Sample analysis
The exact page members saw before kick-off — now with the final score.
This analysis was published before the match on October 17, 2026; it finished 3–2. Members get the same page for every match across 11 leagues, before kick-off.
EredivisieToday FinishedWon
Sparta Rotterdam – Willem II
Sparta Rotterdam3
Willem II2- 1Sparta Rotterdam
- 65%
- Market 1.50 · +3
- XDraw
- 20%
- Market 4.50 · −1
- 2Willem II
- 15%
- Market 5.50 · −2
- Exp. goals (home)
- 2.34
- Exp. goals (away)
- 1.04
- Market implied
- 62%@1.50
- Edge vs market
- +3 pt
Assessment for this match
Model, market and our own record, judged separately for each market. A status describes the evidence, not a new win probability.
- Match result (1X2)Sparta RotterdamEvidence insufficient
- Model
- 65%
- calibrated 72%
- Market
- 62%
- @1.50 · closing
- Record
- 55%
- 6/11 · gap 0…+5 pts · league
Similar picks hit 6/11 (55%), but the 95% range (28–79%) neither clears the 55% bar nor rules the pick out.
- Over/Under 2.5Over 2.5Evidence insufficient
- Model
- 66%
- calibrated 59%
- Market
- –
- no price captured
- Record
- 79%
- 23/29 · model 60–70% · league
Similar picks hit 23/29 (79%), but the 95% range (62–90%) neither clears the 65% bar nor rules the pick out.
No bookmaker price captured.
- Both teams to scoreBTTS YesEvidence insufficient
- Model
- 59%
- calibrated 54%
- Market
- 59%
- Record
- 72%
- 13/18 · model 50–60% · league
Similar picks hit 13/18 (72%), but the 95% range (49–88%) neither clears the 65% bar nor rules the pick out.
Historical evidence analysis
What the stored model, the calibration curve, the bookmaker price and comparable past matches say about each selection, side by side. Figures are computed by code from recorded data; nothing here is a new win probability.
- Match result (1X2)Sparta RotterdamEvidence insufficient
- Model
- 65%
- calibrated 72%
- Bookmaker
- 62%
- closing · Sat, Oct 17, 20:10
- Similar matches
- 56%
- 14/25 · 95% 37–73%
- Sparta Rotterdam at home
- –
- too few matches
- Willem II away
- –
- too few matches
- In 25 comparable matches the selection came in 56% of the time (95% 37–73%); the model's 65% sits inside that range. Because neighbours are chosen by the model's own expected goals, this is a calibration check, not independent evidence.
- Sparta Rotterdam's home record in the pool is thin (6 matches, fewer than 10); no rate is shown.
- Willem II's away record in the pool is thin (2 matches, fewer than 10); no rate is shown.
- Model 65% and bookmaker 62% are within 5 points of each other.
- Over/Under 2.5Over 2.5Evidence insufficient
- Model
- 66%
- calibrated 59%
- Bookmaker
- –
- no price recorded
- Similar matches
- 68%
- 17/25 · 95% 48–83%
- Sparta Rotterdam at home
- –
- too few matches
- Willem II away
- –
- too few matches
- In 25 comparable matches the selection came in 68% of the time (95% 48–83%); the model's 66% sits inside that range. Because neighbours are chosen by the model's own expected goals, this is a calibration check, not independent evidence.
- Sparta Rotterdam's home record in the pool is thin (6 matches, fewer than 10); no rate is shown.
- Willem II's away record in the pool is thin (2 matches, fewer than 10); no rate is shown.
- No bookmaker price was recorded for this match, so the model cannot be compared with the market.
- Both teams to scoreBTTS YesEvidence insufficient
- Model
- 59%
- calibrated 54%
- Bookmaker
- 59%
- latest · –
- Similar matches
- 44%
- 11/25 · 95% 27–63%
- Sparta Rotterdam at home
- –
- too few matches
- Willem II away
- –
- too few matches
- In 25 comparable matches the selection came in 44% of the time (95% 27–63%); the model's 59% sits inside that range. Because neighbours are chosen by the model's own expected goals, this is a calibration check, not independent evidence.
- Sparta Rotterdam's home record in the pool is thin (6 matches, fewer than 10); no rate is shown.
- Willem II's away record in the pool is thin (2 matches, fewer than 10); no rate is shown.
- Model 59% and bookmaker 59% are within 5 points of each other.
Teams' own history (in the pool)
- Sparta Rotterdam at home
- 6 matches · Jul 11–Sep 19 · home 17% / draw 50% / away 33% · over 2.5 50% · BTTS 33%
- model average in these matches: home 34% / draw 25% / away 42%
- Willem II away
- 2 matches · Aug 8–Oct 11 · home 100% / draw 0% / away 0% · over 2.5 100% · BTTS 100%
- model average in these matches: home 60% / draw 22% / away 19%
Comparable past matches
nearest by the model's expected goals (2.34 / 1.04) · radius 0.35 · scope: covered leagues (this league had too few) · 25 of 25 within radius
25 matches · Jul 18–Oct 17 · home 56% / draw 28% / away 16% · over 2.5 68% · BTTS 44%
model average in these matches: home 65% / draw 20% / away 15%
- Sep 14 Inter – Udinese5–3 · λ 2.33/0.98 · d 0.10
- Oct 9 Fluminense – Coritiba4–0 · λ 2.28/1.01 · d 0.10
- Sep 18 FC Groningen – PEC Zwolle3–0 · λ 2.32/1.11 · d 0.11
- Sep 14 Braga – Estoril1–0 · λ 2.25/1.00 · d 0.15
- Oct 10 Hoffenheim – Hamburger SV2–3 · λ 2.45/1.04 · d 0.16
- Sep 5 West Ham United – Derby County3–0 · λ 2.26/0.96 · d 0.17
- Oct 17 Villarreal – Elche0–0 · λ 2.45/1.08 · d 0.18
- Aug 19 Atletico Madrid – Malaga2–0 · λ 2.43/1.12 · d 0.18
Not available
- xG: not part of the live model (dc-2.0-xg is a shadow version); no timestamped per-match history, so not computed.
