Reading Australian Cricket Statistics: Averages, Rates and Traps

Cricket statistics explained properly start with one uncomfortable fact: almost every headline number in the game is a ratio, and a ratio can be moved by changing either the top or the bottom of the fraction. A batting average divides runs by dismissals, not by innings. A bowling strike rate divides balls by wickets, not by overs. An economy rate divides runs by overs, and says nothing at all about whether anyone got out. Once you know which quantity sits underneath each number, most of the arguments that fill Australian talkback radio in December resolve themselves in about ten seconds.

The short answer for anyone who wants it up front: a batting average is runs per dismissal and is inflated by not outs; a batting strike rate is runs per hundred balls and is the more honest measure in white-ball cricket; a bowling average is runs per wicket and a bowling strike rate is balls per wicket, and the two together tell you far more than either alone. Everything else on this page is about the situations where those definitions quietly mislead, and about the Australian context - Sheffield Shield pitches, a home summer played on hard, bouncy surfaces, a Big Bash competition designed to make batters take risks - that changes what a given number is worth.

This is a guide to reading numbers, not a records list. If you want records, the Australian player pages and the history of Australian cricket cover that ground. What follows is the set of habits that separate someone who can quote an average from someone who can tell you whether the average means anything.

What a batting average actually measures

A batting average is total runs divided by the number of completed innings, where "completed" means the batter was dismissed. Innings in which the batter finished not out contribute their runs to the numerator but do not add to the denominator. This is not an accident or an oversight. The convention dates from the nineteenth century and rests on a defensible idea: if a batter was 60 not out when the innings ended, we do not know what they would eventually have scored, so it is unfair to treat 60 as a finished innings.

The trouble is that the convention treats every not out identically. A number eleven stranded on 2 not out and a number five stranded on 140 not out both escape the denominator. Over a career, the batters who benefit most are those who bat where innings tend to end around them: lower-middle order in first-class cricket, and finishers in one-day and T20 cricket. In Australian domestic cricket this shows up most clearly in the Marsh Cup and in Big Bash records, where a specialist finisher can carry an average in the forties on a modest body of work.

None of this makes the average wrong. It makes it specific. The batting average answers the question "how many runs does this player produce between dismissals", and that is a genuinely useful thing to know. It simply does not answer "how many runs will this player score today", which is the question most people think they are asking.

The not-out problem, worked through

The cleanest way to see the effect is to compute both numbers. Runs divided by dismissals gives the average. Runs divided by innings gives runs per innings, sometimes called the arithmetic mean per appearance. The gap between them is a direct measure of how much the not outs are doing.

ESPNcricinfo's long-running analysis of the not-out question makes the point with Test-era comparisons: two batters with similar run tallies can sit far apart on average purely because one played in sides whose innings ended around them and the other did not. Not-out rates are also higher in wins than in losses, for the obvious reason that a side chasing successfully leaves batters unbeaten. A player who spent a career in a dominant Australian side of the late 1990s and 2000s therefore had structurally more chances to finish not out than a contemporary in a struggling side.

The practical habit is simple. Whenever you see an average quoted, look immediately for the innings and not-out columns beside it. If not outs are under about 8 per cent of innings, the average and the runs-per-innings figure will be close and the average is doing honest work. If not outs are above 15 per cent, treat the average as an upper bound and look at the runs-per-innings number as the lower bound. The truth about the player is somewhere between them.

Bradman's 99.94 and why the arithmetic matters

The most famous number in Australian sport is a worked example of everything above. Don Bradman played 52 Tests and batted 80 times. He scored 6,996 runs, with 29 centuries and 12 double centuries. He was not out on 10 occasions, so 70 of his innings ended in a dismissal. Six thousand, nine hundred and ninety-six divided by seventy is 99.94.

Divide the same 6,996 runs by all 80 innings and you get roughly 87.45 runs per appearance. That is still an absurd figure - no other Test batter in history is close to it - but it is not 99.94, and the 12.5-run gap is entirely the not-out convention at work. The 10 not outs represent about 12.5 per cent of his innings, which is a normal rate for a top-order batter in a strong side, not an anomaly.

• 52Tests played
• 80Innings batted
• 10Not outs
• 29Test centuries

Don Bradman's published Test career totals. The average is runs divided by dismissals, not by innings, which is why 80 innings and 70 dismissals give different answers.

