Batting average vs strike rate: what each one measures

Two numbers sit next to almost every batter's name on a statistics page, and they are usually read as if they were two versions of the same judgement. They are not. Batting average and strike rate answer different questions, and a batter can be excellent by one measure and unremarkable by the other without any contradiction at all.

The average asks: how much does this batter score before getting out? The strike rate asks: how quickly? One is a measure of survival converted into runs. The other is a measure of runs converted into tempo. Cricket is the only major sport that routinely publishes both, because it is the only one where the resource being spent changes from format to format. In a Test the scarce resource is wickets. In a Twenty20 it is balls. Every argument about which number matters more is really an argument about which resource is running out first.

What the batting average actually calculates

The batting average is runs scored divided by dismissals. Not by innings. By dismissals.

That single detail is where most of the misunderstanding starts. A batter who has played twenty innings and been dismissed fifteen times has an average with fifteen in the denominator, not twenty. The five not-out innings contribute their runs to the numerator and nothing at all to the denominator.

The logic is sound. An average is meant to answer the question, "if this batter walks out now, how many runs will they make before somebody gets them out?" An innings that ended because the overs ran out, or because the last partner was dismissed at the other end, or because the captain declared, did not end because the batter failed. Counting it as a completed failure would understate the batter every time.

That reasoning holds cleanly for a top-order batter who is occasionally stranded. It breaks down badly for a lower-order batter who is stranded as a matter of routine.

How not-outs inflate an average

Consider a number nine who bats in twelve innings, is dismissed in six of them, and remains unbeaten in the other six because the innings ended around him. Suppose he scores 12, 4, 22, 8, 15 and 3 when dismissed, and 9, 14, 2, 18, 6 and 11 when not out. That is 124 runs from twelve innings, which is a shade over 10 per innings. But the average, dividing by six dismissals, reads 20.67. The number has doubled without a single extra run being scored.

Nothing improper has happened. The formula is being applied exactly as defined. But the resulting figure no longer means what a reader instinctively assumes it means. It does not say "this batter typically makes about 21". It says "for every dismissal this batter costs the side, roughly 21 runs arrive", which is a different and much narrower claim.

The distortion grows with three things:

• The proportion of innings ending not out. A batter not out in half of all innings has an average roughly double the runs-per-innings figure.
• Batting position. The lower a batter comes in, the more often the innings ends while they are still there, and the more often they are shielded by a partner farming the strike.
• Format. In limited-overs cricket the innings has a hard stop, so somebody is always unbeaten at the end. Middle and lower-order white-ball averages are systematically inflated by comparison with Test averages at the same position.

The counter-check is simple and any decent statistics page gives you the pieces for it. Divide runs by innings rather than by dismissals. If the two figures are close, the average is telling you something honest. If runs-per-dismissal is far above runs-per-innings, the gap is being manufactured by not-outs, and the average is describing a role rather than a level of skill.

A specialist finisher who comes in with four overs left, is unbeaten far more often than not, and is picked precisely because he does not get out at the death, will always carry a headline average well above what he typically contributes in a single innings. That is not a lie. It is just a number answering a question nobody asked.

The opposite distortion

Not-outs are the famous problem, but the average has a quieter one running the other way. A batter who is regularly sent in as a nightwatchman, or who is pushed up the order in a chase to have a swing at the new ball, is being asked to do a job in which failure is the expected outcome. Every cheap dismissal in that role lands in the denominator at full weight. The average charges the batter for the tactic.

The same applies to anyone who bats in the hardest conditions of a match by design. A top-three batter faces the new ball, seam movement and the freshest bowlers. A number five arrives when the shine is gone. Their averages are placed in the same column and compared as though they had been set the same examination.

What the strike rate actually calculates

Batting strike rate is runs divided by balls faced, multiplied by 100. A strike rate of 135 means 135 runs per 100 deliveries.

