Goals Saved Above Average: The Stat That Reveals Who Actually Stops Shots

Last updated: October 27, 2026

David De Gea once saved 13 more goals than an average goalkeeper would have across a single Premier League season. That number comes from StatsBomb’s goalkeeper analysis, and a simple save percentage never came close to capturing it. A save percentage treats a tap-in and a top-corner rocket as the same event, so it flatters keepers who face easy shots and punishes those who face daggers. The expected goals framework fixed that imbalance for strikers first, then handed goalkeeping its own correction.

In simple terms: Goals Saved Above Average measures how many goals a keeper prevented compared to what an average keeper would have conceded facing the same shots.

This metric judges a goalkeeper on the difficulty of the chances faced, not the scoreline behind him. Here is how GSAA works, what the 2024-25 Premier League numbers expose, and where the stat quietly misleads.

---

Key Takeaways

• Quality over quantity: Goals Saved Above Average weights every shot by its difficulty, so stopping ten daggers counts for more than stopping fifty tame efforts.
• The league is the baseline: A keeper is measured against what an average keeper would concede facing identical shots, the logic behind StatsBomb and Opta’s goals prevented models.
• Henderson led late in 2024-25: By late April, per Opta, Dean Henderson had prevented 4.4 goals, ahead of Ederson, Guglielmo Vicario and Matz Sels on 4.2.
• Volume distorts the table: Busy keepers on weaker teams can post huge totals simply by facing more shots, so context still matters.
• It is shot-stopping only: GSAA says nothing about distribution, sweeping, or claiming crosses.

---

---

What Is Goals Saved Above Average (GSAA)?

Goals Saved Above Average is a shot-stopping metric that counts how many goals a goalkeeper prevented relative to a league-average keeper facing the same shots. It starts from the difficulty of every shot on target, then compares the goals an average keeper would have conceded with the goals this keeper actually conceded. A positive number means the keeper beat the baseline, and a negative number means he fell short of it.

The metric exists because raw save percentage rewards the wrong thing. A keeper who faces twenty long-range efforts can post a glittering save percentage without ever making a genuinely hard save. GSAA strips that luck out by asking a sharper question: given these exact shots, how many would a typical keeper have stopped?

Think of it as a handicap system for goalkeeping. StatsBomb’s goalkeeper analysis uses the same idea and reports it as goals saved above average. The result is a single number that travels across teams and seasons, which is exactly why recruitment departments lean on it.

---

How GSAA Is Calculated: A Baseline, Not a Save Count

GSAA is calculated by subtracting a goalkeeper’s actual goals conceded from the goals an average keeper would have conceded facing the same shots. The expected figure comes from weighting each shot on target by its difficulty, a process modern data calls post-shot expected goals. Post-shot expected goals reads the placement, pace, and angle of a strike after it leaves the boot, so a shot drilled into the top corner carries a far higher value than one rolled at the keeper.

A worked example makes it concrete. Imagine a keeper faces a season of shots that an average goalkeeper would concede 38 times. If he actually concedes 25, his GSAA is plus-13, the same margin StatsBomb credited to David De Gea in his strongest shot-stopping season.

FBref published the identical concept under the label post-shot expected goals minus goals allowed, often shortened to PSxG plus-minus, until its advanced data was withdrawn in January 2026. The math is the same, and only the branding differs between data providers.

---

What the 2024-25 Premier League Numbers Reveal

The 2024-25 Premier League season showed elite shot-stopping spread well beyond the top of the table. By late April, per Opta’s goals prevented figures, Crystal Palace’s Dean Henderson led on 4.4, with Ederson, Guglielmo Vicario and Matz Sels on 4.2. Brighton’s Bart Verbruggen was bottom on minus-5.3, per the same data.

The table below shows the extremes in Opta’s goals prevented figures as of late April 2025.

