All-rounder value in cricket is a squad-balance question

The argument turns up every time a side names an XI with an extra seamer in it. Somebody says the player at seven is not a real all-rounder. Somebody else says he averages about thirty with the bat and about thirty with the ball, so what more do you want. The conversation stops there, because the two people have never agreed on what the question is.

All-rounder value in cricket gets discussed almost everywhere as a property of the player. It is not. It is a property of the eleven he is standing in. The same cricketer, with the same two averages, is worth a great deal to one side and almost nothing to another, and that is not a paradox or a scouting cliché. It falls straight out of how a team of eleven has to be assembled under rules that cap how many overs any one person may bowl.

This is the case for the difference figure, the case against it, and the argument that actually decides selection.

The difference figure, and what the subtraction is claiming

Take a player's batting average, which is runs scored divided by times dismissed. Take his bowling average, which is runs conceded divided by wickets taken. Subtract the second from the first. A positive result is the whole test.

The subtraction is not arbitrary and it is not a fudge. It rests on a symmetry that genuinely exists in the sport. Every innings contains ten wickets and no more. When the player is dismissed he has spent one of his own side's ten, and the runs he made are what the side received in exchange. When he takes a wicket he has spent one of the opposition's ten, and the runs he conceded are what the opposition received in exchange. Both halves of the comparison are priced in the same two commodities: runs, and one tenth of an innings.

So a difference of, say, plus ten says something clean. Across a career, the wickets this player takes have been ten runs cheaper than the wickets he loses. Every time he completes one exchange in each direction, the match balance moves ten runs his way.

That is a better-founded number than most single-figure cricket statistics, which is exactly why it is worth understanding why you should not stop there.

• 11Players in a side
• 10Wickets available in any innings
• 10Overs one bowler may send down in a 50-over innings
• 4Overs one bowler may send down in a T20 innings

Rule-defined limits, not measured values. The per-bowler cap in limited-overs cricket is one fifth of the innings, which is what makes five the minimum viable number of bowlers.

The subtraction gives you a rate, and a selector needs a total

Here is the flaw that matters most, and it is not the one usually raised.

Batting average and bowling average are both rates. They are runs per event. The difference between them is therefore also a rate, runs per matched pair of events, one wicket lost against one wicket taken. It tells you the exchange rate. It says nothing at all about how many exchanges occur.

A batter is dismissed roughly once per innings. That is close to a constant, and the format sets it rather than the player. A bowler's wicket count is nowhere near constant. It scales with the overs he is given, which is a decision the captain makes on the morning of the match, and which varies from twenty-five overs a game to none.

So the two halves of the subtraction are sampled at wildly different frequencies, and the imbalance runs in only one direction. A player's batting contribution accumulates whether anybody wants it to or not. His bowling contribution accumulates only to the extent that somebody uses him.

Convert the rate into a total and the picture changes completely. Runs contributed with the bat in a typical match is the batting average multiplied by dismissals per match. Runs prevented with the ball is not the bowling average at all. It is wickets per match multiplied by the gap between his bowling average and the bowling average of whoever would otherwise have sent down those overs. A bowler who takes wickets at 28 is worth nothing in a side whose sixth-choice bowler also takes them at 28. He is worth a fortune in a side whose next option goes at 45.

That second sentence is the whole article in one line, and it is why the honest version of this question is a selection question rather than a player question.

Volume is most of the argument, and the difference figure discards it

Three cricketers can post an identical difference figure and be three entirely different assets.

Consider three invented players, each averaging exactly ten runs more with the bat than with the ball. The first bowls the odd over of off spin when the ball is going straight and the captain wants a change. The second is a genuine third seamer who takes the second new ball. The third opens the bowling. The difference figure ranks them equal. No selector on earth ranks them equal.

• Difference figure in runs
• Overs bowled in a typical innings

Occasional bowler104

Third seamer1014

Front-line bowler1022

Constructed illustration. All three players are given a batting average exactly ten runs higher than their bowling average, so the difference figure is +10 in every case. The second bar is the workload that difference is earned across. No real player is being described.

