2.5 Evaluating Programming: Benchmarks, Adjustment & CrossFit's Definition of Fitness
Key Takeaways
- Programme evaluation asks two separate questions: is the athlete improving against CrossFit's definition of fitness, and is the athlete moving toward their own stated goal?
- Benchmarks are the instrument: a fixed, repeatable workout with fixed standards, retested on an 8-12 week cycle under matched conditions.
- Adjust one variable at a time and give it 4-6 weeks, or you cannot attribute the change to anything.
- The three fitness models — the ten general physical skills, the hopper, and the metabolic pathways — are evaluation instruments, not slogans.
- A programme that improves one domain while degrading two others has failed the GPP test even when the athlete is pleased with the result.
Two Yardsticks, Not One
Domain 2 of the content outline devotes an entire task cluster to evaluating results, and it names two different criteria: "evaluate program success relative to athlete goals and needs" and "evaluate athlete progress relative to CrossFit's definition of fitness and their personal goal." A programme can succeed on one and fail on the other, and a CCFT is expected to notice.
- Yardstick 1 — CrossFit's definition of fitness. Increased work capacity across broad time and modal domains. Measured by a benchmark battery that spans modalities and durations.
- Yardstick 2 — the athlete's stated goal. Measured by the specific target agreed at intake: the 5 km time, the first strict pull-up, the return to pain-free hiking.
When these conflict — the athlete's deadlift went up 50 lb but their mile time went up 90 seconds, and their goal was to run a half marathon — the conversation is a coaching conversation, not a data problem. State plainly what improved, what regressed, and what the trade-off costs relative to the goal they told you they had.
Benchmarks as the Measurement Instrument
A benchmark is a workout deliberately held constant so that changes in the score can only be attributed to changes in the athlete. Three properties make one usable:
- Fixed task. Same movements, same loads, same reps, same time cap.
- Fixed standards. Written down, coached identically, and enforced identically each time. Standards drift is the most common cause of fake improvement.
- Fixed sampling position. Retested at a comparable point in the training week, not the day after the heaviest session of the cycle.
A practical retest cadence is every 8-12 weeks per benchmark, staggered so you are not testing everything in the same week. Testing more often produces noise; testing less often means you discover a failed cycle only after three months of it.
The Three Models You Are Evaluating Against
CrossFit's definition of fitness is operationalised by models that a CCFT uses as evaluation instruments:
| Model | What it measures | How you use it in evaluation |
|---|---|---|
| Ten general physical skills | Cardiovascular/respiratory endurance, stamina, strength, flexibility, power, speed, coordination, agility, balance, accuracy | Score the athlete's programme exposure and progress skill by skill; find the untrained skills |
| The hopper | Performance across randomly drawn physical tasks | Ask whether an unfamiliar task would go well; a programme producing narrow athletes fails here |
| Metabolic pathways | Phosphagen, glycolytic, oxidative | Check the time-domain distribution of testing and training; find the neglected pathway |
A fourth model — the sickness-wellness-fitness continuum — evaluates health markers rather than performance, and connects to the lifestyle-education domain: blood pressure, body composition, resting heart rate, and blood markers moving toward the fitness end of the continuum is programme evidence too.
Adjusting: One Variable, Four to Six Weeks
When the evidence says a cycle underperformed, resist the urge to change everything. Change one variable and hold it long enough to read the result:
- Volume (total working sets, total reps, weekly minutes)
- Intensity (percentage of 1RM, pacing prescription, time caps)
- Frequency (sessions per week, or exposures per movement pattern)
- Density (work-to-rest structure)
- Exercise selection (movement substitutions)
Four to six weeks is the minimum window in which a change is readable, because most adaptations are not visible before that. If you change volume, intensity, and exercise selection simultaneously and the athlete improves, you have learned nothing you can reuse.
Worked Example
An athlete's 12-week block prioritised strength: three heavy days a week, conditioning reduced to two short pieces. Retest results: back squat 275 to 315 lb (+15%), "Fran" 4:10 to 4:05 (flat), "Helen" 11:20 to 12:40 (-12%), 5 km run 24:30 to 27:10 (-11%). Athlete's stated goal at intake: complete a local sprint triathlon in the summer.
Evaluation: the block succeeded against its own internal target (strength) and failed both yardsticks that matter here. Work capacity narrowed rather than broadened — two of four domains regressed — and the result moves away from the athlete's stated goal. The correct action is not to declare the block a success because the squat went up. It is to report all four numbers to the athlete, name the trade-off explicitly, and write the next block with the aerobic domain as the priority variable, holding strength volume at maintenance (roughly one heavy session per week) rather than eliminating it.
When to Change the Goal Instead of the Programme
Sometimes the evidence says the goal, not the programme, is wrong. An athlete whose intake goal was a 405 lb deadlift within 12 weeks from a starting 315 lb was given an unachievable target; the programme did not fail, the goal-setting did. Renegotiating a goal is a legitimate outcome of evaluation, provided you do it with the data in front of you and you replace it with a target the athlete can still care about.
Over a 12-week block an athlete's back squat rises 15 percent while their 5 km run and 'Helen' times both worsen by more than 10 percent. Their intake goal was a summer triathlon. How should the CCFT evaluate the block?
A coach responds to a disappointing cycle by simultaneously cutting volume 20 percent, raising working percentages, adding two sessions per week, and swapping most barbell movements for dumbbell versions. What is the primary problem with this response?
Which change would most compromise the validity of a benchmark retest used to evaluate a training cycle?