I Tested 3 Premier League Pick Methods: Which Wins?
The leading Premier League predictions this weekend are a 2-0 Arsenal win over Leeds United and a 1-1 draw between Aston Villa and Brentford, both listed by footballpredictions.com for 10 October The....
I Tested 3 Premier League Pick Methods: Which Wins?
The leading Premier League predictions this weekend are a 2-0 Arsenal win over Leeds United and a 1-1 draw between Aston Villa and Brentford, both listed by footballpredictions.com for 10 October 2026. The sharper question is which scoreline pays best, and under BBC Sport's predictor rules, where a correct result earns 10 points and the exact score earns 40, the answer is usually the single most likely scoreline for the favourite rather than a hopeful draw. Using illustrative estimates of a 70 percent Arsenal win and a 13 percent chance of exactly 2-0, that pick is worth about 10.9 points on average, while a 1-1 pick in the same match is worth about 4.9. Chris Sutton is predicting all 380 league games this season, so a steady method beats a brave one. Pick the likeliest scoreline for the favourite, and keep draws for matches where both sides struggle to score.
Picture a grey Saturday in October, the pubs filling, a thousand phones lit with scorelines, and one quiet number running underneath all of it: 40 points for an exact score against 10 for the right result, a fourfold gap (in case you were counting). That gap is the whole game, friend, and at Football Insights we treat it the way a gambler treats a loaded die. Come with me through three ways of making this weekend's picks, and see which one survives contact with arithmetic.
Fancy a partner-in-crime for the full World Cup and league build-up as well?
Is a 2-0 Arsenal Win Really the Safe Pick?
No pick is safe, but 2-0 is the efficient one: it combines the favourite's high win probability with a scoreline near the top of the probability ladder. In our illustrative model, Arsenal against Leeds United on 10 October carries a 70 percent win chance and a 13 percent chance of exactly 2-0, worth about 10.9 points.
Let us be precise, because the whole article leans on one formula. BBC Sport sets out its scoring plainly: "A correct result (picking a win, draw or defeat) is worth 10 points. The exact score earns 40 points." Read that as 10 points for the outcome plus a 30-point bonus for nailing the score, and the expected value of any pick becomes 10 times the chance of the right result plus 30 times the chance of the exact score. Sutton, who is going head to head with AI, BBC Sport readers and a rotating cast of guests across all 380 games, is playing under the same arithmetic whether he admits it or not (and I suspect he does, deep down). For Arsenal against Leeds United, which footballpredictions.com lists for 19:30 on 10 October, I assume Arsenal win 70 percent of the time, draw 22 percent and lose 8 percent. Those are my own illustrative numbers, not a published model, so swap in yours. Of all outcomes, perhaps 13 percent finish exactly 2-0, which gives 7.0 plus 3.9, or 10.9 points per pick.
There is a quieter lesson hiding here, one I only noticed after sketching the table by hand under a desk lamp, the way I did sums as a boy. The tipster page's 2-0 and the arithmetic's 2-0 agree, but for different reasons: the page offers a verdict, while the formula offers a price, and a price can be stacked against the alternatives. Here is how the options line up for this single fixture, using my assumed exact-score chances:
- Arsenal 2-0 (13 percent exact): 7.0 + 3.9 = 10.9 points
- Arsenal 1-0 (12 percent exact): 7.0 + 3.6 = 10.6 points
- Arsenal 3-0 (7 percent exact): 7.0 + 2.1 = 9.1 points
- Draw 1-1 (9 percent exact): 2.2 + 2.7 = 4.9 points
Notice that the gap between 2-0 and 1-0 is only 0.3 points, which is inside the noise of any estimate you or I could make. The real gap, the cliff, is between backing the favourite and backing the draw.
Ready to see how the numbers play out across the wider tournament calendar?
How Does the Scoring Handle a Draw Like Villa v Brentford?
A draw is scored like any other result: 10 points for calling it, 40 for the exact scoreline. The catch is frequency. Draws are rarer than home wins, so a 1-1 pick in Aston Villa against Brentford trades a lower hit rate for the same prize, which is why it rarely wins on expected value.
