THE PRODUCT — FOUR STEPSeach finding rests on the twelve-step review method below
TWELVE STEPS · EVERY REVIEW
The review method and extraction process are being finalised before evidence collection begins. Findings remain unpublished until each review is complete.
0/3 reviews published12 steps before publication
00Plan the reviewDefine the research question, eligibility rules, outcomes, treatment groups, reviewer roles and extraction forms before any study is assessed.REQUIRED STEPEACH REVIEW
01Search and deduplicateSearch each source with a recorded strategy, retain the search history, connect reports from the same study and remove duplicate records.REQUIRED STEPEACH REVIEW
02Screen titles and abstractsTwo reviewers assess every title and abstract independently. Differences are retained and resolved without overwriting either reviewer’s original decision.REQUIRED STEPEACH REVIEW
03Retrieve full textsRecord every access attempt, keep the exact source document used and distinguish unavailable papers from studies that do not meet the criteria.REQUIRED STEPEACH REVIEW
04Assess full-text eligibilityBoth reviewers apply the same eligibility criteria to the same document. Differences are discussed and independently resolved when needed.REQUIRED STEPEACH REVIEW
05Extract the dataTwo reviewers extract structured data independently, link every value to its source and resolve differences before analysis.REQUIRED STEPEACH REVIEW
06Result-specific risk of biasCochrane RoB 2 is applied to the result being analysed, with independent answers, source links and a reconciled final judgement.REQUIRED STEPEACH REVIEW
07Classify treatments and check completenessApply the predefined treatment rules, retain every study and outcome, and check for missing fields, missing source links or dropped records.REQUIRED STEPEACH REVIEW
08Calculate comparable effectsHandle multi-arm, crossover, cluster and repeated-measure designs correctly, using only source-linked values and predefined conversions.REQUIRED STEPEACH REVIEW
09Build and check the evidence networkDefine the treatment groups and comparison network, then check that the studies are connected and sufficiently comparable for analysis.REQUIRED STEPEACH REVIEW
10Run the analysis and sensitivity checksRun the primary network meta-analysis and check consistency, heterogeneity, influential studies, small-study effects, subgroups and sensitivity analyses.REQUIRED STEPEACH REVIEW
11Assess certainty and publishAssess certainty, complete the report and independently check the final evidence package before publishing the findings in full.REQUIRED STEPEACH REVIEW
Nothing publishes until an answer completes the full review. No preliminary rankings, no teased findings, no “early look” — a partial sample can point the wrong way, and readers can’t un-see a wrong answer. Until then the site shows summary counts only, and says so.
RESEARCH IN PROGRESS: VO₂MAX · STRENGTH · HYPERTROPHY
TRACEABLE BY DEFAULT
Release termsstated with each release
Protocolsversioned · public per answer
Decision trailsevery include/exclude logged
Source codeopen release planned
THE PERIODISATION REVIEWin development · not published
A separate living review asks a different question — not what works, but how to arrange it across timescales, from a single session to a whole season. It follows the same twelve-step method above, and it treats the two kinds of evidence differently: controlled trials carry the causal claims, while descriptions of how elite and recreational athletes actually train are labelled observational, not causal. Where the two disagree, the plan follows the trials and says so. It is still being built; nothing is shown as a finding until it is complete.