Matchweek 31
🧠 This Week’s Best Bets
TL;DR: Week 13 of the £100→£1000 Challenge. No review this week - we’re going straight to the picks. Bankroll sits at £96.38. Three bets. A bread-and-butter underdog, a form-driven value play, and one tiny bet on volatility that makes us long gamma on a low-probability, high-variance outcome. Total exposure: just 4.0% of bankroll.
💰 Portfolio Size: 3 bets 📊 Bankroll Exposure: 4.0% (£3.87 of £96.38) ⚡ Theme: Underdogs, shaky form favourites, and a lesson in convexity.
See for yourself: Odds Calculator
Consider following me on X for live-play updates and betting thoughts
🎯 This Week’s Theme: Shaky Favourites and the Art of Backing the Right Underdog
Three matches caught the model’s eye this week. In each case, the bookmakers are overpricing the favourite - sometimes because of reputation, sometimes because they haven’t fully priced in the weight of European football, and sometimes because they’ve simply missed the form picture.
We’re backing three underdogs. Two of them are bread-and-butter value plays with meaningful edge. One of them is something more exotic - a sixteenth-Kelly convexity trade on a volatile, goalless-or-chaotic Burnley shock. More on that below.
💰 SOLID PICK: Brentford Win @ Leeds
The Bet: Brentford to win @ 2.77 (36.1% implied) Model Probability: 46.0% (fair odds: 2.18) Expected Value: +27.4% Stake: 1.94% of bankroll (Eighth-Kelly) Confidence: Medium-High
This one looks odd on the surface. Brentford, the stronger team in ELO terms, priced as the away favourite at 2.77? The bookmakers have made a mess of this line, and we’re here to clean it up.
The model gives Brentford a 46% chance of winning. The market implies just 36.1%. That’s nearly a 10-percentage-point gap - +27.4% expected value. For context, our threshold to bet is +5%. This is more than five times that.
Yes, Brentford have been struggling. Their form rating of 5.1/10 is middling, and they’ve been inconsistent away from home. Leeds sit marginally better at 5.9/10. But ELO doesn’t lie over the long run, and the bookmakers have overcorrected on Brentford’s recent wobbles. Brentford are a better team than Leeds. When the price says otherwise, that’s value.
At Eighth-Kelly, we’re not betting the house. We’re collecting edge at a sensible fraction.
🔥 PREMIUM PICK: Everton Win vs Chelsea
The Bet: Everton to win @ 3.42 (29.2% implied) Model Probability: 33.9% (fair odds: 2.95) Expected Value: +15.9% Stake: 1.65% of bankroll (Quarter-Kelly) Confidence: Medium-High
This is the pick of the week, and it comes with a qualitative story that the model alone can’t tell.
Everton have been unfortunate. 1-0 to Man Utd. 2-0 to Arsenal - the second goal coming in the final minutes after Everton had held a competitive line for most of the match. On a different day with a different referee decision or a more clinical touch, those are draws. The form rating of 6.0/10 reflects this: Everton are a team that is competing, not capitulating.
Chelsea, on the other hand, are a mess. Form rating 4.5/10. Liam Rosenior’s new-manager bounce has faded - he’s now six games in, and the structural weaknesses are showing. Only two clean sheets in those six games. Consistent conceding from open play and set-pieces. And crucially, Chelsea are playing European football. The travel, the rotation, the late kicks, the mental load - none of it shows up in the ELO number, but all of it shows up on matchday.
The market prices this at 3.42, implying 29.2%. We price it at 33.9%. That gap is where the profit lives.
A note on the goals markets: I looked at the under lines for this match. Chelsea under Rosenior tend to be involved in high-scoring games - averaging close to 2.7 goals per game scored, with defensive instability to match. The BTTS and over 2.5 lines here don’t offer sufficient edge once you account for Chelsea’s attacking output even in bad form. The match result is where the value sits cleanest.
On a personal note: Everton have been one of the most interesting value sources in this challenge. Last week I was watching the Everton match and thought there was value in the Under 2.5 at kick-off. I decided to wait ten minutes to make sure no early goals had come in. By the ten-minute mark, the line had already moved into negative EV territory - the market had caught up before I could get the bet on. The lesson: when the edge is there pre-match, take it pre-match. Live markets are ruthless.
🎰 LONGSHOT: Burnley Win @ Fulham
The Bet: Burnley to win @ 5.95 (16.8% implied) Model Probability: 22.5% (fair odds: 4.33) Expected Value: +39.4% Stake: 0.43% of bankroll (Sixteenth-Kelly) Confidence: Low (intentionally)
Right, let’s be honest. Burnley probably won’t win this.
Fulham have a 54% chance of winning. Burnley 22.5%. This is a longshot - flagged as such, sized as such. We are placing less than half a percent of our bankroll on it, and that is the point.
So why bet it at all?
Because 5.95 for a 22.5% probability is a spectacular price. The expected value is +39.4% - the highest raw EV in this week’s portfolio. If you ran this bet a hundred times, you’d win roughly 22-23 of them and each win returns nearly six times your stake. The maths work.
But here’s the framing I find most useful: this bet makes us long gamma. In options trading, being long gamma means you hold a position that benefits from large moves in either direction - you’re paying a small premium for convexity. Back Burnley at sixteenth-Kelly and you’ve done the same thing: paid a tiny amount for exposure to a low-probability, high-magnitude outcome. If Fulham are flat, Burnley nick it on the counter, and the odds land - the return is disproportionate to the stake. If Fulham win, we’ve lost 43p.
