Caffeine & Exercise Performance: What Dose, Timing and Habit Actually Tell Us

Research snapshot

Research Question: What does caffeine evidence support for exercise performance once task, dose, timing, habituation, variability and sleep/recovery are kept distinct?
Main Finding: Caffeine can improve some adult exercise-performance outcomes on average, but effects remain task-, endpoint- and protocol-dependent; no universal optimal dose or timing interval is established.
Evidence Reviewed: 15 governed human evidence records spanning syntheses, controlled dose/timing studies, habituation/genotype evidence and sleep/recovery trade-offs.
Framework: Task → evidence fit → studied dose → timing → habituation/variability → sleep/recovery → bounded decision.
Protocol: Describe studied context, preserve individual uncertainty and fail closed when requested precision exceeds evidence.
Publication Status: SKP-012 v0.2 production activation for governed release TFS-REL-0012 after G5 scientific approval, RP-12 editorial approval and G7 release readiness.

01 / Executive summary

Executive Summary

02 / Research question

Research Question

In generally healthy adults, what does governed evidence support about caffeine and exercise performance across task, dose and timing contexts, and how should habituation, response variability and sleep/recovery trade-offs constrain interpretation?

03 / Scientific context

Start with the exercise task, then interpret the dose and timing conditions actually studied.

Across represented healthy-adult exercise contexts, acute caffeine is associated with small-to-moderate average performance benefits in several domains, but effects remain task-, endpoint- and protocol-dependent. Low and moderate aerobic time-trial dose strata are represented without establishing a universal optimum or a more-is-better rule. Timing can matter in represented contexts without establishing a universal interval. Habitual intake, individual variability and CYP1A2 do not authorize deterministic personalization, while subsequent sleep/recovery remains a separate trade-off domain.

04 / Evidence review

Evidence Review

05 / Core insight

The useful question is not “What is the best caffeine dose?”

Ask whether the evidence matches the exercise task and endpoint, then inspect the dose and timing conditions actually studied. Preserve habituation and individual-response uncertainty, and surface sleep/recovery separately. Studied condition is not personal recommendation.

06 / Framework

Caffeine & Exercise Performance Decision Context

Canonical framework / 1.0

Caffeine & Exercise Performance Decision Context

Task → evidence applicability → studied dose → timing context → habituation/variability uncertainty → sleep/recovery trade-off → bounded decision output.

01 / Task

Define the exercise task and endpoint before exposure detail.

02 / Evidence fit

Match only governed performance evidence to the task represented.

03 / Dose

Describe studied exposure; do not prescribe an optimum.

04 / Timing

Retain protocol specificity; no universal interval.

05 / Habituation & variability

Do not infer tolerance, guaranteed response or genotype prescription.

06 / Trade-off

Surface sleep/recovery separately from acute performance.

A task-first framework that preserves uncertainty instead of manufacturing a universal caffeine strategy.

07 / Protocol

Caffeine & Exercise Performance Decision Support Protocol

Practical protocol

Objective

Interpret governed adult exercise-performance evidence without converting group evidence into individualized dosing or universal timing.

Audience

Generally healthy adults considering bounded exercise-performance evidence; individualized clinical advice remains outside scope.

A meaningful governed task/outcome match exists

Then

Describe the bounded performance evidence and studied exposure context.

Why

Applicability precedes optimization.

An exact personal or universal optimum is requested

Then

Return INSUFFICIENT_FOR_REQUESTED_PRECISION.

Why

SKP-012 does not establish a universal or individualized optimum.

Habitual use or genotype is used to predict response

Then

Retain uncertainty and stop deterministic personalization.

Why

Current authority does not establish individual tolerance or genotype-based prescription.

Sleep/recovery is decision-relevant

Then

Surface it separately from acute performance.

Why

Both effects may coexist and no individual numerical net effect is established.

Core principles

• Task before exposure detail.
• Applicability before optimization.
• Studied dose context, not universal dosing.
• Protocol-specific timing.
• Habitual exposure is context, not a tolerance diagnosis.
• Group evidence is not individual prediction.
• Sleep/recovery is a separate trade-off.
• Fail closed when requested precision exceeds evidence.

Recommendations

• Describe task-matched studied exposure only.
• Reject universal optimal dose and universal timing.
• Do not escalate dose from habitual-use status.
• Do not prescribe from CYP1A2 genotype.
• Keep acute performance and sleep/recovery evidence separate.

Limitations

• No universal optimal dose.
• No universal timing interval.
• No automatic tolerance diagnosis or dose escalation.
• No deterministic responder/genotype classification.
• No universal athlete sleep cutoff.
• No numerical net-benefit score.

Educational decision support only. Not medical advice, diagnosis, treatment, individualized caffeine prescription or medical clearance.

08 / Limitations

Task, protocol and individual-response boundaries materially limit transfer.

Performance evidence is predominantly from healthy adults and materially male-dominant. Exercise domains and endpoints are heterogeneous; synthesis layers overlap; several direct dose/timing studies are small and protocol-specific. Absence of eligible high-dose aerobic time-trial evidence is not evidence of ineffectiveness. Habitual intake is not equivalent to tolerance or withdrawal. CYP1A2 evidence remains non-deterministic. General-adult sleep evidence is broader than athlete-specific evidence, which is smaller and heterogeneous. Diagnosis, treatment, medication interactions, pregnancy and clinical-population personalization remain outside scope.

• No universal optimal dose.
• No universal timing interval.
• No automatic tolerance diagnosis or dose escalation.
• No deterministic responder/genotype classification.
• No universal athlete sleep cutoff.
• No numerical net-benefit score.

