When debating xG vs xA and which soccer metric should fans trust more, the answer isn’t a binary pick. It’s about knowing whose xG and xA you’re reading. In 2026, fantasy draft rooms and betting apps overwhelmingly rely on these expected numbers, but far too few gamers realize that Opta and StatsBomb—the two giants in soccer data—reach their numbers through very different roads. That divergence can lead to wildly different evaluations of the same player, turning a “must-draft” into a “steer clear” depending on which provider your platform uses. Let’s pull back the curtain on both models and see which one deserves your confidence when the clock is ticking.
Not All Expected Values Are Created Equal
At face value, xG measures the probability that a shot becomes a goal based on a set of historical variables. Expected assists (xA) go one step further: they assign credit to the passer based on the quality of the shot that follows. Simple enough. But the devil hides in the feature engineering. Opta and StatsBomb use different raw data sources, different contextual inputs, and even different definitions of what counts as a shot or a key pass. As a result, the same 25-yard strike can be worth 0.03 xG on one platform and 0.09 on another. That gap might look small, but over a 30-game fantasy season, it shifts player rankings enough to sink a draft strategy.
Opta’s xG: The Dependable Workhorse
Opta has been collecting soccer event data for over two decades. Their xG model is built primarily on shot location, body part, and the angle to goal, with a few extra features like whether the shot came from a counterattack or a set piece. It’s a solid, transparent model that performs well in aggregate, and you’ll find it baked into many mainstream platforms like Sky Sports, the official English Premier League site, and a host of fantasy games. Opta’s approach treats every shot in a similar context, without over-individualizing the situation. That global consistency makes it a fair benchmark across leagues.
However, Opta’s model doesn’t account for the goalkeeper’s exact position or the pressure on the shooter. In a crowded penalty area, a point-blank header from a corner gets roughly the same xG as a free header from a cross, as long as location and body part match. This is fine for most analyses, but it fails to capture the nuance that distinguishes a great chance from a merely decent one. For fantasy managers, that blindspot means xG numbers from Opta are often conservative, especially for strikers who thrive on scrappy rebounds or dynamic counterattacks.
StatsBomb’s xG: High-Resolution Rivalry
StatsBomb shot onto the scene with a promise: no more one-size-fits-all expected goals. Their model feeds on high-resolution data, including micro-event tracking, player positions at the moment of the shot, and the distance and speed of the ball. StatsBomb’s xG is trained on a far richer dataset that can separate a shot from 12 yards with two defenders closing down from the same shot with a clear sight of goal. That granularity makes their values look “sharper” and often more intuitive when you watch the build-up back on video.
Yet that refinement comes with a cost. StatsBomb’s model heavily depends on the quality and scope of their tracking data, which isn’t available for every competition. Lower-tier leagues or older matches might rely on older broadcast feeds, introducing subtle noise. Moreover, StatsBomb deliberately tunes their model to minimize overperformance—it rarely gives out huge xG for penalties, for example, to avoid generating inflated numbers from a single dominant match. This is analytically rigorous, but it can make a player’s xG look lower than what a fan “feels” a chance deserved.
xA: Where the Divergence Sets Fantasy Pitfalls
Expected assists amplify the differences between the two giants because xA is derived from xG—the same chance that follows the pass receives a value from the creator’s model. A fancy pass that slices through a defense and leads to a shot from a narrow angle might get 0.15 xA with Opta but 0.28 xA with StatsBomb if the model rewards the pass for beating two defenders and leaving the shot open. Meanwhile, a routine square ball to an open teammate at the edge of the box could be 0.12 xA in Opta and only 0.05 in StatsBomb, because it gave the shooter little time but a clean sight.
This is exactly where fantasy draft traps multiply. Suppose you’re eyeing a creative midfielder whose xA exploded in 2025, and your draft tool uses Opta. During training you might see that other platforms using StatsBomb have that same player ranked 11 spots lower. Which one is right? The answer is that neither is “wrong”—they’re just measuring different things. Opta’s xA is more about the pass location, while StatsBomb’s xA is more about the receiver’s resulting shot quality. If you blindly trust one, you’ll consistently draft players who excel in only one type of chance creation.
Fantasy Draft Traps to Sidestep
The most common trap is over-relying on a single provider’s xA when building your draft board. Here are three concrete scenarios where the choice of model matters:
- Set-piece takers. A dead-ball specialist who whips in crosses toward a tall striker might rack up high xA under Opta, because the model rates headers from central areas as solid chances. StatsBomb’s model might be more skeptical, especially if the striker is usually surrounded by defenders, dropping the xA value significantly.
- Late-game finishers. Players who frequently come on against tired legs and shoot from close range can inflate their passer’s xA late in matches. Opta’s model treats those opportunities as straightforward on-target efforts, while StatsBomb’s model might reduce the xG because of the high defensive pressure even late in a game.
- Cross-heavy wingers. A winger who piles up low crosses into crowded boxes often looks like a fantasy hero with Opta’s xA, because every cross counts as a key pass. StatsBomb’s xA will penalize those crosses if they don’t result in genuinely open shots, rewarding instead players who cut back to the edge of the box for more deliberate attempts.
None of these discrepancies mean we should throw both models out. Instead, switch your approach: pull up both xA values for any player you’re considering in the top five rounds of your draft. If the gap is larger than 0.05 per 90 minutes, dig into the video. Look at the actual chances that player created. That manual sanity check is worth far more than trusting a single algorithm’s number.
How Smart Drafters Use Both Metrics in 2026
The path forward is not about choosing between xG vs xA or between Opta vs StatsBomb—it’s about triangulating them. In 2026, fantasy tools are starting to overlay possession-based models and player tracking data, meaning the raw xG and xA from any provider will soon be just one layer of analysis. Smart drafters will default to the provider that matches their league’s official scoring system, but they’ll also check the alternative provider for a second opinion. If a forward’s xG is consistently higher in Opta but lower in StatsBomb, that tells you he improves his chances by getting into great positions rather than by creating individual brilliance. Conversely, if a midfielder’s xA is higher in StatsBomb, you know he’s unlocking high-quality shots, not just racking up launch numbers.
Even better is to track how a player’s xG and xA evolve across the season. A footballer who overperforms his xG by a full point in one provider’s model might be due for regression, but if that overperformance is mirrored in the other provider’s model, you can adjust your expectations accordingly. The modern data fan, armed with two lenses, sees the game in 3D rather than a flat spreadsheet.
Conclusion: Trust the Process, Not Just the Number
There is no “better” metric in the absolute sense, only better informed analysts. Opta’s model offers consistency and long-standing reliability, while StatsBomb provides depth and contextual sensitivity. For fantasy drafters, the real winning move is knowing which one your platform runs on and then acting as your own scout. Cross-reference the xG vs xA numbers across both formulas, don’t chase a single glowing stat, and remember that every expected metric is a guess, not a prophecy. With that mindset, you’ll avoid the traps that catch managers who simply trust whichever number is in big text.
