Cricket DRS Technology: How Ball Tracking Works—and Why Umpire’s Call Stays

How cricket DRS technology reaches a decision

To answer ‘how does ball tracking work in cricket DRS?’, start by separating the review into jobs. Replays and audio can help establish the delivery, contact with bat or pad and other observable events. For an LBW review, approved ball tracking then reports three rule-defined questions: where the ball pitched, where it struck the batter, and whether its projected path would have hit the wicket.

The third umpire does not simply accept a graphic and announce out or not out. The ICC’s October 2026 Test playing conditions specify an order of checks and tell the official how to combine the available evidence with the original on-field decision. DRS is therefore a human-and-technology decision process: evidence is gathered, an allowed tool produces an output, and the playing conditions determine the consequence.

• Observed evidence: video and synchronized sound help establish what happened before the ball was intercepted.
• Modelled evidence: ball tracking estimates the path the ball would have continued to follow.
• Rule application: the official maps landing, impact and wicket classifications to the LBW law and the original call.
• Authority: the third umpire communicates the result, while the on-field umpire makes the final signal.

What ball tracking measures—and what it predicts

The official rules expose the outputs that matter to the decision, not the vendor’s entire engineering stack. They require ball tracking to report LANDING, IMPACT and WICKET. Landing asks where the ball pitched. Impact asks where it first struck the batter. Wicket asks whether the projected path would have reached the stumps.

That last output is different from a camera replay of an event that already occurred. The batter has stopped the ball, so the remaining path is counterfactual: the system projects what would probably have happened without that interception. Camera observations constrain the calculation, but the graphic after impact is still a modelled continuation, not footage of the ball travelling through the pad.

Hawk-Eye says its cricket officiating system uses synchronized video and audio and optical tracking of ball movement and rotation. That is a vendor description of its implementation, not independent proof of a universal error rate, and the ICC playing conditions also permit the VirtualEye ball-tracking system.

Why Umpire’s Call can preserve opposite on-field decisions

The ICC defines Umpire’s Call as a marginal ball-tracking result for either impact or projected contact with the wicket. For the wicket test, the October 2026 Test conditions classify the output as Hitting when the centre of the ball is inside the Wicket Zone, Umpire’s Call when the ball touches the wicket but its centre is not inside that zone, and Missing when the ball does not touch the wicket.

This explains the apparent paradox viewers notice. The same marginal projected path can leave an original out decision as out and an original not-out decision as not out. The model has not produced two different paths. The rule has assigned the boundary case to the original call because the evidence has not crossed the threshold required to reverse it.

In 2021, the ICC Cricket Committee said this design reflected two goals: correcting clear errors while preserving the on-field umpire’s role because ball tracking includes a predictive element. That is the governing body’s policy rationale, not an independent scientific validation of the threshold.

• Hitting: the relevant ball-tracking threshold is met and the original call may be overturned if the other LBW conditions are also satisfied.
• Umpire’s Call: the result is marginal, so the original decision remains and the reviewing team does not lose that review.
• Missing: the projected ball does not contact the defined Wicket Zone.

The evidence ledger: signal, interpretation and decision owner

A useful way to read a DRS sequence is to ask what each item can establish and who is allowed to act on it. The broadcast animation is a communication layer; the official decision depends on the underlying approved outputs and the playing conditions, not on how persuasive the television graphic looks.

• Replay and stump-microphone sound — can support delivery and contact questions; the third umpire interprets the evidence.
• Real-Time Snickometer or UltraEdge — can support whether the ball contacted bat; it is evidence inside the review, not the final LBW rule.
• Ball tracking — reports landing, impact and wicket classifications; it mixes observed position with a projected post-impact path.
• Playing conditions — define zones, thresholds and reversal rules; they convert the classifications into a review outcome.
• On-field and third umpires — operate the review protocol, communicate the decision and handle evidence or technology failure.

Current review limits differ by format

Under the ICC men’s playing conditions effective in October 2026, a team begins a Test innings with three unsuccessful player reviews. The ODI and T20I conditions provide two unsuccessful reviews per innings. A successful review is not counted as lost, and neither is a review that finishes on Umpire’s Call.

‘Current’ needs a date attached. On 1 October 2026, the ICC said its revised playing conditions would apply to international series beginning on or after that date, while series already in progress would continue under the previous conditions. A match played in October can therefore operate under an earlier document if its series began before 1 October.

• Tests: three unsuccessful reviews per innings under the October 2026 men’s Test conditions.
• ODIs: two unsuccessful reviews per innings under the October 2026 men’s ODI conditions.
• T20Is: two unsuccessful reviews per innings under the October 2026 men’s T20I conditions.
• Umpire’s Call: the original decision stands and the review is retained.

What happens when the evidence is inconclusive or the system fails

The protocol includes a fallback. If the available evidence cannot support a high-confidence change, the third umpire reports that it is inconclusive and the original decision stands. If a technology component is unavailable or fails during the review, the third umpire must tell the on-field umpire and make the decision from the evidence that remains available.

The conditions also allow an unreliable technology to be withdrawn. They say the ICC may remove a system if match use or further testing shows it is not reliable, with the match referee, technical official and relevant cricket authorities involved in the decision. A failed or unavailable component can also affect whether the team’s review is reinstated. These rules matter as much as the ideal-case animation because real officiating systems must fail visibly and recover predictably.

Is cricket DRS actually AI?

The precise answer is: not as a single category. The current ICC playing conditions name replay, sound, ball tracking, edge-detection tools, Hot Spot and LED wickets. They do not describe the full DRS chain as an AI system. Some components may use computer vision, statistical estimation or machine-learning methods, while others are better described as synchronized sensing, geometry, signal processing or rule automation.

Calling every DRS component ‘AI’ makes the system harder to evaluate. The useful questions are narrower: what was directly observed, what was inferred, what uncertainty boundary applies, and which official owns the decision? That vocabulary transfers to other sports more reliably than the label ‘AI umpire’.

A practical checklist for watching an LBW review

A viewer can follow the logic without reverse-engineering proprietary software. Read the sequence in order and keep the original call visible, because that call determines what a marginal output can change.

• Was the delivery legal, and is there evidence of bat contact before pad contact?
• Where did the ball land: in line, outside off, or outside leg?
• Where was first impact, and does the impact-zone rule apply?
• Is the wicket result Hitting, Umpire’s Call or Missing?
• Was the original decision out or not out?
• Did any evidence source fail or remain inconclusive?
• Was the review successful, retained because of Umpire’s Call, reinstated after a failure, or lost?

Limitations and source notes

This article explains the public decision protocol; it is not a laboratory validation of Hawk-Eye, VirtualEye or any match installation. The ICC documents define authorized outputs and rules but do not publish the proprietary calibration process, model implementation, venue-by-venue error distribution or every operational incident. Hawk-Eye’s page is used only for the vendor’s description of its own system. It is not treated as independent performance evidence.

A 2012 ICC committee release reported agreement across a 14-sequence independent comparison and cited a historical improvement in correct decisions with DRS. That sample was small, the systems and protocols have changed, and the result should not be read as a current universal accuracy figure; it is therefore context, not a headline claim here. Technology availability also varies by competition and match.

The format-specific review counts in this guide come from the October 2026 ICC men’s Test, ODI and T20I playing conditions. The ICC’s 1 October 2026 transition notice means an ongoing series can still use an earlier edition. Evidence was checked on 16 October 2026. This analysis contains no betting advice or prediction.