Skip to main content
← All predictions

J1 League · 17th place

Mito Hollyhock crestMito HollyhockvsV-varen Nagasaki crestV-varen Nagasaki

Kickoff: Saturday, June 6, 2026 at 2:00 AM ET · Best Denki Stadium, Mito

Outcome Probabilities

Confidence 5 out of 5 confidence

Home

Mito Hollyhock

33%

Draw

Draw

34%

Away

V-varen Nagasaki

33%

Lean: Draw · close call (1pp margin)

Expected goals: 1.2—1.2(total 2.4)
Upset risk: 5/10

Prediction Evolution

How the model's confidence changes as new info arrives

HomeDrawAway
25%50%75%33%33%33%?T−48hT−24hPre-matchHalf-time

Home win moved from 33% (T−48h) to 33% (Pre-match) → 0pp · 1 stage still to come

Baseline ensemble is well-calibrated; component models show moderate disagreement (xG outlier at 49.2% away vs Dixon-Coles 35.4%), signaling low conviction. Both teams are mid-table strugglers with inconsistent recent form (Mito 1W–2L–2D; Nagasaki 2W–3L). No injuries to either squad. Home advantage and Elo's modest +50 differential warrant a small +0.8pp lift to pH, with corresponding minor reductions to pD and pA. Fixture is genuinely competitive; stay close to baseline.

SPX AI Quant Prediction Model · DC 0.81 / Elo 0.19 / xG 0 · 55/45 weighted ensemble · stats + LLM

Pre-match stage hash · locked 6/6/2026, 3:00:19 AM

c0a7e21e244ec5f505c7ef4fccc60202e2d2bec781d7e19beec842cd898cfa2a

SPX AI Quant Prediction Model

DC 0.81 · Elo 0.19 · xG 0

H 33%D 34%A 33%

Dixon-Coles (Poisson)

H 33%D 32%A 35%

xG: 1.19 — 1.25

Elo (Davidson 3-way)

H 32%D 44%A 24%

Elo gap: +50 pts

xG-Poisson

H 25%D 26%A 49%

Expected goals: 1.01 — 1.53

LLM contextual review

H 33%D 34%A 33%

Based on recent form, news, injuries

Data quality score: 62/100 · verdict publish

Integrity Lock

This prediction's latest stage was generated at 2026-06-04T00:00:40.859Z and its canonical JSON hashed with SHA256. Each prior stage carries its own hash inside the corresponding tab so you can audit every step independently.

78f4e6bf95f880245522ab1a771753a4b0f355e77a96a58b609c05bfdae73ede
✓ Hash verified — contents unchanged

Verify yourself via the API.

How the model works →