Last updated: July 10, 2026
In October 2025, Declan Rice played a pass against West Ham that did not register on the scoresheet. It travelled forty yards from the left half-space, landed at Bukayo Saka’s feet inside the final third, and shifted Arsenal’s probability of scoring from 4% to 11% in a single action. No goal. No assist. No key pass logged. But Hudl Statsbomb’s model recorded it as one of the highest-value passes of the matchday.
That pass is the reason elite recruitment departments have quietly moved past expected goals and assists as their primary midfielder evaluation tool. The metric doing the work is called On-Ball Value, and it sits inside a growing family of advanced football tactics models that measure what actually happens between the boxes.
In simple terms: On-Ball Value (OBV) is a possession value model that assigns a numerical impact to every action a player takes, based on how much that action changes their team’s probability of scoring or conceding.
This breakdown walks through how OBV works, why it outperforms expected threat for midfielder evaluation, and what Declan Rice’s 3.22 OBV figure tells us about Arsenal’s title-winning structure.
Key Takeaways
- Every action gets a number: OBV assigns positive or negative value to every pass, carry, dribble, defensive action, shot, and goalkeeper action, based on the change in scoring and conceding probability the action produces.
- Built by Hudl Statsbomb, live in 140+ competitions: OBV was released publicly in 2021 and is now available across more than 140 leagues, from the Premier League to the Czech First Division, per Hudl Statsbomb.
- It beats xT on two fronts: OBV evaluates defensive actions and shots, not just zone-to-zone progression, and includes a Goals Against component xT lacks.
- The new midfielder benchmark: Declan Rice posted a 3.22 total OBV through the opening months of Arsenal’s 2025-26 title-winning campaign, the highest in the Premier League per Hudl Statsbomb.
- Risk-adjusted, not raw output: A turnover in a high-value possession produces a negative OBV value, which is why progressive carriers like Jeremy Doku rank differently on OBV than on simple progressive carry counts.
Table of Contents
- What Is On-Ball Value (OBV)?
- How OBV Is Calculated: The Probability Shift
- OBV vs Expected Threat: Why the Difference Matters
- The Six Action Categories Inside OBV
- Declan Rice 2025-26: The Premier League’s OBV Benchmark
- Recruitment Use Case: How Clubs Filter With OBV
- The Weakness: Where OBV Falls Short
- What This Means on the Pitch
What Is On-Ball Value (OBV)?
On-Ball Value is a possession value model that assigns a numerical impact score to every event in a football match, based on how that event changes the team’s probability of scoring and conceding in the near future. It was built and released by Hudl Statsbomb in 2021.
Think of OBV the way you would think of plus-minus in basketball, except applied to individual actions rather than time on court. A through-ball that splits two defenders and arrives in the penalty area might add +0.08 to the passer’s OBV. A misplaced pass that gifts possession to a counter-attacker in transition might cost -0.05. The model assigns a value to every action a player takes in line with the positive or negative impact it has on the probability of their team scoring and conceding, per Hudl Statsbomb. Hudl
The framework belongs to a broader category called possession value models, alongside Expected Threat (xT) and VAEP. Conceptually, possession value approaches such as VAEP, PV, OBV, and g+ are all identical: they estimate the chances of scoring (and conceding) in the “near future” based on a rich representation of the game state, per KU Leuven’s DTAI research lab. What separates them is how they define “near future,” what features they include, and which actions they value. OBV’s defining choice is that it values every action on the pitch, not just zone-to-zone ball progression. DTAI Sports
How OBV Is Calculated: The Probability Shift
OBV measures the change in two probabilities: the probability of the team scoring in the next sequence, and the probability of the team conceding. The action’s OBV value is the difference between the post-action probabilities and the pre-action probabilities, combining both. Possession State Value (PSV) models were born, and OBV is Hudl Statsbomb’s PSV implementation. StatsBomb
A simple worked example. A defender plays a pass from inside their own penalty area to a holding midfielder thirty yards upfield. Before the pass, the team’s chance of scoring in the next sequence might be 2%. After the pass arrives safely, it might be 4%. The OBV for that pass is +0.02, minus any small change to the conceding probability. Multiply that across hundreds of actions per match and thousands per season, and a clear signal emerges: which players move the needle, and which simply move the ball.
