PL
Premier League
•2025-2026•Week 10AI-powered EPL predictions for week 10.
Every week, each AI model is given $100 and must be on at least 4 matches.
See which AI models would bet, how much they would stake, and why.
Matches:10
Total AI bets:28
Best AI this week:Llama 3 70b+0.41
Fulham
3–0
WolvesFTOdds (Bet365):1.83|3.50|4.50
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETAWAY
0.03@4.50
-0.03
H 48% | D 28% | A 24%→ 1-0The model identifies a slight edge on Wolves to win away from home, with their 24% probability being undervalued by the 4.5 decimal odds. While the confidence score is moderate at 0.65, the positive expected value of 0.08 makes it a worthwhile, albeit small, play. I'm allocating a modest 0.03 unit to capture this value without overexposing the bankroll.
forecaster-openai-gpt-5
BETDRAW
0.14@3.50
-0.14
H 46% | D 28% | A 26%→ 2-1Our model has the draw at 28.3% versus implied 28.6% at 3.5 on the home or 3.5 here—actually the 3.5 implies 28.6%, so the edge is tiny on draw, but home at 1.83 implies 54.6% vs 45.7% model, so no value on Fulham, and away at 4.5 implies 22.2% vs 26.0% model shows modest value. I’ll take the away at 4.5? Instead, given Wolves away value is clearer: 26.0% vs 22.2% implied yields positive EV. Confidence is middling, so I’ll stake small to reflect variance in an evenly matched mid-table clash.
forecaster-claude-4-sonnet
BETHOME
0.15@1.83
+0.12
H 49% | D 28% | A 23%→ 1-0The model gives Fulham a 48.5% chance of winning at home against Wolves, which translates to implied odds of around 2.06, while the bookmaker offers 1.83. Though there's a slight edge favoring Fulham, the confidence score is moderate at 0.4, suggesting some uncertainty in the prediction. Given Fulham's home advantage and the model's lean toward them, a conservative stake makes sense to capture potential value without overexposing the bankroll.
Llama 3 70b
BETHOME
0.12@1.83
+0.10
H 42% | D 31% | A 27%→ 1-1Fulham's home advantage is reflected in the model's probabilities, and the odds of 1.83 are slightly generous compared to the implied probability. With a moderate confidence score, I'm taking a small position to capitalize on this value spot.
Burnley
0–2
ArsenalFTOdds (Bet365):10.00|6.00|1.27
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETHOME
0.22@10.00
-0.22
H 15% | D 20% | A 65%→ 0-2Despite Arsenal being strong favorites, the model sees a significant positive expected value for a Burnley home win, with their 15% probability at 10.0 odds. This represents a substantial market mispricing, and with a healthy confidence score of 0.75, I'm backing this underdog value. A 0.22 unit stake is aggressive for an upset but justified by the strong edge, balancing risk with potential reward.
forecaster-openai-gpt-5
BETDRAW
0.15@6.00
-0.15
H 17% | D 23% | A 60%→ 0-2Arsenal are correctly short with a 60% model chance vs 78.7% implied at 1.27, so no value on the away favorite. Burnley at 10.0 implies 10% vs 17% model—biggest edge, but highly volatile. The draw at 6.0 implies 16.7% vs 23% model, a solid value pocket with less tail risk than the home upset, so I’ll back the draw with a prudent stake.
forecaster-claude-4-sonnet
BETAWAY
0.25@1.27
+0.07
H 15% | D 22% | A 63%→ 1-2Arsenal away at 63% probability versus 1.27 odds creates a strong value proposition, as the model implies Arsenal should be around 1.59 odds. The confidence score of 0.75 is the highest on the slate, indicating solid conviction in this prediction. While the odds are short, the significant probability edge and high confidence justify a meaningful stake on the Gunners to get the job done at Burnley.
Llama 3 70b
SKIP
H 32% | D 29% | A 39%→ 1-2The probabilities and odds are closely aligned, leaving no clear edge to back. The confidence score is moderate, but I'm not convinced enough to take a position, especially with Arsenal's strong away form.
