ALGORITHM 1.2.0 · LAST REVIEWED 11 SEPTEMBER 2026
The assessment estimates the gap between what you can currently sustain and what a specific trekking objective is likely to ask of you.
It does this by scoring you across seven physical capacities, then comparing each of those scores against a demand profile — a set of target levels for the objective you selected. The output is not a fitness score. It is a distance-to-target measurement, and the same person will score differently against Everest Base Camp than against a New Zealand Great Walk, because the two objectives ask for different things.
Twelve questions are asked. Ten are scored. One records how long you have until departure, which affects the training recommendation but never the score itself. The other asks about knee comfort; it routes the recommendation and also never changes the readiness score. How ready you are today is independent of how much time is left.
Being fit and being ready for a particular trek are different states. A runner with a strong aerobic base can be genuinely underprepared for consecutive days of descent under a loaded pack. The assessment exists to make that kind of mismatch visible before it is discovered on the trail.
02 · The seven dimensions
Each dimension is scored from 0 to 100 using the questions assigned to it. Answer options carry fixed point values between 20 and 90; where a dimension has two questions, its score is the mean of the two.
— Aerobic Base — two questions. Sustainable cardiovascular endurance, inferred from weekly cardio volume and whether you can walk uphill continuously for 30 minutes.
— Time on Feet — one question. Tolerance for long continuous days, inferred from the longest hike you have completed comfortably in the last 8 weeks.
— Uphill Capacity — two questions. Efficiency and staying power on sustained climbs, inferred from the elevation gain you typically handle and how you are faring halfway up a long climb.
— Downhill Durability — one question. Eccentric-load tolerance, inferred from how your legs respond in the 24 hours after a long or steep descent.
— Strength Base — one question. Lower-body strength supporting the trail, inferred from your current strength training practice.
— Pack Capacity — one question. Comfort and control under load, inferred from the heaviest pack you currently hike with regularly.
— Multi-Day Durability — two questions. The ability to repeat hard days without decay, inferred from consecutive hiking days completed recently and how ready you are the day after a long hike.
The dimensions are deliberately non-overlapping in what they ask about. Aerobic Base asks about training volume and sustained effort; Multi-Day Durability asks about recovery between efforts. A person can score well on one and poorly on the other, and that pattern is exactly what the assessment is built to surface.
Four of the seven dimensions rest on a single question. That limitation is addressed in section 09.
03 · Demand profiles
Each objective carries two sets of numbers: a demand per dimension, being the level the trek asks for, and a weight per dimension, being how much that capacity matters to the outcome. Weights sum to 1.00.
Demands — the level each objective asks for
| Objective | Aerobic | Time on Feet | Uphill | Downhill | Strength | Pack | Multi-Day |
|---|---|---|---|---|---|---|---|
| New Zealand Great Walk | 55 | 60 | 55 | 60 | 45 | 50 | 50 |
| Kepler Track | 55 | 65 | 60 | 60 | 45 | 50 | 55 |
| Routeburn Track | 50 | 55 | 55 | 55 | 45 | 50 | 45 |
| Milford Track | 55 | 65 | 55 | 65 | 50 | 50 | 55 |
| Tongariro Alpine Crossing | 50 | 55 | 55 | 60 | 40 | 30 | 25 |
| Tour du Mont Blanc | 65 | 70 | 70 | 75 | 60 | 50 | 70 |
| Everest Base Camp | 75 | 65 | 65 | 55 | 50 | 50 | 70 |
| Kilimanjaro | 70 | 70 | 60 | 65 | 45 | 40 | 60 |
| Everest Three Passes | 85 | 80 | 80 | 70 | 60 | 55 | 85 |
| Annapurna Circuit | 75 | 70 | 70 | 70 | 55 | 50 | 75 |
| Torres del Paine (W or O Circuit) | 60 | 70 | 55 | 60 | 55 | 65 | 65 |
| Other multi-day trek | 60 | 60 | 55 | 55 | 50 | 50 | 60 |
| General Hiking Fitness | 50 | 50 | 45 | 45 | 45 | 40 | 45 |
Weights — how much each capacity matters
| Objective | Aerobic | Time on Feet | Uphill | Downhill | Strength | Pack | Multi-Day |
|---|---|---|---|---|---|---|---|
| New Zealand Great Walk | .16 | .18 | .12 | .14 | .10 | .12 | .18 |
| Kepler Track | .16 | .18 | .15 | .14 | .09 | .12 | .16 |
| Routeburn Track | .17 | .17 | .15 | .14 | .10 | .12 | .15 |
| Milford Track | .15 | .18 | .13 | .17 | .10 | .11 | .16 |
| Tongariro Alpine Crossing | .18 | .20 | .16 | .18 | .10 | .06 | .12 |