It carries none of the not-out problem, because there is no denominator to distort: every ball faced counts, whether the innings ended in a dismissal or not. It is arithmetically the cleaner of the two numbers.

What it lacks is any sense of duration. A strike rate says nothing about how long the batter lasted, and a very high one over a handful of balls is close to meaningless. A batter who faces four balls and hits two of them for six has a strike rate of 300 and has told you nothing you could rely on. Strike rate needs a volume of balls behind it before it stabilises, in the same way an average needs a volume of innings.

There is also a counting rule worth knowing. Balls faced includes no balls, because the batter had a delivery to deal with, and excludes wides, because the batter could not reach it. Runs scored off the bat from a no ball count to the batter. A batter who is bowled at repeatedly outside the tramlines therefore has a strike rate that is not diluted by deliveries they never had a chance to hit.

Why the average rules in Tests

A Test innings has no ball limit. The side bats until ten wickets fall or the captain declares. The only thing that ends an innings against the batting side's will is the loss of wickets. Wickets are the scarce resource, and the average is the measure of how expensive a batter's wicket is to buy.

Two consequences follow.

First, occupation has independent value. A batter who bats through a long session denies the opposition the thing they need most, tires the bowlers, ages the ball, and hands the next batter easier conditions. None of that appears in a strike rate. Some of it appears in the average, because a batter who occupies the crease that long has usually scored heavily while doing it.

Second, tempo in Tests is contextual rather than constant. There are passages where a batter is right to score at a crawl, because the new ball is talking and survival is the whole job, and passages on the fourth evening where the same batter is right to score at twice the rate of a one-day innings, because a declaration is coming. A career Test strike rate averages those opposite instructions into a single figure that describes neither.

Test strike rate is not worthless. It separates a batter who scores briskly enough to give the bowlers time to take twenty wickets from one whose slow accumulation quietly makes a draw more likely. But it is a secondary reading. The primary question in a format with unlimited balls is how many runs the wicket is worth, and that is the average.

Why strike rate rules in T20

Reverse the resource and the hierarchy reverses with it. In a Twenty20 innings the side has 120 balls and ten wickets, and in most innings the balls run out with wickets to spare. Balls are scarce. Wickets, in the aggregate, are not.

This changes the value of a dismissal. In a Test, getting out costs the side one tenth of its innings. In a T20, getting out in the fourteenth over costs the side considerably less than that, because the batter's replacement is also capable of scoring quickly, and because the innings was going to end at ball 120 regardless of who was standing there. The penalty for dismissal is real but much smaller relative to the penalty for consuming deliveries at a low rate.

A batter who makes 45 from 45 balls in a T20 has not helped very much. The same batter making 45 from 25 has left 20 deliveries for somebody else, and those 20 deliveries are worth roughly two overs of scoring at the innings rate. Slow batting in T20 does not merely score fewer runs. It transfers the pressure to the batters who follow, who must score faster than they otherwise would to make up the deficit, and who therefore take more risk and get out more often. The cost of a low strike rate shows up in somebody else's dismissal column.

The one-day format sits between the two. Fifty overs is enough time for an innings to be rebuilt after early wickets, so an average retains real meaning, but the ball limit is binding and strike rate matters throughout. Most serious assessment of ODI batting reads the two together and refuses to rank on either alone.

Why a low-average finisher can be worth more than the number suggests

Take a middle-order batter in a T20 side who averages in the low twenties and strikes at close to 160. Set beside an opener averaging in the high thirties at a strike rate near 125, the finisher looks like the inferior player. Very often he is not, and the reason is that the two numbers are being generated under conditions the statistics do not record.

The finisher bats in the hardest overs. The last four or five overs of an innings are bowled by the best death bowlers in the side, to defensive fields, with yorkers and slower balls at the batter's toes. The scoring is faster because the batter is taking more risk against better bowling, not because the batting is easier.

The finisher's job description guarantees dismissals. A player instructed to attack from the first ball will get out more often than one instructed to assess. Every one of those dismissals is charged to the average at full weight, and the successful version of the same shot is credited only as runs.

