The Bouldering Lock-Off Test: Measuring Static Hold Time Across Elbow Angles

A climber who trains fingerboard hangs three times a week can still stall on a boulder problem the moment both feet cut and the only option left is holding a bent arm at chest height while the other hand searches for the next edge. The fingers aren't the problem there — the crimp is holding fine. What gives out is the elbow: it can't stay locked at that angle long enough to finish the search. Most testing never catches this, since it tests grip in isolation with a fingerboard hang, or full-body power with a campus board move. Almost nothing measures how long an arm holds a fixed bend against bodyweight, or whether that capacity changes with how bent the arm actually is.

That's the gap this protocol closes. It isn't a finger test and it isn't a pull-up test — it measures how long a climber can hold a static lock at four elbow angles on a large edge that removes finger strength as the limiter, then turns the four numbers into a profile showing exactly where the hold breaks down: a flat decline across all four angles versus a sharp drop at one, most often 90°, where a large share of compression and roof-move failures actually happen.

Ask a climber why they fell off a reachy problem and my fingers gave out is the default answer, even when video shows the hand never actually opened. More often: the elbow angle drifted a few degrees as the shoulder fatigued, the body swung off the wall, and the reach came up short before the grip was ever really tested. A fingerboard test can't see this — it measures a straight or near-straight-armed hang. A dynamometer sees it even less, since it measures a grip that isn't loaded by body position at all.

Isolated elbow flexor torque, measured on an isokinetic dynamometer, generally peaks around 80-110° of elbow flexion and drops off toward full extension and full flexion — so a 90° lock-off actually sits near the strongest point on the muscle's own torque curve, not the weakest. If athletes still fail specifically at 90° on the wall, the limiter usually isn't the elbow flexors in isolation. It's the scapular and lat stabilization holding the shoulder blade down and back while the arm works at that angle, under a body position a bent-arm hang or a fingerboard test never loads the same way. Testing hold time at more than one angle is what separates a general endurance problem from a genuinely angle-specific one.

No lab access is required, but the setup needs a way to check an elbow angle, not just a hang position. Use a large jug or big edge — at least 3cm deep, flat and positive — rather than a fingerboard crimp, so grip is never the limiting factor.

Mark chin height and each target angle before the first trial, not during it — checking angle mid-hold with a phone slows the tester and adds noise to the number being recorded.

1. Warm-up (8-10 min): Easy hangs on the jug, band pull-aparts, then 2 submaximal lock-offs at roughly 70% effort.
2. Set the four target angles: 150°, 120°, 90°, and 60° — the included angle between upper arm and forearm. Mark each against the athlete's arm length with the goniometer app once, so the tester can eyeball future reps without re-measuring.
3. Fix the test order: always 150° first, 60° last. The middle angles are highest-output and most sensitive to carried-over fatigue, so testing them mid-session keeps that even.
4. Full rest between angles: 8-10 minutes — short rest is the fastest way to flatten a real profile into a meaningless one.
5. Run the trial: pull to the marked angle. The tester confirms it's within ±5° before starting the clock — timing begins once the position is locked in, not on the pull.
6. End the trial when any one of three things happens: the angle drifts past 10° for over a second, the chin/shoulder marker drops more than 3cm, or the hand opens off the edge.
7. Trials per angle: one is enough once familiar with the setup; on a first session run two and keep the longer, since unfamiliarity under-reports true capacity.

Total session time runs 45-55 minutes including rest — long enough that most coaches run it as its own testing day.

Plot the four hold times against their angles and look at the shape of the line, not any single number. A roughly even decline from 150° down to 60° describes a climber whose limiter is general upper-body isometric endurance — longer, harder circuits should raise the whole curve together at the next test.

A cliff at one angle looks different: hold times on either side sit close to the general trend, but one angle drops far more than its neighbors. In field testing that cliff shows up at 90° more often than any other, matching how many reachy compression moves and roof lips put the working arm near a right angle while the other hand searches. That angle needs specific work — holding at 90° itself, not a longer fingerboard hang — since strength built at one joint angle transfers only partially beyond roughly 15-20°, a pattern covered in the research below.

Thépaut-Mathieu, Van Hoecke, and Maton (1988), in the Journal of Applied Physiology, trained the elbow flexors isometrically at one fixed joint angle over several weeks, then tested strength across a range of angles away from it. Gains were largest at the trained angle and fell off with distance from it, with meaningful transfer generally limited to roughly 15-20° either side. The limitation: the study used isolated single-joint contractions on a lab rig, not a bodyweight lock-off loading the shoulder and scapula the way a climbing hold does, so the transfer window in a climbing-specific position is inferred, not measured directly. It's still the strongest evidence for why testing, and training, more than one angle matters.

Grant, Hynes, Whittaker, and Aitchison (1996), in the Journal of Sports Sciences, compared elite and recreational climbers across a battery of strength and endurance measures and found upper-body static endurance — including a bent-arm hang — separated the two groups far more clearly than one-off strength measures did, with elite climbers holding roughly one and a half to two times as long on comparable endurance tasks. Grip and finger-flexor strength differences between groups were comparatively modest. The caveat: a cross-sectional comparison with a modest sample per group, testing a general bent-arm position rather than a fixed 90° lock-off — it supports testing endurance over one-off strength without validating this exact protocol's numbers.

The ranges below come from field testing this protocol with boulderers across a range of grades, using the large-jug setup above, not a peer-reviewed norm table — none exists yet for a four-angle lock-off test. Treat them as a starting point for flagging an outlier angle, not a pass or fail cutoff.

Notice the shape inside a column, not just the ranges: hold time rises from 150° to 120°, then falls at 90° and further at 60°, at every level tested. That matches the torque-angle curve for elbow flexion described earlier, with a climbing-specific twist — the drop from 120° to 90° tends to be steeper than isolated dynamometer studies would predict, consistent with the added scapular and lat demand at that angle in a loaded position rather than a limit of the elbow joint itself.

Most of what shortens or inflates these numbers happens in the setup, not in the hold itself.

A flat, even decline across all four angles responds to general upper-body isometric endurance work: longer sustained hangs on the jug at moderate intensity, slow pull-up negatives, and circuit-style climbing that keeps the arms under tension longer than a single problem usually demands.

A cliff at one angle needs training at that angle, not a broader program. If 90° is the outlier, the fix is holding at 90° itself — added time or load in small increments, following the angle-specificity pattern Thépaut-Mathieu et al. describe, since gains built at 120° or on a straight-arm hang carry over only partially to a right-angle lock. Three sets at 90°, working toward the next V-grade band above and retested every 4-6 weeks, closes this gap faster than volume added anywhere else in the program.

Don't skip the retest. An angle-specific weakness that isn't checked again can quietly reopen once training focus shifts elsewhere, and running the same four angles under the same setup is the only way to know it's still closed.