Premier League (EPL) predictions & betting tips – matchweek 18th

Predictions for this week

Manchester United vs Manchester City

Sat 17 Jan 12:30

head to head

1X2

Manchester City 48%

BTTS

BTTS Yes 59%

Over/Under 2.5

Over 2.5 57%

Liverpool vs Burnley

Sat 17 Jan 15:00

head to head

1X2

Liverpool 51%

First team to score

Liverpool 57%

Over/Under 2.5

Under 2.5 53%

Tottenham vs West Ham

Sat 17 Jan 15:00

head to head

1X2

Tottenham 52%

First team to score

Tottenham 58%

Over/Under 2.5

Under 2.5 52%

Chelsea vs Brentford

Sat 17 Jan 15:00

head to head

1X2

Brentford 41%

BTTS

BTTS Yes 61%

Over/Under 2.5

Over 2.5 59%

Leeds vs Fulham

Sat 17 Jan 15:00

head to head

1X2

Leeds 43%

BTTS

BTTS Yes 64%

Over/Under 2.5

Over 2.5 62%

Sunderland vs Crystal Palace

Sat 17 Jan 15:00

head to head

1X2

Sunderland 36%

Over/Under 2.5

Under 2.5 65%

BTTS

BTTS No 57%

Nottingham Forest vs Arsenal

Sat 17 Jan 17:30

head to head

1X2

Arsenal 53%

First team to score

Arsenal 58%

BTTS

BTTS Yes 54%

Wolves vs Newcastle

Sun 18 Jan 14:00

head to head

1X2

Newcastle 52%

BTTS

BTTS Yes 58%

First team to score

Newcastle 57%

Aston Villa vs Everton

Sun 18 Jan 16:30

head to head

1X2

Aston Villa 50%

First team to score

Aston Villa 56%

Over/Under 2.5

Under 2.5 52%

Brighton vs Bournemouth

Mon 19 Jan 20:00

head to head

1X2

Brighton 48%

Over/Under 2.5

Over 2.5 68%

BTTS

BTTS Yes 67%

Premier League (EPL) predictions for this week

This week’s board looks far more comfortable in the goal markets than in clean 1X2 winner calls. Most 1X2 leans live in the 38–57% range — meaning the model is naming the likeliest single result, but not with “true favourite” separation. The loudest signals are about game-state: multiple fixtures sit in the 60–69% band for BTTS Yes and/or Over 2.5, which usually points to open matches, fragile control, and momentum swings.

Where the model is leaning (1X2)

The 1X2 slate is a cluster of preferences rather than convictions. The top lean is Liverpool (57%), followed by Arsenal (53%) and Manchester City (51%). After that, it’s mostly low-to-mid 40s — volatile territory where the “other outcomes” still dominate the probability space.

Editorially, that table is the story: this isn’t a “safe winners” week. The better angle is to frame several fixtures as high-event and let the goal markets carry the strongest narratives.

Manchester United vs Newcastle: the loudest goals profile

If there’s a headline match from a model-shape perspective, it’s Manchester United vs Newcastle. The 1X2 lean is modest (United 45%), but the supporting markets are about as clear as it gets this week: Over 2.5 and BTTS Yes both land at 69%.

• The model is describing an open game more than a confident winner.
• BTTS + Over at the same high level usually implies both teams create and concede chances.
• For previews, this is a strong “tempo + chaos” anchor fixture: expect events.

The strongest model leans: where the stats are loudest

The best confidence tier this week sits around 63–69% in the goal markets, plus a couple of standout “game-state” indicators (like first team to score). These are the calls that most clearly define match tone.

The editorial takeaway: the model expects several fixtures to be high-event. Brentford–Bournemouth and Chelsea–Villa read as classic “open game” profiles (both BTTS Yes + Over 2.5 at 66%), while Palace–Spurs and Forest–City also project goal flow.

Risk radar: 1X2 leans with fragile backing

This keeps the week honest. Several 1X2 calls sit in the 38–45% zone — which means the model is naming a single likeliest outcome, but it’s standing on thin ice. In these matches, “upset” isn’t a shock; the numbers already say the favourite isn’t protected.

Notice the tension: some of these fragile 1X2 calls are paired with strong goals signals (United–Newcastle, Chelsea–Villa, Palace–Spurs). That’s often the model saying it trusts the openness of the fixture more than the identity of the winner.

Predictions to treat with caution

These are markets where the model has an edge, but it’s thinner — mostly in the 53–59% zone. Useful for preview colour, but not the same headline weight as the 66–69% goal clusters.

In content terms, frame these as slight tendencies: Liverpool’s “score first” is a decent angle, Burnley–Everton leans “lower scoring”, and Arsenal–Brighton has mild support for Arsenal striking first plus a modest BTTS lean.

Putting it together for a weekly preview

• Sell the week’s identity as goal-driven: Man United–Newcastle is the flagship (Over 2.5 + BTTS Yes both at 69%), with Brentford–Bournemouth and Chelsea–Villa reinforcing the “open game” theme (both markets at 66%).
• Keep 1X2 honest: the highest winner lean is only Liverpool 57%. Most picks sit in the 40s — better treated as “most likely result” rather than “confident call”.
• Use fragile favourites as tension points: Everton (38%), Villa (39%), and Palace (40%) are exactly the kind of leans that create preview drama — the model is leaning, but not protecting.
• Use the 55–59% band for texture: Under 2.5 in Liverpool–Wolves (55%) and Under 2.5 in Burnley–Everton (59%) help diversify the narrative away from “all goals, all the time”.

In short: this week reads like a round where the model gives you goal-volume clarity in several fixtures — and asks you to be cautious with a clean list of winners.