Show the numbers

Invented example: three players, one difference figure

Item	Difference figure in runs	Overs bowled in a typical innings
Occasional bowler	10	4
Third seamer	10	14
Front-line bowler	10	22

The gap between the second bars is the value. A difference of plus ten earned across four overs an innings moves a match by a rounding error. The same difference earned across twenty-two overs belongs to a front-line bowler who bats, and the side has effectively found a spare place in its XI.

This is why "he averages thirty and thirty" is an incomplete sentence. Thirty and thirty across two overs a game describes a batter with a party trick. Thirty and thirty across eighteen overs describes one of the most valuable cricketers in the country. The two averages are identical. Volume is doing all the work, and the standard figure throws it away before you get a chance to look at it.

A rough correction is easy enough to make in your head. Multiply the difference figure by the wickets the player takes per match. That converts runs per exchange into runs of match margin per game, and it demotes the part-timer instantly.

All-rounder value in cricket is really an argument about the eleventh place

A side has eleven places. One goes to a wicketkeeper. In limited-overs cricket the bowling has to be covered, which sets a floor on how many of the remaining ten must go to people who bowl. Everything left over is batting.

That is the entire budget, and an all-rounder's contribution to it is not measured in runs or wickets. It is measured in places.

Picture two sides with identical playing resources. Side A fields five specialist bowlers, five specialist batters and a keeper. Side B fields four specialist bowlers, six specialist batters and a keeper, and one of those six batters bowls a full quota at a standard the captain trusts. Both sides can get through fifty overs. Side B has one more genuine batter in it than Side A, and it got that batter for nothing.

The extra player is the value. Not the all-rounder's own runs, not his own wickets, but the marginal cricketer who is now in the XI because a place was released. Which means the size of the value depends entirely on who that marginal cricketer is.

If your seventh-best batter is nearly as good as your sixth, releasing a place gains you almost nothing. If there is a cliff between them, releasing a place is worth a great deal. The identical all-rounder is worth double to a side with a strong sixth batter waiting outside the team and a tail that starts too early inside it, and worth very little to a side already carrying batting depth it cannot fit in. This is not a subtlety. It is the main effect, and it is invisible in every statistic computed from the player alone.

9%45%45%

• Wicketkeeper1
• Bowling places the over cap makes compulsory5
• Places left over for batting5

The structure a selector works inside, not a recommendation. The bowling floor is set by the one-fifth-of-the-innings cap, which requires five bowlers to cover fifty overs. Every place not spent on the keeper or the bowling floor is available for batting.

Show the numbers

Where the eleven places go in a fifty-over side

Item	Value
Wicketkeeper	1
Bowling places the over cap makes compulsory	5
Places left over for batting	5

The chart above shows why the argument is so persistent. Five batting places plus a keeper who bats is a thin batting side by modern one-day standards. The only way to get a sixth batter into that XI without dropping below five bowlers is for one body to occupy two of the boxes. That is the job description, and it explains why sides pay a premium for it rather than for two thirds of a good batter.

The wicketkeeper is the same trade wearing different gloves

The clearest way to see that this is a places argument rather than a player argument is to notice that cricket already runs the identical trade in another position, and nobody calls it controversial.

A wicketkeeper occupies a compulsory place. Somebody has to stand behind the stumps, and no amount of batting depth removes that requirement. So a keeper who bats at a specialist standard is doing exactly what an all-rounder does: collapsing two requirements into one body and releasing a place for somebody else. The debate about whether to pick the better gloveman or the better batter is the all-rounder debate with the disciplines swapped, and it turns on precisely the same question, which is how much the released place is worth against how much keeping quality is being given up.

The parallel is useful because keeping quality is far harder to measure than bowling quality, which forces the argument onto honest ground. Nobody tries to settle it with a subtraction. Selectors talk instead about how many chances go down, what the standing-up work is like against spin, and who the alternative is. The craft of keeping and what it actually costs to compromise it sits underneath every one of those conversations, and the discipline of that argument is the discipline the all-rounder argument usually lacks.

The over caps are why the fifth bowler is not optional

In a fifty-over innings no bowler may send down more than one fifth of the overs. Five bowlers at ten overs each covers exactly fifty. In a Twenty20 innings the same one-fifth rule gives four overs each, and five bowlers cover exactly twenty.

