Why Cricket's Most Expensive Stars Often Become the Season's Least Valuable Evidence

There is one number on the IPL auction table and another number on the field. The gap between the two is what I have watched most closely over the last five seasons.

Over its last three matches, one team's powerplay strike rate fell from 142 to 118. The same team's two most expensive overseas batters scored 29 off 34 balls. In the commentary box, the conversation was still about whether those two stars were 'finding form'. I wrote in my notebook: this is not form, this is role conflict.

I have watched matches for many years, and one thing keeps returning. There is no straight line between price and contribution. Auction price is set by one set of information — two seasons of strike rate, six-hitting rate, age, injury history. On-field contribution is set by an entirely different set of information — which over the ball is being bowled in, who is bowling, what the pitch is doing, whether the team is chasing or defending. These two datasets never speak the same language.

The auction model buys the future; the batting order consumes the present.

I have a structure in my spreadsheet from the last five IPL seasons. For every team I have tracked the auction price of its three most expensive players and, separately, their real match-impact score for that season. I keep the impact score simple: match-deciding innings, death-over bowling economy, runs saved in the field, and contribution in matches decided within ten runs.

The picture is uncomfortably clear. Among the three most expensive players, one routinely has less impact than the average impact of the top five non-expensive players. The reason is not complicated. When a team buys an opener at the highest price, he will open — he will bat at a 142 strike rate, because his contract says so. But if the match situation says three wickets have fallen in eight overs and seven an over are needed, that same batter's 'natural' game loses the team the match.

The auction model cannot capture this. The model assumes a good player is good in all situations. In reality, a good player is good in specific situations, and the team is using him in others.

Consider a team that spends three crore on a finisher. For a team in the playoff race, this finisher's real value appears when he walks in at the 16th over and makes 18 off 22 balls... no, when he makes 18 to 22 runs. But if the team has already reached 100+ before the 16th, the finisher's contribution that night looks small on paper. The next match, if the team loses wickets early, the same finisher makes 24 off 9 — the match is saved. Both are the correct job; one shows as 'low-scoring', the other as a 'highlight reel'.

Sitting in the press box makes it clearer. I have seen many times an editor under deadline pressure say — 'He is worth four crore, he has to play the big shot in the big match.' That sentence is where the error begins. Four crore does not force him to play the big shot; four crore reflects the value of the decision to keep him in the side. The match decision is separate.

The information the auction model does not see is dressing-room chemistry.

My second dataset is qualitative, but I try to bind as much of it to numbers as possible. Who forms a good partnership with whom, which bowler gives which fielder confidence, who keeps pulling the group in practice even when out of the XI — none of this has an auction price. And yet in the second half of a season this is what wins matches.

One example from last season sits in my notebook. A team benched a seven-crore overseas all-rounder in favour of a low-priced domestic bowler. Commentary was asking 'why is the expensive player benched'. I pulled the 14-match data: in matches where that overseas all-rounder was used later in the batting order rather than in his preferred slot, the team's powerplay economy worsened, because his mental state was visibly 'I am not being respected as an all-rounder'. And in matches where the domestic bowler played, his nightwatchman role was appropriate, but whenever he was sent as a finisher, his strike rate was 88.

The difference between the two is not skill. The difference is role definition. If a player is used in three different roles, judging him on a white sheet rather than on role-specific data means no one judges him correctly. The auction model does not capture this role-specific information, because it is buying the future.

One thing needs clarifying about my method. This is not a call to abolish the auction. Teams that pay big at auction are not doing anything unproven. The reason is that if you do not bid at auction, you cannot stay in the market. In the IPL, every team must keep a slot open for an overseas player. This rule forces teams to play more low-priced players, otherwise the powerplay bowling combination does not fit. The model is appropriate; I am only saying its limits should be acknowledged.

I have said how I could be wrong. Now let me say where that error is most likely.

First, my impact score itself is debatable. I have set my own weighting for what counts as match-deciding. If someone argues wickets matter more than batting runs, my ranking changes. That I have accepted more format-dependence than Test cricket is also a weakness.

Second, of the matches I have analysed, in some I was in the commentary box and the rest I watched on screen. Watching on screen, can I see field placements outside the camera? Certainly not. Which means what was being discussed at the non-striker's end, what instructions were going out, is not in my data. Without that information, some of my conclusions may be over-simplified.

Third, separating price from match management is my own bias. A coach may be consciously keeping an expensive player in a lower role to protect him. I call that 'role conflict' when it is conservative strategy. That error is possible on my side.

But one thing I say with confidence, and it is what I saw with my own eyes.

At a small ground in Sri Lanka, in a match where I sat beside the commentator decoding the delivery strategy ball by ball, I saw the team's most expensive batter play four dot balls in the powerplay, because at the non-striker's end the captain had instructed him not to take singles and to wait for the six. A fielder was stationed at square leg, meaning to open that shot the team would have to play two more balls. He was eventually out for 19 off 23. In the commentary box after the match it was said 'he is in bad form'. I would say that night's instruction was the main cause of that innings. The player did not play badly; he followed instructions.

My sense is that over the next two seasons, two kinds of teams will form.

One kind buys stars at big auction prices and builds the game plan around that star. The other keeps regular contributors at lower prices and uses role-based data to make decisions match by match. The second kind will win more matches than the first, because across three matches a week a star cannot absorb the mental and physical load.

If you watch a match, next time start with one question. What role is this player in right now, and was his auction price set for that role. If the answer is 'no', then today's innings deserves more sympathy than the paper shows.

I keep everything in my notebook public, because memory designs convenient patterns. I am writing down one prediction for this season in advance: teams that keep the roles of their three most expensive players fixed every match will have at least a 16 percent higher chance of reaching the playoffs than teams that keep changing roles waiting for an expensive player's 'form to return'. For now I am leaving it written here; at season's end we can check.