Why do we add 1 here?

[jysullivan]

AlaskaHerring/2019_forecast/admb/HER_best_condEffort.12_322/her.tpl

Line 1343 in 349a0de

reck = mfexp(log_reck) + 1.0; // recruitment compensation ratio

[smartell]

the parameter log_reck is estimated in log space. But the parameter reck must be > than 1.0, and cannot equal 1. Therefore, if you add 1, then even if log_reck is a negative number, this ensures that the constraint reck > 1 is met.

[jysullivan]

Thanks, @smartell, your explanation makes perfect sense. An important point of clarification: if I'm using the estimated log_reck outside of the model to estimate brps, should I use the transformation? It obviously has a large impact on the results...

AlaskaHerring/2019_forecast/r/brps.R

Line 298 in 349a0de

reck <- exp(pars_s$log_reck) + 1 # FLAG - in her.tpl, he adds 1. Assuming we should do the same.

[smartell]

Yes use the transformations

…

On Aug 23, 2019, at 2:34 PM, Jane Sullivan ***@***.***> wrote: Thanks, @smartell <https://github.com/smartell>, your explanatio makes perfect sense. An important point of clarification: if I'm using the estimated log_reck outside of the model to estimate brps, should I use the transformation? It obviously has a large impact on the results... https://github.com/commfish/AlaskaHerring/blob/349a0deda1a557fd490358877f242e0ef309f557/2019_forecast/r/brps.R#L298 <https://github.com/commfish/AlaskaHerring/blob/349a0deda1a557fd490358877f242e0ef309f557/2019_forecast/r/brps.R#L298> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#15?email_source=notifications&email_token=AAJOESGRPU5JR2YFCDLYN73QGBJYDA5CNFSM4IPCJ6K2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5BMMHQ#issuecomment-524469790>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAJOESFFXVLDGGIK2QGD76LQGBJYDANCNFSM4IPCJ6KQ>.