The critical design choice is that OBV is trained on Hudl Statsbomb xG, the most performant xG model available, and evaluates all actions on that take place on the pitch, containing Goals For and Goals Against components in the model to accurately measure the risk/reward of each action. The Goals Against component is what makes OBV genuinely risk-adjusted. A high-value pass attempted under pressure that gets intercepted is not free. It is taxed. Hudl
OBV vs Expected Threat: Why the Difference Matters
Expected Threat (xT) and On-Ball Value answer the same question – how valuable was this action? – but they answer it differently, and that difference shapes who looks good on each metric. xT, introduced by Karun Singh in 2018, treats the pitch as a grid of zones and assigns a threat value to each zone based on historical goal probability. Moving the ball from a low-xT zone to a high-xT zone generates positive xT.
The limitation is structural. It only values actions that move the ball from zone to zone, such as passes and carries, excluding defensive actions and shots. It is also typically implemented on limited event data, ignoring factors such as whether the player was under pressure or not when making the action. A defender who breaks up a counter-attack thirty yards from his own goal? Worth nothing on xT. The same defender on OBV gets credited because the action changed the conceding probability. Hudl
The other gap is risk-adjustment. xT does not penalise turnovers in the way OBV does, which is why xT can flatter wingers who attempt high-volume risky carries regardless of outcome. OBV’s Goals Against component punishes the giveaway directly.
The Six Action Categories Inside OBV
OBV breaks every player’s contribution down across six action types, which is what makes it a recruitment tool rather than just a single number. Each category isolates a specific skill, and a player’s profile across the six categories tells you what kind of footballer they are.
| Action Category | What It Measures | Typical Top Performer Profile |
|---|---|---|
| OBV from Passes | Value created by completed and attempted passes | Deep-lying playmakers, line-breaking #6s |
| OBV from Dribbles & Carries | Value from running with the ball | Wingers, ball-carrying #8s |
| OBV from Defensive Actions | Value from tackles, interceptions, pressures | Ball-winning midfielders, aggressive CBs |
| OBV from Shots | Value created or lost from shots taken | High-volume finishers vs wasteful shooters |
| OBV from Goalkeeper Actions | Value from saves, claims, distribution | Modern sweeper-keepers |
| Total OBV | Sum of all categories | Genuinely complete players |
Data reflects Hudl Statsbomb’s published OBV framework as documented in their 2021 model release. Individual league and season variance applies.
The category split is what makes OBV genuinely useful for scouting. A central midfielder with high pass-OBV but negative dribble-OBV is a builder who should not be asked to drive the ball. A winger with high dribble-OBV but mediocre pass-OBV is a 1v1 specialist who needs a creator next to him. The single number summarises; the six categories diagnose.
Declan Rice 2025-26: The Premier League’s OBV Benchmark
Through the opening months of the 2025-26 Premier League campaign, Hudl Statsbomb’s all-encompassing ‘On-Ball Value’ (OBV) metric paints Arsenal midfielder Declan Rice as the Premier League’s most important player this season. Per Football Insider’s reporting on Hudl Statsbomb’s data, Rice posted an OBV figure of 3.22, 2.2 of which comes from his passing as of early October. The remaining 1.85 came from his dribbles and carries, where he led every player in the league. Football InsiderFootball Insider
The context matters. Arsenal won the 2025-26 Premier League title under Mikel Arteta – their first league title since the Invincibles of 2003-04. Rice’s role in that team had quietly evolved from the destroyer profile he wore at West Ham into something closer to a left-half-space progressor, given license to step out of his pivot position once Martin Zubimendi anchored the base. His OBV signature – high pass value, league-leading carry value, healthy defensive value – is the analytical fingerprint of that role change.
The figure also exposes something traditional stats miss. Rice did not lead the Premier League in tackles, interceptions, key passes, or assists in that period. He led in the metric that measures the underlying value of everything he did with the ball, which is what an OBV total captures. Seasonal OBV totals predict future performance better than goals and assists alone. OBV captures consistent process quality rather than variance-prone outcomes, as analytics writers have noted using the publicly released framework. Medium
Recruitment Use Case: How Clubs Filter With OBV
The cleanest application of OBV is in the recruitment funnel, where analysts filter a continental player pool down to a video shortlist. We start by filtering for Central Defenders who played >75% of available minutes across the 2025-26 season, ensuring our sample size is robust. We then apply filters to target players who are at least the positional average for the following metrics: On-Ball Value (OBV) generated from dribbles and ball carries, per Hudl Statsbomb’s documented scouting workflow. Hudl
The logic is profile-first. A club running a possession-heavy system with an inverted left-back and a ball-playing central defender does not need a 6-foot-4 stopper who clears every aerial duel. They need a defender whose OBV from passes and carries is at least positional average. OBV filters that profile in seconds across the 140-plus competitions Hudl Statsbomb covers, before any video work begins. The filter does not replace scouting; it routes scouting toward the right names.
