10 réponses à ce sujet

[sergio1992]

I wonder, rolling 100 with a paladin is equivalent to rolling a 98 with a kensai?

Does the game roll the dices depending on the class you are currently using? 

Thanks for help

[ussnorway]

The class does effect the roll, yes.

[Grond0]

The same dice rolls are made in all cases, but there are often minimum scores applied for race and class. This is a particular advantage for a paladin as charisma has a minimum of 17, i.e. the only legal scores are 17 or 18. This does make getting a high total for a paladin a lot easier, but a total of 100 is still exceptionally good!

[sergio1992]

Thanks to both. Is there a way to check how the dice are rolled?

Modifié par sergio1992, 21 janvier 2013 - 11:12 .

[AnonymousHero]

If you mean seeing the actual dice rolled, then no. Someone did a pretty exhaustive analysis using an autoroller a few years ago, but I think that post was lost to the interwebs when the old forums were closed.

I believe the following summarizes the rolling mechanics as they were found to be:

- The game rolls 3d6 for each stat, individually.
- Any stats which are lower than class/race minimum are increased to the class/race minimum. (As Grond0 mentioned.)
- If the sum of all the stats is <75, start over.

I'm not sure of the order of those two last steps.

[Grond0]

I think the game actually rolls 4d6 and then takes the best 3 of them. Along with minimum score requirements that explains why the average scores achieved are rather higher than the 10.5 expected of 3d6.

[AnonymousHero]

You may be right (hazy memory and all), but the 75 minimum explains the higher-than-3d6-should-be, I think.

[Grond0]

I agree that explains the total, but if it were only the impact of the total you would still expect a higher number of low individual characteristic scores than I think are actually obtained - a roll of 3 for instance is incredibly rare and you'll probably have to re-roll quite a few times just to see one below 7.

[AnonymousHero]

Grond0 wrote...

I agree that explains the total, but if it were only the impact of the total you would still expect a higher number of low individual characteristic scores than I think are actually obtained - a roll of 3 for instance is incredibly rare and you'll probably have to re-roll quite a few times just to see one below 7.

The rarity of extreme rolls is because 3d6 has that type of distribution which is heavily clumped around the average, i.e. you're must more likely to get a value close to the average than getting a value at the extremes. (It's basically a multinomial distribution AFAIUI, but the minimum cutoff changes the distribution somewhat.)

That 3 is rarer than 18 is again explained by the "total points must exceed N" cutoff. Very low scores increase the likelihood of hitting the cutoff and so should be observed less often than extremely high individual scores. (Since the game simply hides those results from you.)

Btw, the fact that one does see rolls of 3 from time to time seems to indicate that the game can't be rolling 4dX .

EDIT: Just a quick follow-up: Frabjous' original post is gone, but Camdawg's executive summary is probably accurate: http://forums.gibber...showtopic=12703

In short: We were both wrong on the minimum individual stat thing -- the game rerolls any stat that is below the race/class minimum. (Which further biases results away from very low stats.)

Modifié par AnonymousHero, 22 janvier 2013 - 06:13 .

[sergio1992]

Thanks to everyone.

[Ecoris]

I haven't posted for some time, but recently dusted of the game.
It's as described in the link posted by AnonymousHero. I remember doing the probability calculations to confirm frabjous' numbers. I think I still have his reroller somewhere.

Modifié par Ecoris, 16 février 2013 - 11:24 .