Goal Guru's premium predictions now include a score table: a 6×6 grid with the chance of every scoreline from 0-0 to 5-5, coloured like a heat map, with the three most likely results called out ("1-1: 12%"). It holds more information than any single prediction on the screen, once you know how to read it.
The Layout
Home goals run down the side, away goals run across the top. Each cell is one exact final score, and its number is the chance the model gives that score. The darker the cell, the more likely the score. The three most likely scores are outlined and called out with their chances, so you don't have to hunt for them.
Lesson One: The Most Likely Score Is Still Unlikely
The single most likely score in a football match rarely gets much more than 10–15%. Football has low scores and lots of possible outcomes, so the probability is spread across many cells. When the top cell says "1-1: 12%", the model isn't predicting a 1-1. It's saying 1-1 is the likeliest of many unlikely outcomes, and that roughly 88 times in 100 the match ends some other way.
That's the most common misreading of a score table, and it's why the grid is more useful as a shape than as a single answer.
Lesson Two: The Result Is Hiding in the Grid
Draw a line down the diagonal, from 0-0 to 5-5:
- The diagonal (0-0, 1-1, 2-2 …) is every draw.
- Below the diagonal, where home goals are higher, is every home win.
- Above the diagonal is every away win.
Add up the cells in each region and you get, roughly, the match-result prediction. When the dark cells cluster on one side of the diagonal, one team is a clear favourite. When they hug the diagonal, the model expects a tight game.
Lesson Three: So Are the Goal Lines
- Over or under 2.5 goals: every cell whose two numbers add up to 3 or more is "over". The rest is "under".
- Both teams to score: every cell outside the top row and the left column (where one side has 0).
- Clean sheets: the 0 row is an away clean sheet, and the 0 column is a home clean sheet.
This is why the score table is a good sanity check. If the model gives over 2.5 at 65%, the dark area of the grid should sit in the high-scoring corner, and you can see whether it does.
Lesson Four: Look at the Spread, Not Just the Peak
Two matches can have the same most likely score, say 1-0, and look completely different. In one, the probability is packed into a few cells around 1-0 and 0-0. In the other, it's smeared across 2-1, 2-2, 3-1, and 1-3. The first is a low-scoring, predictable game. The second is open, and anything could happen. The peak is the same, but the spread tells you more.
Goal and Corner Ranges
Next to the score table, the goal and corner cards give each total as a range: the band the final count is likely to land in, and how likely that band is. A card might show 1 to 3 goals at 70%, with the expected number of goals beside it. There's a card for the match, for each team, and for each half, and corners get the same treatment. Expected corners are shown for both teams.
Ranges answer a different question from the grid. The grid asks "what exactly?", which is hard to answer. A range asks "roughly how many?", which is easier and more useful. If you're looking for corner-heavy matches, the corner range says far more than any single number.
During the Match
Once a match is live, the live predictions recount the ranges from the current score and corners, and a Next Goal card shows how likely the home side, the away side, or nobody is to score next. A 1-0 lead with a strong away Next Goal reading is a match the scoreline isn't telling you much about.
Check It Against the Record
A heat map is only as good as the probabilities in it. The Track record shows how often the model's chances actually came true, market by market, so you can see whether its 70% behaves like 70%. More on that in How We Measure Prediction Accuracy. For everything else Guru Vision predicts, see Guru Vision.