Why does Base not have an overall fitness score?

Because strength and endurance are different capacities measured in different units, and pooling them produces a number that cannot answer either question. A rising blended score could mean your squat improved or your threshold improved or both moved in opposite directions by different amounts — and the last case is the one a hybrid athlete most needs to see.

Are fitness scores like strain and readiness useful?

Readiness-type scores that combine recovery inputs are genuinely useful, because their inputs measure the same underlying thing and the output is one decision: how hard to go today. Scores that combine performance capacities are different — they average things that are not commensurable, and the averaging is where the information is lost.

What should I track instead of a fitness score?

Strength as an estimated one-rep-max trend across your main lifts, endurance as threshold pace or FTP with its own confidence label, and both against shared budgets: recovery score and measured maintenance calories. Four numbers instead of one, each of which points at a specific action.

Does Base score recovery?

Yes, because that is a case where combining inputs is legitimate. HRV, resting heart rate and sleep are three measures of the same underlying state, and the output is a single decision about today's session. Combining is appropriate there and inappropriate for performance capacities.

Why Base Has No Single Fitness Score

What a blended score would actually mean

Suppose your fitness score goes from 71 to 74. What happened?

Your squat e1RM rose 4% and your threshold pace was flat.
Your threshold improved 3% and your squat was flat.
Your squat rose 8% and your threshold degraded 4%.
Both improved slightly.
You lost 6 lb, so your power-to-weight improved while absolute strength fell.

Every one of those produces the same +3, and they call for completely different responses. The third case is the specific thing a hybrid athlete is trying to avoid, and it is indistinguishable from the fourth in the summary.

A number that cannot distinguish "both improved" from "one ate the other" is not a summary. It is a loss of information wearing a summary's clothes.

The unit problem is not solvable by normalising

The usual proposal is to z-score each capacity against your own history and average them. That does make the units comparable, and it introduces a worse problem: the weighting is a value judgement pretending to be arithmetic.

Is a 1% gain in threshold pace worth a 1% gain in squat e1RM? For a marathoner training for a race, no — threshold is worth ten times as much. For someone in a strength block, the reverse. For a HYROX athlete, it depends which stations are limiting.

Any fixed weighting encodes one answer to a question only the athlete can answer, and then hides it inside a number that looks objective. That is not a technical limitation. It is a category error.

When combining IS legitimate

Base does combine inputs into a score in one place: recovery. That is not inconsistent, and the difference is worth stating precisely.

	Recovery score	A blended fitness score
Inputs	HRV, resting HR, sleep	Strength, endurance
Relationship	Three measures of one underlying state	Two different capacities
Output	One decision: how hard to go today	No single decision
If it moves	Act on today's session	Cannot tell what to act on
Weighting	Defensible from evidence on which signal predicts readiness	A value judgement about your goals

HRV, resting heart rate and sleep are noisy proxies for the same thing — autonomic state and accumulated fatigue. Averaging noisy measures of one quantity reduces the noise. That is exactly what averaging is for.

Squat strength and threshold pace are not noisy measures of one quantity. They are two quantities. Averaging them does not reduce noise; it destroys signal.

What replaces it

Four numbers, each pointing at a specific action.

Strength: e1RM trend across your main lifts. The slope over a block, not any single session. It converts sets at different weights and rep counts onto one axis, which is what makes a block's worth of messy training data readable.

Endurance: threshold, with its confidence. Pace per mile or watts, carrying a high/medium/low label based on how recent and how long the effort behind it was. Never blended with the strength number and never blended between run and ride, because a combined "endurance score" across two disciplines has the same problem one level down.

Recovery: scored against your own baselines, producing a RIR offset for today and a deload trigger.

Energy: measured maintenance, and the surplus you actually averaged rather than the one you set.

Four numbers is more to read than one. It is also four numbers you can act on, versus one you cannot.

The thing a single score would be hiding

For a hybrid athlete the interesting information is almost never the level of either capacity. It is the relationship between them and the two budgets they share.

Rising e1RM, rising threshold, flat recovery: the block is working, hold everything.

Rising e1RM, falling threshold, rising endurance volume: the endurance work is going in at the wrong intensity. Check whether the easy days are easy.

Flat e1RM, rising threshold, falling recovery: endurance load has taken over, whether or not you decided that.

Rising scale weight, flat e1RM, falling recovery: this is not a bulk. It is an overreach with extra calories on it.

Every one of those is a different diagnosis with a different fix, and every one of them is a pattern across several numbers. A single score collapses all four into the same reading.

What to do next

If you currently steer by a blended score in another app, try this for one block: write down the two component capacities separately at the start and the end. If they moved in the same direction, the score was telling you the truth. If they moved in opposite directions, look at what the score said, and consider what it cost you not to know.

Why every number here shows its math is the same argument from the other end.