- Elo: research only; no per-match pre-match snapshot in the database, so not computed.
- Comparable matches were drawn from all covered leagues because this league had too few within the radius.
- Home team's home record: 6 matches, below 10.
- Away team's away record: 2 matches, below 10.
Model breakdown
The pick is Sparta Rotterdam win at a calibrated 72% (raw model 65%). Expected goals 2.34–1.04. The model's probability sits 3 points above the market's implied price at closing odds (Bet365_default). Odds movement (4 snapshots): Sparta Rotterdam 1.62 → 1.57 (+2 pts). No notable movement since opening. Sparta Rotterdam won 0 of their last 5; Willem II won 0 of their last 5.
Where this match sits in our record
For each market, the bucket this match falls into and our past hit rate in that bucket. The marker is the past hit rate; the 50 and 65 lines are thresholds. Not a guarantee, a record of what happened in the same situation.
edge vs market 0…+5 pts · Eredivisie: 6/11 hit (55%)
model level 60–70% · Eredivisie: 7/10 hit (70%)
model level 60–70% · Eredivisie: 23/29 hit (79%)
1X2 disagreement <5 pts · Eredivisie: 8/12 hit (67%)
model level 50–60% · Eredivisie: 13/18 hit (72%)
edge vs market 0…+5 pts · Eredivisie: 21/25 hit (84%)
Bucket edges and the full tables are on the record page. Open the record
Match result (1X2)
| Outcome | Model | Fair odds | Market odds | Market | Edge |
|---|---|---|---|---|---|
| 1Sparta Rotterdam | 65% | 1.53 | 1.50 | 62% | +3 |
| XDraw | 20% | 5.00 | 4.50 | 21% | −1 |
| 2Willem II | 15% | 6.87 | 5.50 | 17% | −2 |
Pick: Sparta Rotterdam. Calibrated probability 72% (raw model output 65%). Raw model output and calibrated probability are different things: the raw figure is what the model produced, the calibrated one is what past results say that raw figure has actually delivered. Calibration curve: covered, 618 matches, fitted Oct 12. Edge = model probability minus the bookmaker's margin-free probability, in percentage points. In our record a positive edge has not been in our favour: when the model sat 5+ points above the market, picks landed about 30%; within −5…+1 points, 50–56%. The market's number is the better-calibrated one. Double chance 1X: 85%.
Goals
| Line | Over | Under |
|---|---|---|
| 1.5 | 85% | 15% |
| 2.5 | 66% | 34% |
| 3.5 | 44% | 56% |
| Both teams to score | Yes59% | No41% |
Most likely scores
model · bookmaker- 2–19.7%5.9%
- 2–09.3%10.5%
- 1–18.3%11.1%
- 1–08.0%10.0%
- 3–17.6%9.1%
- 3–07.3%8.3%
Bookmaker column: implied probability from the listed correct-score price, margin included (only the prices the book lists).
stored: Calibrated calls: Over 2.5 59% · BTTS Yes 59%. The table above is the raw Poisson grid from the expected goals.
Asian handicap
| Sparta Rotterdam line | Home covers | Push | Away covers |
|---|---|---|---|
| -1.5 | 44% | – | 56% |
| -1 | 44% | 23% | 34% |
| -0.5 | 66% | – | 34% |
| 0 | 66% | 18% | 15% |
| +0.5 | 85% | – | 15% |
| +1 | 85% | 10% | 5% |
| +1.5 | 95% | – | 5% |
Whole and half lines only. The fair line is where both sides are closest to even after pushes; it is information, not a recommendation.
Other markets vs bookmaker
| Market | Selection | Model | Fair odds | Market odds | Market | Edge |
|---|---|---|---|---|---|---|
| Asian handicap | Sparta Rotterdam -1 | 56% | 1.77 | 1.83 | 52% | +4 |
| Asian handicap | Willem II +1 | 44% | 2.30 | 1.98 | 48% | −4 |
| Draw no bet | Sparta Rotterdam | 82% | 1.22 | 1.20 | 78% | +3 |
| Draw no bet | Willem II | 18% | 5.49 | 4.33 | 22% | −3 |
| Both teams to score | Yes | 59% | 1.71 | 1.57 | 59% | 0 |
| Both teams to score | No | 41% | 2.41 | 2.25 | 41% | 0 |
Two-way markets with the bookmaker's margin removed. Asian handicap uses the book's main line; on quarter lines the model figure averages the two neighbouring lines. Draw no bet is the model's home/away split with the draw removed. Information, not a recommendation. The edge column is model minus market: +5 points or more is shown in red — in our record, picks landed less often when the model sat this far above the market (1X2 30%, BTTS 48%, O/U 36%).
Absences & line-ups
No absence or line-up data for this match yet; it arrives within 48 hours of kick-off.
Source: API-Football. Absences are informational; the model does not account for them yet.
Model
- Version
- dc-1.0
- Poisson 1X2
- 66% / 18% / 15%
- Published
- Sat, Oct 17, 23:20
- Status
- Finished · Settled
The headline 1X2 figures come from the Dixon-Coles fit, which corrects low-scoring draws. The Poisson grid on this page does not apply that correction, so its 1X2 split can differ by a point or two.
The same analysis is ready for today's matches.
7 days of full access. No card, cancel anytime. Every pick stays on the record, win or lose.
Probabilities are model estimates and are not guarantees. This page is not betting advice. 18+.