Both numbers are exact, and exactness is the problem. A five-bowler attack carries no slack whatsoever. If one bowler is taken apart in his first spell, the captain cannot simply stop using him, because the overs have to come from somewhere and everybody else is already at the limit. The choice narrows to bowling a man who is being hit, or handing overs to somebody who is not a bowler at all.

A sixth option changes the shape of an innings rather than just its arithmetic. It lets a captain break a partnership with a match-up, hide four overs of a bowler having a bad day, hold a strike bowler back for the death, and answer a left-hand and right-hand pair without burning a specialist's quota. Anyone who has watched a captain reset a field while quietly counting overs on his fingers knows what running out of legal options looks like.

Test cricket has no per-bowler cap, so the constraint stops being legal and becomes physical. A four-man attack asked to bowl ninety overs a day, twice, in heat, will break, and the breakage tends to arrive on the fourth afternoon when the match is being decided. The all-rounder in a Test side is a workload valve first and a batting bonus second, and he changes what a captain can do with a declaration and with the second new ball, because both of those decisions are really questions about how much bowling is left in the shed.

A genuine all-rounder changes the team sheet. A bowler who bats does not

Here is the test that actually separates the categories, and it is not a test of quality.

Would removing the player's second discipline change who else is in the side?

If the answer is yes, the second discipline has selection value. If the answer is no, it has match value and no selection value, and the two get confused constantly.

A front-line seamer who bats usefully at eight is picked for his bowling. Delete every run he has ever scored and he is still picked, and the other ten names are unchanged. His runs are worth having. They are not worth a place, because no place was ever at stake. He is a bowler who bats.

A batter who sends down a few overs of gentle medium pace is picked for his batting. Delete the bowling and he is still picked, though his captain now has one fewer escape hatch. Same structure, opposite discipline.

The genuine all-rounder is the case where both answers come back yes, which means both disciplines clear the bar that the next available specialist sets. That is rare, and the rarity is why the label gets argued about so fiercely.

There is a fourth category, and it is the one that ruins sides. A player whose batting would not displace the next batter and whose bowling would not displace the next bowler is not filling two holes. He is standing between two holes. Fifty-over cricket spent a long period building sides out of these cricketers on the theory that flexibility beat quality, and the theory kept losing to teams that simply picked the best five bowlers and the best five batters available to them.

1. 1Start from the bowling floorDecide how many overs the format forces you to cover, and therefore how many bowlers you must find. Fifty overs at a ten-over cap means five. Twenty overs at a four-over cap means five. Test cricket sets no legal floor, so set a physical one instead.
2. 2Pick the bowling on bowling aloneChoose the best attack available for the surface, ignoring how anybody bats. A bowler picked partly for his batting is a bowler you have downgraded on purpose, and the downgrade is paid on every ball of the innings.
3. 3Pick the batting on batting aloneFill the remaining places with the best batters available, again ignoring the second discipline. You now have a provisional XI and, usually, a problem: too few bowling options, or a tail that starts too early.
4. 4Look for the overlap playerAsk whether any cricketer in the squad clears both bars, meaning his bowling would make the attack on merit and his batting would make the top seven on merit. If somebody does, he collapses two of your places into one.
5. 5Name the player he releasesThe value is not the all-rounder's own record. It is the extra specialist who now fits. Write that player's name down, because if you cannot, the all-rounder has bought you nothing.
6. 6Price the dropCompare the bowling you lose by using the all-rounder's overs instead of a specialist's against the batting you lose by using his place instead of a specialist's. If the released player is worth more than both losses combined, pick him.
7. 7Check availability, not just abilityConfirm the player can bowl a full quota and bat in his position on the same day, in the same conditions, for the length of the series. A part-fit all-rounder weakens two positions at once.

The sequence a selection meeting runs through, set out step by step. Player names and standards are deliberately left out because the answer changes with every squad. The order of the questions does not.

Why the ICC multiplies the two ratings instead of subtracting them

The ICC does not publish an all-rounder rating the way it publishes batting and bowling ratings. It publishes an index, and the construction repays a careful read because it makes a different judgement from the difference figure.

The index is the batting rating multiplied by the bowling rating, divided by 1,000. Both inputs sit on a scale that tops out at 1,000, so the index lands on the same nominal scale.

