Expected Threat (xT) Explained
Expected Threat, usually shortened to xT, is one of the most useful ideas in modern football analytics because it asks a question traditional event data often misses: how much closer did that action move a team towards scoring?
Goals and assists capture the end of an attacking move. Expected Goals measures the quality of shots. Expected Assists estimates the value of passes that directly create shots. But football is full of valuable actions that happen before the obvious moment: the centre-back who breaks the first line, the full-back who carries into the half-space, the midfielder who switches play into isolation, or the winger who receives between defenders and forces the opposition to retreat.
Expected Threat is designed to measure that hidden attacking value. It assigns value to possession based on where the ball is on the pitch and how likely that location is to lead to a goal. A pass, carry, dribble or reception can then be evaluated by how much it increases or decreases the attacking value of the possession.
In simple terms: xT measures how threatening a team’s possession becomes as the ball moves across the pitch.
What Is Expected Threat?
Expected Threat is a possession value model. It estimates the probability that a team will eventually score from possession in a given area of the pitch, assuming the possession continues from that location.
The core idea is intuitive. Having the ball near your own corner flag is usually not very threatening. Having the ball at the edge of the opposition box is much more threatening. Having it in the six-yard box is extremely threatening. xT converts those differences into a value framework.
If a player passes from a low-value zone into a higher-value zone, the action receives positive xT. If they move the ball backwards into a less threatening area, the action may receive negative xT. If they complete a safe sideways pass that barely changes attacking potential, the xT value may be close to zero.
This does not mean backward passes are bad or sideways passes are useless. Football context matters. A backward pass can reset pressure, open a new passing lane or protect possession. xT simply measures the immediate change in territorial attacking value, not the full tactical purpose of the action.
That distinction is important. Expected Threat is not a complete player rating system. It is a model for valuing ball progression and possession development.
Why xT Exists: The Gap Between xG and Possession Value
Expected Goals is essential because it tells us about shot quality. But many players influence attacking output without taking shots or playing the final pass.
Consider a midfielder who repeatedly receives under pressure, turns away from a marker and passes into the attacking midfielder between the lines. Those actions may not become assists. They may not even directly precede shots. But they can be the reason a team sustains pressure and eventually creates high-quality chances.
Similarly, a centre-back who plays vertical passes through the opposition press can add enormous value without appearing in goals, assists or xG totals. A full-back who carries the ball from the middle third into the final third may force defensive rotations that shape the rest of the attack.
Traditional output metrics often under-credit these players. xT helps by measuring contribution earlier in the possession chain.
The model answers questions such as:
Which players consistently move the ball into more dangerous areas?
Which teams generate threat through central progression rather than crosses?
Are a winger’s carries actually increasing attacking value, or simply moving the ball into low-probability wide zones?
Does a midfielder progress possession into zones that lead to shots, or only accumulate safe pass volume?
For recruitment, tactical analysis and prediction modelling, this matters. Goals are rare. Shots are less rare, but still limited. Possession actions are abundant. xT gives analysts a larger sample of attacking behaviour to study.
How Expected Threat Is Calculated
The most common xT framework begins by dividing the pitch into zones. Each zone is assigned a value based on how likely a possession starting in that zone is to eventually produce a goal.
The pitch can be divided into a regular grid, such as horizontal and vertical cells, or into more sophisticated zones that reflect football geography. The model then learns from historical possessions.
For each zone, the model asks:
How often does a team shoot from this zone?
How likely are shots from this zone to become goals?
If the team does not shoot, where does the ball typically move next?
How often does possession continue or break down?
This produces a value for every area of the pitch. Zones close to goal usually have higher xT values. Central zones near the penalty area tend to be more valuable than wide zones at the same distance because they more often lead to higher-quality shots. Deep defensive zones usually have low values because possessions starting there have many steps to travel before becoming dangerous.
Once the pitch has values, each action can be scored.
If a pass starts in a zone worth 0.01 xT and ends in a zone worth 0.05 xT, the pass has added 0.04 xT. If a carry begins in a zone worth 0.03 and ends in a zone worth 0.06, it has added 0.03 xT. If a player loses the ball, the action may be treated as ending the possession, removing future attacking value.
