(Discussion) Character Points Change Issue

[orsal]

No, your missing the elegance of a simpler program design. You don't need to run 6^24 binary comparisons to check each total, you need to run 15^6 by simplifying the algorithm.

OK, agreed, although to nitpick that should be 16^6. (Fenceposts again.)

And taking just 1 ability scores average points value for all results will not yield you the results for two reasons. 1) You can't take criteria like "one ability must be greater than 13" into consideration,

That was exactly what I was referring to above when I wrote

If you get into a "reroll the whole character if you don't get something at least this good overall" formula, it becomes more complicated, because you can't consider each of the six attributes in isolation anymore.

My spreadsheet as it stands does not take that rule into effect. I could fix it to throw out characters with no score at least 14 (takes a little thinking, if you want I'll explain how and why it works, but mathematics is already boring most of our readers, and this would only make it worse), but I don't think I could fix it to require a net +1 or higher. However, if scores are rolled with 4d6, and the character is guaranteed at least one +2, it is very unlikely that the sum of the ability modifiers will be 0 or less, so that won't alter the average very much.

and 2) Since point buy is non-differential, you can't break it down into a linear equation that breaks up evenly, so since 18s and 17s are worth more than 14s and 9s, you get a total which is related to the most common values 12-14 that will dominate the single instance, but not all instances that exist for 6 arrays.

That doesn't matter. So long as I compute the average of the point buy counts, rather than the ability scores, of the individual abilities, I'll get the average of the sum of the point buy counts when I multiply by six. The reason is this: if you have any number of random variables (let's say six, call them S, D, C, I, W, X), the expected value (i.e. mean average) of (S+D+C+I+W+X) is the sum of the individual expected values. That's a theorem from probability theory.

 

[Creamsteak]

Heh... I've never been one to listen when the naysayers say nay to calculating stuff...

That doesn't matter. So long as I compute the average of the point buy counts, rather than the ability scores, of the individual abilities, I'll get the average of the sum of the point buy counts when I multiply by six. The reason is this: if you have any number of random variables (let's say six, call them S, D, C, I, W, X), the expected value (i.e. mean average) of (S+D+C+I+W+X) is the sum of the individual expected values. That's a theorem from probability theory.

Yeah, sorry I didn't say that right. I guess I meant to say that since point buy jumps between values at different rates, you can't use an equation instead of comparisons to get the point buy value for a number. Either way it doesn't make any reasonable difference in cpu time I guess.

Those are easy to do with a program though, all you need are the comparisons (reduce them to single binary comparisons though, I originally used a series of binary comps and it more than tripled the time to calculate). I've got the efficiency down to 5 minutes or less on my pc, with an error margin of about 0.00000001%, which is pretty close, and it calculates the number to be very close to 30.46 when using the 1 number greater than 13 and total ability modifier of +1 or better (and calculates that they comprise about 86% of all rolls or something like that IIRC).

 

[orsal]

Yeah, sorry I didn't say that right. I guess I meant to say that since point buy jumps between values at different rates, you can't use an equation instead of comparisons to get the point buy value for a number.

If the ability score is n then the point count is either

(n-8) + min(n-14,0) + min(n-16,0)

or

min(n-8,0) + min(n-14,0) + min(n-16,0)

The former counts scores below eight for negative points; the latter, for zero points. Both are easy to do in any spreadsheet program.

 

[Creamsteak]

And easy to do using any computer programming language. Except in a language, you can choose whether to make points a data-type returned by a method, or you can store it in a parallel array. Both easy, and use less time than mins and maxes on excel (though it's not like anybody would notice).

 

[Kahuna Burger]

Not to interupt to dance of compeditive geek-fu....

As a DM I like to see durable characters, and as a player, I like somewhat specialized characters. A 32 or 33 point buy system allows a specialized character with some tradeoff, while still having the minimums in certain survival stats that you need for a 1st level start. I wouldn't go a lot higher, since I like to see some stats as average or below, but I'm concerned that at lower pointbuys (and again, esp a 1st level start) you end up with somewhat homogeounous characters who have the "minimum to survive" in a few stats and no real areas to shine and say "this is what I'm one of the best at."

Thats why I support a slightly higher point buy than we have now, but I fear not the slippery slope argument that it can only get better with more. If everyone is "one of the best" at two or more things, its just homogenous characters at a higher level, and I'm not fond of that either.

Kahuna Burger

 

[Creamsteak]

I don't think this is arguing a slippery slope simply by the definition... because I think we ALL know that there are relevant disimilarities between a 28 and a 32 point buy, and a 32 and a 36, and a 36 and a 42. Perhaps one point doesn't create a big gap at any place, but the increase in power grows rapidly.

