Before the next IPL broadcast, a commentator will describe a batsman as exceptional on the basis of his strike rate. The number will be cited as evidence. The audience will nod.
Strike rate is runs scored divided by balls faced, multiplied by one hundred. It is easy to calculate, easy to communicate, and routinely misapplied.
Here is the problem. Strike rate has no memory. It does not know when in the innings the runs were scored, how many wickets were in hand, what the par score at that venue looks like, or who was bowling. It measures output without context. In T20 cricket — a format where context determines almost everything — that is a significant limitation.
The same number, two very different performances
Consider two batsmen, both finishing an IPL match with a strike rate of 135.
Player A batted in the powerplay. He came in at the fall of the first wicket, with nine wickets in hand and the fielding restrictions in place. The conditions were favourable for aggressive batting. The bowling was the opening attack, fresh but constrained by placement rules. He scored 40 off 30.
Player B batted at number five. He came in at over 14, with five wickets in hand, chasing a total of 178. The field was up. The bowling was the death specialists. He scored 27 off 20.
Both produced identical strike rates. Their resource contributions to the team were not equivalent.
Player B's innings was, by any meaningful measure, harder to produce. He scored at the same rate under conditions that demand it — against better bowling, with fewer safety margins, with a specific target to chase. Player A's innings had value, but the context was materially easier.
Strike rate cannot tell this story. Runs per resource can.
What runs per resource measures
The concept of resources in cricket was formalised by Duckworth and Lewis for the purpose of rain-interrupted matches — combining wickets in hand with overs remaining into a single resource percentage. The same framework, adapted for batting analysis, produces a metric that adjusts scoring contribution for the conditions under which it was produced.
A runs per resource figure above the team or league average, measured consistently, indicates a batsman who adds value beyond what his raw numbers show. A batsman who posts high strike rates in low-pressure, high-resource situations may not sustain that output when the game demands it of him.
This matters for team selection, and it matters for prediction.
Why this creates pricing inefficiencies in the market
Bookmaker pricing for T20 matches draws, in part, on public perception. Public perception of T20 batting quality is shaped almost entirely by raw numbers: average, strike rate, recent scores. These are the figures that appear in previews, in broadcaster graphics, in the pre-match analysis that informs casual betting decisions.
Resource-adjusted metrics do not appear in mainstream coverage. They are not communicated in pre-match graphics. The general market participant who is following a batsman on the basis of his 145 career strike rate has almost certainly never seen his resource efficiency broken down by innings type.
This creates a systematic gap between how the market prices a batting line-up's contribution and what the underlying data suggests about likely first-innings totals or chase capability. Our machine learning models work with resource-adjusted metrics as a core input. The divergence from bookmaker pricing, when it exists and is consistent, is where the statistical edge tends to live.
What we actually measure
Our models assess batting performance across five analytical dimensions, each weighted by how predictive it has proven to be over multiple seasons of T20 data. Raw scoring rate is a component of that — it carries information. But it sits alongside bowling quality faced, resource state at entry, venue-specific norms, and seasonal trend data.
No single metric determines a probability. What matters is how multiple measures align. When resource-adjusted batting data, bowling matchup analysis, and venue history all point in the same direction, that alignment is more informative than any one number viewed in isolation.
The result is that a recommendation from EdgeXI may sometimes disagree with public narrative — the team with the bigger names, or the batsman with the higher profile, does not always hold the statistical edge. When the models see a genuine probability gap relative to bookmaker pricing, that is when a recommendation is issued.
When they do not, no recommendation is issued. The discipline of not betting is as important as the analysis itself.
The stat you see on television
There is nothing wrong with strike rate as a starting point. It communicates something real. The problem is treating it as a finishing point — as though the number on the graphic tells you what you need to know about likely match performance.
It tells you what happened. It does not, by itself, tell you what is likely to happen. In a sport where conditions, contexts, and matchups shift match by match, that distinction is where the analytical work lives.
Past performance does not guarantee future results. What it does, when properly contextualised, is inform probability.
EdgeXI's IPL 2026 recommendations are published on Tipstrr before each match and tracked publicly. Follow on Telegram for daily updates during the season.