Determining abilities for cohorts?

[howandwhy99]

As a DM it's 3d6 in order. Of course, special cases may apply and modify those rolls as well as all the racial and potential class abilities.

If I am playing, well, I guess I post a notice or something. Then do interviews.



EDIT:
What makes cohorts decidedly different from anyone else who might travel with you is they are basically journeymen apprentices. You can train more than one at a time, but you make a separate commitment to each. In that they are sort of like wives (or husbands) if you like.

 

[Dwimmerlied]

Okay, I have now checked the MM and DMG, which has cleared up a few things.

Firstly, on the topic of ability scores, the MM recommends that NPCs built with "PC classes" (which would generally include cohorts) should probably use the elite array* - 15/14/13/12/10/8 (or 4d6-drop-lowest). And, per the DMG on NPCs**, such 'elite' characters would receive max hit points for their first die.

On the topic of gear, the MM says to determine the creatures ECL*** (that is, level plus level adjustment, if any; if it's a monster with monster hit dice, add those too), and use that as the "level" for the table on p.127.

* MM p.290
** DMG p.110
*** MM p.291

Fantastic! Thanks heaps!

Seems I've been doing some stuff wrong...

I think that's the way it often goes! I find lots of situations like this where I'm sure (or kinda sure) I've read the guidelines for this or that before, but finding where can be tricky because sometimes the rules that you are need are recorded in disparate sources. Those page numbers are going on my cheat sheet. Once again thanks everyone for the help!

 

[Dwimmerlied]

If I am playing, well, I guess I post a notice or something. Then do interviews.

In fact, I suppose RAW (if a little metagamey), you could interview potential cohorts by submitting them to fixed attribute tests such as arm wrestles. You can calibrate against a character of known strength because with these tests higher ability scores automatically win without roles