A subscription earns its place by removing a recurring obstacle.
After a Saturday match, two free sites can show different xG totals for the same team. That is not necessarily an error: models and shot-inclusion rules vary. For a quick match check, either figure may be enough. The calculation changes when a published report needs a consistent, citable source, or an automated spreadsheet needs shot-level data without manual copying. A paid feed is worth considering if it fixes that specific bottleneck—not merely because its numbers look more precise. Before subscribing, check coverage, export or API access, update timing, and whether the license permits publication.
Free match pages often show team and player xG alongside scores, shots and lineups. Some add shot maps, making it easier to see whether a high xG total came from several close-range chances or many unlikely attempts. For occasional match reviews, the free websites for xG analysis may provide enough detail without a subscription.
Downloadable tables can help with basic trends: comparing a team’s xG for and against over several matches, or checking whether a player is getting chances despite a quiet scoring run. Before paying for more data, test a free source against the exact work it needs to support:
Check a handful of recent and older matches. Free data is less useful if the figures exist on screen but cannot be gathered consistently.
A paid xG subscription does not automatically provide a more accurate number. The xG figure comes from a model; the fee often pays for access to the data and tools around it. Providers can assign different values to the same chance because their models differ, not because one charges more.
The useful question is what the free source cannot provide:
Someone reviewing a weekend match may need none of these extras. Someone publishing weekly charts may value dependable exports and a licence that permits publication, even if the free figures look similar. Before subscribing, check a sample for missing shots, update delays and usable fields. Then confirm that the access terms allow the intended use.
Verify the exact leagues, cups and seasons needed. A large league count can conceal short archives or gaps in particular competitions.
Check for shot coordinates, match time, player and team IDs, penalty flags and xG values. Definitions and notes on model changes help make comparisons across seasons defensible.
Compare promised refresh times with recent matches. Check whether corrections and backfills are documented; silent revisions can disrupt a recurring report.
A dashboard suits occasional research. Repeatable work needs exports or an API, documented limits and terms that permit the intended use; test a sample before buying.
Pick a handful of matches from the same seasons that both sources cover, including a penalty, a busy match and one with a late correction. Familiarity with how xG is calculated helps distinguish different model estimates from missing data.
Compare shot counts before comparing xG totals. Check whether penalties, blocked attempts and disallowed goals appear in each feed, and whether their treatment is documented.
Match several shots by minute and player, allowing for small timestamp differences. Check whether player IDs stay stable across seasons and whether name changes or transfers break a simple join.
Save a sample export and check it again after corrections have had time to arrive. A provider that revises historical shots should make those changes traceable.
Different models can assign different xG to the same shot; identical totals are not the test of quality.
FotMob and Understat can show different xG totals for the same fixture without either figure being a mistake. Before combining them, check the metric label and methodology notes: is the figure pre-shot xG or a post-shot measure? Are penalties included, and which attempts count as shots?
A side-by-side comparison of FotMob and Understat xG is a useful starting point, but the definitions attached to the figures matter more than which total looks plausible. Also check whether displayed totals are rounded and whether historical figures can change. In a spreadsheet, record the source, metric name, inclusion rules, and retrieval date. If a rule remains unclear, keep the series separate rather than averaging them.
A monthly price is only the starting point. Annual billing can lower the advertised rate while making a short trial expensive. Check when a plan renews, whether cancellation stops the next charge or ends access immediately, and whether downloaded data remains usable afterward.
For an API or bulk export, the practical cost depends on the work being done:
Access rights deserve a separate check. Permission to view xG on a dashboard does not necessarily include permission to store its underlying records, publish charts made from them, or use them in a commercial product. A hobby blog, a public download and a client report may fall under different terms. Current licence terms—not an old pricing-page screenshot—should settle what is allowed. If the intended use is unclear, written clarification from the provider is more useful than an assumption.
Estimate the requests or exports needed for one normal month, then add any overage charges and the time required to clean and join the results.
The deciding question is how often xG figures must be collected and what happens after collection. A paid plan saves little time if someone checks a few matches each month; it can matter much more when the same research is repeated every week.
A simple threshold helps: estimate the hours spent gathering and correcting data in a typical month, then compare that time with the full cost of the plan. If collection is infrequent or the workflow is still changing, staying with free sources avoids paying for capacity that may never be used. Upgrade when a recurring bottleneck is clear—not merely because the paid dataset has more features.
Choose a league and period typical of the intended work, then define an output: for example, a month of match reports with a team xG chart. Fix the matches and deadline before testing either source.
Record the time spent finding figures, copying them, matching team names and checking missing matches. Keep a note of any figures that cannot be obtained under the free source’s access rules.
Use the same matches and produce the same output. Separate initial setup time from work that would recur each week; a fast first download matters less if every update needs manual repairs.
Count gaps filled and corrections needed, and check whether the paid data improves the chart or report in a way readers would notice. Different xG estimates alone are not evidence that either source is wrong.
Before an annual commitment, request a sample export, read the field documentation and confirm the licence covers the intended publication or automation. If the trial saves little recurring time or leaves key gaps, a shorter plan—or free data—may be enough.
Is paid xG inherently more accurate?
No. A subscription may buy better coverage or access, but accuracy depends on the shot data and model design. Claims of greater accuracy need published validation or a test against a suitable sample.
Should two sources report the same xG total?
Not necessarily. Models can assign different values to the same shot, while sources may also differ on penalties, excluded events, rounding or later corrections. Large gaps are worth checking against shot records before treating either total as wrong.
Does a small project need an API?
Usually not if occasional exports or permitted manual checks cover the work. An API becomes useful when figures must refresh automatically or many matches need repeatable processing; its rate limits and licence terms still matter.
How can competing accuracy claims be checked?
Look for documentation on included shots, model inputs and validation methods, then compare matched matches for missing or revised events. A few differing totals alone cannot show which model predicts outcomes better.
Free xG sources are enough when they cover the required matches and let the work get done without awkward workarounds. A subscription makes sense only when a specific gap keeps recurring—and the time saved or output gained is worth the full cost, including preparation, usage limits and licence restrictions.
Before subscribing, test one must-have capability on one real task: for example, export the shot records needed for a regular report and check whether they can actually be used in the finished work. If that test removes a persistent bottleneck, the smallest suitable paid plan may be worthwhile. If not, keep the free workflow.
Expected goals, or xG, estimates the quality of scoring chances a team created. A shot…
An xG value of 0.05 means that 100 comparable attempts would be expected to produce…