| Tour du Mont Blanc | .15 | .14 | .18 | .17 | .12 | .10 | .14 |
| Everest Base Camp | .20 | .16 | .15 | .10 | .10 | .12 | .17 |
| Kilimanjaro | .22 | .17 | .14 | .08 | .10 | .10 | .19 |
| Everest Three Passes | .20 | .16 | .15 | .10 | .10 | .10 | .19 |
| Annapurna Circuit | .20 | .15 | .15 | .12 | .09 | .10 | .19 |
| Torres del Paine (W or O Circuit) | .14 | .17 | .12 | .13 | .11 | .16 | .17 |
| Other multi-day trek | .15 | .16 | .14 | .14 | .12 | .12 | .17 |
| General Hiking Fitness | .15 | .15 | .14 | .14 | .14 | .14 | .14 |
Provenance of these numbers. The demands and weights are expert judgement — set from trek profiles, route data and coaching experience, then reviewed for internal consistency. They are not derived from outcome data, because no such dataset exists yet. Kilimanjaro weighting Aerobic Base at .22 while Tour du Mont Blanc weights Downhill Durability at .17 reflects a considered view of what each route punishes. It does not reflect a measured correlation with anyone completing either trek. See section 9.3.
04 · Producing a result
Four steps run in order. Every one of them is deterministic — the same answers against the same objective always produce the same result.
Step 1 — Dimension scores
score[dimension] = round( mean( points of each answered item in that dimension ) )
Step 2 — Ratio against demand
Each dimension is expressed as a ratio of what the objective asks for. A ratio of 1.00 means the demand is met exactly.
ratio[d] = score[d] ÷ demand[d]
Step 3 — Two weighted totals
The model computes two separate weighted sums, and they are used for different purposes. This split is deliberate and is the substance of the 1.2.0 revision.
weighted = Σ ( ratio[d] × weight[d] ) uncapped — drives bands
cappedWeighted = Σ ( min(ratio[d], 1.10) × weight[d] ) capped — drives the index
The cap means a dimension can contribute at most 110% of its demand toward the index. Beyond that, additional capacity in that dimension stops buying points. Excess capacity is not discarded — it is reported separately as headroom, so it is visible without being mistaken for readiness elsewhere.
Step 4 — Index, band, status, recommendation
The index comes from the capped total. The band comes from the uncapped total, then passes through the priority-gap floor. Per-dimension status comes from the raw score against the raw demand. The training recommendation is a function of the band, the number of priority gaps, whether the objective is at altitude, and how many days remain.
05 · Dimension status
Each dimension is labelled by its distance from that objective's demand — in absolute points, not as a percentage.
— Strong — score is at least 15 points above the demand. Comfortably above what this trek asks for.
— On Track — score is within 5 points below the demand, or above it. At or effectively at the required level.
— Building — score is between 5 and 20 points below the demand. Below the mark but within reach with focused work.
— Priority Gap — score is more than 20 points below the demand. Far enough below to warrant deliberate attention before departure.
The 5-point tolerance on On Track exists because the underlying scores are coarse — a single answer choice can move a dimension by 20 points or more, so treating a 3-point shortfall as a failure would imply precision the instrument does not have.
These bands are fixed in absolute points across all objectives. A 20-point shortfall is a priority gap whether the demand is 45 or 75, even though it represents a larger proportional deficit in the first case. This is a known simplification; see section 9.4.
06 · The readiness index
A single number from 1 to 100. It answers one question: how close is your current preparation to this objective's full demand profile?
index = clamp( round( cappedWeighted × 100 ), 1, 100 )
then, if any dimension is below its demand → hold at 99
100 means every demand is met. That is the whole meaning of the scale, and the second line above enforces it: a weighted average can round up to 100 while one dimension still sits below its target, so the top of the scale is reserved for results where none does.
Headroom
Capacity above the demand profile is reported alongside the index rather than inside it:
headroom = round( (weighted − 1) × 100 ) 0 when weighted ≤ 1
Someone reading “96 · 12% above the demand profile” can see both that they are close to fully prepared and that they carry surplus. Folding that surplus into the index would have let strength in one area conceal a deficit in another — which is precisely the failure mode the assessment exists to prevent.
