Field Notes
Primer

Readiness scores tell you today. They do not tell you the lever.

A readiness score can help a coach notice the athlete's current state. It still does not answer the harder question: what should change in training?

A readiness score is not useless. It can be a fast way to say, "something looks different today." That is valuable in a roster review.

The mistake is treating that score as the training decision. A single recovery number can tell a coach that the athlete is carrying fatigue. It does not automatically tell the coach whether to reduce volume, move intensity, protect the benchmark, change sleep targets, or ignore the signal because it has not mattered for this athlete before.

The score is a state. The coaching decision needs a lever.

What readiness gives you

Current state

Low, normal, or high relative to some baseline. Useful for attention, triage, and conversation.

What the coach still needs

Actionable lever

The factor that likely changes the athlete's outcome, how strong the evidence is, and what adjustment follows.

01Do not confuse state with cause

Readiness is usually a bundle. HRV, sleep, resting heart rate, soreness, strain, recent training load, and subjective wellness may all feed the score. That bundle can be useful, but it also hides the reason behind the change.

If readiness is low after a hard block, the obvious answer might be "pull back." Sometimes that is correct. Sometimes the athlete performs well in that exact state because the low score reflects recent productive load, not a harmful pattern. Sometimes the score is low because sleep was poor, and training load is not the lever at all.

The causal question is not "is readiness low?" It is: which input appears to move the outcome for this athlete, under this condition?

02Use readiness as a filter, not a verdict

A readiness score is good at changing what a coach checks first. Low readiness can raise the priority of fatigue, sleep regularity, soreness, travel, or acute load. High readiness can make the coach more willing to test, press, or hold the plan.

That is a filter. It narrows attention. It should not promote itself into a verdict unless the athlete's own history supports that move.

  • Attention: this athlete looks different today.
  • Candidate: sleep, load, HRV, soreness, or decoupling may explain the change.
  • Claim: one candidate has enough athlete-specific evidence to guide the plan.
  • Action: the training adjustment follows the claim, not the score alone.
Readiness contextNot the final claim
State
Readiness is below this athlete's rolling baseline.
Candidate
Recent aerobic decoupling and missed sleep regularity are both elevated.
Gate
Decoupling has a supported link to benchmark performance; sleep is exploratory for this athlete.
Decision
Protect the benchmark by trimming the final aerobic block if decoupling rises again.
Agent wording
Mention readiness as context, not as the reason. The reason is the supported decoupling relationship.

03The lever has to connect to an outcome

A useful training lever is tied to something the coach cares about: benchmark output, late-session pace, completion rate, soreness response, or next-session quality. Without an outcome, the tool is only ranking feelings.

AthDash keeps the outcome attached. If the candidate driver is sleep regularity, the question becomes whether sleep regularity has been associated with the athlete's benchmark sessions after accounting for recent load and fatigue. If the driver is decoupling, the question becomes whether late-session drift has preceded weaker outcomes for this athlete.

That is why the article what should I change for this athlete this week? starts with the coaching decision. A score is easy to display. A lever has to earn its place.

04Agents need the same distinction

This matters even more when an AI coaching agent is helping draft check-ins or athlete replies. If the agent sees "readiness is low" and writes the whole message from that fact, it will sound more certain than the evidence allows.

The better pattern is simple: pass readiness as context, then make the agent read the evidence packet before it writes. The packet should say which driver is supported, what outcome it affects, how wide the interval is, and what the agent is licensed to claim.

That saves tokens because the model does not need to re-read every raw export. More importantly, it prevents the model from turning a state signal into a causal explanation.

The rule

Readiness can tell the coach where to look. AthDash helps decide what the evidence allows the coach to say and change.

05The best score is still a starting point

The practical workflow is not anti-readiness. It is readiness plus a harder second step.

First, use the score to orient. Second, test the candidate drivers against the athlete's own history. Third, keep the caveat visible. Fourth, let the coach make the final call. If an agent writes afterward, it repeats only the licensed version of that call.

That is the difference between "they are red today" and "here is the training adjustment the evidence supports." The first is a status. The second is coaching.