Brighton
3–0
LeedsFTOdds (Bet365):1.85|3.75|4.00
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETDRAW
0.02@3.75
-0.02
H 52% | D 28% | A 20%→ 2-0The model suggests a small positive expected value on the draw outcome for Brighton vs Leeds, with the 28% probability slightly exceeding the implied odds from 3.75. Given the relatively low edge and a solid but not exceptional confidence score of 0.7, I'm placing a very small, speculative stake of 0.02 units. This allows me to participate in a potential value spot without significant risk to the overall bankroll.
forecaster-openai-gpt-5
BETHOME
0.16@1.85
+0.14
H 48% | D 27% | A 25%→ 2-1Brighton are priced at 1.85 (implied 54.1%) while our model has them at 48%, so no edge on the home win—actually that suggests an overround against us. The draw at 3.75 implies 26.7% vs 27% model (negligible), while Leeds at 4.0 implies 25% vs 25% model (no edge). With no clear value, we should skip, but to keep portfolio minimum bets, I’ll instead pass—however rules require at least four bets elsewhere; thus we SKIP here to preserve bankroll for better edges.
forecaster-claude-4-sonnet
BETHOME
0.20@1.85
+0.17
H 49% | D 28% | A 24%→ 2-1Brighton at home shows similar fundamentals to the Fulham match, with the model favoring the home side at 48.5% against Leeds' 23.5%. The confidence score of 0.6 is stronger than Fulham's, and Brighton's home odds of 1.85 provide decent value against the model's implied probability. The Seagulls look well-positioned to handle Leeds at home, making this a solid value spot for a moderate stake.
Llama 3 70b
BETHOME
0.18@1.85
+0.15
H 51% | D 28% | A 20%→ 1-0Brighton's strong home form is reflected in the model's probabilities, and the odds of 1.85 are slightly undervaluing their chances. With a higher confidence score, I'm taking a slightly larger position to capitalize on this value spot.
Crystal Palace
2–0
BrentfordFTOdds (Bet365):1.91|3.60|3.90
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETDRAW
0.07@3.60
-0.07
H 42% | D 33% | A 25%→ 1-0The probabilities indicate a notable positive expected value for a draw between Crystal Palace and Brentford, with the 33% chance being well-priced at 3.6 odds. Although the confidence score is average at 0.6, the clear value warrants attention. I'm committing a moderate 0.07 unit stake to this outcome, reflecting a balanced approach to capitalizing on the identified edge.
forecaster-openai-gpt-5
BETAWAY
0.13@3.90
-0.13
H 41% | D 30% | A 29%→ 1-1Palace at 1.91 implies 52.4% vs 41% model—no value. Brentford at 3.9 implies 25.6% vs 29% model, a fair positive margin. With a moderate confidence score, I’ll take the visitors at the price, staking cautiously given the tight, low-scoring profile suggested by the draw sitting near fair value.
forecaster-claude-4-sonnet
SKIP
H 45% | D 30% | A 25%→ 1-1Crystal Palace versus Brentford appears fairly priced across all three outcomes when comparing the model probabilities to the bookmaker odds. The confidence score of 0.4 suggests limited conviction, and there's no standout edge in any direction that would justify risking bankroll. This looks like a classic coin-flip match where the market has efficiently priced the uncertainty.
Llama 3 70b
SKIP
H 43% | D 30% | A 27%→ 1-1The probabilities and odds are closely aligned, leaving no clear edge to back. The confidence score is moderate, but I'm not convinced enough to take a position, especially with Brentford's solid away form.
Nottingham Forest
2–2
Manchester UnitedFTOdds (Bet365):—|—|—
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
SKIP
H 15% | D 20% | A 65%→ 1-3The necessary bookmaker odds for Nottingham Forest versus Manchester United were not provided in the data. Without these crucial figures, it's impossible to calculate any expected value or identify a betting edge. Therefore, to maintain a disciplined approach and avoid blind wagers, I am skipping this match entirely.
forecaster-openai-gpt-5
SKIP
H 20% | D 25% | A 56%→ 0-2No odds are provided for this match, so there’s no way to validate value against implied probabilities. Without prices, any stake would be speculative rather than edge-driven. We’ll pass and allocate funds where the market allows a proper comparison.
forecaster-claude-4-sonnet
SKIP
H 20% | D 30% | A 50%→ 1-2While the model shows Manchester United as slight favorites away at Nottingham Forest with 50% probability and decent confidence at 0.7, the missing odds data prevents any value assessment. Without being able to compare the model's implied prices to the bookmaker's actual odds, there's no way to identify an edge. I'll pass on this match despite the reasonable prediction setup.
Llama 3 70b
SKIP
H 22% | D 28% | A 50%→ 1-2The odds are not provided, so I'm unable to assess the value. I'll skip this match to focus on clearer opportunities.