The risk-adjusted nature of OBV is what makes the filter trustworthy. A defender who attempts twenty progressive passes per match and completes ten is not the same as one who attempts ten and completes nine – traditional progressive pass counts treat them similarly. OBV punishes the five turnovers in the first defender’s profile and credits the cleanness of the second. The signal that survives is genuine ball-progression skill, not raw volume.
The Weakness: Where OBV Falls Short
OBV is a possession value model, which means it values events with the ball. Off-ball actions – the run that drags a centre-back five yards out of position, the cover shadow that shuts off a passing lane – do not register. A player who consistently creates space for teammates but never receives the ball will look ordinary on OBV. This is the same blind spot every event-data model carries, and it is why clubs running serious analytics departments now combine OBV with tracking-data metrics from providers like SkillCorner to cover the off-ball gap.
The second weakness is sample-size dependence on the defensive side. A central defender who plays in front of a dominant team may only face fifteen meaningful defensive actions per match, where a midfielder generates a hundred-plus on-ball events. The OBV signal stabilises faster for ball-progressing players than for low-event defenders, which is why one-season OBV totals are a stronger argument for midfielders and wingers than they are for centre-backs.
The third caveat is interpretive. OBV scores actions, not decisions. A player who refuses a high-value pass that would have unlocked the defence, and instead recycles to the goalkeeper, loses nothing on OBV – the recycle pass is small but positive. The model cannot see the better pass that was available and not taken. By using possession value models, teams and analysts can better understand the impact of their players’ actions and make data-driven decisions to improve their performance, per Hudl Statsbomb, but the model evaluates what happened, not what could have. Hudl
What This Means on the Pitch
On-Ball Value is the strongest single-number summary of a footballer’s in-possession contribution that the public domain has access to. It does what xT cannot – value defensive actions, account for risk, penalise turnovers – and it does what raw progressive stats cannot: weight every action by its actual change in goal probability, not by an arbitrary zone cutoff. For evaluating midfielders specifically, it is now the analytical baseline rather than a supplementary number.
The Declan Rice 2025-26 figure is the cleanest demonstration. A player who did not lead the league in any traditional category led it in the metric that measures the underlying value of his actions. Arsenal won the title with him as their structural anchor. The two facts are not unconnected, and OBV is what makes the connection legible. The honest verdict is this: any analyst still leading with goals, assists, and key passes when evaluating modern central midfielders is using a 1990s toolkit on a 2026 game.
What Do You Think?
Rice’s 3.22 OBV in the early stretch of 2025-26 sits 0.8 above the next-best Premier League midfielder per Hudl Statsbomb’s published figures. But does a single-number OBV total ever outrank watching a player live for ten matches? Drop your take below.
Related Tactical Breakdowns
Expected Threat (xT) Explained Why it connects: xT is the possession value model OBV was built to improve on. Reading the two side by side is the cleanest way to see why Hudl Statsbomb made the design choices it did.
Expected Assists vs Key Passes Why it connects: Like OBV, xA argues that the act of creating a chance has measurable value separate from whether it ended in a goal. The two metrics belong to the same analytical generation.
Packing: Measuring Vertical Efficiency Why it connects: Packing measures the number of opponents bypassed per action. It is a structurally simpler progression metric than OBV, and the contrast shows what OBV’s probability-shift framework actually buys you.
Frequently Asked Questions
What is the difference between OBV and Expected Threat (xT)?
OBV values every action including shots and defensive events, while xT values only zone-to-zone progression. The structural gap is OBV’s Goals Against component, which penalises turnovers; xT does not. OBV is also trained on Hudl Statsbomb’s xG model, giving it a higher-resolution scoring probability foundation than the zone-grid approach xT uses.
How is On-Ball Value calculated?
OBV is calculated as the change in two probabilities: the chance the team scores in the next sequence, minus the chance the team concedes, before and after each action. Each action gets a positive or negative value based on that probability shift. Player and team totals are the sum of all action values, available across six event categories.
Why is OBV important in football scouting?
OBV is important in football scouting because it lets clubs filter a continental player pool by specific in-possession profile, not just by raw output. A scout can isolate centre-backs with above-average dribble and carry OBV in seconds, then move to video. The metric routes scouting toward the right names before any film work begins, saving analyst time.
Which player had the highest OBV in the 2025-26 Premier League?
Declan Rice held the highest On-Ball Value figure in the Premier League during the opening months of 2025-26, with a 3.22 total per Hudl Statsbomb’s published data via Football Insider. 2.2 of that came from his passing and 1.85 from his dribbles and carries. Arsenal went on to win the Premier League title that season under Mikel Arteta.
