Role of statistics in predicting Manchester United games

Modern football analysis is no longer built only on opinions or form tables. Statistics now play a major role in understanding how matches unfold, especially in leagues like the Premier League, where margins are extremely small. Nowadays, analysts use metrics such as expected goals (xG), historical scoring patterns and probability models to predict the likelihood of specific outcomes.

For football fans and bettors, predicting draws and exact scores has always been one of the hardest challenges. But statistics have made the process far more structured. Instead of guessing, analysts now use data models that translate attacking strength, defensive stability, and shot quality into probability distributions for final scorelines.

Statistical models help turn match narratives into measurable probabilities. When used correctly, they provide a powerful framework to understand why certain matches are more likely to finish 0-0, 1-1 or 2-2, especially in close competitions like the Premier League. This is really helpful if you are into correct score prediction.

Why Draws in Football Are So Difficult to Predict

Football is a low-scoring sport compared to many others. A single goal can decide a match. This makes predicting draws particularly tricky.

Unlike basketball or cricket, where scoring happens frequently, football matches often hinge on very few events. Because of this, randomness plays a larger role.

However, patterns still exist.

Draws tend to occur more frequently when:

• Two teams have similar attacking and defensive strength
• Both sides create chances but struggle to convert them
• Tactical setups prioritise defensive structure
• Teams sit close to each other in the league table

Mid-table clashes in the Premier League often produce these conditions.

For example, matches between teams like Brighton, Crystal Palace or Brentford often produce tight games with low scoring margins. Until last season, the Wolves were also a part of this list. When neither side has a clear attacking edge, statistical models begin to show a higher probability for 1-1 or 0-0 outcomes.

This is where football analytics become useful.

How Expected Goals (xG) Changed Football Analysis

One of the most important modern statistics in football is Expected Goals (xG).

Expected goals measure the probability of a shot resulting in a goal. The calculation considers factors such as:

• Shot distance
• Angle to goal
• Type of assist
• Defensive pressure
• Body part used

The value ranges between 0 and 1, where 0.50 means the chance has a 50% probability of being scored.

When all shots from a match are combined, analysts can estimate how many goals a team should have scored based on chance quality.

Expected goals provide a better picture of performance than the final scoreline alone because the outcome of a match often contains randomness. A team may dominate chances but lose due to poor finishing or an exceptional goalkeeper’s performance.

Let’s take the example of Manchester United.

Manchester United have experienced several matches in recent seasons where their xG suggested a different story from the final score.

For example:

• United may produce an xG of 2.1 but score only once.
• The opponent might generate 0.8 xG but score a late equaliser.

From a statistical perspective, that match profile increases the probability of a 1-1 draw even though the flow of the game might suggest otherwise.

Last season, United scored just 44 goals across 38 matches. This season, they’ve already scored 51 goals in 29 games so far. After the opening three matches of the 2025/26 season, United had registered the highest expected goals (xG) in the Premier League with a total of 6.78, underlining the quality and quantity of chances created.

By March 2026 (after 29 matches), they are fourth on the list. A total xG of over 50 suggests that their tactical approach is successfully generating high-quality chances. They are averaging around 1.66 to 1.91 xG per 90 minutes.

This is why analysts often start with xG numbers when projecting scorelines.

Turning Expected Goals into Scoreline Probabilities

Once the expected goals for both teams are calculated, the next step is to translate them into scoreline probabilities.

This is usually done using a mathematical model called the Poisson distribution.

The Poisson distribution estimates how often a certain event occurs over a fixed period of time. In football analytics, that event is the number of goals scored by a team.

If a team is expected to score 1.4 goals, the Poisson model can calculate the probability of scoring:

• 0 goals
• 1 goal
• 2 goals
• 3 goals
• or more

By running this calculation for both teams, analysts create a full scoreline probability table.

Scoreline	Probability
0-0	7%
1-0	13%
0-1	11%
1-1	15%
2-1	10%
1-2	9%

This table gives a realistic view of what the most likely results are.