Scores before and after 1.2.0 are not comparable. Under algorithm 1.1.0 the index was divided by a ceiling of 1.25, so a person meeting every single demand scored 80. Under 1.2.0 the same answers score 100. Any index recorded before 11 September 2026 sits on a different scale and should not be compared with one recorded after. Stored results carry their algorithm version for this reason.
07 · Readiness bands
The band is a qualitative summary derived from the uncapped weighted total, then constrained by a floor rule.
— Strongly Prepared — weighted total of 1.10 or above. Meets or exceeds the profile across the board.
— Trek Ready Soon — weighted total of 0.92 or above. Close to the demand profile; refinement rather than rebuilding.
— Moderately Ready — weighted total of 0.72 or above. A real base is present, with meaningful work still to do.
— Building Foundation — weighted total below 0.72. Foundational capacity is the priority before objective-specific training.
The priority-gap floor
A weighted average can be pulled up by strong dimensions while one capacity remains far below what the trek demands. Two rules prevent that from reading as readiness. First: any priority gap prevents a result being called Strongly Prepared — it is held at Trek Ready Soon. Second: a priority gap in one of the objective's three highest-weighted dimensions additionally prevents Trek Ready Soon — it is held at Moderately Ready.
Whenever this rule changes a band, the results page states it explicitly and names the dimension responsible — for example: “Held at Moderately Ready because aerobic base is still a priority gap for this objective.” The floor never promotes a band; it can only hold one down.
Effect, measured across all possible responses. When this rule was introduced (version 1.2.0), every combination of answers was evaluated against the six objectives offered at the time: 7,864,320 scenarios. It reduced results labelled “ready” that carried at least one priority gap from 34.3% to 14.5%, and none of the remaining results held a gap in a primary dimension. Peaksted now offers more objectives and the demand profiles have since been revised, so these two figures have not been re-measured. No result is labelled Strongly Prepared while carrying any priority gap at all.
08 · Worked example
A real profile scored against Everest Base Camp, showing every intermediate value.
| Dimension | Score | Demand | Ratio | Capped | Weight | Status |
|---|---|---|---|---|---|---|
| Aerobic Base | 65 | 75 | 0.87 | 0.87 | .20 | Building |
| Time on Feet | 65 | 65 | 1.00 | 1.00 | .16 | On Track |
| Uphill Capacity | 63 | 65 | 0.97 | 0.97 | .15 | On Track |
| Downhill Durability | 65 | 55 | 1.18 | 1.10 | .10 | On Track |
| Strength Base | 62 | 50 | 1.24 | 1.10 | .10 | On Track |
| Pack Capacity | 65 | 50 | 1.30 | 1.10 | .12 | Strong |
| Multi-Day Durability | 65 | 70 | 0.93 | 0.93 | .17 | On Track |
Under algorithm 1.1.0 this same profile returned an index of 79. The person had not changed; the scale had been misleading them. Three dimensions still need work, and the index reflects that they are close — not that they are 79% of the way to a pass mark.
09 · Known limitations
These are the reasons a result should be read as a structured starting point rather than a measurement. Each is stated with what would be required to remove it.
9.1 — Every input is self-reported
Nothing is measured. The assessment records what a person believes about their own training and recovery, and people are systematically imprecise about both — recalling their best week rather than their typical one, or answering aspirationally. Questions are worded to ask about recent, concrete, verifiable behaviour to narrow this, but it cannot be closed by wording alone.
What would remove it: objective inputs — connected activity data, a timed uphill test, or a supervised field assessment.
9.2 — Four dimensions rest on a single question
Time on Feet, Downhill Durability, Strength Base and Pack Capacity are each inferred from one answer. A single item cannot be checked for internal consistency, and one misread question moves that dimension by 20 points or more — enough to change its status and, through the floor rule, potentially the band. The three two-item dimensions are proportionally more stable, but two items is still a thin instrument.
What would remove it: a second item per dimension, chosen so that the pair correlates in real responses. This lengthens the assessment, so it is being held until enough completed assessments exist to select the right questions rather than guess at them.
9.3 — The demand profiles are unvalidated
The single largest limitation. Every demand and weight in section 03 is expert judgement. No study establishes that Aerobic Base 75 is the right threshold for Everest Base Camp, or that weighting it at .20 predicts anything about how that trek goes. The profiles are internally coherent and defensible from route characteristics; they have not been tested against outcomes.
