh odds and context before relying on any displayed edge. Avoid acting when the monitoring plan is blocked, expired, or unresolved.
Review loop
Statusblocked
Recommended actionavoid
Score delta-28
Shiftslower/raise
Review loop blocks the recommendation until failed gates or abstention rules clear.
thesis builder - challenge - medium confidencePrimary thesis is TPS Turku: edge +8.0%, EV +19.3%.
Evidence: TPS Turku is the leading thesis because the model price beats the no-vig market by +8.0% with +19.3% EV per unit. Belief is moderate: TPS Turku at 44% with +8.0% edge, +19.3% EV, uncertainty 31/100, expires in 60 minutes.
red team - block - high confidenceRed team found 2 blocking issue(s).
Evidence: Monitoring state: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve. Runtime holdout economics: Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive. Confidence and risk: Confidence is medium and risk is medium. Context coverage: Missing signals: Injury and suspension news, Confirmed lineups. Belief freshness: Belief is moderate: TPS Turku at 44% with +8.0% edge, +19.3% EV, uncertainty 31/100, expires in 60 minutes.
Required: Monitoring state: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve.
data gap checker - repair - medium confidence2 missing signal(s) remain: Injury and suspension news, Confirmed lineups.
Evidence: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve. Refresh bookmaker odds: Current edge is +8.0% and EV is +19.3%; a price move can erase value. Injury and suspension news: This missing signal can move the probability estimate or downgrade confidence before kickoff. Confirmed lineups: This missing signal can move the probability estimate or downgrade confidence before kickoff.
Required: Fetch missing context, refresh probabilities, and rerun actionability.
repair planner - repair - medium confidenceRepair plan has 6 required check(s).
Evidence: Keep as analysis-only watchlist unless confidence improves or risk falls. Fetch the highest-priority missing context signals and rerun the decision. Refresh odds, context, and model state before trusting this belief. Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Refresh bookmaker odds: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Injury and suspension news: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
Required: Keep as analysis-only watchlist unless confidence improves or risk falls.
final reviewer - block - low confidenceFinal reviewer recommends avoid; review loop status is blocked.
Evidence: Actionability is 70/100: block a public recommendation until failed gates are fixed. Decision can be shown with normal responsible-use language, while still requiring fresh odds before action. Risk shift: raise; confidence shift: lower.
Repairs: Keep as analysis-only watchlist unless confidence improves or risk falls. Fetch the highest-priority missing context signals and rerun the decision. Refresh odds, context, and model state before trusting this belief. Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Refresh bookmaker odds: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Injury and suspension news: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
Unresolved: Monitoring state: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve. Runtime holdout economics: Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive. Confidence and risk: Confidence is medium and risk is medium. Context coverage: Missing signals: Injury and suspension news, Confirmed lineups. Belief freshness: Belief is moderate: TPS Turku at 44% with +8.0% edge, +19.3% EV, uncertainty 31/100, expires in 60 minutes. Missing signal: Injury and suspension news. Missing signal: Confirmed lineups. Monitoring is blocked.
Release criteria: Fresh odds must keep no-vig edge and expected value positive. Belief state must be unexpired before the decision remains visible. Monitoring plan must be active or explicitly cleared. Actionability must stay actionable or the product should downgrade to watch/avoid. Keep as analysis-only watchlist unless confidence improves or risk falls. Fetch the highest-priority missing context signals and rerun the decision. Refresh odds, context, and model state before trusting this belief.
Robustness stress test
Statusfragile
Score10/100
Survival rate0%
Worst caseActionability downgrade
Robustness is 10/100: only 0/6 stress tests survive, so the recommendation is fragile.
Edge -4.0%; EV -13.1%. Actionability status is blocked with score 70/100.
Repair: Keep as analysis-only watchlist unless confidence improves or risk falls.
Hedges: Goals: Over 2.5 Goals at model 58%. Both Teams To Score: BTTS Yes at model 58%. Double chance: Turku PS or Draw at model 69%.
