The Goalkeeper and the Penalty: definition and limits

A penalty is the only event in football where the goalkeeper is expected to fail. The metrics that surround it are thin, noisy and frequently misunderstood, and they are worth treating separately from everything else.

The arithmetic of a penalty

A penalty carries a pre-shot value near three-quarters of a goal, which makes it the single most valuable standard event in the sport. The keeper's baseline expectation is to concede, and a save is a large positive deviation rather than a routine act. Any metric that treats a penalty save as an ordinary save will misstate the keeper's contribution.

That baseline is the reason penalties are excluded from post-shot keeper metrics on most careful treatments. Including a handful of near-constant high-value events adds a large random term to a season total and makes the resulting figure much harder to interpret than the shots it is meant to describe.

What actually predicts a save

The evidence points towards preparation and timing rather than reflex. Keepers who save penalties disproportionately do so by holding their ground slightly longer and by committing to a side based on studied patterns rather than reacting to the ball, since the flight time from the spot leaves almost no margin for a reaction save.

The practical consequence is that penalty saves cluster by keeper in a way that ordinary saves do not, which is unusual in a sport where most overperformance regresses. A keeper with a genuinely strong penalty record may be exhibiting a durable skill, though the number of penalties any keeper faces keeps the sample permanently small.

The sample problem

A keeper who faces five penalties a season and saves one has a twenty per cent save rate. Another who faces five and saves none has zero. The difference between them is one event, and the resulting statistics are quoted as though they described abilities. Nothing in the metric family corrects for a sample this small.

This is why penalty save rates should be reported with the number of penalties faced and should never be used to rank keepers. Over a career a keeper may face fewer than fifty penalties, which is a sample small enough that a ranking by save rate mostly ranks the luck of the draw.

The shoot-out distinction

Penalty shoot-outs are a separate event class and are often reported together with in-match penalties, which muddles both. Shoot-outs have different pressure, different preparation and a different scoreline context, and their inclusion in a career tally of penalty saves produces a figure that describes two unlike situations.

The clean treatment reports in-match penalties, shoot-out penalties and total separately. It is more columns than anyone wants, and it is the only presentation that allows two keepers to be compared without an invisible difference in how their events were classified.

The shooter's side of the ledger

Penalty taking has its own metrics and its own regularities, including the observation that conversion falls when a team is relying on the kick to stay in major competition. The keeper is the smaller half of the story, and a table that records only saves describes one participant in an event that has two.

The balanced presentation therefore reports penalties faced, penalties saved, penalties conceded and the shoot-out record separately. The four figures together describe a keeper's history from the spot, and each in isolation describes nothing much.

Reporting honestly

The honest caption says how many penalties are included, over what period, and whether shoot-outs are counted. It also says the sample is small, because it always is. A penalty save rate without those qualifications invites exactly the over-reading that the number cannot support.

The deeper point is that goalkeeping metrics in general describe a low-event position, and penalties are the extreme case of that position. Every conclusion drawn from penalty data is provisional, and stating the provisionality is not hedging; it is the accurate description of what the numbers can bear.

Key reference points

• A penalty is worth about three-quarters of a goal, so a save is a large positive deviation.
• Most careful keeper metrics exclude penalties and report them separately.
• Penalty saves depend more on preparation and timing than on reflex.
• Samples are tiny; a season's rate can hinge on one event.
• In-match penalties and shoot-outs are different event classes.
• Always state the count, the period and whether shoot-outs are included.

Penalty statistics are the clearest case in football of a number that is arithmetically correct and analytically almost empty, and reporting them with their sample is the only honest treatment.