What would remove it: outcome data — assessment results paired with what actually happened on the trek. This is why completed assessments are retained, and it is a multi-season undertaking, not a near-term fix.
9.4 — Status thresholds are absolute, not proportional
A 20-point shortfall triggers a priority gap regardless of whether the demand is 45 or 75. Against a demand of 45 that is a 44% deficit; against 75 it is 27%. The absolute rule is simpler to explain and behaves predictably, but it is stricter on low-demand objectives than on high-demand ones.
What would remove it: proportional thresholds. Deliberately deferred — the change would move a large number of existing results, and there is no evidence yet as to which rule better reflects reality.
9.5 — Altitude tolerance is not modelled
For the high-altitude objectives (Everest Base Camp, Everest Three Passes, Annapurna Circuit and Kilimanjaro), the demand profiles describe the physical work of trekking at altitude. They do not estimate how any individual will acclimatise. Susceptibility to altitude illness is not reliably predicted by fitness, and a strong result on this assessment carries no implication about it whatsoever.
What would remove it: nothing available here. This belongs with a doctor and a properly graded ascent profile, and the assessment defers to both.
9.6 — No screening for health, injury or history
The assessment asks nothing about medical conditions, medication, injury history, age or prior altitude experience. All of these bear materially on whether a trek is advisable, and none is captured.
What would remove it: a screening layer separate from the readiness score. Kept separate deliberately — health screening is a clinical function and should not be folded into a training tool.
9.7 — A point-in-time snapshot
The result describes today. It does not project forward — it will not tell you where you will be in twelve weeks, because that depends on the training done between now and then. The timeline question shapes the recommendation, never the score.
What would remove it: retesting. Results are stored against a stable assessment identifier so a person's trajectory can be tracked across repeat assessments.
10 · What this is not
Stated plainly, because the failure modes matter more than the features.
— Not a medical assessment. It does not screen for, detect, or account for any health condition. It is not a clearance to trek, and it is no substitute for advice from a doctor — particularly for high-altitude objectives.
— Not a prediction of success. Weather, altitude, illness, group pace, guiding quality and luck all bear on how a trek goes. None is modelled here.
— Not an altitude tolerance test. Fitness and acclimatisation are distinct, and the assessment says nothing about the second.
— Not a validated psychometric instrument. It has no published reliability or validity statistics, because the data required to compute them does not yet exist.
— Not a pass or fail. A low index against a demanding objective is an accurate description of distance, not a verdict on the person or the plan.
What it is good for. Making the shape of someone's preparation legible — which capacities are carrying them, which are lagging, and what a specific objective is actually going to ask. For that purpose the model is coherent, transparent and consistent: identical answers always produce an identical result, and every threshold that produced it is written on this page.
11 · Data and retention
What a completed assessment records, and why.
— Stored: the answers given, the computed dimension scores, the objective, the resulting index and band, the training recommendation, the algorithm version, and a unique assessment identifier.
— Contact details are stored separately from the assessment record, so a deletion request can remove the identifying fields while the anonymous result stays available for improving the model.
— Algorithm version is written into every record. A result scored under 1.1.0 is never silently compared with one scored under 1.2.0.
— Marketing consent is recorded as its own explicit field. Completing an assessment is a service interaction; it is not treated as consent to marketing.
— Purpose: retained results exist to test the demand profiles against outcomes and to select second items for the single-item dimensions.
Retention periods and deletion requests are covered in the privacy policy.
12 · Version history
Scoring changes are versioned. Any change that moves results is recorded here.
— 1.2.0 — 11 September 2026. Per-dimension surplus capped at 110% of demand; index rescaled so 100 means the full demand profile is met; index held at 99 while any dimension is unmet; headroom reported separately; priority-gap floor applied to bands with an on-screen reason. Not comparable with earlier versions.
— 1.1.0 — August 2026. Corrected the ranking of strongest area and biggest gap to compare capacity against each objective's demand rather than against raw score, so results reflect what the chosen trek asks for.
— 1.0.0 — Initial seven-dimension model, demand profiles and banding.
Product routing, meaning which training pathway is recommended, was verified unchanged across all 7,864,320 possible results (six objectives) when 1.2.0 was introduced. Nobody’s recommendation moved as a side effect of the scoring revision.
