Not the Powerplay: The Overs 7–15 Boundary Rate Decides Who Touches the Trophy

Over the last three matches, that team's powerplay run rate has fallen from 9.4 to 7.1, and yet its position in the points table has not moved a single place. I was not looking at the scorecard. I had the ball-by-ball sheet open. Overs 7 to 15: fifty-four deliveries, four boundaries, a strike rotation rate of 68 per cent. Runs arrived; risk did not. The table says the team is in form. The data says the team has hit its own ceiling. That gap is the great deception of the regular season.

I opened the Expected Notes, and the match began to confess. The pre-match projection gave that side 178. The innings finished at 214. The error was not in the powerplay, where the model lost only six runs. The error sat in overs 7 to 15, where the boundary rate came out 41 per cent above expectation. The model was not wrong. The model had simply answered an old question perfectly.

During a regular season we carry a simple arithmetic in our heads: points, net run rate, winning momentum. A league season runs 70 to 74 matches, and at least 30 of them finish with a margin above 20 runs. Those 30 matches inflate net run rate and hoist teams up the table on the strength of destroying weak opponents, with no evidence gathered against opponents of their own weight. The table then stops being a record of merit and becomes a misreading of structure.

Across the last two IPL seasons I have built a four-metric baseline: powerplay run rate, overs 7–15 run rate, death-over economy, and win ratio in matches decided inside 10 runs. Line them up and the correlation is plain. Powerplay rate has a weak relationship with titles; middle-over run rate and death-over economy have a much stronger one. The logic is straightforward. In the powerplay the fielding side is forced to keep men inside the circle, two fielders are stranded, and the ball is new. Boundaries are comparatively cheap there. From over 7 to 15 the field spreads, the boundary lengthens, spinners bowl through the middle. That is where the real negotiation between run rate and wickets happens.

My method is simple and unforgiving. I do not use Expected Notes as a prediction. I use it as a falsifiable hypothesis. Before the match the model gives a score and draws an over-by-over expectation curve. After the match I look at where the curve broke, and in which phase it broke. That habit took shape in 2026 at the Mumbai City data desk: the scoreline does not tell the truth, the phase pattern inside the scoreline does. In cricket this reads even more clearly, because an innings divides into three unmistakable phases.

Sunrisers Hyderabad's 2026 season is the cleanest proof. On 15 April 2026, at the M. Chinnaswamy Stadium in Bengaluru, they scored 287 for 3 against Royal Challengers Bengaluru, then the highest team total in IPL history. The powerplay was a hockey stick: Travis Head and Abhishek Sharma kept clearing the ropes inside the first six overs. Yet in the final, on 26 May 2026 at Chepauk, they lost to Kolkata Knight Riders by 8 wickets, bowled out for 113 after batting first. A side that can make 287 in 20 overs can fold for 113 in 17.2, because on that evening the ball turned, the pace came off, and their data sheet held no plan for manufacturing boundaries between overs 7 and 15.

Kolkata won that trophy in a different currency. Sunil Narine and Varun Chakravarthy together pushed the opponent's boundary rate down through the middle overs; Varun took 21 wickets in the 2026 season, most of them in that phase. What the scoreboard calls a turning point, the data calls a strike-rate differential.

The death-over arithmetic is harsher still. In the 2026 season Jasprit Bumrah took 20 wickets at an economy of 6.48, in a year when the overall scoring rate touched the sky. That is where the real question sits. Does a franchise want an opener striking at 200, or a death bowler going at 6.48? The market pays for the first. The model weights the second. A four-over death spell can directly save 30 to 35 runs, and those 30 runs usually decide the result.

In the 2026 season Royal Challengers Bengaluru won their first title on 3 June in Ahmedabad, beating Punjab Kings by 6 runs. The match rolled into the final over, and that is not accidental. Sides that win matches decided inside 10 runs can make decisions under pressure, and that capacity replicates in the play-offs. Sai Sudharsan's 759 runs won the Orange Cap that season, yet his value should be priced not as top-order speed but as the ability to find boundaries between overs 7 and 15. That middle block is now the most expensive land in the league.

This is where market inefficiency shows itself. Auction money flows towards openers and finishers, because their work is easy to catch in a metric: strike rate, sixes, highlight reels. But the cricketer who delivers four dot balls and a maiden-like over between overs 7 and 15 quietly moves the table and stays cheap. From my forty-two years of watching the game, I can say the crowd gets bored in those overs, and the selectors lose the trophy in exactly those overs.

Which is why I have to distrust my own model. Correlation is not causation. Is a strong overs 7–15 run rate the cause of a title, or the product of a solid top-order platform? If openers leave 50 for 1, the middle order buys the freedom to take risk, which means the independent variable sits above, not below. A 14-match league is also a small sample. Home advantage, travel and pitch classification shift the arithmetic. It is worth remembering that in the 2026 empty-stadium season home wins fell from 46 to 38 per cent. Change the conditions and the numbers change, and we habitually leave the conditions out of the count.

Matchup data carries a sharper danger. If a batter's good record against spin rests on eight balls and two sixes, that is not information, it is noise. I read matchups only when the delivery count crosses 60. Otherwise the number drags you towards a story. Expected Notes is not an oracle. It is a test.

Over the next three weeks I will watch two things. One, a side's economy in overs 16 to 20 on a flat deck: below 9 and they go deep, touching 11 and the table position will not save them. Two, the win-loss ratio in matches decided inside 10 runs, because a side sitting at 5-1 in those games is not described at all by the number printed beside its name. The numbers were never the story; they were the trail. The rest is annual baseline work, and that work refuses to wait until the last day of the season.