I'm trying to generate a range of 1-7 for attributes for a game.
What would be better, rolling 1d6+1 (which skews upwards a bit) or 2d4-1 (which bell curves around an average)?
This is a scifi game with average people as protagonists, so I generally want average attribute stats for PCs, but nobody likes getting gimped PCs with 1s in their stats either.
>>45882404
How much the stats matter - that is, is there a significant difference between 4 and 5?
Most games with "average people" PCs generally tend to skew things up just a tad to avoid gimpy characters. They're still well within average, just the upper end of it.
2d6, divide by 2, round up
2d4-1 actually produces 1-7, which d6+1 produces 2-7. Getting a 1 only has a 6.25% chance of happening with 2d4-1. If you want things skewed a bit higher, there's always 3d4-1 and drop the lowest dice.
>>45882404
(1d10+1d12+1d6)/4
http://anydice.com/program/7d9a
No skew.
>>45882831
This can't produce a 7.
>>45882404
Roll a d8. Reroll on 8s.
Rolling for stats is awful
>>45883186
any good example system of point buy ?
>>45883375
I'm not quite sure what you mean. Pointbuy is pretty self-explanatory. The only real choice involved in designing it is whether you assign points linearly or have a scaling cost.
>>45883375
Nope.
though the best way to do random rolls is always gonna be something like Eclipse Phase or Traveller's "life path" system, where the random rolling involves also explaining chunks of the setting to the players as they go through the algorithm.
>>45882516
1-7 for attributes out of a total roll potential of 32 for all the other stuff.
>>45882837
Yes, that's I'm considering.
>>45883110
A bit too complicated for a system that's supposed to be quick roll.
>>45883165
That's fast but essentially the same as d6+1.
>>45883186
I think it's neat but YMMV. It's a game where you roll a bunch of things randomly for quick character creation.
>>45882404
How many attributes total and do you roll straight down the line, or do you get to distribute your scores to whichever attributes you want?
>>45883685
6 attributes.
Distribute rolled scores as desired among them.
>>45882404
Use a seven sided die.
>>45882404
Roll six d6s. Count up the number of dice that read 4-6 and add them to a base of 1.
>>45883775
Not sure if I was clear, but I'm saying that you get +1 for each die that comes up with a result of at least 4.
Using this method, there's a 50% chance that any given character will have at least one score of 6 or 7 (and an equal chance of at least one score of 1 or 2).
If you want to be nicer to PCs, you could have a roll of 3-6 indicate +1. That gives you the results in this pic. That means that there is a 42% chance that any given character will have at least one 7, and a 68% chance of having 2 or more scores of 6 or more.
>>45883983
That seems a bit too generous.
>>45883647
>That's fast but essentially the same as d6+1.
No, it isn't?
Roll d8, reroll on 8 gives out a result from 1 to 7
d6+1 gives out a result from 2 to 7
>>45883700
Maybe do a random array? For instance, roll a d20 on the table below to get the stats to distribute. That way nobody gets screwed.
1) 7, 6, 5, 3, 2, 2
2) 7, 6, 4, 3, 3, 2
3) 7, 6, 3, 3, 3, 3
4) 7, 5, 5, 3, 3, 2
5) 7, 5, 5, 4, 2, 2
6) 7, 4, 4, 4, 3, 3
7) 6, 6, 6, 3, 2, 2
8) 6, 6, 5, 3, 3, 2
9) 6, 6, 5, 4, 2, 2
10) 6, 6, 4, 4, 3, 2
11) 6, 6, 4, 3, 3, 3
12) 6, 5, 5, 5, 2, 2
13) 6, 5, 5, 4, 3, 2
14) 6, 5, 5, 3, 3, 3
15) 6, 5, 4, 4, 4, 2
16) 6, 5, 4, 4, 3, 3
17) 6, 5, 4, 4, 4, 2
18) 5, 5, 5, 5, 3, 2
19) 5, 5, 5, 4, 4, 2
20) 5, 5, 4, 4, 4, 3
>>45884260
I actually agree.
One thing that would be helpful for this thread is to determine an ideal mid-point for PCs and for regular folks. I think it's helpful to have more space at the top end than the bottom, so I wouldn't set the midpoint for average folks any higher than 3. That probably means that PC stats should average somewhere around in the 3.5 to 4 range, and no higher.
>>45882404
For your own game system, I'm guessing?
