Re: [Crawl-ref-discuss] New skill training

[David P. <dp...@mi...>]

On Tue, 5 Jul 2011, Raphael Langella wrote:

> There's still a lot of things to address. First is obviously testing and
> bugfixing.


Thank you for the new skill system!

> Manuals: giving an instant skill bonus is no good. It would be a no-brainer
> to use any you find unless you really don't need the skill and are aware of
> the details of the skill system, acquirement and other spoilery stuff like
> that.
> Instead, they could provide a passive bonus which doubles the amount of
> skill points gained. It's charged from the beginning with an amount of skill
> points. So no matter if you train fast or slow, now or later, in the end
> you'll get the same bonus anyway. Just activate it, and as long as you're
> carrying it with you, each time your skills is trained, you get double the
> skill points.

Isn't this still a no-brainer to use? In order to make it a genuine 
choice, there should be some drawback to go with the bonus. I.e. manuals 
should reduce experience otherwise channeled into other skills and amplify 
the target skill. So you pay a bit (reduced training for skills different 
from the manual's) and you gain something (skill bonus for the manual's 
skill). Then you have to weigh those two against each other.

Another option is to scrap manuals.

Still another option is to give manuals a new role: they could allow 
players to get a head start into a new skill, by transferring a fixed 
amount of xp into the skill, but asking for another skill to pay with. In 
a sense, they'd be a poor man's Reskilling.

> Sage: same as manual, but all the XP goes into the skill. Commands are
> disabled in the skill menu and the sage skill is at 100%.

This is the radical version of what I propose above. (The way you 
explained it, manual use has a bonus but no malus.)

> Draining: Eronarn's idea: drain your skills but temporarily. Allows us to
> have a stronger tactical effect without permanently screwing the player. I
> think it shouldn't drain all your skills. Something like 1d3 skills maybe.
> They are regenerated only when trained not by waiting. They train twice as
> fast to catch up.

It's a good idea. It should drain from the top skills, I think. (Or use 
the actual skills as weights: with Axes 5, Shield 2, Armour 3, you would 
have 50% axes drained etc.)

> Implementation should be easy. We just have to put it in you.skill(), so
> stuff like HP from fighting, spell slots, acquirement are not affected. Just
> have to decide on the numbers. How many skill points to drain? We probably
> want to start by converting the XP drained into skill points, but then maybe
> we need to multiply it by some factor.

Yes.

> Also, we need a cap. We don't want to leave the player crippled without 
> any skill, or he wouldn't be able to regenerate them. Don't drain more 
> than 1/3 skill level?

Perhaps we can afford 1/2. If you use the weighted system I propose above, 
then for subsequent draining the current, already drained skills should be 
used. It can also take into account how much drainage is left for a skill 
(if it reached the cap, it wouldn't work as a weight anymore).

If there is desire, I can come up with the formula. The basic question is 
whether skill draining should be reliable (deterministic) or chaotic 
(random).

> Now, with an increase tactical effect, we can also reduce or remove the
> permanent one to compensate. I suggest the XP drained can be regenerated
> just like the skill points, so we have a consistent system with skills, XP
> and monsters (see below). Or only half can be regenerated, the other half is
> still permanent loss.

I am fine with full regeneration.

> While we're at it, we can do the same for monsters. Save the base HD and HP
> of the monster, use it for XP calculation (I think we need to implement that
> for zombies anyway). The draining brand decrease HD and HP, but only
> temporarily. If you flee, the monsters will have fully regenerated when you
> come back.

This is probably a nerf to the draining ego, but it can be very 
interesting.

Cheers,
David