Stat Tests

[MishkaL1138]

So, if I get it straight…

Nowadays, RPer's do fights by rolling /roll <stat_lvl> (for example, to dodge a hit, /roll AGI). But, since people can lie with that, you propose the system does it for you? With player-set stats? (I don't use maxed stats, even thought I do have them maxed with Mish xP)

Well, good luck, you got a dev interested. I think this wish can be granted Soon™.

[Sangwa]

Normal 'RPer fights' should be done by using /challenge or whatever the command is. Though some players enjoy scenic fights, where they can use whatever system they prefer.

The purpose here is not just for fights, but for different types of actions. For instance (since pickpocketing isn't implemented) one player could test his agility against the other's agility [or will, depending on what you consider to involve that type of perception]. In the system Rlydontknow speaks of, you'd take roll turns until one wins and the other loses. And yes, these rolls would accurately involve the stats the player has, therefore leaving no space to lies.

[Sarras Volcae]

Everything is based on luck Sarras. Two Arm Wrestlers in the same level can go at it and witness different results each time they arm wrestle. Like in any other game.

doesn't this system allow the weaker opponent to win, though, even against a much stronger opponent? it's like a child having the possibility of beating an adult powerlifter. doesn't seem very realistic.

[Sangwa]

Not in the system I mentioned with opposite tests. If you make luck just 25% of the max value, a maxed character will win most of the times, unless he is paired against someone around his level.

To illustrate:
Let's say we have a 100 str character and a 400 character. 400 character will always win if we consider that 100 is the luck throw (100/400 = 0.2 = 20% of max value is considered luck)

Stat + Roll  =   Sum
100 + 100    = 200
400 + 1        = 401

Only a character with 302 or more could win a character with 400 and then it would be pretty hard. Considering dice rolls are independent (i.e. one doesn't influence the other, which is what happen in PS) this makes a probability of 0.01*0.01=0.001=0.1%. Very hard to get:

302 + 100 = 402
400 + 1 = 401

This also means luck has more weight for weaker, less trained characters. Which makes sense. If you're not as trained, this means that luck has more influence in your results. Either way, if it's done like Rlydontknow or Sarva says, you can even adapt to the situation. You can decide the amount of luck involved with the "cap" parameter, or by choosing the roll.

[RlyDontKnow]

Well, good luck, you got a dev interested. I think this wish can be granted Soon™.

heh, indeed should be expected "soon™"
got the c++ part mostly done, however I'll have to extend the mathscript functionality a bit to get it working properly (as we just decided to make the actual test script customizable by the server maintainer)

anyway, there'll just be the possibility to test yourself for now, so for the arm wrestling I think each player could test itself against a certain difficulty or something like that, not sure. maybe there'll be testing others later, but it's a lot more complicated than just testing yourself :x

[Sangwa]

Yup, both players test themselves at the same difficulty and with the same cap, until one of them is successful and the other isn't. It's pretty good!

[RlyDontKnow]

Yup, both players test themselves at the same difficulty and with the same cap, until one of them is successful and the other isn't. It's pretty good!

or just one wins by a bigger margin (will go for the >+10%, >+0%, =0%, >-10%, <-10% for now)

[Sangwa]

Yup, if the throw results are shown. That would fail in keeping the ranks secret, but that's a minor thing.

[RlyDontKnow]

Yup, if the throw results are shown. That would fail in keeping the ranks secret, but that's a minor thing.

not the exact results, but some lingual representation of them (like xyz almost did it instead of "had 95%" or whatever)

[Talad]

I like it, I think we should implement this.

I was thinking of two cases.

1) One person (potentially the organizer of the event) can ask for a test, for example "check on strength against 150 value". The player will then issue a command like /test strength 150 . The server will roll a dice on strength and see if the player manages to roll equal or higher than the value specified (this means than the higher his strength is the easier the roll will be for him). The resulting message will be: "Rly tested his strength against 150 and succedeed".

2) Test your stat against another player. In this case the target will receive a confirmation message if he wants to perform the test. He can refuse to do so. The command will be "/test strength Talad", this will ask the server to roll two dices, one for you and one for Talad and compare the results. The higher wins (again also in this case the one with higher strength has more chances to win).

[verden]

I proposed something close to this in tracker report 2767 or 2768 quite a while ago. Stats contest as well as a saving throw command for testing against skills. I believe I was accused of wanting to implement a completely parallel and redundant combat system. But commands to enable roleplay are needed on the system.