If you want something similar to a bell curve, then (2d4-1). If you want a continuous distribution (same odds of each result), then d8, re-rolling 8's.
Now, if you're trying to set a floor (hard or soft), then start with one of those options above and then give them a certain number of chances to drop results and reroll them. Either a certain number of chances, or allow them to reroll any stat below a certain value.
>>45883375
I don't know how important attributes are to your game, or what kind of distribution you're looking for, so I can't get you a fine-tuned point-buy, but here's an example of one. Players get somewhere in the range of 10 to 12 points to spend (there should probably be something you can do with a single leftover point or maybe two--boost hit points, get bonus XP, increase starting wealth, get connections or whatever).
Score = Cost
1 = -3
2 = -2
3 = 0
4 = 2
5 = 4
6 = 7
7 = 11
>>45885032
Here's another with slightly smoother scaling. If the other one gives you 10 points to allocate, this one should give you 15. If you use this one, you *definitely* need for there to be something else you can spend a leftover point or two on.
Score = Cost
1 = -5
2 = -3
3 = 0
4 = 3
5 = 7
6 = 12
7 = 18
Note that if you want to make 7s rarer, add 2 to their cost on this table (7 = 20) or 1 to their cost in the previous one (7 = 12),
>>45885032
The problem with point buy, at least in the groups I frequent, is that people take fucking forever to assign their points. They're constantly fiddling with the stats, changing their minds, whining about how they wish this stat was 1 cheaper, how they've got 1 point left and nothing to do with it, etc. It's so annoying.
With rolling it's 60 seconds and you're done, let's start playing.
>>45884799
Here's a significantly improved version. Roll a d6 to determine which table to use and roll a d10 on that table. Arrange stats as desired.
>>45885180
Yeah, there are definitely issues. Some of those you can iron out by building them into a larger system (with, for instance, other shit to spend a leftover point or two on). Some of those are inherent in the system, itself. But rolling attributes has problems to. Somebody can get unlucky, and if you're only rolling one or two dice for each score, you aren't going to see the same sort of moderating effect as you do in a 3d6 bell curve (which means that discrepancies are going to be larger and more common).
>>45886221
On the other hand, something like >>45886032 is quick, easy and nobody ends up getting screwed. You do, however, have to reference a table, which is less than ideal, but hardly a deal breaker.
>>45886032
That's not bad, thanks.
But why roll 2x when the distribution is flat on both rolls? Why not just have 1 table and 1 roll? Because then you'd have to roll d30?
>>45888148
>Because then you'd have to roll d30?
Yeah. Lots of people don't have d30s, and even if they do, they might always want to cart the golf balls around. Plus, most d30s are pretty shitty. If those things aren't precision, they roll around like oversize marbles.
>>45888472
Do you have a version of the table that's a single d20 roll?
>>45891394
I made that one by including all possible combinations that were at the same basic power level, with a few constraints (no more than a single 1 or 7). I could cut it down, but then I'd have to decide which third of the table to leave out. I could drop all arrays that have five of the same number under the excuse that they're a bit boring. Quickly eyeballing things, it looks like that gets us halfway there. But it also increases the number of percentage of characters who have a 7, which I think is already a tad on the high side. Is it okay to have almost half of characters to have a 7 (rather than almost a third)? And what criteria do I use to trim things the rest of the way?
>>45891710
Dunno, sorry, but thanks for your effort.
7d2 where 2 counts as zero.
Admittedly this is between 0 and 7 but I don't think it veers as much as the other +/- things you're considering if you just count 0 as 1 or 7.
>>45891710
>I could drop all arrays that have five of the same number
I meant four of the same number.
>>45891834
Normally, I'd find having to roll two dice to get a certain result to be obnoxious, but here I don't think it's a big deal for several reasons. First, they're different size dice, so you can roll them at the same time. Second, the tables are pretty compact, so you don't have to search around to figure out where your result is. Third, you only have to do this during character creation, when time isn't a factor, and you're actually rolling far *fewer* dice than you would normally have to. I mean, even if you generated each attribute with a single die roll, that would still be 6 rolls instead of 2 (the standard D&D method takes 18 dice rolls... or 24, depending on whether you're talking old or new school).
In any case, I've already closed Excel, so to give you another table, I'd have to copy over the values in the pic, which wouldn't be particularly difficult, but it's something you could do as well as I could.
>>45883375
Shadowrun has a good attribute purchase system.