Goalkeeper (2024-25)	Post-Shot xG Minus Goals Allowed	Shot-Stopping Read
Ederson (Manchester City)	+4.2	Saved around four goals more than an average keeper facing the same shots
Dean Henderson (Crystal Palace)	+4.4	The league’s best shot-stopping return at that point
Matz Sels (Nottingham Forest)	+4.2	Finished the season on +4.3, fourth in the league, per The Analyst
Bart Verbruggen (Brighton)	-5.3	Conceded roughly five more than the post-shot model expected

Opta goals prevented (expected goals on target minus goals conceded, excluding own goals) as of 23 April 2025, per the Premier League.

The spread tells a story raw tables hide. Ederson and Bart Verbruggen both played behind possession-heavy sides, yet by late April one keeper had added more than four goals of value and the other had given away more than five. That swing of almost ten goals between two keepers in broadly similar systems is precisely the gap this metric was built to surface.

---

Why Volume Quietly Inflates the Table

Shot volume is the biggest confounder in any Goals Saved Above Average ranking. A keeper who faces 150 shots on target has far more chances to bank value than one who faces 70 behind a dominant team. A busy keeper on a relegation side can therefore out-score a serene keeper at a title contender, even when the second keeper is the better shot-stopper on a per-shot basis.

Rate metrics solve half of this problem. Goals prevented per shot faced, or per 90 minutes, strips out the volume effect and reveals efficiency rather than raw accumulation. Good goalkeeping datasets let analysts toggle between the totals and the rates for exactly this reason.

The lesson is to read the counting stat and the rate together. A high total with a weak rate signals a keeper propped up by workload, while a strong rate on low volume signals quality waiting for a bigger stage.

---

GSAA vs Save Percentage and Clean Sheets

GSAA corrects the two numbers fans still quote most: save percentage and clean sheets. Save percentage ignores shot difficulty, so a keeper facing soft chances looks elite while doing very little. Clean sheets reward the whole defence and the fixture list, not the keeper’s hands.

Consider two keepers who share an identical 75% save percentage. If one faced a procession of close-range chances and the other faced long shots, their shot-stopping was not remotely equal, yet the percentage hides it completely. The metric separates them instantly, because it knows what every shot was actually worth.

This is why analysts now treat save percentage as a starting point rather than a verdict. It tells you the outcome, but never the difficulty of the work behind that outcome.

---

The Weakness: What GSAA Cannot Measure

GSAA captures shot-stopping and nothing else, which makes it a dangerous stat to use on its own. A modern goalkeeper is judged on distribution, on sweeping behind a high line, and on claiming crosses, none of which appear anywhere in the number. A keeper can post a glowing shot-stopping figure while quietly gifting goals through poor passing under pressure.

The metric also leans entirely on the quality of the underlying model. Post-shot expected goals cannot see a defender’s deflection, a slip, or a screened sightline, so noise enters every single figure. Small samples make this worse, which is why a brilliant ten-game run can evaporate over a full season.

This is where the goalkeeper distribution profile matters as much as the save count. Manchester City signed Gianluigi Donnarumma in 2025 for his shot-stopping pedigree, then asked him to match Ederson’s footwork, a demand no shot-stopping metric could ever have predicted. Judging that transfer on goalkeeper shot-stopping metrics alone would miss the entire reason it was a gamble.

---

The Final Word

Goals Saved Above Average is the closest football has to an honest shot-stopping scorecard, and it earns its place in every serious recruitment file. It does what save percentage and clean sheets never could, by pricing the difficulty of the work rather than the outcome that followed. The danger is treating it as a verdict on a goalkeeper rather than a verdict on one of his jobs.

Read it next to volume, next to the distribution numbers, and next to your own eyes. A keeper is not the sum of his saves, but the saves are finally measured properly, and that alone is worth a great deal.

---

What Do You Think?

Manchester City let Ederson leave in 2025 and replaced him with Gianluigi Donnarumma, a keeper better known for his shot-stopping than for his feet. Was that swap a tactical upgrade or an expensive mistake? Make your case below.

---

Related Tactical Breakdowns

---

Frequently Asked Questions