Multiplication and subtraction behave very differently, and the choice is not cosmetic. Subtraction is compensatory: a large enough score in one discipline offsets a poor score in the other, because the two are simply netted off. Multiplication is not compensatory in the slightest. A near-zero factor drags the product towards zero however large the other factor is.

Work it through with four pairs that all add up to the same total.

900 and 10090

700 and 300210

600 and 400240

500 and 500250

Pure arithmetic applied to the ICC's published construction, which multiplies the batting rating by the bowling rating and divides by 1,000. Each invented pair below sums to 1,000. The pairs exist only to show the shape of the function.

Show the numbers

Equal totals, very different indexes

Item	Index score
900 and 100	90
700 and 300	210
600 and 400	240
500 and 500	250

Every pair in that chart sums to 1,000, and the index ranges from 90 to 250. The function rewards balance and punishes lopsidedness, which is exactly the property you want from a measure trying to establish whether both skills are real. A player who is a world-class batter and a negligible bowler is a world-class batter. The index refuses to call him an all-rounder, and it is right to refuse.

The construction has a second effect worth noticing. Because both inputs are ratings rather than averages, they already carry adjustments for opposition strength, match result and recency that the raw difference figure has no way of applying. The trade-off is that ratings are opaque where averages are transparent. Anybody can check an average. Almost nobody can reproduce a rating, which is why the full mechanics of the ratings tables are worth reading before you use one to win an argument.

The same player is worth different amounts in each format

The over caps make this concrete rather than vague.

In Twenty20 the sixth bowling option is close to the most valuable thing in a squad, because four-over quotas make match-ups the primary tactical weapon and a captain with only five bowlers cannot use any of them. The all-rounder who can send down two overs in the powerplay and two at the death, or who can be withdrawn entirely on a day when his match-up is wrong, is selling his captain optionality rather than overs. That value is enormous and completely absent from any difference figure, and it is closely related to the reason bowling at the death is treated as its own specialism.

In fifty-over cricket the value is arithmetic. Ten-over quotas, fifty overs, five bowlers exactly. A sixth option releases a batting place, and a batting place in a format where sides regularly bat a long way down is worth a lot. The powerplay structure across the formats matters here too, because a part-time bowler is far more exposed inside the first ten overs than in the middle, so a captain with a genuine sixth option can spend his weakest bowler's overs where they hurt least.

In Test cricket the value is endurance. No caps, no quotas, and a fifth bowler is not a legal requirement but a physical one across five days and, more to the point, across a series of them. An all-rounder who bowls fifteen overs a day for four days is not releasing a batting place so much as extending the working life of the four bowlers around him, which is why he tends to be valued most by teams touring in heat with short gaps between matches.

That variation is also why cross-format comparisons of all-rounders are close to meaningless. The same two averages are answering three different questions, and the differences between the formats run deeper than the number of overs, as the piece on what actually separates Tests, one-day cricket and Twenty20 sets out.

What an all-rounder costs: one body, two ways to break

Every argument for the all-rounder is an argument for concentrating two jobs in one person, and concentration carries a cost that team sheets rarely price.

If a specialist bowler is injured, the side loses a bowler. If an all-rounder is injured, the side loses a bowler and a batter at once, and the replacement almost never covers both. A squad built around one cricketer doing two jobs has a single point of failure sitting at number seven, and the failure mode is not subtle. It shows up as a side that looks balanced on paper for two years and then cannot field a functioning XI for eighteen months.

There is a quieter version of the same cost that operates while the player is perfectly fit. Bowling ten overs and then batting in the top six on the same day, repeatedly, across a long season, is a workload no specialist carries. Bodies respond by degrading one discipline to protect the other, usually without anybody deciding to. A bowling all-rounder whose batting quietly declines, or a batting all-rounder who loses a few miles an hour, has not lost form. He has been paying the concentration cost, which is the same reason managing a fast bowler's workload has become a selection discipline in its own right.

The practical consequence is that all-rounder value has to be discounted for availability, and the discount is larger than most people apply. A cricketer available for two thirds of a season in both disciplines is not two thirds of an all-rounder. For the third of the season he is missing, the side is short in two places at once and has to break its own balance to cover him.

Batting at seven is not the same job as batting at four

An all-rounder's batting average is almost always earned in a different environment from a top-order batter's, and comparing the two raw figures is a category error that even good analysis commits.

