Research finding 02: the path was not a continuous decline
The published yearly July series includes increases and decreases, rather than a straight line. In the ten selected observations, the highest UK Regular strength was 149,230 in July 2021; the lowest was 136,030 in July 2025. The size of the gap reflects a selected set of annually observed snapshots, not a claim about an unobserved daily minimum or maximum.
Between July 2025 and July 2026, the published Regular total moved from 136,030 to 137,480, an increase of +1,450 (+1.1%). This recent direction should not be interpreted as proof of a particular recruitment campaign's effectiveness: personnel stock reflects both entering and leaving, and there can be revisions to source numbers.
Researchers should use the rolling 12-month intake/outflow archive for movement statistics, keeping those periods distinct from this exact-date strength series. A new recruit, an exit, a transfer or a training status change cannot be uniquely reconstructed from two aggregate headline numbers.
Research finding 03: service-specific outcomes matter
Headline tri-service figures conceal different experiences for the Army, Royal Navy/Royal Marines and RAF. The archive therefore retains the three services separately instead of dividing a total by three or reporting standalone Royal Marines personnel numbers where the MoD table combines Navy and Marines.
The distinction also matters for users searching for military recruitment advice. The number serving does not reveal vacancies in a specific trade, the relative competitiveness of an officer board, eligibility for a role or the current requirements for a particular assessment.
See our detailed Army versus Navy/Marines versus RAF ten-year comparison for service-by-service changes and differences in reserve populations. The service table is also available in machine-readable source-vintaged JSON.
Methodology: why every point has a dated source
We transcribed July rows from three official Ministry of Defence personnel-statistics reports. July 2017–2018 are linked to the July 2020 report's historical Table 1; July 2019–2022 to the July 2022 report's historical Table 1; and July 2023–2026 to the revised July 2026 report's Table 1. This is a deliberately bounded historical compilation, not automated extraction of every quarter ever published.
Older publications can be revised. Some numbers differ across editions, and historical counts may not have identical revision histories. We preserved each printed source vintage rather than quietly mixing revised values or presenting a falsely precise continuous series. All included counts are rounded to the nearest ten; the components can sum to a slightly different number from the independently rounded total.
The Army's trained-strength definition changed in 2016. For this particular longitudinal analysis we compare total personnel, UK Regular Forces and Volunteer Reserves within their own series. We do not retrofit modern trained/trade-trained measures onto earlier populations or assert that Regular strength is equivalent to the number of trained people.
What this research cannot tell a candidate
It cannot calculate an individual applicant's selection probability. The Armed Forces do not use identical applicant definitions, the Navy and Royal Marines are combined in these headline tables, and historic personnel stock is not a set of applicant-cohort outcomes. Dividing applications by entrants, or military personnel by applicants, would produce a misleading “acceptance rate”.
If your purpose is preparation rather than research, start with the independent military pathway finder, compare official role-specific fitness requirements, and confirm all recruitment-stage information directly with the service. An independent website must not imply official Armed Forces affiliation.
How to reuse and cite the original dataset
Use our versioned long-format CSV or JSON data with source metadata. Suggested attribution: “Project Selection, UK Armed Forces July strength research edition 1.0.0 (16 October 2026), adapted from Ministry of Defence Quarterly Service Personnel Statistics, OGL v3.0.” Always include the official MOD report links when republishing data and do not imply MOD endorsement of our analysis.
Share corrections with the exact July date, population and source Table 1 link via our contact page. A proposed correction goes through an editorial review; it does not silently update the current published dataset.