The reason to work through Bradman rather than a modern player is that the extremity makes the mechanism visible. Nobody disputes that he was the best batter who ever lived. What the arithmetic shows is that even an unambiguous number carries a convention inside it, and that a reader who does not know the convention is reading something slightly different from what is written.

• Innings ending in a dismissal87.5%
• Innings ending not out12.5%

A structural split of a published career record. It counts how many innings ended in dismissal versus not out, and says nothing about the quality of those innings.

Show the numbers

How Bradman's 80 Test innings ended

Item	Value
Innings ending in a dismissal	70
Innings ending not out	10

The mean hides the shape of a career

An average is a single number standing in for a distribution, and distributions in cricket are heavily skewed. Most innings are short. A small number are very long. The mean is dragged upward by the long ones, so it sits well above the score a batter most commonly makes.

This matters when you are trying to predict rather than to summarise. A batter averaging 45 does not typically make 45. They more often make 10 or 15, occasionally make 120, and the mean lands where it lands. Two batters with identical averages can have completely different shapes: one consistent, grinding out 35 to 55 most weeks, the other alternating single figures with hundreds. For a Test side that needs someone to survive an hour against the new ball at the Gabba, those two players are not interchangeable, even though the statistic says they are.

Australian selectors and state coaches deal with this by looking at innings-by-innings sequences rather than aggregates, and by counting things the average cannot show - fifty-plus scores, conversion of fifties into hundreds, and how often a batter survived the first 20 balls. None of that is visible in a career line. All of it is available on any full scorecard archive, and reconstructing it takes a few minutes.

Strike rate means three different things in three formats

Batting strike rate is runs per hundred balls faced. The definition is constant across formats. Its meaning is not.

In Test cricket a strike rate is mostly a description of temperament and match situation. A strike rate of 45 is unremarkable for an opener asked to see off a new Kookaburra ball on a green first-morning surface. A strike rate of 75 from a middle-order batter usually indicates a player who counter-attacks. Neither is inherently better; what matters is whether the tempo suited the state of the match.

In the Marsh Cup and other 50-over cricket, strike rate becomes a genuine trade-off against average, because the resource being spent is balls and there are only 300 of them. In the Big Bash, with 120 balls, strike rate is the dominant batting measure and average is close to secondary. A Big Bash opener who scores 25 off 12 balls has done more for the side than one who scores 40 off 40, and only one of those innings improves an average. This is why comparing a Shield batting average to a Big Bash batting average is meaningless, and why the Big Bash format guide is worth reading before drawing conclusions from T20 numbers.

Bowling average, strike rate and economy are one triangle

The three standard bowling numbers are arithmetically linked. Bowling average is runs per wicket. Bowling strike rate is balls per wicket. Economy rate is runs per over. Given any two, the third follows: average equals strike rate divided by six, multiplied by economy rate.

That relationship is the most useful thing a casual reader can learn, because it tells you which number is doing the work. A bowler with a good average and a poor strike rate has bought that average with economy - they are hard to score off but do not take wickets quickly. A bowler with a good strike rate and a poor economy is a wicket-taker who leaks runs. In a Test match at the MCG on a fifth-day pitch, the second bowler is more valuable. In a Big Bash final, the first is.

Australian conditions push the triangle in a specific direction. Hard, bouncy pitches with true carry generally produce lower economy rates and better strike rates for fast bowlers than subcontinental surfaces do, while spinners in Australia typically show worse strike rates than they would in Asia. That is a property of the environment, not of the bowler, and it is one of the reasons the Australian pitch conditions discussion belongs in any statistical comparison.

Reading Nathan Lyon's Test line

A worked example makes the triangle concrete. As of the middle of 2026, Nathan Lyon's published Test record stood at 143 matches, 571 wickets at an average of 30.20, with best innings figures of 8 for 50, 24 five-wicket hauls and 5 ten-wicket matches, from 35,117 deliveries.

From those figures the rest follows by arithmetic. Dividing 35,117 balls by 571 wickets gives a strike rate of about 61.5 balls per wicket - roughly a wicket every ten overs. Multiplying 571 wickets by an average of 30.20 gives about 17,244 runs conceded, and dividing that by 5,852 overs gives an economy rate of just under 3.00. Those three numbers together describe a bowler who is genuinely hard to get away, who takes a wicket at a workmanlike rate rather than an explosive one, and who has done it across a volume of overs that very few spinners anywhere have matched.