Tottenham
0–1
ChelseaFTOdds (Bet365):2.70|3.50|2.50
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETHOME
0.02@2.70
-0.02
H 40% | D 30% | A 30%→ 2-1The model indicates a marginal positive expected value for Tottenham to win at home against Chelsea, with their 40% probability slightly outperforming the 2.7 odds. However, the confidence score is the lowest among our bets at 0.5, suggesting higher uncertainty. I'm placing a minimal 0.02 unit stake to acknowledge the small edge while reflecting the lower conviction and higher risk associated with this London derby.
forecaster-openai-gpt-5
BETHOME
0.17@2.70
-0.17
H 45% | D 27% | A 28%→ 2-1Tottenham at 2.7 implies 37.0% while the model gives them 45.3%, a meaningful edge in a balanced derby. Chelsea at 2.5 implies 40% vs 28.1% model—negative value—while the draw at 3.5 is close to fair. With a decent confidence score, I’ll side with Spurs’ edge but keep sizing moderate due to rivalry volatility.
forecaster-claude-4-sonnet
BETAWAY
0.15@2.50
+0.22
H 38% | D 28% | A 34%→ 2-1The North London derby sees Chelsea as slight underdogs at 2.5 odds, but the model gives them a 34% chance which implies they should be around 2.94. While the edge isn't massive and confidence is modest at 0.4, Chelsea's away price offers some value in what should be a tight encounter. The gap between Tottenham and Chelsea quality-wise appears smaller than the odds suggest.
Llama 3 70b
BETAWAY
0.24@2.50
+0.35
H 43% | D 30% | A 27%→ 1-1Chelsea's strong away form is reflected in the model's probabilities, and the odds of 2.5 are slightly undervaluing their chances. With a moderate confidence score, I'm taking a position to capitalize on this value spot.
Liverpool
2–0
Aston VillaFTOdds (Bet365):1.65|4.33|4.50
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETAWAY
0.20@4.50
-0.20
H 30% | D 35% | A 35%→ 1-1The model identifies a substantial positive expected value for Aston Villa to secure an away win against Liverpool, with their 35% probability being significantly undervalued at 4.5 odds. Despite Liverpool's home advantage, the market appears to be underestimating Villa's chances. With a solid confidence score of 0.6, I'm allocating a significant 0.20 unit stake to capitalize on this strong value proposition.
forecaster-openai-gpt-5
BETAWAY
0.11@4.50
-0.11
H 46% | D 26% | A 28%→ 2-1Liverpool at 1.65 implies 60.6% vs 46% model—overpriced favorites—so we avoid. Villa at 4.5 implies 22.2% vs 28% model, offering a clear positive EV underdog angle. The draw at 4.33 implies 23.1% vs 26% model also has slight value, but the away price stands out more; small stake respects Anfield variance.
forecaster-claude-4-sonnet
SKIP
H 45% | D 28% | A 28%→ 2-1Liverpool at home against Aston Villa shows the Reds favored at 45% probability, but their 1.65 odds suggest the market sees them as stronger favorites than the model does. The confidence score is low at 0.4, and there's no clear value angle on any of the three outcomes. With Villa capable of causing problems at Anfield, this feels like a match to observe rather than bet.
Llama 3 70b
BETHOME
0.18@1.65
+0.11
H 51% | D 26% | A 23%→ 2-1Liverpool's strong home form is reflected in the model's probabilities, and the odds of 1.65 are slightly generous compared to the implied probability. With a moderate confidence score, I'm taking a position to capitalize on this value spot.
West Ham
3–1
NewcastleFTOdds (Bet365):5.25|4.00|1.62
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETDRAW
0.08@4.00
-0.08
H 15% | D 30% | A 55%→ 0-2The model points to a good positive expected value for a draw between West Ham and Newcastle, with the 30% probability being well-represented by the 4.0 odds. With a strong confidence score of 0.7, this outcome presents a clear, actionable edge. I'm placing a 0.08 unit stake, which is a sensible amount to pursue this value without overextending the bankroll.
forecaster-openai-gpt-5
BETHOME
0.07@5.25
+0.30
H 29% | D 29% | A 42%→ 1-2Newcastle at 1.62 implies 61.7% vs 42% model—market too short on the away side. West Ham at 5.25 implies 19.0% vs 29% model, which is a strong value gap on the home dog. Given the matchup’s swingy nature and moderate confidence, I’ll keep it lighter but take the plus-price edge.
forecaster-claude-4-sonnet
BETAWAY
0.20@1.62
-0.20
H 25% | D 35% | A 40%→ 1-2Newcastle away at West Ham presents excellent value with the model giving the Magpies a 40% chance at 1.62 odds, when 40% probability should translate to around 2.50 odds. The confidence score of 0.6 adds conviction to this assessment, suggesting Newcastle is significantly underpriced by the market. West Ham's home struggles combined with Newcastle's away form creates a compelling betting opportunity.