The finisher faces few balls per innings. With a small number of deliveries in each appearance, the innings-by-innings scores are naturally low and volatile, which drags the average down independently of quality. A batter averaging 22 from innings that are typically fifteen balls long is performing at a rate the opener would have to strike at 145 to match over the same span.

The runs arrive when they are worth most. Twenty runs in the nineteenth over are not interchangeable with twenty in the fifth. They are scored against a side that has already committed its plans, they are the runs the chasing team must then find under direct pressure, and in a chase they are the difference between winning and not. Neither average nor strike rate is weighted by match situation.

The honest way to compare the two is not to argue about which number to trust but to ask what each player is being asked to do, and then to look at the measure relevant to that job. For a top-order batter, average and balls faced. For a finisher, strike rate and the proportion of innings finished. For an anchor in a chase, both, plus the shape of the required rate while they were in.

What both measures hide

Neither figure knows anything about circumstance, and the gaps are large.

Neither knows who was bowling. Runs scored against the fifth bowler on a flat surface and runs scored against a genuinely fast new-ball attack land in the same column. Nothing in either number distinguishes them.

Neither knows the conditions. A seaming pitch under cloud, a turning fourth-day surface and a flat batting deck in a high-scoring competition all produce runs that are counted identically. Averages are not adjusted for the era or the standard of the ground.

Neither knows the match situation. A hundred scored with the match already won reads exactly like a hundred that saved a match. Nor is there any credit for coming in at 20 for 3, which is precisely when batting is hardest and averages are lowest.

Neither knows about luck. A batter dropped early and going on to a big score, and a batter caught at slip off an equally good ball, are separated by the fielder's hands and by nothing in the arithmetic.

Neither records the shape of the innings. A strike rate of 130 might come from a batter who scored steadily throughout, or from one who crawled for twenty-five balls and then hit four sixes. The second version leaves a much bigger hole for the rest of the order if the acceleration never arrives.

Neither counts the balls consumed by a partner. A batter farming the strike to protect a tail-ender improves their own strike rate by taking more of the deliveries, and the number gives no hint that a teammate was standing at the other end doing nothing.

Modern statistics answer some of these with more granular measures: dot ball percentage, boundary percentage, control percentage, runs by phase of the innings, and average and strike rate broken down by bowling type. Those are the tools that begin to separate a batter from their circumstances, and the more of them a page offers, the less weight the two headline numbers have to carry.

Reading the two together

A short set of habits gets most of the value out of both figures without being fooled by either.

• Check the innings count before believing anything. Both measures are noisy over small samples, and a spectacular figure over a handful of innings is usually noise.
• Check the not-out count against the innings count. A high ratio means the average is being lifted by role rather than by run-scoring.
• Compare like with like. Compare openers with openers, finishers with finishers, and Test figures with Test figures. Cross-format and cross-position comparisons of raw averages are close to meaningless.
• Look at balls faced per innings alongside strike rate. It tells you whether a high strike rate is being generated over a meaningful span or over cameos.
• Ask what the format is short of. Where wickets are scarce, weight the average. Where balls are scarce, weight the strike rate. Where both bind, as in the fifty-over game, insist on both.
• Treat a big gap between the two as information, not as a contradiction. A high average with a low strike rate describes an accumulator. A low average with a high strike rate describes a risk-taker. Which one a side needs depends entirely on the other ten names on the card.

Neither number is a verdict. Both are summaries, and a summary discards exactly the detail that makes an innings what it was. The reader who knows what has been discarded gets far more out of the two columns than the reader who treats them as a scoreline.

If the columns themselves are unfamiliar, the guide to how to read a cricket scorecard takes the card apart field by field. The equivalent question on the other side of the game is covered in bowling economy, average and strike rate, and the terminology used throughout is set out in the cricket statistics glossary. You can also browse player records to see how the two figures diverge across formats and positions.