CMP registry

Candidate Management Procedure Registry

1 Meaning of the status labels

Status	Meaning
Test candidate	Defined sufficiently for testing; not shortlisted or adopted.
Comparison case	Included to isolate an important assumption or design choice; not an advanced or adopted procedure.
Sensitivity case	Retained to test an alternative specification; not the focal case or an adopted procedure.
Focal case	Principal case in the current analytical comparison; not by itself a recommendation or adoption decision.
Advanced candidate	Selected in the current analysis for robustness testing; formal status not yet confirmed.
Shortlisted	Retained by the working group for formal comparative evaluation.
Recommended	Supported by a documented working-group recommendation.
Adopted	Formally adopted through the applicable SPRFMO process.
Retired	Kept in the historical record but no longer under active consideration.

The SC14 release field distinguishes the eight procedures in the named SC14 package from earlier comparison cases. HS-20 (MP43) and PR-20 (MP44) are the focal cases for the current analysis. HS+20 (MP29) and PR+20 (MP32) are sensitivity cases. HSsym, PRsym, HS-30, and PR-30 complete the common 500-draw comparison. The focal designation records the current analytical emphasis; it is not a formal recommendation or adoption decision.

2 Working-label convention

The label column uses the concise convention from the SC14 MSE report when a CMP belongs to the retuned annual-change comparison. HS means hockey-stick and PR means power ramp. +20 denotes the base 20% maximum increase, -20 the reverse asymmetry with a 20% maximum decrease, sym symmetric 15% limits, and -30 a 30% maximum decrease. The immutable cmp_id column retains the legacy identifier used by code and saved results.

CMPs outside this convention keep their existing MPxx label. Thus the registry and downstream exports do not invent a descriptive name where no agreed convention exists.

Table 1: The eight CMPs in the named SC14 release.

Legacy ID	Working label	Role	Rule	Minimum	Limit	Trigger	Target	Decrease multiplier	Increase multiplier
MP29	HS+20	sensitivity case	hockeystick.hcr	270	0.1	2.125000	2000	0.85	1.20
MP43	HS-20	focal case	hockeystick.hcr	270	0.1	1.937500	2000	0.80	1.15
MP45	HSsym	comparison case	hockeystick.hcr	270	0.1	2.062500	2000	0.85	1.15
MP47	HS-30	comparison case	hockeystick.hcr	270	0.1	2.000000	2000	0.70	1.20
MP32	PR+20	sensitivity case	powerramp.hcr	270	0.1	1.640625	1500	0.85	1.20
MP44	PR-20	focal case	powerramp.hcr	270	0.1	1.601562	1500	0.80	1.15
MP46	PRsym	comparison case	powerramp.hcr	270	0.1	1.625000	1500	0.85	1.15
MP48	PR-30	comparison case	powerramp.hcr	270	0.1	1.578125	1500	0.70	1.20

The complete machine-readable registry retains six earlier comparison cases so their historical identifiers and settings are not lost. They are marked sc14_release = no and are not part of the current SC14 result set, Slick file, figures, scorecard, or annual-use example.

3 Candidate pages

MP43: focal all-index hockeystick case (HS-20).
MP44: focal all-index power-ramp case (PR-20).
MP29: earlier +20 hockeystick sensitivity specification.
MP32: earlier +20 power-ramp sensitivity specification.
Trial and robustness run reference: the full set of tuning trials and the operating models used to test MP29 and MP32.

The current Slick comparison contains exactly the eight rows in the table above: four hockey-stick and four power-ramp procedures, each evaluated using the same 500 draws.

4 How these candidates were developed

The jmMSE analysis tested alternative target catches, minimum catches, limit values, harvest control rules, and ways of calculating the stock indicator.

Run	Role in development	July WUR presentation interpretation
MP16	Hockeystick, target 2000, no minimum/limit	Potential candidate before minimum/limit refinement
MP25	Powerramp, target 1500, no minimum/limit	Potential candidate before minimum/limit refinement
MP29	Hockeystick with minimum 270 and limit 0.1	Earlier +20 design; retained as a sensitivity case
MP32	Powerramp with minimum 270 and limit 0.1	Earlier +20 design; retained as a sensitivity case
MP35-40	CPUE-only, shortcut-assessment, and perfect-information variants	Tests of how the source of stock information affects results
MP41-42	500-replicate versions of the MP29/MP32 designs	Test of whether the number of simulation replicates affects results; not separate CMPs
MP43-44	Retuned MP29/MP32 structures with a 20% maximum decrease and 15% maximum increase	Focal HS-20 and PR-20 cases for the current analysis

The earlier SCW17/Paper-09 rows are retained in source history as workshop test proposals. They are not the primary candidate list for this site.

5 Points to keep in mind

min, target, and initial are catch quantities in the current code and presentation. lim and trigger are values on the relative-index scale. Final catch units must be stated before adoption.
dlow and dupp limit the decrease and increase in catch advice from one year to the next.
Index names match model_tune.R exactly.
Each trigger was tuned so that the probability of being in the Kobe green zone (SB0green: spawning biomass above $SB_{MSY}$ and fishing mortality below $F_{MSY}$) was 0.60 over 2041-2050 in the reference analysis. The resulting trigger is part of the candidate definition and should not be recalculated during annual application.
In the completed 500-draw refinement, MP43 reached 0.592 at trigger 1.9375 and MP44 reached 0.598 at trigger 1.6015625, within the tuning tolerance of 0.01.

6 Requirements before changing a candidate’s status

A candidate can move beyond test status only when it has a completed specification, repeatable performance results, a calculation test using saved inputs, and a decision record identifying the approving body and date.