Required rechecks: Bookmaker price shortens: Refresh bookmaker odds and remove the pick if edge or EV is no longer positive. Adverse team news or weather: Fetch lineups, injuries, suspensions, weather, and news before keeping the same action. Data quality decays: Improve provider coverage or downgrade confidence until real data fills the gap. Belief expires before refresh: Rerun the belief state and monitoring plan before showing the candidate again. Review-loop repair pressure: Keep as analysis-only watchlist unless confidence improves or risk falls. Actionability downgrade: Keep as analysis-only watchlist unless confidence improves or risk falls. Refresh bookmaker odds: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value. Injury and suspension news: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
Evaluation plan
Statusno action
SettlementTPS Turku
Model prob.44%
Break-even37%
Quoted odds2.70
Target CLVN/A
Evaluation plan records the abstention so future calibration can learn whether avoiding the market was correct.
required - resultSettled match and market result
Settle whether TPS Turku won for Turku PS vs Ilves.
required - closing-oddsClosing odds snapshot
Capture closing odds for TPS Turku and compare them with the quoted 2.70.
required - contextLate context resolution
Record whether monitoring tasks changed the thesis before kickoff: Refresh bookmaker odds, Injury and suspension news, Confirmed lineups.
required - calibrationCalibration outcome row
Link the settled outcome to this decision run so confidence, health, Brier score, ROI, and CLV can be measured.
Success: TPS Turku settles as correct for the chosen market. Closing-line value is at least positive or the closing no-vig probability confirms the edge. Settled outcome improves calibration for medium-confidence medium-risk decisions. No unresolved review-loop release criterion would have blocked the pick at kickoff.
Failure: TPS Turku loses or pushes against the selected market settlement rules. Closing odds move against the thesis enough to erase the pre-match value edge. A required recheck was missed: Bookmaker price shortens: Refresh bookmaker odds and remove the pick if edge or EV is no longer positive. The outcome joins similar stored cases that later discount this pattern.
Learning questions: Was 44% model probability calibrated against the binary settlement result? Did the no-vig market probability of 36% underprice the selection at decision time? Did closing odds validate the edge or expose stale market data? Did the unresolved monitoring or review-loop checks predict the final risk?
Post-match actions: Store the settled outcome through the decision outcome endpoint with the linked decision_run_id. Recompute calibration by confidence and decision health after settlement. Compare quoted odds with closing odds for closing-line value. Keep learned thresholds inactive until enough real historical fixtures and odds are imported.
Case memory
Statusready
Similar cases4
Adjustmentnone
Avg similarity21%
Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
similarity 21% - avoid - stableHome
api-football:1513227; score 79; reliability 93/100; different value/no-value state, quality and reliability proximity
similarity 21% - monitor - stableHome
api-football:1547591; score 55; reliability 93/100; different value/no-value state, quality and reliability proximity
similarity 20% - avoid - stableHome
api-football:1532827; score 74; reliability 98/100; different value/no-value state, quality and reliability proximity
Learning profile
Training statusshadow only
Real fixtures6114/1000
Learned edge+4.0%
Backtest yieldNo data
Latest real-data backtest is available for shadow comparison only; live guardrails require an active model-bound calibration promotion.
Decision score
Score60
Actionavoid
Confidencemedium
Riskmedium
Agent stages
Fixture and market intake - passed78/100
Turku PS vs Ilves; data quality 92%; missing signals 2.
Model integrity check - passed100/100
Model produced 1.89-1.12 expected goals and winner total 1.000.
Market edge search - passed100/100
TPS Turku is the current best edge at +8.0%.
Risk and variance gate - warning58/100
TPS Turku is classified as medium risk.
Contradiction and abstention review - warning64/100
1 watch/conflict checks and 1 triggered abstention gates.
Final arbitration - passed100/100
Weighted decision score 60; the agent will only surface the selection if guardrails agree.
Decision committee
Consensusleaning
Final actionavoid
Consider votes4
Avoid votes0
The public-action invariant downgraded the committee candidate from consider to avoid. Decision committee recommends consider with leaning consensus: stable decision: surface TPS Turku as inspectable value, but require fresh odds and context checks before action.
model advocate - neutral - votes considerModel advocate
TPS Turku is mathematically live because model probability is 44% against no-vig 36%.