These numbers are not literal guarantees. A value of 0.05 does not mean the team will score exactly five times from 100 identical situations. It is an estimate derived from a large set of similar possession states.
The Difference Between xT, xG and xA
Expected Threat is often confused with Expected Goals and Expected Assists, but each metric measures a different part of the attacking process.
Expected Goals measures the quality of a shot. It is only assigned when a shot occurs.
Expected Assists usually measures the xG value of shots created by a player’s final pass. It is only credited when the recipient shoots.
Expected Threat measures the change in scoring probability created by moving the ball from one area to another. It can be assigned to passes, carries, dribbles and other possession actions, even when no shot follows immediately.
Imagine a sequence:
A centre-back passes through midfield into a pivot.
The pivot turns and finds a number ten between the lines.
The number ten slips the winger into the box.
The winger cuts back for a striker.
The striker shoots from close range.
xG values the striker’s shot. xA credits the player who made the final pass, if the cut-back led directly to the shot. xT can value the earlier actions: the centre-back’s line-breaking pass, the pivot’s forward pass and the winger’s carry into the box.
This is why xT is particularly powerful for identifying players who help attacks mature before the final action.
What Good xT Looks Like on the Pitch
A high-xT action is not always spectacular. It is often a pass or carry that changes the defensive problem facing the opponent.
A vertical pass into the half-space can be valuable because it breaks a line and gives the receiving player access to central or box-entry options. A switch of play can generate xT when it moves possession from a congested area into a wide zone where the attacker has time to advance. A carry into the penalty area can create major threat because it changes the defending team’s risk profile: defenders must either engage, concede space, or risk a foul.
By contrast, a long diagonal to a winger standing deep and wide may look progressive but add little threat if the receiving position rarely leads to valuable shots. A cross from a poor angle might enter the box but still carry limited value if historical outcomes from that location are weak.
This is where xT improves analysis. It separates movement from meaningful progression. Not every forward pass is valuable. Not every carry is dangerous. The location, direction and resulting possession state matter.
Passes, Carries and Receptions: Different Ways to Create Threat
xT can be applied to different action types, but interpretation should vary.
For passes, xT measures the value gained by moving possession from the passer’s location to the receiver’s location. A midfielder who repeatedly completes passes into central final-third areas will usually accumulate positive xT.
For carries, xT measures the value gained as the ball is moved by the same player. This is important for dribblers, ball-carrying centre-backs and midfielders who break pressure. A carry from the touchline into the box can be more threatening than several low-value passes around the back line.
For receptions, analysts can evaluate how often players receive in valuable areas. This is different from creating threat with the ball. A forward who constantly receives in high-xT zones may be excellent at movement and positioning, even if the pass itself is credited to another player. At team level, reception value can reveal whether a structure consistently puts players between the lines or behind the defence.
The best analysis separates these layers. A player may be an elite passer into threat, an elite carrier into threat, or an elite receiver of threat. Those are not the same skill sets.
How Football Data Lab Applies Expected Threat
Football Data Lab treats Expected Threat as part of a possession-value pipeline rather than a standalone leaderboard metric. The objective is not simply to rank players by total xT, but to understand how teams create repeatable attacking advantage and how that advantage translates into future chance quality.
The process begins by converting event data into possession sequences. Each action is aligned to a consistent attacking direction, so left-to-right and right-to-left possessions are comparable. Open-play sequences are separated from set-piece phases because the value of pitch locations differs when defensive structure is reset or disrupted. We then assign each on-ball action a start location, end location and possession outcome.
The pitch is valued using historical possession outcomes. Zones are not treated as tactically equal just because they are geographically similar. Central lanes near the box, wide crossing corridors, deep build-up areas and penalty-area pockets all behave differently in terms of shot creation. The model estimates how possession value changes as the ball moves through those areas.
