PDO Calculator: Shooting Plus Save Percentage
PDO Calculator: Shooting Plus Save Percentage
Ice hockey’s PDO adds an on-ice shooting percentage to an on-ice save percentage. League-wide the two are complementary by construction — arithmetic, not the discovery about luck it is usually quoted as.
PDO — shooting plus save
S% + Sv%50 on-ice goals from 600 shots on goal for, and 42 conceded from 580 against, on the 100 scale
Formula
- the NHL’s own name for it
- “On-ice shooting percentage plus save percentage (5-on-5)”, whose glossary entry adds, in the league’s own words, “Created by Tim Barnes, it is also known as PDO” — the governing league crediting a blogger with a statistic it publishes
- the two components
- goals for over shots on goal for, and one minus goals against over shots on goal against. The NHL’s S% is “goals divided by shots” and “does not take missed shots or blocked shots into consideration”; its Sv% is SV / SA, where “Only shots on goal are considered”. The save side is the goaltender’s work attributed to whoever was on the ice
- the scale
- 1.000 in the NHL’s own data, where the field is skaterShootingPlusSavePct5v5; 100 in most analytics writing. Nothing else differs, and the acronym does not stand for anything
Worked example
50 on-ice goals from 600 shots on goal for, and 42 conceded from 580 against, on the 100 scale
Shooting 100 × 50 ÷ 600 = 8.33%, save 100 × (1 − 42 ÷ 580) = 92.76%, PDO = 101.092 on the 100 scale or 1.011 on the NHL's
Now the identity. Put the SAME complete set of shots on both sides — 48 goals from 620 shots each way — and the percentages are 7.74194% and 92.25806%, summing to exactly 100.000 because 48/620 and 1 − 48/620 are complementary. No luck was involved and nothing regressed
An INDIVIDUAL's need not be 100, and the defaults above show it: 101.092, because the 600 shots taken and the 580 faced are different sets. The components also move independently — drop the goals against to 30 and PDO rises to 103.161, drop the goals for to 40 and it falls to 99.425 — so read the two numbers under the headline, not the sum
Enter 60 goals from 50 shots and the calculator returns no value: goals cannot exceed shots on goal. Nor can zero shots on either side
The NHL’s own PDO, and the sum reproduced
| Skater, 2025-26 NHL regular season, 5-on-5 | On-ice shooting % | On-ice save % | Published PDO | The two added |
|---|---|---|---|---|
| Zach Werenski | 0.099 | 0.906 | 1.0050 | 1.005 |
| Chris Kreider | 0.096 | 0.884 | 0.9800 | 0.980 |
Separating the identity from the empirical claim
| The claim | What kind | Evidence, and its limits |
|---|---|---|
| League-wide shooting % plus save % equals 100 | An arithmetic identity | Provable, with no data and no exceptions: over a complete set of shots the two are complementary. |
| An individual’s PDO tends toward the league mean | An empirical claim about regression | Supported, not quantified. Charron, 60 team-seasons: the highest first-half group fell from 102.2 to 100.0 and the lowest rose from 98.2 to 99.5. Bucket averages, no coefficient. |
| Deviations are mostly luck, but not entirely | An empirical claim about skill | Partly supported. Of 138 skaters above an 11% on-ice shooting percentage over ten seasons, 3.6% repeated; at 10% it was 15.1% and at 9%, 26.5%. The rising rate as the bar falls is itself evidence of a small skill component, which nobody this author read sizes. Independent analytics writing; no standard errors or tests. |
| A specific PDO figure is unsustainable | Not supportable from anything read here | Charron treats a 5-on-5 PDO outside roughly 98.5 to 101.5 as suspect — one writer’s rule of thumb from 60 team-seasons, not a published threshold. |
An identity, an empirical claim, and the sentence that confuses them
This page is about ice hockey — the game played with a puck on an ice rink, whose international rules are the IIHF’s and whose North American professional rules are the NHL’s own. It is not about field hockey, the eleven-a-side game on a turf pitch governed by the FIH, whose periods, set pieces and statistical definitions are different; this site keeps the two sports in separate sections for that reason.
PDO adds an on-ice shooting percentage to an on-ice save percentage. The NHL publishes it — under the name “On-ice shooting percentage plus save percentage (5-on-5)”, with its own glossary recording that it was “Created by Tim Barnes, it is also known as PDO” — and the sum reproduces exactly: Zach Werenski’s 2025-26 figures of 0.099 and 0.906 give the published 1.0050, and Chris Kreider’s 0.096 and 0.884 give the published 0.9800. The acronym stands for nothing.
PDO is not a governing-body statistic in origin. It was named on a hockey blog — the Irreverent Oilers Fans site, active from 2006 to 2008 — and the NHL’s own glossary credits it to Tim Barnes, who wrote there as Vic Ferrari. The acronym stands for nothing at all.
Now the part that is almost always got wrong. PDO is routinely described as regressing to 100, or to 1.000, as though that were a discovery about luck. Across a whole league it is not a discovery at all — it is arithmetic. Every shot on goal is either a goal or a save, so for any COMPLETE set of shots, goals ÷ shots and 1 − goals ÷ shots add to exactly 1. A league at 5-on-5 is such a set, because every shot for one team is a shot against another. The two percentages are therefore complementary by construction, their sum is exactly 100, and no season’s data could show otherwise: take 48 goals from 620 shots on both sides and they come out 7.74194% and 92.25806%, which is 100.000 exactly.
