MC-System: Idea to smooth out the 'Jerkyness'

[Q-wert]

Eyo. I read through the "Illarion skilling vs traditional rpg skilling"-thread and from what I observed the general complaint was that the gameplay feels like we have an immediate reward for 'not skilling'. Which does make skill related actions seem wrong for the people in the angry crowd who did not read or understand the code.

What do we want to encourage people to?
-To play the game, I figure.

The MC-System in its general idea is doing just that. The more you play, the more you gain skill. Simple as that. Only problem with the current implementation is, that learning speed responses noticeably to the actions a player does within one play session, which apparently does upset many people. So here an idea that should give a more consistent learning experience:

I: Calculate the learning rate by over all actions points and time spent in game

• [skill gain modifier]=([all action points spent on skill related actions]/[minutes spent ingame])

Should work fine once a consideable time was spent inagme. Once that time is spent rping, skilling or doing whatever ceases more and more to have any immediate effect noticeable for the player while one still would gain skill in relation to the time spent ingame.
Although, as pointed out by Estralis: Initially such a formula reacts 'very' jerky to any input given, due to the division by a small number. Which is very undesirable as a system should not take 30 hours of playing to work fine.

II: Smooth out the staring value with a decreasingly influential constant.

• [t] = time until the formula in I shows 'stable' behaviour
  [tp] = percentage of that time already passed (in 100 intervals)

Until [t] has passed:

• [skill gain modifier]=
  ([all action points spent on skill related actions]/[minutes spent ingame])*[tp]
  +[average player skill gain modifier at character creation or first log in]*1/[tp]

Save the difference between [average player skill gain modifier] and ([all action points spent on skill related actions]/[minutes spent ingame]) for the interval of each percentage and multiply it with the concurring tp.

• Sum that all up and name it [player credit].

Once [t] has passed, for another 2*[t]:

• [skill gain modifier]=
  ([all action points spent on skill related actions]/[minutes spent ingame])
  +([player credit]/100)/2

Once that is over, we can finally just use the formula of I.

I haven't run a simulation on this (am not really familiar with the control-stuff matter, to be honest), but as I am sick at home I might give it a try sometime next week.


Opinions, ideas, rotten tomatoes on the formula? (And please keep general ranting about mc to the respective threads.)

[Estralis Seborian]

That's generally the direction I also evaluated. My approach was a little more simple, though:

[skill gain modifier]=([Constant_A+all action points spent on skill related actions]/[Constant B+minutes spent ingame])

This means that you start out with

[skill gain modifier]=Constant_A/Constant_B

If Constant_A and Constant_B are large enough, you would not notice any jerkyness beyond the "normal". And in the long run, their influence would vanish by itself as you are approaching

[skill gain modifier]=([all action points spent on skill related actions]/[minutes spent ingame])

as desired. The transition phase would be a little tricky and also, finding good values for Constant_A and Constant_B is not trivial. I would basically tune them so that the current "jerkyness" of the MC system is the baseline.

The big downside of this approach is that players would virtually lose the option to "reset" their skill gain modifier. But perhaps this is just what is desired, as many struggle with understanding all this MC stuff instead of playing the game. Thanks a lot for the input! I will run a simulation anytime soon and report the result.

[Uhuru]

I believe someone asked this before... is there no way to shut off learning skills in certain categories? Just because my player has to fish to survive, doesn't mean she wants to continue to learn every time she does so.

Just wondering if there is any way to be selective. Or if this is completely out.

[Estralis Seborian]

If skilling up in a skill is not rewarding, we should make skilling up in a skill rewarding. Not introduce a bypass to "deactivate" learning that particular skill. My opinion on this.

[Uhuru]

Not sure it is about rewarding. Fishing and cooking are fine, but a character, playing a fighter, may prefer to RP being a bad fisherperson and cook, even though he has to fish and cook quite regularly. Just because a character can learn to fish to level 100 doesn't mean it is practical from an RP perspective or even desired. (sorry to pick on fishing.)

Someone asked about it, I was just trying to pursue it.

[Q-wert]

Deactivating learning for certain skills might be worth its own thread as its not really connected to the general formula or how to smooth the bouncing learning curve.

Personally I had a 'focus' on a skill group come to mind. Something that gives a very slight increase in skillgain (and MC-cost) on the focused group but decreases the same significantly for everything else. But hey, maybe in its own thread to keep things clean and easy?