Your training log says the max lift is up eight percent this block. Your climbing says you're pumped stupid by the third attempt on a boulder you have half worked out. Both things are true at once, and they're not in conflict, because they're answers to two different questions.
Max strength asks how much force you can produce once, at your peak, with everything fresh. Critical force asks something closer to what actually decides your session: how much force you can keep producing, over and over, without your forearm shutting down.
What critical force is
Critical force is the intensity of intermittent finger loading you can sustain more or less indefinitely, the point below which you recover as fast as you fatigue and above which the clock is running. It gets expressed as a percentage of your maximum strength. Our post on CARC-style off-wall endurance training uses it as the metric that experiment lived or died on, and it gives a useful sense of scale: climbers at the 9a level tend to sit around 45 to 50 percent of their max, while the subject in that self-experiment started at 31 percent.
That gap is the interesting part. Two climbers with identical max lifts can have very different critical forces, and the one with the higher percentage is the one still trying hard on attempt five.
The published test
The version researchers use is an all-out test, and it is unpleasant in a specific way. Giles and colleagues established the protocol on a 20mm edge in a half crimp: 24 repetitions of a seven-second maximal pull with three seconds of rest between them, which works out to four minutes of continuous effort. Every rep is a maximal effort, not a paced one. Your force falls off a cliff in the first minute and then flattens out, and critical force is read from where it flattens, taken from the force produced in the final contractions.
Later work has run variations. Baláš and colleagues used a 23mm hold and calculated critical force from the mean force over the last 30 seconds. The shape of the test is the same in both: pull maximally, on a fixed rhythm, until your force stops dropping. Seven on, three off, twenty four times is the version we run.
Warm up first and stop on pain. Full safety guidance.
What it actually takes
Here is the part that gets skipped in most summaries. This test is a measurement of force during every one of those 24 pulls. That means a load cell, a device that reads force directly and logs the trace. There is no version of it with plates and a timer, because the whole point is watching the force curve decay, and a plate stack gives you one binary answer per attempt: it came off the floor, or it didn't.
In practice that is a force dynamometer paired with an app that records the trace and does the arithmetic, which is how we run it here. There are several; Frez is the one we use. The device and the software matter less than the fact that something is logging force continuously rather than you watching a number and guessing.
If you're weighing up whether a force gauge belongs in your setup, our honest case for and against a dynamometer covers that purchase properly. This test is one of the genuine arguments in the "for" column, alongside rate of force development. It's also one of the few finger-training questions where plates simply cannot answer it.
How much to trust the number
More than a vibe, less than a lab value. The original work reported excellent test-retest reliability. Baláš and colleagues put the standard error of measurement at roughly 4.9 lb, and, more usefully for anyone planning training off the result, found that the all-out test tends to overestimate the intensity a climber can genuinely sustain in hangboard work.
So treat the output the way our training log post tells you to treat any single session: as one data point in a trend, not a verdict. A critical force that moves five percent between tests has probably not moved. A number that climbs steadily across three test days over a season has told you something real.
If you don't own a load cell
What you can build is a repeatable benchmark. Pick a fixed edge, a fixed added load somewhere in the submaximal range you can hold for a while, and a fixed work-to-rest rhythm. Do repetitions on that rhythm until your position gives out, not until your hand opens on a rep you have already lost. Record the number of reps and every condition around it. Repeat it under the same conditions in six or twelve weeks.
You can still track endurance. You can't call it critical force, and the difference is worth keeping straight.
That number is not critical force. It's a personal endurance benchmark with your own noise floor, which is exactly what a training log is for. All the hygiene from our testing post transfers without modification: same edge, same grip position, warm fingers, and the same warm-up every time, or you're comparing two different tests and calling the difference progress.
Testing versus training, since they look identical
Repeaters show up here as a test and in our hangboard protocols comparison as a training protocol. Same movement, opposite intent. A test is run to failure under standardised conditions and costs you a session. A training set stops with something in the tank and gets repeated for weeks. Run your test twice a season. Run your training as often as recovery allows.
What to do with the answer
A low percentage relative to your max says your ceiling is high and your ability to work near it isn't, which is an endurance problem rather than a strength one. That's what the CARC post is for, including its caveats about how early the evidence is. A high percentage with a modest max says the opposite, and your time is better spent on the max lift.
Either way you need a setup you can reproduce. Same edge, same platform, same assembly, every test. Our Isometric Training Bundle is one way to get there: the Dual Edge Block on the Traveler platform, packed down and rebuilt the same way each time. That repeatability is doing more for your data than another edge would.
The bottom line
Critical force is the number that explains why the strong climber pumps out and the climber who can keep producing force keeps trying. The validated way to measure it is four brutal minutes on a rhythm with a load cell reading every pull, and even then the number carries real measurement error and tends to flatter what you can sustain. If you own a gauge, run it twice a season and read it as a trend. If you don't, build a benchmark you can repeat and stop calling it critical force. The honest version of a number is worth more than a precise-sounding one you can't reproduce.