explore effect of Bmark_year ranges on variance of derived quantities

[Rick-Methot-NOAA]

While spending hours trying to find a code bug I stumbled into realizing something that I once knew and had forgotten. The problem I saw was a change in the variance in some derived quantities. In particular, I noticed that the variance of:
#_Mgmt_Quantity
SSB_unfished 49909.7 1353.54
was less than the variance of:
#_Spawn_Bio
SSB_Virgin 49909.7 1384.54
SSB_Initial 49909.7 1384.54

and several other variances also were different because they depend on SSB_unfished.

I tracked this down to a change in Bmark_years in my input file. If Bmark_years select just the initial equilibrium year, then variance of the two quantities are the same, but if a range of years are selected, for say biology, then the variance is reduced because the variance of a mean is always less than the variance of an individual year. This will happen even if biology is not time-varying. It is just basic stat 101.
Users have been routinely selecting Bmark_years without having any awareness of this effect.
I think we should create a testbed in which we demonstrate and document the effect of changing Bmark_years (and Fcast_years).

[k-doering-NOAA]

Thanks, Rick, this is helpful to know! Would the goal of the testbed here be to inform users about the potential differences in uncertainty between derived quantities? Or some other goal?

[Rick-Methot-NOAA]

Step one is to inform us!
Then inform users.
Then provide guidance on good practices or to insert warning. For example, SS3 could determine that it is illogical to select a range of years for biology if none of the biology parameters are time-varying.
Test would look like this:
a. create three versions of a model that estimates growth: one with no time-varying parameters, one with growth time-varying, one with selectivity time-varying.
b. run each model with bmark_yrs all set to just start year
c. do runs with varying combinations of ranges for Bmark_years
d. similarly for Fcast_yrs, but here only selectivity years and relF years needs to be evaluated

[k-doering-NOAA]

Makes sense - do you mind if I open this as an issue in the ss-documentation repo?

[Rick-Methot-NOAA]

Go for it

…

On Wed, Dec 1, 2021 at 12:49 PM Kathryn Doering ***@***.***> wrote: Makes sense - do you mind if I open this as an issue in the ss-documentation repo? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/orgs/nmfs-stock-synthesis/teams/developers/discussions/10/comments/3>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABPV4IBIHWJQSJZVKHVBAGDUO2C3RANCNFSM5JCO5E3Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.