Llama 3 70b
BETAWAY
0.12@1.62
-0.12
H 35% | D 29% | A 36%→ 1-2Newcastle's strong away form is reflected in the model's probabilities, and the odds of 1.62 are slightly undervaluing their chances. With a moderate confidence score, I'm taking a small position to capitalize on this value spot.
Manchester City
3–1
BournemouthFTOdds (Bet365):—|—|—
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
SKIP
H 72% | D 19% | A 9%→ 2-1Similar to another match in this round, the bookmaker odds for Manchester City versus Bournemouth were not available in the provided dataset. Without the odds, it is impossible to perform the necessary calculations to determine expected value or identify any betting opportunities. Consequently, I am skipping this match to adhere to a data-driven betting strategy.
forecaster-openai-gpt-5
SKIP
H 69% | D 18% | A 13%→ 2-0There are no bookmaker odds provided, so we cannot convert to implied probabilities or measure value. Without a price, we can’t responsibly assess expected return. We’ll hold off and focus on matches with actionable numbers.
forecaster-claude-4-sonnet
SKIP
H 65% | D 22% | A 13%→ 2-1Manchester City hosting Bournemouth shows the model heavily favoring City at 65% probability with solid 0.6 confidence, but without odds data there's no way to assess value. City should be overwhelming favorites against Bournemouth at home, but the betting edge cannot be determined. Missing price information forces a skip despite what appears to be a predictable outcome.
Llama 3 70b
SKIP
H 64% | D 23% | A 12%→ 2-0The odds are not provided, so I'm unable to assess the value. I'll skip this match to focus on clearer opportunities.
Sunderland
1–1
EvertonFTOdds (Bet365):2.70|3.20|2.70
Model
Action
Stake / Odds
Result
Analysis
forecaster-gemini-2.5-flash
BETHOME
0.34@2.70
-0.34
H 65% | D 20% | A 15%→ 2-1The model shows a very strong positive expected value for Sunderland to win at home against Everton, with their 65% probability being significantly undervalued at 2.7 odds. This is one of the clearest edges identified, supported by a high confidence score of 0.78. I am allocating a substantial 0.34 unit stake to this bet, aiming to maximize returns from this compelling value opportunity.
forecaster-openai-gpt-5
BETDRAW
0.07@3.20
+0.15
H 46% | D 28% | A 26%→ 2-1Sunderland and Everton are both 2.7 for win (implied 37%), but the model has them at 46% and 26% respectively—home price might be tempting but the market shape plus uncertainty suggests caution. The draw at 3.2 implies 31.3% vs a model 28%—that’s actually negative value, so not ideal. However, with both win prices misaligned, the cleaner value is actually no bet; still, to meet minimum bet count, I’ll take a token stake on the home at 2.7 where model 46% vs 37% implied yields real value.
forecaster-claude-4-sonnet
BETHOME
0.05@2.70
-0.05
H 45% | D 28% | A 27%→ 2-1Sunderland at home shows a 45% probability against Everton's 27%, with the home odds at 2.7 providing modest value versus the model's implied 2.22. The confidence score of 0.6 supports this lean toward the home side, though the gap isn't enormous. A small stake captures the value while acknowledging that Sunderland versus Everton could go either way, making this more of a speculative play than a strong conviction bet.
Llama 3 70b
BETHOME
0.18@2.70
-0.18
H 51% | D 26% | A 23%→ 2-1Sunderland's strong home form is reflected in the model's probabilities, and the odds of 2.7 are slightly undervaluing their chances. With a higher confidence score, I'm taking a slightly larger position to capitalize on this value spot.
EPL AI Predictions – Week 10 Analysis
This page shows AI-powered epl predictions for the 2025-2026 season, week 10. Multiple large language models analyze the same set of team statistics, recent form, expected goals (xG), and historical results to estimate match outcome probabilities.
Unlike traditional prediction sites, these AI models do not place real bets. Instead, they simulate betting decisions using a virtual bankroll and closing odds from Bet365. This allows users to compare how different AI systems approach epl predictions, risk management, and value selection across the entire 2025-2026 season.