--- title: "CMP registry" --- {{< include ../_includes/sprfmo-banner.qmd >}} <div class="paper-title">Candidate Management Procedure Registry</div> ## Meaning of the status labels | Status | Meaning | |---|---| | Test candidate | Defined sufficiently for testing; not shortlisted or adopted. | | Comparison case | Included to isolate an important assumption or design choice; not an advanced or adopted procedure. | | Sensitivity case | Retained to test an alternative specification; not the focal case or an adopted procedure. | | Focal case | Principal case in the current analytical comparison; not by itself a recommendation or adoption decision. | | Advanced candidate | Selected in the current analysis for robustness testing; formal status not yet confirmed. | | Shortlisted | Retained by the working group for formal comparative evaluation. | | Recommended | Supported by a documented working-group recommendation. | | Adopted | Formally adopted through the applicable SPRFMO process. | | Retired | Kept in the historical record but no longer under active consideration. | The `SC14 release` field distinguishes the eight procedures in the named SC14 package from earlier comparison cases. **HS-20 (MP43) and PR-20 (MP44) are the focal cases for the current analysis. HS+20 (MP29) and PR+20 (MP32) are sensitivity cases.** HSsym, PRsym, HS-30, and PR-30 complete the common 500-draw comparison. The focal designation records the current analytical emphasis; it is not a formal recommendation or adoption decision. ## Working-label convention The `label` column uses the concise convention from the SC14 MSE report when a CMP belongs to the retuned annual-change comparison. `HS` means hockey-stick and `PR` means power ramp. `+20` denotes the base 20% maximum increase, `-20` the reverse asymmetry with a 20% maximum decrease, `sym` symmetric 15% limits, and `-30` a 30% maximum decrease. The immutable `cmp_id` column retains the legacy identifier used by code and saved results. CMPs outside this convention keep their existing `MPxx` label. Thus the registry and downstream exports do not invent a descriptive name where no agreed convention exists. ```{r} #| label: tbl-cmp-registry #| tbl-cap: "The eight CMPs in the named SC14 release." #| echo: false registry <- read.csv("../data/cmp-registry.csv", check.names = FALSE) sc14_registry <- registry[tolower(registry$sc14_release) == "yes", ] display_registry <- sc14_registry[c( "cmp_id", "label", "status", "hcr", "min", "lim", "trigger", "target", "dlow", "dupp" )] names(display_registry) <- c( "Legacy ID", "Working label", "Role", "Rule", "Minimum", "Limit", "Trigger", "Target", "Decrease multiplier", "Increase multiplier" ) knitr::kable(display_registry) ``` The complete [machine-readable registry](../data/cmp-registry.csv) retains six earlier comparison cases so their historical identifiers and settings are not lost. They are marked `sc14_release = no` and are not part of the current SC14 result set, Slick file, figures, scorecard, or annual-use example. ## Candidate pages - [MP43](mp43.qmd): focal all-index hockeystick case (HS-20). - [MP44](mp44.qmd): focal all-index power-ramp case (PR-20). - [MP29](mp29.qmd): earlier +20 hockeystick sensitivity specification. - [MP32](mp32.qmd): earlier +20 power-ramp sensitivity specification. - [Trial and robustness run reference](control-reference.qmd): the full set of tuning trials and the operating models used to test MP29 and MP32. The current Slick comparison contains exactly the eight rows in the table above: four hockey-stick and four power-ramp procedures, each evaluated using the same 500 draws. ## How these candidates were developed The `jmMSE` analysis tested alternative target catches, minimum catches, limit values, harvest control rules, and ways of calculating the stock indicator. | Run | Role in development | July WUR presentation interpretation | |---|---|---| | MP16 | Hockeystick, target 2000, no minimum/limit | Potential candidate before minimum/limit refinement | | MP25 | Powerramp, target 1500, no minimum/limit | Potential candidate before minimum/limit refinement | | MP29 | Hockeystick with minimum 270 and limit 0.1 | Earlier +20 design; retained as a sensitivity case | | MP32 | Powerramp with minimum 270 and limit 0.1 | Earlier +20 design; retained as a sensitivity case | | MP35-40 | CPUE-only, shortcut-assessment, and perfect-information variants | Tests of how the source of stock information affects results | | MP41-42 | 500-replicate versions of the MP29/MP32 designs | Test of whether the number of simulation replicates affects results; not separate CMPs | | MP43-44 | Retuned MP29/MP32 structures with a 20% maximum decrease and 15% maximum increase | Focal HS-20 and PR-20 cases for the current analysis | The earlier SCW17/Paper-09 rows are retained in source history as workshop test proposals. They are not the primary candidate list for this site. ## Points to keep in mind - `min`, `target`, and `initial` are catch quantities in the current code and presentation. `lim` and `trigger` are values on the relative-index scale. Final catch units must be stated before adoption. - `dlow` and `dupp` limit the decrease and increase in catch advice from one year to the next. - Index names match `model_tune.R` exactly. - Each trigger was tuned so that the probability of being in the Kobe green zone (`SB0green`: spawning biomass above $SB_{MSY}$ and fishing mortality below $F_{MSY}$) was 0.60 over 2041-2050 in the reference analysis. The resulting trigger is part of the candidate definition and should not be recalculated during annual application. - In the completed 500-draw refinement, MP43 reached 0.592 at trigger 1.9375 and MP44 reached 0.598 at trigger 1.6015625, within the tuning tolerance of 0.01. ## Requirements before changing a candidate's status A candidate can move beyond test status only when it has a completed [specification](specification-template.qmd), repeatable performance results, a calculation test using saved inputs, and a decision record identifying the approving body and date.