Compare that with the two great Australian bowlers of the previous generation. Shane Warne finished with 708 Test wickets and Glenn McGrath with 563, both at averages in the low to mid twenties, and both at strike rates meaningfully better than Lyon's. The gap is real, and no amount of context erases it. But the volume figure - Lyon's 35,117 Test deliveries - is the number that explains why he has been picked for a decade and a half, and it is the one that never appears in a headline.

Sheffield Shield averages are not Test averages

The single most common statistical error in Australian cricket conversation is comparing a Shield average with a Test average as though they were measured on the same scale. They are not.

The Shield is played in October, November, February and March, largely outside the deep summer, on pitches that are frequently green and often at grounds prepared with less traffic than a Test venue. Cricket Australia's own end-of-season assessment of the 2025-26 competition observed that only a handful of players averaged better than 40 with the bat while a large number of bowlers averaged under 20. A Shield batting average of 40 in that environment is a strong season. A Test batting average of 40 is a serviceable career.

This works in both directions. A bowler who averages 24 in the Shield is not necessarily a better bowler than a Test bowler averaging 27, because the Shield bowler has been operating in conditions that suit him. Selectors know this, which is why Shield form is weighed rather than counted - a point explored at length in the companion guide to how Australian selection actually works.

The 2025-26 Shield: a season where 40 was a lot

The most recent completed Shield season gives a usable snapshot of that scoring environment. Victoria topped the table with 60.86 points from 10 matches, and South Australia finished second on 44.81 before winning the final at Junction Oval by 56 runs from 26 to 30 March 2026 - South Australia's fifteenth Shield title and their first back-to-back pair.

Cricket Australia's team-of-the-season selection gives the individual aggregates.

Player	State	Runs	Batting average	Wickets	Bowling average
Cameron Bancroft	Western Australia	674	33.70	-	-
Sam Harper	Victoria	645	37.94	-	-
Matthew Renshaw	Queensland	499	49.90	-	-
Peter Handscomb	Victoria	688	36.21	-	-
Jordan Silk	Tasmania	586	34.47	-	-
Jake Lehmann	South Australia	561	40.07	-	-
Liam Scott	South Australia	496	41.33	23	25.52
Sam Elliott	Victoria	-	-	33	17.18
Corey Rocchiccioli	Western Australia	-	-	38	28.10
Liam Hatcher	New South Wales	-	-	44	19.84
Cameron Gannon	Western Australia	-	-	42	24.38

Read that table with the average-versus-volume distinction in mind. Peter Handscomb scored the most runs of the group at 688, but at 36.21. Matthew Renshaw had much the better average at 49.90, from 499 runs - fewer innings, higher rate. Which of those was the better season depends entirely on whether you are asking who produced most for their side across a whole summer or who was hardest to dismiss.

Liam Hatcher44

Cameron Gannon42

Corey Rocchiccioli38

Sam Elliott33

Season wicket tallies for the four specialist bowlers named in Cricket Australia's team of the season. It is not the complete competition wicket list, only the bowlers selected in that XI.

Show the numbers

Leading wicket takers, 2025-26 Sheffield Shield team of the season

Item	Value
Liam Hatcher	44
Cameron Gannon	42
Corey Rocchiccioli	38
Sam Elliott	33

58%42%

• Peter Handscomb runs688
• Matthew Renshaw runs499

Peter Handscomb's and Matthew Renshaw's 2025-26 Sheffield Shield run tallies. The two players had similar-quality seasons by different routes, which is exactly what a single average cannot show.

Show the numbers

Runs against average, two Shield batting seasons

Item	Value
Peter Handscomb runs	688
Matthew Renshaw runs	499

Home and away splits change the picture

Australian pitches reward pace, bounce and back-foot play. That environment flatters some players and punishes others, and because Australian domestic cricket is played almost entirely within Australia, a player's entire first-class record can be built in a single set of conditions.

For international comparison this matters enormously. An Australian batter's record in Australia and their record in India, England or the West Indies can differ by 15 runs or more per dismissal, and the direction of the gap tells you something real about technique. The same applies to bowlers: Australian quicks generally have superior home numbers, and Australian spinners generally have superior away numbers, for reasons that have nothing to do with talent.

