sensitivity
26 stays first as weights change within the tested space.
ROMSILVA / AUDIT / 27 AUG 2026
19 became law.
26 ranked first in the model.
HG 123/2026 set a target of 19 regional forestry directorates on 5 March 2026. Romsilva's organisational-audit presentation published on 27 August says the 26-region scenario ranked first across the tested weighting configurations; 19 appeared in the top three in about 74% of them.
What you get: separate what became a legal target from what ranked first in a model, and see why neither number is a verdict on its own.
Next: the calendar, scenarios, 14 criteria, result, and what remains unknown.
ROMSILVA / OFFICIAL AUDIT19 LEGAL TARGET26 MODEL RANK 1
The decisive sources are HG 123/2026 and the official audit presentation. This page makes no claims about individuals.
TO TWO QUESTIONSthe stamp is not the score
LEGAL TARGET19regional forestry directorates set by government decision
5 MAR 2026
MODEL RESULT26the scenario ranked first across tested weight configurations
AUDIT PUBLISHED 27 AUG
First answer: law and model have different jobs.
01 / the chronology
The decision came before the summary
05 MAR 202619
Government adopts HG 123/2026: territorial reorganisation targets 19 regional directorates.
27 AUG 202626
Romsilva publishes the audit presentation: the 26-region scenario ranks first in the tested model.
What the sequence shows: the legal target existed before the audit summary was published. It does not by itself show why the target was chosen or what will be implemented.
02 / one figure, another question
19 is not 26 with a new label
LEGAL DOCUMENT19
a target for regional structure set by a government decision.
AUDIT MODEL26
a scenario compared across criteria and weights in an organisational analysis.
IT IS NOTa vote
the model ranking is not a public consultation and does not automatically change the law.
| source | claim | does not automatically follow |
|---|---|---|
| HG 123/2026 | the legal target is 19 | that the audit must rank 19 first |
| Romsilva audit | 26 ranks first in the published model | that 26 became law |
| both | a documented sequence exists | individual blame or measured financial outcome |
03 / the audit mechanics
Seven scenarios enter one grid
The official presentation tests seven regional-directorate counts and 14 active criteria. Each criterion is normalised on a 0–5 scale, then different weight configurations are tested.
SCENARIOS10 · 13 · 19 · 20
four of the counts compared in the audit.
SCENARIOS24 · 26 · 41
three more counts; 26 is a scenario, not the law's result.
COUNT DIFFERENCE26 − 19 = 7scenarios
The numeric difference describes two compared configurations. It is not an estimate of staff, cost, or savings.
04 / the model result
26 wins the grid. 19 stays competitive
| scenario | result | how to read it |
|---|---|---|
| 26 regions | rank 1 in 100% | across all tested weight configurations |
| 19 regions | top 3 in about 74% | strong in the model, but not first in every configuration |
| 7 scenarios | compared | 10, 13, 19, 20, 24, 26, and 41 |
“First” has a scope: first in the audit model, with the criteria, normalisation, and tested weights. It does not by itself mean correct, mandatory, or implemented.
05 / what does not follow
The model does not rewrite the decision
a position
19 enters the top three in about 74% of configurations.
legality
that the 19-region target is unlawful or that 26 is already adopted.
financial outcome
the 924,000-lei audit cost is not a saving and does not prove a realised gain.
Interpretation risk: this is an organisational analysis, not an investigation of people. The page publishes no names and assigns no motives.
06 / the evidence boundary
Rank one has a footnote
WE CAN SAY19 is the legal target; 26 ranks first in the published model.Two claims from two documents, joined by a checkable chronology.
WE CANNOT SAYthat 26 is mandatory or that 19 failed.The presentation does not publish every input, score, or later decision.
The value of this object is to hold decision and model together without confusing them. A region count can be a legal choice; the same count can be a model assumption.
07 / the check sheet
Return when a new decision appears
ROMSILVA / 27 AUG 2026LOCAL ARCHIVE
Legal text: HG 123/2026.
Audit: official Romsilva page and PDF presentation.
The local archive stores the PDF and official page with hashes. The complete model scores are not reconstructed from the published material.
Recheck after implementation or correction
Earlier if the government decision changes, the audit presentation is revised, or complete score tables appear. Keep legal status, model rank, and actual outcome separate.
A decision can say 19. A model can prefer 26. Chronology keeps them on one page.