What is genuinely empirical is a different question: how quickly an INDIVIDUAL team’s or player’s PDO moves toward the league mean, and how much of a deviation is repeatable skill. An individual is not a complete set of shots — the shots you take and the shots you face are different events — so nothing forces an individual PDO to 100, and the defaults here give 101.092. On that question the best evidence this author could reach is independent analytics writing rather than a governing body or a journal, and the table below sets out what it shows with that label on it. No source reports a correlation coefficient, a standard error or a significance test, or sizes the skill component. Saying so is more useful than repeating the folk version as settled.
The situation matters as much as the arithmetic. The NHL’s own glossary says of its shot-attempt statistic that “Only 5-on-5 shot attempts are currently counted”, and it publishes the shooting-plus-save sum only as a 5-on-5 figure. A number computed over every situation in the game — power plays, penalty kills, empty nets — is a different number, usually a worse-behaved one, because special teams have systematically different shooting and save percentages. This page does the arithmetic on whatever you give it, so state the situation beside any figure you quote from it.
Two practical consequences. Read the components rather than the sum, because a PDO of 103 driven by the save side says something quite different from one driven by shooting. And expect no band: a threshold would be drawn either around an identity, where it means nothing, or around a noise distribution, where it would be one author’s sample sold as a norm. For the shot attempts behind both see Corsi and Fenwick, and for the NHL’s definitions of the shots and saves see save percentage.
Frequently asked questions
What is PDO in ice hockey?
On-ice shooting percentage plus on-ice save percentage, conventionally at 5-on-5. The NHL publishes it and its glossary notes it was “Created by Tim Barnes, it is also known as PDO”. The acronym stands for nothing.
Why does PDO average 100?
Because it has to. Every shot on goal is a goal or a save and every shot for one side is a shot against another, so league-wide the two are complementary and sum to exactly 100 — an arithmetic identity, not a finding about luck, and no season’s data could contradict it.
Does PDO regress to 100?
An individual’s does tend toward the mean, and THAT part is an empirical claim rather than arithmetic. Charron, using 60 team-seasons, found the highest first-half PDO group fell from 102.2 to 100.0 in the second half of the same season. How fast it moves and how much is skill is not quantified in anything this page could cite: the usual one-line summary merges a proof with a measurement.
Is a high PDO just luck?
Mostly, not entirely, and the evidence is weaker than the confidence with which this is usually asserted. Of 138 skaters above an 11% on-ice shooting percentage over ten seasons, 3.6% repeated; at 10% it was 15.1%, at 9%, 26.5%. The rising rate as the bar falls is itself a sign of a small skill component, which no source read here sizes.
Why does this page not tell me whether my PDO is good?
Because any line would be drawn either around an identity, where it has no meaning, or around a noise distribution, where it would be a sample presented as a standard. One writer suggests treating a 5-on-5 PDO outside roughly 98.5 to 101.5 as suspect — a rule of thumb from 60 team-seasons, not a published threshold.
Related calculators
References
- National Hockey League. NHL Stats Glossary, served as JSON at api.nhle.com/stats/rest/en/glossary (read 12 Oct 2026; 103,774 characters). The league’s own definitions and formulas, quoted verbatim wherever this site states one. nhl.com itself is robots-disallowed, and edge.nhl.com/en/glossary redirects to a page with no definitions on it.
- National Hockey League. Skater percentage summaries, 2025-26, as JSON at api.nhle.com/stats/rest/en/skater/percentages (read 12 Oct 2026; 940 skaters). Publishes shootingPct5v5, skaterSavePct5v5 and skaterShootingPlusSavePct5v5 — the league’s own PDO, on the 1.000 scale.
- McKenzie B. “The real story of how Corsi got its name.” TSN, 6 October 2014, excerpted from Hockey Confidential. A television reporter’s account, not a governing body’s: Corsi was named by the blogger writing as Vic Ferrari after Sabres goaltending coach Jim Corsi, who did not invent it — “I just liked his moustache.” Names the Irreverent Oilers Fans site, active 2006 to 2008, as the birthplace of Corsi, Fenwick and PDO, and records Elliotte Friedman identifying Vic Ferrari as Tim Barnes for CBC in August 2014.
- jfresh. “Percentage Luck in Hockey, Explained.” Substack (read 12 Oct 2026). Independent analytics writing, not a governing body or a journal: no correlation coefficients, standard errors or significance tests, and its charts arrived as images, so only its prose figures are used. Supplies the on-ice shooting percentage repeat rates quoted here, and calls team PDO from season to season almost random with a small skill component.
- Charron C. “Analytics mailbag: Save percentages, PDO, and repeatability.” The Leafs Nation, 27 May 2014. Independent analytics writing, and descriptive bucket averages over 60 team-seasons rather than a regression; no correlation coefficient is reported. Supplies the first-half to second-half PDO figures quoted here. The piece is internally inconsistent once, calling a gap its own table puts at 2.3 percentage points “23-point”.