Any serious comparison therefore has to be filtered. ESPNcricinfo's statsguru lets you split a career by country, ground, opposition and date range in a few clicks, and doing so is the difference between a defensible claim and a pub argument. When Ashes selection debates start each summer, the home-and-away split is usually the number that decides them internally, and almost never the one quoted publicly.

Batting position changes everything

Runs at number three are not the same as runs at number six. The number three bats against a hard ball and fresh bowlers, often within a few overs of the start. The number six bats against an older ball, tired bowlers, and frequently with a platform already built.

Every major statistics service allows filtering by batting position, and the exercise is worth doing whenever an average looks surprising. A player whose career average is boosted by a stretch in the lower middle order will look different once the number is restricted to the top four. Conversely, a top-order batter with a modest average may look substantially better once you remove the innings they played out of position.

In Australian domestic cricket this is especially relevant because states move players around the order to cover injuries and international call-ups. A Shield batter may have batted at three, five and seven in a single season. The season average combines those roles into one figure that describes none of them.

Sample size and the minimum-innings trap

Small samples are where cricket statistics do their worst damage. A batter with 8 innings and one big hundred can carry an average in the sixties that will not survive contact with a full season.

Most record lists apply a qualification threshold - a minimum number of innings, or a minimum number of balls bowled - specifically to keep these cases out. When you see a "leading averages" list that has no threshold, treat it as decoration. When you see one that does, note where the threshold sits, because a list qualified at 10 innings and one qualified at 20 innings will produce very different names.

A rough working rule for batting: below about 20 completed innings, the average is dominated by one or two scores. Below about 40, it is still moving noticeably. Bowling averages stabilise a little faster in terms of matches, because a bowler accumulates 30 or 40 wickets in a Shield season while a batter accumulates perhaps 15 dismissals.

Balls faced is the real unit in white-ball cricket

In a Test, a batter can in principle bat all day; time is the constraint and dismissals are the currency. In a 50-over or 20-over match the constraint is balls, and the balls are shared. Every delivery one batter uses is a delivery another cannot.

This is why runs per ball rather than runs per dismissal is the correct primary measure in the Big Bash and the WBBL. A batter with a strike rate of 118 and an average of 22 has usually contributed more than one with a strike rate of 105 and an average of 30, because the extra 13 runs per hundred balls compound across an innings and free up deliveries for whoever bats next.

The corollary is that a T20 batting average should almost never be quoted alone. Paired with a strike rate and a balls-faced total it becomes informative. On its own it is closer to misleading, and it is the number most often cherry-picked in a selection debate.

Economy rate and the phase problem

Economy rate is straightforward arithmetic and treacherous in interpretation, because in white-ball cricket the value of an over depends heavily on when it is bowled. Powerplay overs and death overs are far more expensive than middle overs, by a wide and consistent margin.

A Big Bash bowler who bowls exclusively in the middle overs will have a better economy rate than one who bowls the nineteenth and twentieth, and the difference has almost nothing to do with skill. Analysts adjust for this by comparing bowlers against the competition average for the phases they actually bowled. Public scorecards do not do that adjustment for you, but the over-by-over record on any full scorecard lets you check which overs a bowler was given.

The same logic applies to bowling averages in the Shield: a first-change bowler who is brought on when the shine has gone is bowling in an easier phase than the opener who took the new ball, even though the season summary treats their wickets identically.

What the numbers do to tail-enders and all-rounders

Batting averages break down at the bottom of the order. A number eleven who scores 6 not out most weeks accumulates a small number of runs and very few dismissals, so their average is volatile and largely meaningless. Nathan Lyon's Test batting record - 1,705 runs at 12.81 with a top score of 47 - is a reasonable illustration: the average is real, it just tells you very little beyond "occasionally hangs around".

All-rounders present the opposite problem. A single number cannot express whether a player is worth a place, because the value comes from the combination. The convention is to compare batting average against bowling average: if the batting average is comfortably higher, the player is contributing on both sides. Liam Scott's 2025-26 Shield season - 496 runs at 41.33 and 23 wickets at 25.52 - clears that bar by a wide margin, and that combination is why he was named in the team of the season and, subsequently, in Australia A squads.