Challenge: Injury and suspension news: No live injury/suspension provider is connected yet. Confirmed lineups: Lineups are not available in the MVP mock provider.
Checks: Recompute probabilities after fresh odds and context updates.
Reliability is 84/100 with stable health and trust calibration action.
Challenge: Runtime holdout economics: Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive.
Checks: Honor triggered abstention gates before any public recommendation.
Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
Challenge: Compared against 40 stored football decisions. Average similar-case reliability is 96/100.
Checks: Store this decision and settle outcomes so memory can learn from real results.
final arbiter - neutral - votes considerFinal arbiter
stable decision: surface TPS Turku as inspectable value, but require fresh odds and context checks before action.
Challenge: Injury and suspension news Confirmed lineups Fresh bookmaker price
Checks: Downgrade to monitor or avoid if adverse lineup, injury, weather, live-state, or data-quality signals oppose the selected side. Recalculate immediately; if the no-vig edge falls below zero or EV turns negative, remove TPS Turku from value picks.
Open disagreements: Final public action is avoid; the pre-invariant committee candidate was consider. Market skeptic voted monitor: The price still has to survive market scrutiny; odds shortening would change the edge on TPS Turku. Context scout voted monitor: Context is incomplete for Turku PS vs Ilves; 2 missing signals remain.
Agent deliberation
Primary thesisTPS Turku is the leading thesis because the model price beats the no-vig market by +8.0% with +19.3% EV per unit.
Dissenting thesisThe market may be pricing context the MVP cannot fully see yet, especially Injury and suspension news, Confirmed lineups.
Synthesisstable decision: surface TPS Turku as inspectable value, but require fresh odds and context checks before action.
supported - medium confidenceValue thesis: TPS Turku
TPS Turku carries model probability 44%, no-vig implied 36%, odds 2.70, and EV +19.3%.
Impact: Supports showing the pick as inspectable value.
needs data - high confidenceCounter-thesis: market may know more
The selection still depends on a live bookmaker price. Market movement scenario projects consider at score 53.
Impact: Requires odds refresh before the recommendation can be trusted.
needs data - high confidenceContext-risk thesis
2 missing context signals remain, led by Injury and suspension news, Confirmed lineups.
Impact: Controls whether the engine considers, monitors, or abstains after late provider data arrives.
supported - high confidenceFinal arbitration
Base case decision score is 60; calibration says trust with 84/100 reliability.
Impact: stable decision: surface TPS Turku as inspectable value, but require fresh odds and context checks before action.
supported - medium confidenceCase-memory thesis
Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
Impact: Memory stays neutral.
What would change the decision
high priority - team newsInjury and suspension news
This missing signal can move the probability estimate or downgrade confidence before kickoff.
If confirmed: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
high priority - lineupsConfirmed lineups
This missing signal can move the probability estimate or downgrade confidence before kickoff.
If confirmed: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
high priority - oddsFresh bookmaker price
A shorter price can erase the +8.0% edge and +19.3% EV.
If confirmed: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and EV.
medium priority - trainingHistorical learning profile
Without enough real historical fixtures and odds, thresholds stay conservative defaults.
If confirmed: Import real fixtures and odds, run backtests, then activate learned thresholds only after the corpus is large enough.
medium priority - trainingSimilar stored decisions
Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
If confirmed: Keep memory neutral while continuing to store and settle outcomes.
Bad-data path: Downgrade to monitor or avoid if adverse lineup, injury, weather, live-state, or data-quality signals oppose the selected side. Market-move path: Recalculate immediately; if the no-vig edge falls below zero or EV turns negative, remove TPS Turku from value picks.
Context adjustment
Home shift+0.6%
Away shift0.0%
Total shift0.0%
Signals2
Context layer applied residual probability effects after reviewing 2 structured signals.
standings - acceptable - home positiveLeague standings context
Turku PS rank 7, 22 pts; Ilves rank 9, 16 pts.
weather - acceptable - neutralWeather check
Turku, Finland forecast near kickoff: partly cloudy, 21C, wind 6.8 m/s, precipitation probability 0%.
