Football statistics can tell the truth and still give the wrong impression. A team can dominate possession, take far more shots, and finish with the stronger-looking statistical profile while still having played the less effective match. The numbers may all be accurate. What they do not automatically explain is how those numbers were produced.
That distinction becomes especially important with advanced metrics. Opta's explanation of expected goals defines xG through the probability of individual shots becoming goals, which illustrates a broader point: every football statistic measures something specific rather than the entire match.
Learning to read football analysis properly therefore means asking what circumstances produced a number before deciding what that number means. Keeping track of the matches worth reviewing is easier with free football match calendars by tips.gg, but interpreting those matches still requires more than comparing the totals in a box score.
Possession Tells You Where the Ball Was, Not How Dangerous It Was
Possession percentage is one of the easiest football statistics to overinterpret.
The instinctive assumption is that the team with more of the ball controlled the match. Sometimes that is true. But possession can also accumulate through long spells of harmless circulation while the opposition deliberately defends deeper and waits to attack in transition.
ChaseYourSport has already highlighted a useful modern example in its analysis of Xabi Alonso's Chelsea, where the team opened the 2026/27 Premier League season with wins despite averaging only about 32 percent possession across its first two matches.
This is why analysts also use territorial measures such as field tilt. Rather than treating activity everywhere on the pitch equally, field tilt focuses on which side is spending more time or completing more actions in advanced areas.
A team can therefore have 60 or 65 percent possession without consistently threatening the opponent's goal. Another side may see much less of the ball while controlling the spaces from which the most dangerous attacks develop.
Possession answers who had the ball. It does not automatically answer who used it better.
Game State Changes the Meaning of the Numbers
One of the first questions to ask when reading match statistics is simple: what was the score when those numbers accumulated?
A team that falls behind has a reason to commit more players forward, press higher, take greater risks, and shoot more often. A team protecting a lead can concede some possession or territory in exchange for defensive security.
The trailing side can therefore finish with stronger attacking totals partly because the score forced it to attack for longer.
Sixty percent possession while chasing a two-goal deficit is not equivalent to sixty percent possession while controlling a level match. The same applies to shots and xG.
A final box score compresses ninety minutes of changing incentives into one set of totals. Reading the match properly means restoring that sequence.
More Shots Do Not Automatically Mean Better Chances
Shot count is useful because it tells us how often attacks ended with an attempt. What it does not tell us is how dangerous those attempts were.
Ten speculative efforts from distance can produce a larger shot total than three clear chances inside the penalty area. A blocked effort through several defenders and an uncontested shot from close range both add exactly one to the total.
The 2020 Champions League quarter-final between Manchester City and Lyon is a useful example. City finished with 71.6 percent possession and attempted 18 shots to Lyon's seven. Lyon won 3-1.
That does not prove possession or shot statistics were misleading. They accurately described City's greater volume. What they could not establish by themselves was which team converted the decisive situations more effectively.
Shots on target add information, but they still do not measure chance quality. A weak shot directly at the goalkeeper and an excellent attempt heading toward the corner are both recorded in the same category.
That gap is one of the reasons xG became so useful.
xG Measures Shots, Not Every Dangerous Situation
Expected goals improves on raw shot count by estimating how likely an actual attempt was to become a goal.
The crucial limitation is that standard xG begins with the shot.
Imagine a forward receiving the ball in a highly dangerous position. He has the opportunity to shoot immediately but takes another touch, tries to beat the goalkeeper, and forces himself toward a much tighter angle.
If he eventually shoots, the xG model evaluates the shot he actually took from the worse position. It does not award him the value of the better opportunity he chose not to use.
If the move ends without a shot at all, perhaps through hesitation, a poor final pass, or a defender recovering, that dangerous situation contributes no standard xG.
That is why team xG should not be treated as the total amount of attacking threat a side created. More precisely, it describes the accumulated goal probability attached to the shots that actually happened.
Mason Greenwood provides a useful example of a different xG mistake: treating finishing results as if they must immediately reflect a permanent level of finishing ability. He scored 10 Premier League goals from 3.6 xG in 2019/20. By the midpoint of the following league season, StatsBomb had him at just one goal from 2.5 xG. The point is not that either short stretch revealed his "true" finishing level. It is that actual goals can move dramatically above or below expected output over relatively small samples.
Other analytical models address another part of the gap. Expected threat, or xT, and other possession-value approaches can assign value to actions that move the ball into more dangerous areas before a shot occurs.
That does not weaken xG. It defines what xG can and cannot tell us.
Defensive Numbers Can Measure Workload as Much as Quality
The same contextual problem appears with defensive statistics.
Tackles, interceptions, blocks, and clearances describe genuine defensive actions, but a high total can mean different things. A centre-back recording eight tackles may have defended superbly. He may also have spent most of the match dealing with attackers reaching dangerous areas.
Paolo Maldini's famous line captures the other side of the idea: "If I have to make a tackle then I have already made a mistake."
The point is not that tackling is bad. Elite defending often starts earlier through positioning, anticipation, body orientation, and control of passing lanes. A defender who prevents dangerous situations from developing may need fewer visible interventions than one repeatedly forced into emergency challenges.
The mistake is assuming that more defensive actions automatically mean better defending.
Passing and Touch Numbers Depend on the Player's Job
Pass completion and touch counts can create the same false certainty.
A player completing 95 percent of his passes may be excellent in possession, but the number alone does not show whether those passes broke lines, moved the ball forward, or repeatedly travelled sideways under little pressure.
Erling Haaland is an extreme example of why involvement statistics need tactical context. In Manchester City's 4-0 win over Bournemouth early in his first Premier League season, he recorded only eight touches of the ball. The low involvement became a talking point, but Haaland later dismissed the concern after scoring against Borussia Dortmund, joking that his dream was to touch the ball five times and score five goals.
The point behind the joke was serious. A striker can influence a match through movement, positioning, occupation of defenders, and finishing without accumulating the kind of touch or passing volume expected from a midfielder.
The same principle applies to passing accuracy. A creative midfielder attempting difficult line-breaking passes may complete fewer passes than a teammate who mainly recycles possession. The safer percentage is not automatically the more valuable performance.
Statistics need to be judged against what the player was actually being asked to do.
The Number Is the Start of the Analysis
Football statistics become useful when they lead to another question rather than an immediate conclusion.
Possession needs territory. Shot totals need location and quality. xG needs an understanding of what happened before the shot and what dangerous situations never became shots. Defensive actions need workload and positioning. Passing and touch numbers need tactical role. All of them need game state.
None of this makes the statistics unreliable. The problem appears when a narrow measurement is treated as a complete description of the match.
A box score records what happened. Analysis begins when those events are placed back into the circumstances that produced them.
