The BPL's xG Mirror: What 1,248 Shots Taught the League About Itself

Hook

February 2026, Sher-e-Bangla National Cricket Stadium, Mirpur. In the press box I was holding one scorecard while another ran inside my head. The side that posted 178 in 20 overs was being praised on air as a masterclass in finishing. My shot log disagreed. Of those 178 runs, 39 came off edges and mishits, and shot quality alone produced an expected total of only 142. That is 36 runs manufactured from balls that would have dismissed the same batter on any other day.

The scorecard tells you what happened. It never tells you why, or whether it will happen again. In post-match talk we remember the 178 and forget the 142. That memory bias is the most expensive habit in Bangladesh cricket.

Context

My first xG model was built for football, not cricket. In November 2026, aged twenty-four and working from my flat in Rajshahi, I joined Dhaka-based new-media outlet Golpo Sports as a junior data analyst. I treated data as scripture. I coded 1,248 shots from the 2026-17 Bangladesh Premier League myself. The result was uncomfortable: Abahani Limited Dhaka scored 34 goals from just 27.6 xG, while Sheikh Jamal Dhanmondi scored 29 from 31.2 xG. The team at the top of the table converted fewer chances into more goals; the team at the bottom wasted more.

Back then I wrote about what teams deserved. The xG table became a weekly fixture, the series ran twelve parts, and the outlet's traffic doubled. But the real gain was a realisation: the word deserved is not analysis, it is the absence of analysis. I stopped writing about merit and started writing xG differential.

That raised the next question. Football shot quality can be measured because shot location and angle form a consistent geometry. Can cricket be modelled the same way? What are the conditions before you import a football metric into cricket?

In June 2026, after the BPL series was noticed by StatsBomb, I worked the Russia World Cup as a remote event-data analyst. In Germany versus Mexico, Germany took 26 shots for 1.3 xG; Mexico's 12 shots produced 1.1 xG. Germany's PPDA was 6.9, leaving 18 transition chances open behind them. I published before the final whistle: Germany would not escape Group F. Germany finished bottom. PPDA showed me Germany, because the number told the team's story rather than its name.

Back in Bangladesh I wanted one thing. In Bangladesh, I taught a league to see its own xG. Now I wanted the league to see its own pressure numbers.

Core Analysis

1. Three phases, three languages

A cricket xG model cannot be a universal number. Boundaries and overs cannot be counted in isolation; shot quality must be counted. A cover drive where the batter middles the ball is worth far more than an edge, even though the scorecard records four for both. Coding 1,248 shots from 2026-17, I broke each ball into four variables: length, line, batter balance and timing, then separated phases — powerplay (1-6), middle (7-15), death (16-20). Expected shot value differs so sharply across phases that one merged number simply hides the error.

Venue is an even larger variable. Mirpur pitches keep low and spinners dominate overs 7 to 15. Chattogram offers more pace onto the bat and rewards powerplay stroke-play. Sylhet brings dew, making the second innings easier. Using one xG threshold for all three is not understanding the game. In my log, the same shot is worth six at Mirpur and eight at Chattogram, yet television punditry collapses the two.

In the powerplay the field is restricted and the gaps beyond thirty yards are wide. Any low-risk shot already generates seven or eight an over. Yet we routinely watch the first four overs filled with dot balls to protect wickets. The middle overs are worse: from 7 to 15 the ball is merely rotated and the scoreboard goes quiet. The arithmetic shows the most expensive passage of a match is overs 7 to 15, where nobody looks like losing while the run rate quietly dies.

2. From PPDA to a pressure index: the mapping conditions

Football's PPDA measures how many passes an opponent completes before a defensive action — tackle, interception or foul. A lower figure means more aggressive pressing. Germany's 6.9 meant they pressed high after losing the ball and left space behind.

In cricket I built an equivalent, the Ball per Pressure Event, or BPE. The definition must be written down first, otherwise this is just football cosplay. In my mapping, a pressure event is any delivery where the bowler hits a planned length, forces the batter into defence, and the strike rate on that ball stays under 100. Dot balls, singles and broken plans all count. A lower BPE means an aggressive bowler; a higher BPE means a bowler merely releasing the ball.

Good powerplay units in the BPL sit near a BPE of 4; average units sit at 6 or 7. The best spinners in the 7-15 phase hold BPE below 5, while the league average is close to 8. The spinners are not creating pressure because nobody is asking them to. Where Germany's PPDA exposed a team creating space behind its own press, the BPL shows the reverse: a league that sits back and manufactures the gaps for itself.

3. What the auction buys versus what the field rewards

The most uncomfortable finding is that auction price and xG contribution are barely related. Auction platforms need excitement and familiar faces. When names like Shakib Al Hasan or Tamim Iqbal come to the table, value is set by career memory and popularity, not recent xG contribution. A young seamer holding an economy near nine at the death gets a base price, while an older name conceding fourteen an over gets double. When a league buys reputation, it turns its own error into capital. The biggest opacity in Bangladesh franchise cricket is not the table; it is the auction room.

I am not telling selectors that a model outranks a name. I want the league to have a mirror in which it can see why a pressure bowler in overs 7 to 15 is priced below reputation. A model works as a mirror, not as a verdict.

4. Empty stadiums, cold T20 tables

During the 2026 shutdown I consulted for Brentford FC. Analysing 306 behind-closed-doors matches across the Bundesliga, Championship and Serie A, I found home win rate fell from 43.1 per cent to 33.8 per cent, home xG differential dropped 0.21, and distance covered in the final fifteen minutes fell 5.2 per cent. That produced the CrowdNull adjustment; Brentford altered set-piece routines and used it during the 2026-21 promotion push.

Empty stadiums taught me that home advantage is a variable, not a law. In cricket the lesson applies directly. In T20, home advantage comes largely from two sources: familiarity with conditions and the cushion effect on umpiring decisions. The second matters more in cricket than football, because LBW and catch adjudication need human judgement. When crowds return, nobody in the league records which way that cushion leans.

Contrarian Angle

The biggest enemy of xG is its own fan base. In league coverage, xG has become a weapon: a losing side's xG is cited to argue it really won. That is not using a metric; it is abusing one. xG can explain shot quality but not why an experienced batter refused to rotate strike in the 17th over. Decisions, form and umpiring standards — the three things football xG cannot touch — are equally out of cricket xG's reach.

Correlation is not causation. A bowler with more yorkers at the death has a lower economy, which invites the conclusion that yorkers cause it. The real cause may be a captain setting a better field, or batters failing their own plans that day. Miss that distinction and you buy a bowler who collapses in a different field.

Bangladesh's data reality imposes another condition. We cannot demand European-style tracking data; here, ball-by-ball stroke type, line, length and field mapping are largely hand-coded. Before building anything I had to design collection with scorers, coaches and video analysts — who records what, under which definition, at which venue. Bring in a metric before fixing definitions and that metric becomes no better than a guess.

One more lesson from another domain: rushing players back from ACL injuries destroys second careers. The body heals; the mental block does not. In cricket it shows up as a shortened run-up or a batter avoiding the cut shot. Data misses it, because data measures outcomes, not courage.

Takeaway

For the next round I will watch three things. First, strike rate in overs 7 to 15, where league run rates quietly die. Second, for every spinner, how many balls they hold a strike rate under 100 on pressure events. Third, the gap between reputation and xG contribution in the auction room, and when the league intends to close it.

If the league is willing to look into its own mirror, next season's most expensive cricketer may be the seamer with no star logo and the lowest BPE at the death. He does not chase revelations; he calibrates until they appear. An ESTJ builds the pipeline first and the poetry second. The question now belongs to the league.