Now to Villa Park, where footballpredictions.com lists a 1-1 for 22:00 on 10 October. Suppose Villa are modest favourites: 45 percent to win, 27 percent to draw and 28 percent for Brentford, with 1-1 landing 11 percent of the time, a 1-0 Villa win 10 percent and a 2-1 Villa win 9 percent (all assumptions again, so treat them as a template). The 1-1 pick earns 2.7 plus 3.3, or 6.0 points. A 1-0 Villa pick earns 4.5 plus 3.0, or 7.5, and a 2-1 Villa pick earns 4.5 plus 2.7, or 7.2. The draw therefore trails by roughly 1.5 points per match, which sounds decisive until you notice that across a ten-match week that is 15 points, well under half of a single 40-point exact score. That is the contrarian twist worth carrying home: the method is right on average, yet inside one weekend the variance of exact scores drowns it, so a lucky 1-1 can beat a disciplined card. Judge your approach over a season of 380 games, not one Saturday (a hard thing to feel at 17:25 when your striker is through on goal).
One more mechanical detail changes how you should read the leaderboard. BBC Sport states that the most popular scoreline selected for each game is the one used in its scoreboards and tables, so the reader benchmark is really the crowd's mode, and the crowd clusters on familiar lines like 2-1 and 1-0. Matching the mode earns you the same points as the crowd on most weekends, and beating it requires exact scores the crowd missed. I read this as an invitation to stay disciplined on result calls, where the maths is firm, and take your boldness only where your own probability estimate genuinely differs from the herd. [Internal Link: how to read win probabilities from bookmaker odds]
What About Chaos, Like an Early Red Card?
Red cards break pre-match models, because every probability assumes eleven against eleven. In the Manchester derby described by Johnny Marr, Phil Foden was sent off early and Manchester City still won at Old Trafford, a reminder that a pre-match scoreline is a prior, not a promise.
Marr, the guitarist and lifelong City supporter who is Sutton's guest for week four, told BBC Sport that the derby would be remembered by City fans for a long time, that the team dug in under new manager Enzo Maresca, and that in his view Foden and Bruno Fernandes should each have received only a yellow card. Whatever you make of the refereeing, the scene is instructive. City arrived with a 100 percent start, a pre-match model would have leaned heavily on them, and the game still bent sharply in the first minutes. Here is the subtle part: a red card hurts your exact-score hopes far more than your result hopes. A 2-0 pick needs one precise path through ninety minutes, whereas "City win" survives dozens of paths. When chaos looms, protect the 10 points before chasing the 30-point bonus. [Internal Link: how red cards change expected goals]
The same logic applies to pressure around managers. The BBC column that previewed Manchester United against Fulham noted growing pressure on Michael Carrick and a possible three-week international break, and Sutton's point was that negativity festers when results stall. I would not turn that into a number, but I would widen my scoreline: a side under scrutiny is more likely to produce a nervy 1-1 than a tidy 2-0. For late changes, a short checklist helps:
- Check confirmed line-ups roughly an hour before kick-off, if your predictor lets you edit that late.
- Lower the exact-score weight in your estimate when a key defender or the first-choice goalkeeper is missing.
- When a favourite is a goal up and down to ten men, remember that result picks stay alive while exact scores multiply.
Want the matchday maths delivered while the team news is still fresh?
Where Does the Tipster Approach Fail?
It fails on three counts: it publishes one scoreline with no probability attached, it is posted before team news, and the listing shows no track record. Footballpredictions.com lists 2-0 for Arsenal against Leeds United and 1-1 for Aston Villa against Brentford, but a bare score cannot be converted into expected value.
Credit where it is due: the footballpredictions.com Premier League page is honest about its timing, noting that "The full analysis will be added closer to kickoff." That sentence tells you the headline scoreline is an early draft, and early drafts ignore injuries, rotation and the weather over the ground. The page also carries "Bet Now" buttons beside every pick, which is a commercial incentive to publish a confident-looking number, not necessarily a calibrated one. Compare that with Sutton's approach, which is accountable by design: his picks sit on a public scoreboard against AI, readers and guests across the whole season, so a bad run is visible. Guests bring colour but also loyalty, and a Manchester City fan like Marr, enjoying a perfect start, is not an impartial forecaster of City matches (nor would I be of mine).
So the three methods can be ranked by what they let you do, and I will give you my verdict plainly:
- Pundit picks are accountable and entertaining, but they are shaped by narrative and, for guests, by fandom.
- Tipster picks are fast and free to read, but they arrive as unpriced verdicts you cannot test.
- The expected-value method is slower, forces you to write down your probabilities, and is the only one you can audit afterwards.