In simplified terms, if a team has the ball in a deep central build-up zone, the possession carries modest attacking value. If the next pass reaches a midfielder facing forward between the opposition midfield and defensive lines, the possession value rises. If that midfielder then plays into the box, the value rises again. Each player in the chain receives credit for the increase they create.
A simplified Football Data Lab-style example might look like this:
A centre-back starts possession in a low-value build-up zone.
He passes into the central midfielder, increasing possession value slightly because the ball has moved beyond the first pressing line.
The midfielder receives on the half-turn and plays into the right half-space near the box, producing a larger xT gain.
The winger carries into the penalty area, creating another increase.
The final cut-back leads to a shot, which is then evaluated separately through Expected Goals.
The important point is that xT and xG are connected but not identical. xT describes how the possession became dangerous. xG describes the shot that eventually resulted. In Football Data Lab’s modelling pipeline, those layers are analysed together to distinguish teams that merely circulate the ball from teams that reliably move possession into zones where high-quality shots become more likely.
We also use xT profiles to examine attacking style. Two teams may generate similar shot totals but arrive there differently. One may rely on repeated wide entries and crosses. Another may progress centrally through the half-spaces. Their xG totals might converge over a short period, but their xT maps can reveal different underlying attacking mechanisms.
For prediction work, this matters because stable territorial threat can provide information before it fully appears in goals. A team consistently moving the ball into dangerous zones may be creating the platform for future shot quality, even if recent finishing has been poor. Conversely, a team scoring from limited threat may be riding unsustainable execution unless the underlying possession value improves.
Football Data Lab therefore evaluates xT alongside shot quality, tempo, opposition strength, venue effects and game state. A high xT total against a passive opponent protecting a lead is not interpreted in the same way as high xT generated against an organised defence in a level match. Context determines how much signal the model should take from the possession value.
Using xT for Team Analysis
At team level, Expected Threat helps explain attacking identity.
A possession-heavy side may dominate the ball but produce limited xT if most circulation happens in safe defensive and midfield zones. This is the classic sterile possession problem: high pass volume, low penetration.
A transition team may produce less total possession but generate high xT per action because its attacks move quickly into dangerous spaces. Such teams may not control territory for long periods, but when they win the ball, their first few actions can be extremely valuable.
A crossing side may accumulate xT through wide progression, but the quality of that threat depends on where the ball is delivered from and whether the team can access cut-back zones rather than low-probability crossing angles.
A central combination team may show strong xT in the half-spaces and zone 14, reflecting a structure that repeatedly finds players between the lines.
This makes xT maps particularly useful. Instead of asking only how much threat a team creates, analysts can ask where that threat is created. A heat map of positive xT actions can reveal whether the team’s danger comes from left-side overloads, right-side switches, central carries or penalty-area entries.
For coaches, this can support tactical planning. If an opponent creates most of its xT through one full-back and winger combination, defensive preparation can focus on blocking that progression route. If a team struggles to generate xT through midfield, recruitment may prioritise a player who can receive under pressure and pass forward.
Using xT for Player Analysis
Player xT analysis is powerful, but it must be handled carefully.
Total xT can favour players who simply have more touches, especially players in dominant teams. A full-back at a possession-heavy club may accumulate more threat than an equally talented player in a weaker side because he has more opportunities.
That is why analysts often look at xT per action, xT per 90 minutes, xT from passes, xT from carries, and xT relative to team share. Role is critical. Centre-backs, pivots, number tens, wingers and strikers operate in different zones and face different risk-reward trade-offs.
A centre-back with strong xT numbers may be exceptional at breaking lines. A midfielder with high xT from carries may be important against pressing systems. A winger with high xT from receptions may be benefiting from intelligent movement and good service. A striker with low passing xT may still be tactically valuable if his role is to occupy centre-backs and finish moves.
The best use of xT is comparative and contextual. Compare players in similar roles, similar team structures and similar leagues where possible. Look at how they generate threat, not only how much.
Limitations of Expected Threat
Expected Threat is valuable, but it is not a perfect measure of football value.
The first limitation is defensive pressure. A pass into a high-value zone is not equally valuable if the receiver is tightly marked, facing his own goal, or about to be tackled. Basic xT models based only on location may miss this.