Weighted factors
Value edge - score 72 - weight 0.3223
Positive no-vig edge of +8.0% versus margin-adjusted implied probability.
Expected value - score 100 - weight 0.1616
Expected return is +19.3% per unit at decimal odds 2.70.
Confidence guardrail - score 55 - weight 0.211
medium confidence after edge, probability, and data-quality checks.
Data quality - score 47 - weight 0.188
Data quality is 92%.
Outcome variance - score 5 - weight 0.141
medium risk based on confidence and odds level.
Missing context - score -8 - weight 0.1-1
Lineups, injury/suspension news, and weather are not connected yet.
Live state - score 0 - weight 0.060
Pre-match state avoids in-play recalibration risk.
Historical learning - score 0 - weight 00
Latest real-data backtest is available for shadow comparison only; live guardrails require an active model-bound calibration promotion.
Case memory - score 28 - weight 0.072
Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
Historical discipline - score 0 - weight 00
Run the public historical evidence proof before using history to discipline raw model edges.
Self-critique checks
Probability normalization - clear
Home/draw/away probabilities sum to 1.000.
Confidence versus data quality - clear
Confidence is medium with 92% data quality.
Recommendation versus missing context - clear
Missing context does not contradict the current action.
Risk versus action - watch
TPS Turku is medium risk while action is consider.
Live state versus model - clear
Fixture is not live, so pre-match modeling is coherent.
Public reasoning trace
Read fixture context, team strength, recent form, odds, and data quality.
Audited data coverage: Data coverage is 55/100 with partial input support and 6 production gap(s).
Simulated scorelines from expected goals: 1.89-1.12.
Applied Dixon-Coles low-score correction with rho -0.0488 before deriving football winner, totals, BTTS, and scoreline probabilities.
Context layer applied residual probability effects after reviewing 2 structured signals. Missing context still tracked: Confirmed lineups, Injury and suspension news.
Core model context signal: Provider football context xG: Poisson expected goals consumed 1 provider/context injury, suspension, lineup, player-form, weather, or news signal(s). Goal adjustments 0.00-0.00..
Applied market prior to 1 priced market with average weight 58% and average bookmaker margin 10%.
Compared model probabilities with no-vig market probabilities and EV; strongest reviewed edges: TPS Turku +8.0% edge, +19.3% EV, Draw -1.4% edge, -7.3% EV, Ilves Tampere -6.6% edge, -20.0% EV.
Ran odds intelligence: Odds intelligence found 1 actionable value candidate(s) across 1 market(s); best is TPS Turku.
Audited market movement: TPS Turku has a resilient market buffer: current odds 2.70, fair odds 2.26, and 16% shortening tolerance before EV reaches zero.
Scored decision factors at 60; selected TPS Turku only because edge and confidence passed the guardrail.
Audited tool execution: Tool execution audit is blocked: Load team/player historical results, Fetch injuries, suspensions, lineups, and news must run before the current decision can be trusted.
Applied control policy: Control policy blocks public display: 8 blocker gate(s), primary blocker Value selection.
Synthesized deliberation: stable decision: surface TPS Turku as inspectable value, but require fresh odds and context checks before action.
Ran decision committee: 4 consider, 2 monitor, 0 avoid; final action avoid with leaning consensus.
Built monitoring plan: Monitoring is blocked with critical priority until guardrails, missing signals, or market value improve.
Audited actionability: Actionability is 70/100: block a public recommendation until failed gates are fixed.
Ran review loop: Review loop blocks the recommendation until failed gates or abstention rules clear.
Compiled research brief: Turku PS vs Ilves is blocked from a public recommendation. Block public recommendation until failed gates, missing data, or invalid memory conditions are repaired.
Opened decision notebook: Notebook is blocked: 2 falsifier or abstention condition(s) prevent public trust.
Stress-tested robustness: Robustness is 10/100: only 0/6 stress tests survive, so the recommendation is fragile.
