One Spell, Three Different Truths: Cricket Data Is Blind Without Format

I was watching a first-class match from a nearly empty gallery at Delhi's Arun Jaitley Stadium last season. The scoreboard showed a right-arm seamer's spell: 11 overs, 3 maidens, 23 runs. Beneath it, in small type, "Economy 2.09". What I wrote in my notebook at that moment was not the economy. It was the zones of line and length, the number of balls outside fourth stump, and how many times the batter got trapped on the front foot.

At the break, a colleague in the next seat was hunting an answer on his phone: "This bowler's economy is 2.09 — would he be worth buying in the IPL?" I told him that number does not answer that question. The reason is simple, and almost every television graphic in cricket avoids it — 2.09 is a Test-format number, and the IPL is an entirely different scale.

That evening I understood that cricket's oldest error is not in the statistics. It is in the empty box beside the statistics. When the format field is blank, we fill it with the number, and then we build the story around it.

Cricket's three principal formats — Test, ODI, T20 — are not separated merely by overs. Inside each, the behaviour of the ball, the fielding regulations, the wear on the pitch and the time-structure of the match differ; the same cricketer is therefore forced into three different jobs.

In Test cricket a bowler's value is set by wickets, and wickets arrive through patience. Holding one length for four or five overs with the new ball, turning the ball behind the bat on a fifth-day surface — the returns come late, sometimes an hour later, sometimes a session later. A pace bowler's average near 30 is elite at Test level; an economy under 3 there means he is controlling the ball and refusing to let the opposition grow.

In T20 the arithmetic inverts. In the six-over powerplay the fielding circle is up; in overs 16 to 20 the batter is forced to take risk; the bowler's core job is the dot ball and the yorker. An economy of 8.0 there is middling, under 7 is good, 6.5 is a match-winning spell. The middle overs of an ODI — overs 7 to 40 — are another thing again: an economy near five or six is valuable there because the licence to chase wickets is limited.

These three numbers are not the same unit. One has time hidden inside it, another has a risk calculation, the third has the constraint of fielding law. Placing them side by side without naming the format means drawing two differently measured things on the same line.

Batting strike rates fall into the same trap. A strike rate of 130 in 2026 is not 130 in 2026 — the baseline of T20 batting has climbed year on year, as boundary sizes, the two-bouncer restriction and power-hitting training methods have changed. A strike rate is a number, but its meaning is an era-dependent position.

This is where the first condition of data analysis arrives: fix the box before you measure — format, phase, role, era. Drop one of those four and the rest of the arithmetic stays mathematically valid while becoming meaningless.

On the first page of my notebook, a rule has been written since 2026: "Give the zone number first, then the opinion." That year, at the Under-17 World Cup final in Delhi, I measured England's pressing triggers by hand, mapping every entry into the half-spaces across a 96-page book. Coming to cricket, I found the same discipline harder, because cricket does not have football's single pitch geometry — three formats mean three separate geometries. I do not chase patterns; I build cages strong enough to test them.

Problem one: the unit error. A Test average describes a bowler's wicket-taking rate; a T20 economy describes his run-saving rate. Neither is ever a proxy for the other. A seamer can take wickets at 27 in Tests and simultaneously concede at 10.5 an over in the T20 death phase — because in the first he uses an outside line to force the batter onto the back foot, and in the second that same line comes back as sixes. Both are true, both tell us something about him, but one cannot be used to predict the other.

Problem two: the phase. T20's powerplay, middle and death are three different games. In the powerplay the new ball swings and the field is in, so wicket probability rises; at the death there is no swing and the field is spread, so economy becomes the only measure. A bowler who holds 7.2 in the powerplay can go at 11 at the death. A single aggregate economy averages the two together and hides which one is his strength and which his weakness. What usually follows: the team uses him in the wrong phase, the results turn, and we say "he has lost form". The variable was phase, not form.

Problem three: the role. Opener, anchor, finisher — these three cannot be measured on one scale. The finisher's job is 25 off 12; the anchor's is 45 off 40 to give the side a base. Put their strike rates side by side and the anchor will always look "slow", even though his role is structurally essential. In Tests the role distinction is finer still: the number five's job differs from the number three's, because he usually faces the second new ball on a broken pitch.

Problem four: the era. An economy of 8.0 in 2026 is not 8.0 in 2026. Average T20 scores have climbed every few years, so economies have risen even where bowling quality has not changed. Comparing bowlers across two decades without era-adjusted numbers means judging old figures on today's scale.

Read those four problems together and one conclusion follows, which for me is the central rule of cricket analysis: a number is meaningful only when format, phase, role and era are written beside it. Drop one and the number becomes a guess.

And here the question of the empty box arrives. Suppose a match report names no format, no player, no information points at all. What is the analyst's job then? The easy answer is to guess and fill the gap. The difficult and correct answer is: "insufficient information, cannot assess."

I know that is not attractive to publish. Readers want numbers, not uncertainty. But I learned the value of the empty box from a different experiment. In 2026, when stadiums emptied, I used Bundesliga behind-closed-doors matches and calculated that the home win rate fell from 43.3 per cent to 33.3 per cent. The number is not what matters; the method is — empty stadiums gave me the control group I never dared to request. The crowd is a variable; the noise is a confound; the silence was data. The same holds in cricket: information that is absent is itself information. Filling the empty box is not analysis, it is invention.

Problem five: the market. A high IPL auction price is not evidence of international success. The price is a market signal generated in one format, for one role, by one team's specific need. If a side is shopping for a death-overs specialist, the price rises for that role; nothing in it says anything about the player's Test capability. Right-to-Match and retention arithmetic work on the same logic — a franchise matches a price on the basis of an existing relationship, not of overall playing quality.

That market signal also blurs the relationship between domestic and international cricket. The seamer with fifty wickets in first-class cricket and the seamer who succeeds in Tests may be the same man, but the two measures sit apart on pitch quality, on the batting depth of the opposition, and on workload. Those differences are measurable, if anyone wants to measure them. Most of the time nobody does, because the story is easier.

Now the part where I disagree with the consensus. The received view is that cricket's data has grown, and therefore its analysis has improved. My reading is the reverse: data has grown and confidence has grown; accuracy has not kept pace, because the discipline of writing "absent" into a missing field has not grown with the data.

The press box taught me that consensus is often just a missing variable. A number takes two seconds to place on a graphic; writing the format beside it takes four — and four seconds is expensive on broadcast. So the number stays and the format goes. The viewer remembers the number and forgets the context. At that exact moment the number begins to live independently.

The second blind spot is technical, not personal. An empty data set is often not analyst negligence but a failure upstream — a paywall, a blocked crawler, a schema mismatch, a misclassification. If that defect is not caught, it passes downstream as "analysis", where every sentence sounds confident while being unsourced. Cricket has seen the most dangerous form of unsourced confidence in its corruption cases — Hansie Cronje in 2026; Salman Butt, Mohammad Amir and Mohammad Asif in 2026. There, an accusation without evidence was dangerous, and so was a denial without evidence. The lesson is single: evidence first, verdict after.

So the next time we watch a match, we can run a small test. Before accepting the number that floats onto the screen, ask three questions: which format, which phase, which era. If all three answers are missing, accept the number but do not believe it.

In Delhi, I learned that a notebook can outlast a broadcast — because the broadcast forgets the context, and the notebook does not. Over the next series, watch which analysis fills the format box first and which one starts with the number. That difference tells you which is analysis and which is only noise.