Three things differ. The bowling is usually older, which suits some methods and ruins others. The field is frequently spread, because a lower-middle-order batter tends to arrive either when the innings is being rebuilt or when it is being accelerated. And the not-out rate is much higher, because a player batting at seven is far more likely to be stranded by the innings ending than a number four is.

That last one distorts the figure more than people expect. A not out adds runs to the top of the batting average fraction and nothing to the bottom, so a player who is regularly stranded ends up with an average that describes the cost of his wicket rather than the size of his typical innings. The full mechanics of that distortion, and why it is a feature of the definition rather than a bug in it, are set out in the piece on what a batting average and a strike rate each divide by.

The correction is not to adjust the number. It is to change the comparison. An all-rounder's batting should be judged against other players batting in the same position, in the same format, in the same conditions, and above all against the specific alternative your own squad would otherwise put there. A player averaging in the low thirties at seven may be the best number seven available anywhere. The same average at four would be a problem waiting to be solved.

The same logic governs his bowling. An all-rounder's overs arrive at particular times, usually in the middle phase, usually when the ball is not new and the field is out. Judging those overs against a new-ball specialist's is the bowling version of the same mistake, and the three bowling figures and how they interact are the right tools for unpicking it.

What the market pays for all-rounder value in cricket, and how a rule change moved it

Player auctions are the only place in the sport where somebody is forced to put a number on this, and the numbers have been consistent in shape even where the amounts vary wildly. Sides pay a premium for cricketers who fill two boxes, and the premium is largest in the formats with the hardest bowling caps.

The reason is exactly the place-releasing argument, and an auction makes it unusually explicit. A franchise buying a squad is buying a set of eleven slots plus overseas restrictions plus a bowling floor, and every constraint it can satisfy with one purchase rather than two leaves budget for a better player somewhere else. That is why the structure of an IPL auction tends to produce bidding wars over the two-skill cricketers long before the specialists are settled.

Which makes what happened to that premium a useful natural experiment. When a competition lets a side substitute one player for another after the toss, so that a specialist bowler can be swapped in for a specialist batter partway through a game, the arithmetic that created the premium is partly dismantled. A side no longer needs one body doing two jobs, because it can field twelve cricketers across the match and use each of them in a single discipline. The substitution rule that brought this into franchise cricket attracted a great deal of comment about entertainment and very little about what it does to squad building, which is where its real effect lands.

The lesson generalises. All-rounder value is created by a constraint. Loosen or remove the constraint and the value falls without a single thing changing about the players. Any competition that expands squads, permits substitutions or relaxes bowling caps is quietly reducing the price of two-skill cricketers. Any competition that tightens them is raising it. The same effect shows up across the franchise circuit, where the leagues with the most restrictive overseas and substitution rules consistently value the same players differently.

How to judge an all-rounder in about four minutes

None of this needs a database. It needs five questions asked in the right order, and most published arguments about all-rounders never get past the first one.

• Check the overs, not just the averages. Overs bowled per match tells you whether the second discipline is a plan or a decoration. Below roughly five overs an innings in limited-overs cricket, treat the player as a specialist who occasionally bowls.
• Read the batting position and the not-out rate together. An average earned at seven with a high proportion of not-outs describes something different from the same number earned at four, and the difference is not small.
• Name the player who gets in because of him. If you cannot say who takes the released place, the all-rounder is buying your side nothing, whatever the two averages say.
• Ask whether either discipline would be picked alone. Not both. Either. A player whose bowling would not make the attack and whose batting would not make the top seven is filling two gaps at the standard of neither.
• Discount for availability across a whole season. Two disciplines mean two injury exposures and a workload no specialist carries, and a side short at seven is short in two places simultaneously.
• Use the difference figure last. It is a screening test with genuine logic behind it, it is unstable over small samples, and it misleads most reliably at the exact moment it looks most impressive.

The thing to carry away is that this is a selection argument wearing the costume of a statistical one. The difference figure asks a real question and answers part of it. The eleven places question asks the same thing properly, and it cannot be answered without knowing who else is in the squad. That is why two well-informed people can look at one player and disagree completely, and why both of them can be right about different teams. More on how sides are assembled, and on the rest of the game's mechanics, sits in the cricket archive and across the wider collection of explainers.