Why there is no readiness score
Contrast produces no single readiness, recovery or strain number. A composite hides which input moved it, so it cannot be checked or acted on; it also has to weight inputs that nobody has established weights for. The app reports each measure separately with its own interval and its own evidence tier instead.
Every recovery product ships a number between 0 and 100. Contrast does not, and the reason is not minimalism.
What a composite actually does
A readiness score takes several measures — typically HRV, resting heart rate, sleep duration, maybe respiratory rate — normalises each against your own history, weights them, and sums them.
The weighting is the part that does not survive inspection. To produce one number you have to decide that a one-standard-deviation drop in HRV is worth some specific fraction of an hour of lost sleep. Nobody has established that exchange rate for athletes, and certainly nobody has established it for someone who sat in a sauna last night.
So the weights are chosen. They may be chosen sensibly, and they are still a product decision presented as a measurement.
The decomposition problem
Say your score is 61 today and 78 yesterday. What changed?
You cannot tell from the number. It might be four hours of sleep, or a 12 ms HRV drop, or a small decline in three inputs at once. Those have completely different implications, and the score collapses them into one direction and one magnitude.
A number you cannot decompose is a number you cannot check. And a number you cannot check is one you either trust blindly or ignore, both of which are worse than reading the three inputs.
The noise problem, which is worse for composites
Each input carries its own noise, and combining them does not average it away when the inputs are correlated — which these are.
Resting heart rate is a good illustration of the scale. In a large wearable study it was seasonal, lowest in July and highest in January, and one person in five had at least one week that swung by 10 bpm or more. HRV is noisier still and varies enormously between people, which is why only a within-person comparison means anything.
A composite built from those inputs inherits all of it and reports a single confident-looking integer. The spurious precision is the problem: 61 looks like a measurement to three significant figures, and the underlying data does not support two.
What Contrast reports instead
Each measure on its own card, with four things:
- the difference, in its own units
- a 90% interval from a 1,000-iteration bootstrap
- the number of nights on each side
- a label for how strong the evidence behind the metric is
And a verdict that is allowed to be about the same — which most of them will be. A comparison whose entire likely range is too small to matter says so, rather than being rolled into a score where a null result becomes a small nudge in some direction.
The cost is honest to state: four cards take longer to read than one number, and several of them will say nothing conclusive for the first month. A composite would always have something to show. That is its appeal and also the objection to it.
Where a composite is the right answer
This is a positioning difference rather than a fault in the other approach.
If you want one glanceable answer each morning, and you are willing to delegate the weighting to whoever built the app, a readiness score is a reasonable compression. Athletes with a coach interpreting it use them productively. The autonomic monitoring literature supports the general idea that heart rate measures track training status, even where the specific composite is proprietary.
It fits badly here for a specific reason. Contrast’s whole position is that every number shows its method, its limits and its evidence tier. A composite is the one figure that cannot do that — its method is a weighting nobody can justify, its limits are the union of its inputs’ limits, and its evidence tier would have to be theoretical at best.
Shipping a score would have meant either abandoning the labels for one metric, or labelling the app’s headline number as its least defensible one.
The related refusals
Two others follow from the same reasoning.
No calorie count for the heat. Heart rate in a sauna reaches exercise levels by a mechanism that does not consume the oxygen an exercise formula assumes. Contrast reports the minutes and refuses the conversion, with a citation for refusing.
No streaks or badges. A streak is a mechanism for making someone do a thing on a day they would otherwise skip. Applied to getting into ice water, by people some of whom are managing cardiovascular risk, that is not a neutral piece of interface.
Three refusals, one principle: do not produce a number whose derivation you would be embarrassed to publish next to it.
Common questions
Does Contrast have a recovery or readiness score?
No, deliberately. It reports HRV, resting heart rate, sleep and the rest separately, each with the size of the difference and the range it likely falls in.
What is wrong with a composite score?
It hides which input moved it, so you cannot check it or act on it. It also requires weighting inputs against each other, and nobody has established what those weights should be for sauna and cold exposure.
Is a readiness score useless then?
No. It is a reasonable compression when you want one glanceable answer and are willing to trust the weighting. It is a poor fit for a log whose whole purpose is showing where each number came from.
What does Contrast show instead?
Each measure on its own card, with the difference, a 90% bootstrap interval, the number of nights on each side, and one of four labels for how strong the evidence behind it is.
Sources
- Shaffer F, Ginsberg JP. An overview of heart rate variability metrics and norms. Frontiers in Public Health, 2017.
- Quer G, Gouda P, Galarnyk M, Topol EJ, Steinhubl SR. Inter- and intraindividual variability in daily resting heart rate. PLoS One, 2020.
- Bellenger CR, Fuller JT, Thomson RL, Davison K, Robertson EY, Buckley JD. Monitoring athletic training status through autonomic heart rate regulation: a systematic review and meta-analysis. Sports Medicine, 2016.