Registered evaluation plan: Evaluation plan records the abstention so future calibration can learn whether avoiding the market was correct.
Compared with stored case memory: Case memory found 4 similar stored decisions and did not find enough historical pressure to downgrade.
Historical learning profile: shadow-only; Latest real-data backtest is available for shadow comparison only; live guardrails require an active model-bound calibration promotion.
Applied historical discipline: No 10-year public historical evidence is attached to this decision run. Instruction: Run the public historical evidence proof before using history to discipline raw model edges.
Checked missing context such as confirmed lineups, injuries, weather, and live events before final verdict.
Evidence reviewed
model - acceptable - positivePoisson expected-goals model
Turku PS expected goals 1.89, Ilves expected goals 1.12; top scoreline 1-1.
market - acceptable - positiveNo-vig market probability comparison
TPS Turku has 44% model probability versus 36% no-vig implied probability. Raw implied is 37% and market margin is +9.7%. Expected value is +19.3% per unit.
form - acceptable - neutralRecent form signal
Turku PS: W-W-L-W; Ilves: D-L-L-L.
data quality - strong - neutralData quality guardrail
Data quality is 92%; low quality downgrades confidence.
team news - missing - unknownInjury and suspension news
No live injury/suspension provider is connected yet.
lineups - missing - unknownConfirmed lineups
Lineups are not available in the MVP mock provider.
model - acceptable - neutralCore model context adjustment
Model 69% - fair odds 1.44 - needs bookmaker market
Risks and next checks
Sports outcomes are uncertain; the decision engine is an analysis layer, not a guarantee.
Bookmaker prices can move quickly, so stale odds can erase a calculated edge.
Refresh bookmaker odds and recalculate no-vig market probabilities before showing a final value edge.
Check injuries, suspensions, and confirmed lineups close to kickoff.
Add weather for outdoor football totals and tempo-sensitive markets.
Re-check after 02:51 PM; this belief expires in 60 minutes.
Refresh bookmaker odds by 01:54 PM; Current edge is +8.0% and EV is +19.3%; a price move can erase value.
Keep as analysis-only watchlist unless confidence improves or risk falls.
Fetch the highest-priority missing context signals and rerun the decision.
Refresh bookmaker odds: Recalculate implied probability, bookmaker margin, no-vig probability, value edge, and expected value.
Injury and suspension news: Refresh the model, rerun the no-vig EV calculation, and re-check abstention gates.
Refresh bookmaker odds: Complete this check, then rerun the decision engine before trusting the current posture.
Injury and suspension news: Complete this check, then rerun the decision engine before trusting the current posture.
Abstention gate triggered without mutating probability: Runtime holdout economics.
Review loop: Review loop blocks the recommendation until failed gates or abstention rules clear.
Data coverage: Data coverage is 55/100 with partial input support and 6 production gap(s).
Injuries: No provider-backed injury feed is connected.
Fetch or verify Injury and suspension news before raising confidence.
Context-shock tolerance breached: Worst-case stress still leaves minimum edge/EV margin at -13.1%.
Runtime holdout economics: Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive.
Fetch lineups, injuries, suspensions, and news: Injuries: No provider-backed injury feed is connected.
Load historical training profile: Latest real-data backtest is available for shadow comparison only; live guardrails require an active model-bound calibration promotion.
Next tool task historical-results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Load team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Control policy: Load team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Load team/player historical results: Historical evidence is incomplete. Turku PS: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches); Ilves: provider-backed via api-football-recent-fixtures-form-rating-v1 (4 matches). Player history: missing; no leakage-safe player-performance form was attached.
Fetch injuries, suspensions, lineups, and news: Injuries: No provider-backed injury feed is connected.
Model probability: Runtime probability trace is blocked: final probability 44% is preserved, but it does not clear the active action gates.
Data coverage: Data coverage is 55/100 with partial input support and 6 production gap(s).
Remove or downgrade if odds shorten more than 16% from the current quote.
A 5% odds shortening does not fully break the thesis, but still requires a refresh.
Fresh odds must keep no-vig edge and expected value positive.