The second limitation is player quality. A zone may be dangerous on average, but the value of receiving there depends on who receives the ball. Giving the ball to an elite one-v-one winger in the box is different from giving it to a defensive midfielder under pressure.
The third limitation is tactical context. A backward pass may reduce xT immediately but improve the possession by drawing pressure or creating a later switch. xT can undervalue patience and rest defence.
The fourth limitation is game state. Teams chasing a goal often generate territorial threat against deeper opponents. Teams protecting a lead may concede low-quality territorial progression while defending the box. Raw xT totals can therefore mislead without context.
Finally, xT does not replace xG. A team can move the ball into threatening zones but fail to convert that threat into shots. That gap is analytically important. It may indicate poor final-third decision-making, weak penalty-box presence, or an opponent that allows entries but blocks central shots.
Expected Threat in Prediction Modelling
xT is not a direct match prediction metric on its own, but it can improve understanding of team strength.
Football matches are ultimately decided by goals, and goal prediction models often rely heavily on Expected Goals because xG is closely tied to shot quality. However, xT provides earlier information about the attacking process that leads to xG.
A team that consistently creates high xT but low xG may be reaching promising areas without producing shots. Analysts then need to examine whether this is a finishing-phase problem, a tactical issue, or simply variance. A team that creates low xT but high xG may be relying on unusually efficient attacks or individual moments.
In practical modelling, xT can help separate repeatable attacking territory from noisy short-term shot outcomes. It can also help profile upcoming matchups. If one team is strong at generating threat down the right and the opponent is vulnerable to progression on that side, that interaction may influence expected chance creation.
The key is integration. xT should sit alongside xG, shot volume, defensive quality, pressing data, venue effects, injuries where available, and tactical context. Used alone, it tells only part of the story. Used properly, it adds an important layer between possession and shots.
FAQ: Expected Threat
What does Expected Threat measure?
Expected Threat measures how much an on-ball action increases or decreases the attacking value of a possession based on where the ball moves on the pitch. It estimates how likely possession in different areas is to eventually lead to a goal.
Is xT the same as Expected Goals?
No. Expected Goals measures shot quality. Expected Threat measures possession value before a shot happens. xT can credit actions that help create danger even if they are not the final pass or shot.
Can a player have negative xT?
Yes. If a player moves the ball from a more threatening area to a less threatening area, or loses possession in a valuable area, the action can have negative xT. This does not always mean the action was tactically wrong, but it does mean immediate attacking value decreased.
Why do wide players often score well in xT?
Wide players frequently receive in advanced areas and make box entries, crosses or carries. These actions can increase possession value. However, analysts should check whether the threat comes from genuinely dangerous areas or from repeated low-quality wide deliveries.
Does xT account for defensive pressure?
Basic xT models are mainly location-based and may not fully capture pressure, body orientation or defensive shape. More advanced approaches can incorporate contextual information, but interpretation should always consider match situation and opponent behaviour.
How should xT be used in player recruitment?
xT can identify players who consistently progress possession into dangerous areas. Recruitment analysis should separate passing xT, carrying xT and receiving value, then compare players within similar roles and tactical contexts.
Can xT predict future goals?
xT can support prediction by measuring whether teams regularly move the ball into dangerous zones, but it should not be used alone. It is most useful when combined with xG, shot data, defensive metrics and contextual modelling.
Conclusion
Expected Threat gives analysts a way to value the parts of attacking football that happen before the shot and before the assist. It measures how possession becomes more or less dangerous as the ball moves across the pitch, helping identify players and teams that create the conditions for chances.
Its strength is not in replacing Expected Goals, but in complementing it. xG tells us about the quality of shots. xT tells us how the ball arrived in threatening areas in the first place.
Used carefully, with context for role, pressure, game state and tactical structure, Expected Threat is one of the clearest tools for understanding possession value. For modern football analysis, it turns ball progression from a vague idea into something measurable, comparable and tactically meaningful.
Want to see this in action? Check our live match predictions for daily updated probabilities on every fixture.