The 15-degree rule: what makes a bowling action illegal

The 15-degree rule at a glance

The rule comes from the ICC’s Illegal Bowling Regulations, which modify the Laws of Cricket for international matches. The table sets out what it measures and how it is applied; how a report is handled, from match officials to suspension, is covered in how chucking is policed.

What the 15 degrees is actually measured between

The limit is not a measure of how bent your arm is. It measures how much the elbow CHANGES across a specific window: from the moment the bowling arm reaches the horizontal in the upswing, to the moment the ball is released. Extension greater than 15 degrees across that window is an illegal action.

This is why a bowler can deliver with a permanently bent arm and be perfectly legal. If the elbow is flexed 20 degrees when the arm passes horizontal and still flexed 20 degrees at release, the extension is zero. Nothing straightened, so nothing was thrown. Some bowlers cannot fully straighten the arm at all for anatomical reasons, and that is not an offence.

The reverse also holds, and it is what catches people out. An action that looks smooth can still extend more than 15 degrees inside a window lasting a fraction of a second, far too fast for anyone standing at the other end to judge reliably.

Why the line is drawn at 15 degrees

Almost nobody bowls with a perfectly locked arm. The ICC’s regulations say scientific research found some straightening at the elbow in almost all bowling actions, and that human-movement specialists found it became noticeable to the naked eye, and started to look like a throw, once it exceeded 15 degrees. That is the stated basis for the limit. How it replaced the earlier tiered tolerances is covered in our history of chucking.

Match data bears out the first half. In footage of 21 elite fast bowlers in Tests, tour matches and one-day internationals, 31 of 34 deliveries straightened at the elbow, by 3 to 22 degrees with a mean of 9. Deliveries that straightened by more than 15 degrees were also faster on average, 39.5 m/s against 37.1 m/s, roughly 142 km/h against 134 km/h, though that compares deliveries rather than testing cause and effect (Portus, Rosemond and Rath, 2006).

For finger spinners, the authors of a study of 48 pathway and elite bowlers linked the gain to speed and revolutions: pathway bowlers with illegal actions were similar to elite legal bowlers, with no significant difference, for ball speed and revolutions at release (Spratford and colleagues, 2018). Correcting an action can cost some of that: in a 2026 case report on two off-spinners, a ball fitted with gyroscopes recorded spin rate, torque and angular acceleration about 10% lower after their suspect actions were corrected, while its skill measures improved (Ferdinands, Doljin and Fuss, 2026).

The angle is not the only possible test. A laboratory study of 83 bowlers found they generally stayed within about 15.5 degrees of extension, but the limit did not separate bowlers previously suspected of throwing from the rest. How fast the elbow was straightening at release separated them better, and the authors recommended that researchers and legislators examine whether a limit on that is feasible (Ferdinands and Kersting, 2007). The ICC’s current regulations still use the 15-degree angle.

Hyperextension and the wrist: what the 15 degrees ignores

The ICC’s regulations define elbow hyperextension as the elbow extending beyond straight, and say any hyperextension is discounted when deciding whether an action is illegal. Before an official test the bowler is also assessed for joint hypermobility on the Beighton scale. Separately, the Law itself (Law 21.2 in the current Laws) allows a bowler to flex or rotate the wrist in the delivery swing.

Some fast bowlers hyperextend. In the match study above, five deliveries from three of the 21 bowlers hyperextended, by about 19 degrees on average (Portus, Rosemond and Rath, 2006). It may help: a Loughborough computer simulation estimated that a fast bowler with 20 degrees of hyperextension and an optimal recoil releases the ball around 5% faster than a similar bowler who cannot hyperextend (Felton and King, 2016). The rule does not count it.