Belief state must be unexpired before the decision remains visible.
Bookmaker price shortens: Refresh bookmaker odds and remove the pick if edge or EV is no longer positive.
Adverse team news or weather: Fetch lineups, injuries, suspensions, weather, and news before keeping the same action.
Injuries: No provider-backed injury feed is connected.
Suspensions: No suspension provider is connected.
Settled match and market result: Settle whether TPS Turku won for Turku PS vs Ilves.
Closing odds snapshot: Capture closing odds for TPS Turku and compare them with the quoted 2.70.
Import enough real historical fixtures and odds before learned thresholds tune live decisions.
Attach public historical evidence or a persisted provider-backed learning profile.
Keep learned thresholds and public picks locked until provider-enriched backtests pass.
Backtest this market against historical closing odds before raising confidence in production.
Sensitivity checks
Odds shorten by 5%: keeps verdict. Edge would shrink to about +9.3%.
Adverse lineup or injury news: keeps verdict. A moderate negative availability or context adjustment would reduce the decision score by about 18 points.
Data-quality upgrade: keeps verdict. Confirmed availability, lineups, injuries, weather, or sport-specific context can improve confidence if they support the existing model signal.
Scenario matrix
Base case - consider60 (+0)
Current model, odds, data-quality, and missing-context state.
Odds shorten by 5% - consider53 (-7)
Projected edge would be +9.3%.
Adverse team news - consider42 (-18)
A moderate adverse injury, suspension, or lineup shock is applied to the selected side.
Confirmed context supports model - consider63 (+3)
Lineups, injury/suspension news, and weather arrive and support the existing model signal.
Abstention gates
Positive edge and EV required - clear
TPS Turku clears the positive-edge filter at +8.0% and EV +19.3%.
Data quality floor - clear
Data quality is 92%; below 62% the agent abstains.
High-variance edge - clear
TPS Turku risk is medium; high-risk picks need a wider edge before consideration.
Live model requirement - clear
Pre-match fixture; in-play event model is not required yet.
Thin context around edge - clear
Missing context is noted but does not force abstention at the current edge/data-quality level.
Future fixture synthetic market - clear
Fixture has no future-season seed or synthetic market flag.
Runtime holdout economics - triggered
Public value picks remain blocked: exact-runtime holdout yield is not positive; exact-runtime closing-line value is not positive.
Learned minimum edge - clear
No real-data training profile is active; default minimum-edge guardrail is used.
Case memory abstention - clear
Case memory did not find enough similar stored pressure to force abstention.
Historical market-prior discipline - clear
No public historical discipline evidence is attached, so this gate stays neutral.
Prediction agent report
Verdict: no clear value
low uncertainty
The agent does not see enough separation between OddsPadi probabilities and no-vig market probabilities for Turku PS vs Ilves.
Reasons
The strongest available edges do not meet the confidence threshold.
Expected goals are 1.89 for Turku PS and 1.12 for Ilves.
Top scoreline estimate: 1-1.
Cautions
Sports outcomes remain uncertain; this is statistical analysis only.
Do not treat positive edge as a requirement to place a bet.
Model football-poisson-v5; data quality 92%.
Model mathematics
Home expected goals1.89
Away expected goals1.12
Total expected goals3.01
Winner probability sum1.000
Top projected scores
1-111%
2-110%
1-09%
2-09%
3-16%
Signals
Home rating edge: 0.11 Positive values favor the home side before market comparison. Turku PS: api-football-recent-fixtures-form-rating-v1, 4 match sample; Ilves: api-football-recent-fixtures-form-rating-v1, 4 match sample.
Recent form edge: 0.631 Computed from W/D/L form with exponential recency weighting; the latest result carries the most weight.
Expected goals total: 3.014 Derived from attack, defense, rating, registry home advantage (1.10), and league goal-rate assumptions.
xG blend adjustment: 0 No provider xG inputs were available, so expected goals used rating, form, attack, defense, league rate, and home advantage only.
Finland
Turku PS🇫🇮Finlandvs
Ilves🇫🇮Finland