09 / Practical takeaways

Practical Takeaways

• Match the exercise task and endpoint before interpreting exposure.
• Treat these as studied exposure conditions, not a personal or universal optimum.
• Do not infer more caffeine = more performance or treat the >6 mg/kg evidence gap as ineffectiveness.
• Keep time trial, time to exhaustion, strength, muscular endurance and power distinct.
• Do not transform a protocol-specific interval into a universal 60-minute rule.
• Habitual use does not establish tolerance and does not authorize automatic dose escalation.
• Keep population evidence separate from individual prediction.
• Do not classify responders or prescribe dose/timing from genotype alone.
• Surface sleep/recovery separately and do not create a universal athlete cutoff.
• Do not fabricate a numerical performance-versus-recovery score.

10 / Scientific references

Scientific References

EV-SKP012-001 — Grgic J et al. Caffeine supplementation and physical performance: an umbrella review of 21 published meta-analyses. Br J Sports Med. 2020. Year: 2020. Study Type: Umbrella review. DOI: 10.1136/bjsports-2018-100278. PMID: 30926628.
EV-SKP012-002 — Grgic J et al. Caffeine supplementation and physical performance: effects on muscle strength and power. J Int Soc Sports Nutr. 2018. Year: 2018. Study Type: Systematic review / meta-analysis. DOI: 10.1186/s12970-018-0216-0. PMID: 29527137.
EV-SKP012-003 — Warren GL et al. Effect of caffeine ingestion on muscular strength and endurance: a meta-analysis. Med Sci Sports Exerc. 2010. Year: 2010. Study Type: Meta-analysis. DOI: 10.1249/MSS.0b013e3181cabbd8. PMID: 20019636.
EV-SKP012-004 — Carvalho A et al. The effect of caffeine supplementation on exercise performance and the influence of habitual caffeine consumption: a systematic review and meta-analysis. Sports Med. 2022. Year: 2022. Study Type: Systematic review / meta-analysis. DOI: 10.1007/s40279-022-01685-0. PMID: 35536449.
EV-SKP012-005 — Graham TE, Spriet LL. Metabolic, catecholamine, and exercise performance responses to various doses of caffeine. J Appl Physiol. 1995. Year: 1995. Study Type: Controlled trial. DOI: 10.1152/jappl.1995.78.3.867. PMID: 7775331.
EV-SKP012-006 — Grgic J et al. Caffeine supplementation and physical performance in resistance-trained men. Int J Sports Physiol Perform. 2020. Year: 2020. Study Type: Crossover study. DOI: 10.1123/ijspp.2019-0433. PMID: 31575825.
EV-SKP012-007 — Harty PS et al. Caffeine timing and lower-body muscular performance. Front Nutr. 2020;7:585900. Year: 2020. Study Type: Randomized crossover trial. DOI: 10.3389/fnut.2020.585900. PMID: 33330586.
EV-SKP012-008 — Wilk M et al. Caffeine intake and bench press performance in strength-trained men. Nutrients. 2019;11(7):1465. Year: 2019. Study Type: Randomized crossover trial. DOI: 10.3390/nu11071465. PMID: 31252655.
EV-SKP012-009 — Barreto G et al. CYP1A2 genotype and caffeine effects on exercise performance: systematic review and meta-analysis. Med Sci Sports Exerc. 2024. Year: 2024. Study Type: Systematic review / meta-analysis. DOI: 10.1249/MSS.0000000000003313. PMID: 37844569.
EV-SKP012-010 — Gardiner C et al. The effect of caffeine on subsequent sleep: a systematic review and meta-analysis. Sleep Med Rev. 2023;69:101764. Year: 2023. Study Type: Systematic review / meta-analysis. DOI: 10.1016/j.smrv.2023.101764. PMID: 36870101.
EV-SKP012-011 — Bodur M et al. Caffeine and sleep in athletes: systematic review. Clin Nutr ESPEN. 2025;65:76–85. Year: 2025. Study Type: Systematic review. DOI: 10.1016/j.clnesp.2024.11.007. PMID: 39551351.
EV-SKP012-012 — Kocak A et al. Caffeine intake before late-day training or competition and sleep in athletes. Sports. 2025;13(9):317. Year: 2025. Study Type: Systematic review / meta-analysis. DOI: 10.3390/sports13090317. PMID: 41003623.
EV-SKP012-013 — Santos et al. Caffeine dose and aerobic time-trial performance. Nutrients. 2026;18(12):1989. Year: 2026. Study Type: Systematic review / meta-analysis. DOI: 10.3390/nu18121989. PMID: 42356375.
EV-SKP012-014 — Xue R et al. Caffeine strategies and time-trial performance: network meta-analysis. Nutrients. 2025;17(23):3792. Year: 2025. Study Type: Systematic review / network meta-analysis. DOI: 10.3390/nu17233792. PMID: 41374083.
EV-SKP012-015 — Taghizadeh Bilondi H et al. Caffeine and muscular strength and endurance: umbrella/meta-analysis. Heliyon. 2024;10(15):e35025. Year: 2024. Study Type: Umbrella/meta-analysis. DOI: 10.1016/j.heliyon.2024.e35025. PMID: 39170391.

How to cite

TasteFromSoul. Caffeine & Exercise Performance: What Dose, Timing and Habit Actually Tell Us. SKP-012, version 0.2. Publisher: TasteFromSoul.

Version history

v0.2 — RP-14 production activation for governed release TFS-REL-0012.
v0.2 — RP-13 release package created and G7 RELEASE_READY granted.
v0.2 — RP-11 bounded scientific re-review and RP-12 editorial approval completed after correction.
v0.1 — RP-09 non-public publication candidate built from the governed SKP-012 chain.