Once you have expected goals for each side, you can translate that into a full scoreline probability table (0-0, 1-0, 1-1, 2-1, etc.).

The Most Common Scorelines in the Premier League

Premier League matches tend to follow a small number of recurring scorelines.

Over long periods, the most common results are:

Scoreline	Typical Frequency
1-1	Very common
1-0	Very common
2-1	Common
2-2	Moderate
0-0	Occasional

The 1-1 draw consistently ranks as one of the most frequent outcomes in European football leagues.

10% of the matches in the Premier League finish 1-1, followed by 1-0 at 9.3%.

Why does this happen?

Because when two teams produce similar xG numbers, the Poisson model naturally clusters probability around one goal each.

Case Study: Manchester United & Tight Matches

Manchester United are an interesting case study because of how their trajectory has changed this season since the arrival of Michael Carrick.

Over several Premier League seasons, United have frequently played games where the xG margin between the two teams remained extremely small.

Consider a typical match profile:

Metric	Manchester United	Opponent
Shots	14	12
xG	1.6	1.4
Possession	53%	47%

In such situations, statistical models will often show high probabilities for:

• 1-1
• 2-1
• 1-2

But rarely for extreme scorelines.

Matches like these are exactly where draw predictions become statistically viable.

In Manchester United’s last 10 matches, 7 matches have seen 3 or fewer goals being scored. One match saw 4 goals being scored, and 2 matches finished with 5 goals scored.

Out of the 10 matches, 5 matches finished with scorelines of 1-1, 2-1 or 1-0. And this is a common scoreline for most of the EPL matches.

Now, let’s look at the goals that have come from set pieces.

Manchester United’s analyst Kaita Hasegawa has been with the team since November 2022. He was appointed during Erik ten Hag’s stint. Since then, managers have changed, but Hasegawa has been retained each time.

According to United analyst’s data, set pieces account for approximately 28-30% of goals in modern football. The Red Devils have scored 15 set-piece goals this season, averaging 8.0 goals every 100 set-pieces, the highest efficiency rate in the division. United’s corner conversion rate stands at 14.3%.

CentreDevils often analyse United games through underlying data such as shot maps, xG trends and possession structure. Those numbers help explain why some fixtures look balanced even when the league table suggests otherwise.

Defensive Teams and the Probability of Draws

Another strong indicator of draws is defensive stability combined with limited attacking efficiency.

Teams that:

• concede very few chances
• generate moderate xG numbers
• avoid high-risk attacking play

often produce more draws.

Examples in recent Premier League seasons include teams like:

• Wolves (now relegated)
• Crystal Palace
• Newcastle (during defensive phases)
• Brighton in certain tactical setups

These teams frequently play matches where total expected goals remain below 2.2 combined.

>Low combined xG almost always pushes the model toward scorelines such as:

• 0-0
• 1-0
• 0-1
• 1-1

And when both teams have similar xG values, 1-1 becomes the most probable outcome.

Limitations of Statistical Models

Even the best models cannot perfectly predict football matches. There are several reasons:

1. Random Situations – Deflections, red cards and penalties can dramatically change results.
2. Tactical Changes – Managers may adjust strategy during matches.
3. Player Form – A striker in exceptional form may outperform xG.
4. Game State Effects – Teams leading the match often change their style.

Statistics have completely changed how analysts evaluate football matches.

Instead of relying on intuition alone, modern analysis combines expected goals, historical patterns and probability models to estimate match outcomes.

Draw prediction remains one of the hardest tasks in football analytics. However, statistics provide clear signals when two teams are closely matched.

In leagues like the Premier League, where competitive balance is extremely high, those signals appear frequently.

The future of football analysis will likely continue moving in this direction. As data collection improves and models become more refined, predicting outcomes will become increasingly sophisticated.

But one thing will always remain true in football.

Even the best statistical model cannot fully remove the unpredictability that makes the sport so compelling.