Where on my own video should I look for it?

Not at impact. Stop the clip from down the line with your lead arm about horizontal on the way down, and read the angle between that forearm and the shaft. Compare it with the same frame from your own earlier swings, because camera height and distance move that angle on their own.

Should I stop trying to hold the angle, then?

Stop treating it as something to grip. Nesbit traced the delay in his scratch golfer to hand path and the wrist angle at the top, with an uncocking torque applied throughout. In Chen's model, holding was worth 1.4 percent of clubhead speed against 5.6 percent for the same torque spent uncocking.

Why does it get worse when I swing harder?

TPI puts limited lower-body contribution near the top of its list of causes: when the lower body does not lead the downswing, the upper body overworks and the club gets thrown from the top. Effort applied at the hands is the version of speed you feel first.

Can a phone read the angle?

From down the line it lies flat in the picture, so a camera can measure it, and Pebble Swing takes joint angles out of the frames along with the backswing-to-downswing tempo ratio. Which wrist movement produced the angle is a separate question, and one lens cannot settle it.

What goes wrong, and why

Casting in the golf swing, and what lag really is

Lag is the angle between your lead forearm and the shaft, and most of it is not a wrist joint at all. Here is what the measurements say about where it goes, and where the standard fix goes wrong.

A golfer in an orange shirt partway through the downswing on an open course, hands still above waist height. U.S. Air Force photo by Senior Airman Haiden Morris, Public domain

Lag is a line angle, not a joint angle

9 degreesLead-wrist radial deviation at the top, skilled golfers

Casting is the angle between your lead forearm and the shaft opening too early on the way down. The Titleist Performance Institute files it as an early release that arrives at the ball with the lead wrist cupped and loft added to the face. On a down-the-line still, though, that angle is a pair of lines in a flat photograph, not a reading off any joint.

Matthew Sweeney and colleagues marked up skilled male golfers, mean handicap 2.7, and filmed them at 400 hertz. At the top of the backswing the lead wrist averaged 27 degrees of extension and 9 degrees of radial deviation, the second with a standard deviation wider than the mean. The trail wrist sat at 58 degrees of extension. Whatever produces the ninety-odd degrees you can see, lead-wrist hinge is a small share.

What the trail wrist is really carrying

1,531 °/sPeak trail-wrist flexion velocity through the downswing

Follow the same swings down. The lead wrists went through an average of 46 degrees of flexion and 44 degrees of ulnar deviation; the trail wrists through 66 degrees of flexion, peaking at 1,531 degrees per second. That was the largest range and the fastest motion of the four tracked, which led the authors to suspect trail-wrist flexion carries most of what coaches compress into the word un-cocking.

The same paper reports something the standard cue does not survive. None of the players held a flat lead wrist at the top, and none held a lead-wrist flexion angle steady for any stretch of it. What a casting golfer is accused of losing is not a position the good ones sit in and defend. It is a relationship coming apart in everybody, at different rates.

The point where the angle actually starts to go

30 N·mLargest wrist torque measured on a low-handicap amateur

Steven Nesbit's three-dimensional study walks through a scratch golfer's downswing torque by torque. The club is pulled along the shaft while a positive swing torque is applied; from about the point where the club is vertical, every torque component rises smoothly until the shaft is roughly parallel with the ground. From there to impact they fall away rapidly, a fall Nesbit notes coincides with the increase in swing motion of the wrists. By parallel, the angle is already going.

Chen, Inoue and Shibara's double-pendulum model begins at the natural release point, where the club's own centrifugal torque turns the wrist joint freely with no torque from the golfer. Delay past it and the required wrist torque climbs. Their figure for a late hit runs past 30 newton metres, the ceiling they took from Neal and colleagues, measured on a low-handicap amateur by inverse dynamics, and they conclude humans cannot supply it.

Hear it

Three clicks: takeaway, top, impact. Start at the tour beat and move to whichever one you can actually swing to.

What casting costs, and what it does not

Casting gets priced in instruction at fifteen or twenty yards a club, and no measurement sits behind that figure. Kai-Jen Cheng and co-authors sorted 84 skilled male golfers into four release styles by the timing of the uncocking: two delayed, one early, one that re-cocks the wrists. The delayed styles reached higher peak uncocking velocities. Clubhead speed did not differ significantly between them.

That is a finding about good players, not a clean bill of health. The cost that holds up is the one TPI describes at the ball: a cupped lead wrist and loft added to the face, a strike and launch problem before a distance one. Forward shaft lean and low point are measured in guides of their own.

The impact frame is a poor place to catch a cast. One golfer throws the club from the top, runs out of arc and still shoves the hands into a passable position. Another arrives at waist height looking correct and undoes it in the last few inches, a separate fault with a separate page.

Why holding the angle is the wrong instruction

Golf Distillery puts the usual advice plainly: keep the wrists fully hinged for as long as possible into impact. Nesbit's data say the delay is not made that way. Delayed wrist motion, he writes, is not achieved by applying a hindrance torque; the mechanism is the path of the hands and the wrist angle the downswing starts with. His scratch golfer ran a positive swing torque from the first instant.

What that golfer did have was a hand path with a large initial radius that decreased continuously, and the deepest initial wrist cock of the four subjects. Together those held down the early centrifugal acceleration, so the club had less tendency to move outward despite the torque pushing it. Near the ball the radius was cut sharply, the acceleration jumped and the club was pulled through. The result, Nesbit writes, gives the impression of a consciously delayed wrist motion.

Chen's simulation prices the two strategies. At 15 newton metres of wrist torque, holding the angle was worth 1.4 percent of clubhead speed against a swing applying none, while the same torque spent uncocking was worth 5.6 percent. The instruction aims your effort at the joint doing the least with it.

Questions

Where on my own video should I look for it?: Not at impact. Stop the clip from down the line with your lead arm about horizontal on the way down, and read the angle between that forearm and the shaft. Compare it with the same frame from your own earlier swings, because camera height and distance move that angle on their own.
Does casting really cost fifteen yards a club?: No published measurement supports that figure. In the one study that grouped skilled golfers by release timing, clubhead speed did not differ significantly between the early and delayed styles. The cost TPI describes sits at the face instead, as a cupped lead wrist and extra loft.
Should I stop trying to hold the angle, then?: Stop treating it as something to grip. Nesbit traced the delay in his scratch golfer to hand path and the wrist angle at the top, with an uncocking torque applied throughout. In Chen's model, holding was worth 1.4 percent of clubhead speed against 5.6 percent for the same torque spent uncocking.
Why does it get worse when I swing harder?: TPI puts limited lower-body contribution near the top of its list of causes: when the lower body does not lead the downswing, the upper body overworks and the club gets thrown from the top. Effort applied at the hands is the version of speed you feel first.
Can a phone read the angle?: From down the line it lies flat in the picture, so a camera can measure it, and Pebble Swing takes joint angles out of the frames along with the backswing-to-downswing tempo ratio. Which wrist movement produced the angle is a separate question, and one lens cannot settle it.

Sources

Every number on this page comes from one of these.