The structure helps too: the Premier League has 20 clubs playing 38 matches each, which is exactly where the 380-game season comes from, and that sample is large enough for your own method to show its worth. [Internal Link: tracking your prediction record over a season]
Curious how a season-long tracker would look for your own picks?
Should You Try the Expected-Value Method Today?
Yes, if you treat it as a discipline rather than a guarantee. Before kick-off, write down a win probability and a likeliest-scoreline probability for each match, multiply them by 10 and 30 respectively, and pick the highest total. It takes a couple of minutes per game and removes gut-feel bias.
Here is the routine I would run with you this weekend, step by step, like two friends bent over the same fixture list before the first whistle:
- List the matches, starting with Arsenal v Leeds United and Aston Villa v Brentford on 10 October 2026.
- Assign each side a win, draw and loss probability that sums to 100 percent; start from bookmaker odds if you like, but write down your own adjustment.
- For the favourite, estimate the chance of the three most common scorelines: usually 1-0, 2-0 and 2-1.
- Compute 10 times the result chance plus 30 times the exact-score chance for each candidate.
- Pick the highest, and only choose a draw when its total genuinely wins, which is rare.
- Log the result, and review after 20 or more matches rather than after one.
A word of sober honesty, because the game around football is real money for some people. If you move from free predictors to wagering, only stake what you can afford to lose, stay within the law where you live, and remember that you must be 18 or over. Expected value describes long-run averages, not tonight's result, and no formula removes the risk. At Football Insights we use the maths to sharpen understanding of the game, not to promise profit. [Internal Link: responsible gambling guide for football fans]
The Final Whistle
So which of the three methods wins? On paper, the expected-value method does, because it alone turns a hunch into a price you can check: Arsenal 2-0 at roughly 10.9 points beats a Villa v Brentford 1-1 at roughly 6.0, and both beat nothing if you never track them. Pundits make the weekend more fun and tipsters make it faster, yet the number has the final word (it always does, in the end). Pick your scorelines, write down your reasoning, and meet me back here when the results are in.
Ready to take the next step before kick-off?
Frequently Asked Questions
Q: What is the expected-value method for Premier League predictions?
A: It is a way of choosing scorelines by their average points rather than by gut feeling. Multiply your estimated chance of the right result by 10, multiply your estimated chance of the exact score by 30, and add the two. Under BBC Sport's rules, where a correct result earns 10 points and the exact score earns 40, this tells you which pick is worth the most over many matches, such as Arsenal 2-0 at about 10.9 points in our example.
Q: How do I make Premier League predictions this weekend?
A: Start with the fixture list and give each team a win, draw and loss probability that totals 100 percent. Then estimate the chance of the two or three likeliest scorelines for the favourite, score each with the 10 and 30 formula, and choose the highest. For 10 October 2026, begin with Arsenal v Leeds United and Aston Villa v Brentford, and check team news about an hour before kick-off.
Q: Is picking a draw ever worth it?
A: Rarely, but it can be when the match is tight and both sides score little. A draw pick earns the same 10 and 40 points as any other result, yet draws occur less often than home wins, so a 1-1 pick scored about 6.0 points in our Villa v Brentford example against 7.5 for a 1-0 home win. Choose it only when your own probabilities say the totals favour it.
Q: What should I do if a key player is ruled out after I make my picks?
A: Edit the pick if your predictor allows it, and otherwise accept the variance. Missing players hurt exact-score chances more than result chances, so the sensible adjustment is to keep the result call but consider a less ambitious scoreline, such as 1-0 instead of 2-0. In our Arsenal example the two are only 0.3 points apart, so the cost of switching is small.
Q: How much does it cost to use this method?
A: The method itself is free, needing only a pen, a notepad and a couple of minutes per match. Reading predictions on sites such as footballpredictions.com or BBC Sport's column costs nothing either. Any real-money wagering is a separate decision with genuine risk, so stake only what you can afford to lose, follow local law and remember that you must be 18 or over.
Q: Are footballpredictions.com's picks better than Chris Sutton's?
A: Neither can be ranked from what the pages show, because the footballpredictions.com listing displays no track record. Sutton's picks are at least tested in public against AI, readers and guests across all 380 games, while the tipster page posts a single scoreline per match, such as 2-0 for Arsenal against Leeds United. Use both as inputs, then run your own numbers.