So, I'll say again that my vote is for 32 or 31. If the majority of support is for 33, I can deal with a point difference.

 

[Kahuna Burger]

I don't think this is arguing a slippery slope simply by the definition... because I think we ALL know that there are relevant disimilarities between a 28 and a 32 point buy, and a 32 and a 36, and a 36 and a 42. Perhaps one point doesn't create a big gap at any place, but the increase in power grows rapidly.

I was thinking more of the comment

That same reasoning would support raising the count from 33 to 36, and then to 39. Will it ever stop?

I know that the response applied to a specific point of reasoning, but I just wanted to make clear that I don't agree with that assessment. Being able to reach a certain minimum in your non concept stats (to survive a single hit at first level or to get a feat for your character concept) while still being noticable specialized in the stats that you feel define your character can be accomplished (IMHO) much better at 33 than 30, but going to "36 then 39" isn't justified by the same logic. At that point you've covered the minimums and the standouts and are just going for power creep.

For instance, when I designed Sara, the gladiator, I wanted her to be tough and strong to deal damage, but wasn't so interested in her being very graceful. However, my mental concept of her was dual wirlding shortswords, purely for style points. (she has the doubletrouble thing going and eventually wants matching frost and flame enchantments). With the dex requirement for TWF, needing at least a 10 int for the three skills she was to be specialized in and a merely positive Cha to pretend to be a performer, she turned out to be pretty good at several things, but not specialized in the way I would have prefered. I don't need a 39 point buy so that she can be the strongest and toughest in the land and also far above average in her speed and charm, but I would like to meet a prereq for a useful feat without having to redesign my character around it. (heavens forbid I had wanted her to be good at tripping and disarming for the amusement value of that fighting style. She'd be a rogue with no con by the time I was done...)

Anyway, the power creep argument had been raised, and I alluded to my rejection of it. Not saying everyone who objected was arguing slippery slope, merely that some seemed to be.

Kahuna Burger

 

[Creamsteak]

Unstuck at 25 votes. My opinion of it is that this is still a bit inconclusive. We will have a proposal vote down when we start tackling the major issues like ECL races and item crafting.

 

[Guilt Puppy]

Looks like I missed the vote, but I'm more interested in arguing, anyway.

- For one, while the statistical argument is interesting from a math geek stand-point, I don't know if it's really applicable. While your average 4d6 character is worth between 31-32 points in point-buy, a 31-32 point character is still going to be better (than the average), at least; point buy allows more customization, meaning more optimization... In other words, you get more bang for your buck. Of course, how many points to you knock off to even it out? It's a tough call.

- Going to 33 to "allow" people to have odd stats is, IMO, backwards. If someone wants odd stats (for that "I got a cookie" feeling at 4th/8th et cetera, which I can appreciate as much as anyone), then by all means, they should make some odd stats. You can do that with 30, you know. Really, the difference is: With 30 points, you'll create a character with an even number of odd stats; with 33, you'll create a character with an odd number of odd stats. And with 33 the character will be more powerful.

- Low stats are good. In my experience, a character with no stat lower than 14 is just as boring to play as a character with no stat higher than 14. (Actually, those low-ball characters can be fun too -- trying to be a utility player, making productive outs and all that, can be pretty satisfying.)

- I'm currently guilty (yay!) of playing one of the most min-maxed characters in ENWorld (see sig)... That was as much in order to preserve the low stats as it was to load up the high stats. I would have been quite happy with 16's or 17's where the 18's are now -- but my plans to make a grumbling, bumbling, foul-mouthed dwarf left me with too many points to play with.

- Re: Survival stats, I've played plenty of penalty-to-con characters from first level, and the vast majority of them lived to see the light of higher hit dice. Generally, they've outlived the gutsy dwarf barbarians who'd charge right into anything thinking their 17 hit points made them invincible.

- I'd also like to add that "I want to make my character more interesting" isn't really a better argument than "I'd like to make my character more powerful." If every character is really cool in a lot of ways right from the beginning, what do they develop into? Instead, why not start off with that charismatic, dual-wielding gladiator, except maybe she isn't so rough-and-tough yet; maybe she's downright weak to begin with. But she works hard, eats right, exercises daily, and most importantly picks up a few magic items, and suddenly she's earning her style points.

In short: I'd vote to reduce the starting point-buy to 28 or so, long before I'd even raise it to 31.

 

[Quirhid]

In short: I'd vote to reduce the starting point-buy to 28 or so, long before I'd even raise it to 31.

I'm with GP! Atleast don't raise it.