It is also one of several reasons a legal action can look illegal. Helen Bayne, a member of the ICC’s panel of human-movement specialists, lists hypermobility, extra side-to-side movement at the elbow, a large carrying angle, a fixed bend in the elbow and unusual wrist alignment as things that can make a legal action look suspect (Bayne, 2021). Measuring it takes care. For Jenny Gunn, who has hypermobility in both elbows and whose action had been repeatedly queried by umpires, Loughborough researchers developed a six-marker method that calculates elbow extension and side-to-side movement separately; at release its extension angles agreed with high-speed video with a mean difference of 0 ± 2 degrees (King and Yeadon, 2012).

Why the naked eye is the wrong instrument

The window from horizontal to release lasts roughly a tenth of a second for a fast bowler. Within it the arm rotates through most of a semicircle while the whole body is also rotating and moving forward. Asking anyone to estimate a 15-degree angular change inside that, from side-on, in real time, is asking for a guess.

One illusion behind false accusations is hyperextension. An elbow that bends slightly backwards past straight gives the arm a visibly kinked silhouette at release that reads as a throw, even when the measured extension across the window is well inside the limit. A low arm slot and shoulder rotation produce the same effect.

This is precisely why the ICC moved to instrumented testing rather than umpire judgement. The rule is a measurement, so it needs measuring.

What an illegal action typically looks like

Bayne describes the usual pattern. The elbow is already bent as the upper arm passes horizontal. It stays bent, or bends further, through roughly the first half of the window, then straightens into release. Seen from in front, the inside of the elbow often leads once the upper arm is about 45 degrees above horizontal, and the straightening comes after that (Bayne, 2021).

She treats the pattern, not its size, as the warning sign: if a coach sees it, the technique is worth working on whatever the measured range. In Spratford’s study, finger spinners with illegal actions were more front-on at back-foot contact, rotated the pelvis through a smaller range and the trunk more slowly, which left the arm to supply the revolutions (Spratford and colleagues, 2018, as summarised by Bayne, 2021).

What your own video can and cannot show

Elbow angle through the delivery can be estimated from ordinary footage if the shoulder, elbow and wrist are visible across the window. Frame rate comes first: at 30fps the entire window is three or four frames, so the peak is easily missed. Treat 60fps as the minimum. Bayne recommends more than 100fps, which many phones’ slow-motion modes exceed (Bayne, 2021).

Camera angle matters next. Put the main camera side-on to the bowling arm, level with the release point. A view from behind or in front is a useful second angle (Bayne recommends filming from in front, behind, side-on and about 45 degrees in front), but do not rely on it alone. Even side-on, a single camera gives an estimate, not a measurement to the degree: the bending plane turns during the swing, so no fixed camera stays square to it throughout (Bayne, 2021). Our step-by-step guide to checking a bowling action covers filming, reading the angle frame by frame and the size of those errors.

What video cannot give you is a ruling. Only the ICC's own process can rule on an action; in Tests, ODIs and T20Is that means an assessment at an ICC-accredited testing centre under controlled conditions with calibrated multi-camera systems. What your own footage tells you is whether your action sits comfortably inside the limit, sits near it, or is worth having formally tested before a match official raises it.

What to do if your action has been questioned

Film it properly first: side-on to the bowling arm, 60fps or better, arm fully in frame from horizontal through release, and several deliveries rather than one, because a single ball says nothing about whether the pattern is consistent.

Look at the change across the window, not the shape at release. A bent arm at release is not the question. The question is whether it was less bent at release than when it passed horizontal, and by how much.

Check every delivery you bowl, not just the stock ball. In an official test every delivery must stay within 15 degrees; it is not an average. A single stock delivery over the limit makes the action illegal, a variation that goes over is dealt with separately, and bouncers and yorkers count as the stock ball, not as variations. Our guide to checking a bowling action sets out how many of each to film, and whether fatigue changes the action. A session shows the pattern; one ball shows an anecdote.

If the extension sits anywhere near the limit, seek formal testing rather than self-diagnosing. The point of measuring your own footage is to know whether that conversation is necessary, not to replace it.