Future Weariness improvements

[actual-nh]

Now that #47653 is merged (😀 🎉 🚀 ), as have #46906 and #45316, there are some future potential improvements in the Weariness system that I am thinking about, in approximate increasing order of programming plus testing time/effort involved:

1. Activity levels between NO_EXERCISE and LIGHT_EXERCISE should allow some recovery, particularly if part of the time is spent sleeping. (low_activity_ticks will need to become a float; Reduce weariness instability (includes tests); clarified #47653 did some setup for this.) EDIT: I am not sure how to allow for sleep time getting better recovery in this regard. Just checking at the 5-minute intervals whether the character is asleep is rather, hrm, erratic and prone to exploits. EDIT2: See below.
2. Weariness still fluctuates oddly at times (see [CR] Catch weariness fluctuations #47273). The weary intake level, added to when calories are absorbed, opposes the weary tracker, which is added to with exertion. Currently, the weary intake level fluctuates quite a bit, in a way that I have to say looks like hypoglycemia, with how fast the character's body is reacting to caloric absorption. (I admittedly may be biased on this; while my fasting blood glucose is a bit high at 102, my hemoglobin a1c indicates that I average 103 - as in with meals. I've also done every-other-day fasting without problems.) An exponentially decaying moving average (see Remove randomness from metabolism #46906 for some other discussion) should help with this. (This is assuming that caloric intake should be affecting weariness in the first place, which I have my doubts about.) This should also remove its current decay speed depending on WEARY_RECOVERY_MULT; currently, this results in a higher WEARY_RECOVERY_MULT speeding up decreases in both weary tracker and weary intake, and thus having opposing effects.
3. Mutations affecting the need for sleep should be allowed for in weary recovery rates, both while sleeping and while awake - see The Elfa mutate "Very little sleep and tireless" are not fit well with weariness #48232. If the character sleeps more efficiently, then they should have enhanced recovery while asleep. If they accumulate less fatigue, then they should be recovering more while awake.
4. As per the original Weariness proposal (Weariness: partial solution for too much calorie burn #42343), look into allowing LIGHT_EXERCISE (or, now, between LIGHT_EXERCISE and MODERATE_EXERCISE) to give a bit of weariness recovery, possibly dependent on other factors such as stamina level and various mutations/enchantments.
5. The above will need to be coordinated with needed metabolism (and weight gain/loss) improvements. For instance, it is desirable to unify mutation and enchantment effects on BMR. Currently, weariness uses Character::base_bmr() - which includes mutation effects via metabolic_rate_base() (in consumption.cpp) but not enchantment effects; the latter are done after multiplying for activity level in Character::get_bmr(). With regard to this - Ping: @Ramza13.

Ping: @anothersimulacrum, @I-am-Erk (and probably other people once I get a chance to do more git "blame" searches).

P.S. BTW, do pings actually work for discussions? I don't recall seeing notifications for any.

[robotic-gamer]

one problem with weariness increasing with calorie consumption is it means using the metabolic interchange increases weariness. awhile ago I found that the load from using it to power the Leukocyte Breeder System caused excessive weariness to the point I just turned the breeder system off.
perhaps either:
-weariness should only be based on stamina use
-some types of calorie consumption should not cause weariness since weariness is mainly muscle fatigue to my understanding

[actual-nh]

I believe that some alterations have been done with regard to the metabolic interchange and stamina/weariness, but I'll check on it.

[Ramza13]

I got pinged so they seem to work. I have a PR right now dealing with the enchantments around bmr that I'm not totally satisfied with so I might take a stab at some of that but the little I looked at scared me a bit with its complexity.

[actual-nh]

(Thanks for confirming the ping!) I'll be sure to take another look at that PR. EDIT: #47637?

[actual-nh]

I've not done as much as I would like on this due to work + other events, but I'm hoping to get a bit more done on it now, probably for 0.F-dev.

[I-am-Erk]

I like the ideas here particularly around recovering weariness slowly at lighter activity levels. I think this might be best set to happen faster the higher your weariness level, so if you're mildly weary, walking around and crafting won't undo it, but if you're dog tired you can relax to a lower level of work and gain at least a little juice back. Thinking here eg of running, then breaking to walk for a bit, then continuing.

I haven't checked on the state of things in a while, but also haven't seen as many complaints. Has the problem with having to sit perfectly still for long periods and not moving at all or losing your rest progress improved?

[actual-nh]

I like the ideas here particularly around recovering weariness slowly at lighter activity levels. I think this might be best set to happen faster the higher your weariness level, so if you're mildly weary, walking around and crafting won't undo it, but if you're dog tired you can relax to a lower level of work and gain at least a little juice back. Thinking here eg of running, then breaking to walk for a bit, then continuing.

Hmm... point.

I haven't checked on the state of things in a while, but also haven't seen as many complaints. Has the problem with having to sit perfectly still for long periods and not moving at all or losing your rest progress improved?

On the one hand, activity level is now tracked per turn (#45316), so fixing this is more possible than it was. On the other hand, weariness is still currently only decreased (every 5 minutes) if your total activity level for the past 5 minutes is (in essence) NO_ACTIVITY; see the first idea above and https://github.com/CleverRaven/Cataclysm-DDA/blob/master/src/activity_tracker.cpp.

[actual-nh]

Ah. I think I see how to handle sleep in a more continuous fashion: add sleep as an activity level, with SLEEP_ACTIVITY equalling something like 0.5f. (Note to self: Will require activity_tracker::new_turn and activity_tracker::activity to take a boolean is_sleeping argument. In character::set_activity_level, if input is NO_ACTIVITY, convert to SLEEP_ACTIVITY if in_sleep_state.) Will have to check metabolism studies to see what the (mean!) sleeping rate of caloric expenditure actually is (should be lower than the "no activity" level!), and use it for SLEEP_ACTIVITY.

EDIT: Hrm. Since current caloric expenditure values are using 24 hours at NO_ACTIVITY consuming BMR calories, may need to adjust NO_ACTIVITY to above 1 to account for sleep hours being less... with the difficulty of mutations enabling less sleep altering the number of sleep hours from 8 per 24. Alternatively, use max( activity_history.average_activity(), NO_ACTIVITY ) for character::get_bmr(). EDIT2: And for character::update_stomach() for log_activity_level?