Fulham aren’t in sparkling form either. The match feels like it could end 0-0 or 1-0 either way. The volatility is real. The price reflects a market that has Fulham as a comfortable favourite when the reality is messier than that.
We’re not predicting a Burnley win. We’re buying optionality at a price the market has mispriced.
❌ Bets We’re Skipping
| Bet | Why We’re Passing |
|---|---|
| Bournemouth Win vs Man Utd | +6% EV - real but thin. Carrick’s bounce is still active and United have genuine momentum. Not worth the allocation when better edges exist. |
| Newcastle Win vs Sunderland | The EV is there, but this is a derby, and derbies don’t respect models. See below. |
| Spurs Win vs Forest | A 7% edge on Spurs is almost insulting. Not enough edge, and it’s Spurs. Congealed butter. |
On Newcastle vs Sunderland: This deserves more than a table entry. The model shows Sunderland as a sizeable underdog at 28% with the bookies offering 5.11 against our 3.57 - that’s legitimate +EV. But you asked the right question: are Sunderland an attacking threat, or do they just contain?
Looking at Le Bris’s system, Sunderland are a counter-attacking team built on wide triangles and Xhaka’s tempo control, with Wilson Isidor’s pace as the primary outlet. They’re third-best in xGA in the league - defensively organised and disciplined. Against a Newcastle side playing all guns blazing (Eddie Howe under serious pressure, Tyne-Wear derby at home, European legs), Sunderland have the shape to absorb and hit on the counter. Isidor vs a patchwork Newcastle backline has real threat.
But here’s the problem: Newcastle 47% is actually the model’s honest assessment of a team playing at home in a high-stakes derby, bruised from Europe, with enormous motivation. That’s a competitive probability. The EV is driven by the bookmakers underpricing Sunderland, not by us having a strong Sunderland conviction. In a derby, I want conviction. We skip.
🎯 Super 6 Predictions
For the Sky Sports Super 6 jackpot game - optimising for exact correct score (5 points) over correct result (2 points)
| Match | Prediction | Confidence | Reasoning |
|---|---|---|---|
| Fulham vs Burnley | 1-0 | Medium | Poisson peak at 1-0 (15%). Fulham grind it out, Burnley’s attacking output too low away from home to expect more than one. |
| Everton vs Chelsea | 1-1 | Medium | Model probability is split. Chelsea can score but can’t keep clean sheets. Everton competitive at home. Poisson peak at 1-1 (10.9%). Aligns with our betting thesis - this is a tight game. |
| Leeds vs Brentford | 1-1 | Medium | Band 1 or 2 fixture. Brentford the stronger side but Leeds motivated. Poisson peaks at 1-1 (13%). Backing our Brentford bet but the draw is the most common exact score. |
| Newcastle vs Sunderland | 1-0 | Medium | Newcastle overwhelming home pressure in a derby context. Howe needs a result badly. Sunderland defensively solid but Poisson model peaks at 1-0 (12.1%) for Newcastle. |
| Spurs vs Forest | 1-0 | Medium | Dyche’s Forest travel well but score infrequently. Forest’s low possession style means Spurs dominate the ball. 1-0 at 14% is the Poisson peak. |
| Arsenal vs Man City | 2-1 | Medium | The title race fixture. Arsenal at home, set-piece dominance, Gyökeres in form. City will score - Pep’s side always do. But the Emirates fortress and Arteta’s dead-ball machine edges it for Arsenal. Poisson joint-peak at 2-1 and 1-1 (both ~9.3%). |
Super 6 Strategy Note: These predictions prioritise the single most likely exact score, typically in the 10-15% probability range. Different goals, different strategies - we’re hunting points, not value.
📋 Portfolio Summary
| Bet | Odds | EV% | Kelly Sizing | Bankroll % | Type |
|---|---|---|---|---|---|
| 🔥 Everton Win vs Chelsea | 3.42 | +15.9% | Quarter-Kelly | 1.65% | PREMIUM |
| 💰 Brentford Win @ Leeds | 2.77 | +27.4% | Eighth-Kelly | 1.94% | SOLID |
| 🎰 Burnley Win @ Fulham | 5.95 | +39.4% | Sixteenth-Kelly | 0.43% | LONGSHOT |
| TOTAL | - | +27.6% avg | - | 4.01% | - |
Bankroll: £96.38 Total Exposure: 4.01% Correlated Risk: None - three separate fixtures
Final Thoughts
Light week. Three bets, four percent of bankroll, no drama.
The disciplined weeks matter as much as the exciting ones. The model found edge. We’re taking it at appropriate fractions. The Brentford play and the Everton play are both well-grounded in the mathematics. The Burnley play is a tiny gamble on optionality - the long gamma trade. If it lands, it covers the other two losses many times over. If it doesn’t, we’ve lost less than the price of a coffee.
More next week on how the challenge is progressing. For now: three bets, trust the process.
All probabilities calculated using 5+ seasons of Premier League data (2,000+ matches). ELO bands updated weekly.
Don’t be an idiot. In case you’re feeling like being an idiot today, here is a page wherein I tell you not to gamble your mortgage away. Hey, that rhymes.
Odds correct at time of writing