Wicketkeeping and fielding statistics are the weakest data in cricket

A keeper's dismissal count is a function of how often the bowlers find the edge. A keeper standing up to spin on a turning deck and a keeper standing back to a four-pronged pace attack are doing different jobs with different opportunity rates, and the scorecard treats both as "caught".

Byes conceded is marginally more informative about technique, but it is still confounded by how much the ball is moving. The genuinely useful keeping metrics - take rate on chances, movement efficiency, glove work standing up - are collected inside state and national programs and are not published.

Outfield statistics are worse. Catches and run-outs are recorded; drops, misfields, boundary saves and pressure created are not. Anyone quoting a fielding statistic in public is quoting the small fraction of fielding that happens to be countable, which is why the discussion around a great fielder is almost always qualitative.

Career-to-date versus a rolling recent window

The number printed beside a player's name is career-to-date, which mixes their first summer with their most recent one. For a player 12 years into a career, the career figure is a historical document and the recent figure is the selection-relevant one.

The habit worth building is to look at the last two or three seasons separately. Statsguru's date-range filter does this in one step. A batter whose career average is 42 but who has averaged 28 over three seasons is in trouble, and a batter whose career average is 33 but who has averaged 48 over two seasons is in form. Both facts are invisible in the headline.

Australian selectors work this way as a matter of course, weighting the current summer heavily and the previous one moderately. It is one of the few areas where the public number and the internal number are genuinely different, and it explains a good deal of the gap between what talkback expects and what the panel does.

Where Australians actually find the numbers

Cricket Australia's own site carries full scorecards, ladders and player pages for the Sheffield Shield, the Marsh Cup, the Big Bash and the WBBL, and is the primary source for domestic figures. ESPNcricinfo carries the same scorecards plus statsguru, which is the tool that makes filtering practical. Wikipedia season articles are reliable for standings, finals results and season summaries, and are usually the fastest way to confirm a table position.

For historical Australian material, the State Library of South Australia's Bradman collections and the various state association archives fill gaps that the commercial databases do not cover. If you are writing something that will be quoted, confirm any figure in two independent places, because transcription errors propagate quickly through secondary summaries and social media graphics.

The related guides on this site - Australian cricket in general, the Australia hub and the cricket section - link out to the competition explainers that give the structural context those numbers sit inside.

A five-minute routine for reading any stat line

Start with innings and not outs. Compute runs per innings alongside the average and see how far apart they are. If the gap is large, the average is being carried by unbeaten scores and should be treated as a ceiling.

Second, check the date range. Split the record into career-to-date and the most recent two seasons. If they disagree, the recent one is more predictive.

Third, check the venue and opposition mix. Count how much of the record was made in Australia. For an Australian player this is often most of it, and that is worth knowing before comparing them with an overseas peer.

Fourth, for a batter check strike rate and balls faced; for a bowler compute balls per wicket and economy from the published totals. Those derived numbers take ten seconds each and reveal what the headline average was hiding.

Fifth, and last, ask what question the number was built to answer. A batting average answers "runs between dismissals". A strike rate answers "runs per ball". Neither answers "is this player any good", and no single statistic in cricket ever will. That is not a flaw in the statistics. It is a reason to read several of them, and to read the selection process and the coaching pathway that surround them, before deciding what any of it means.

A short glossary for Australian scorecards

Average (batting): runs divided by dismissals. Average (bowling): runs conceded divided by wickets. Strike rate (batting): runs per 100 balls. Strike rate (bowling): balls per wicket. Economy rate: runs conceded per over.

Not out: an innings that ended without dismissal, excluded from the batting denominator. Retired hurt: treated as not out for averaging purposes unless the batter later resumed and was dismissed. Maiden: a wicketless, runless over, still counted in the economy denominator.

Bonus points: in the Sheffield Shield, additional competition points awarded for batting and bowling performance in the first innings, which is why Shield ladder points appear as decimals rather than whole numbers. Percentage and net run rate: ladder tie-breakers used in limited-overs competitions rather than in the Shield.

Learn those twelve definitions and the great majority of Australian cricket statistics become readable without any further study. The remaining difficulty is not arithmetic. It is context - conditions, position, phase, era and sample - and context is the part no scorecard will ever print for you.