Random Stat Rolling Within Point Buy

[JNAProductions]

Edit the second: Use the 21-23-25 system, as explained below by Knaight. Much easier to use.


I remember the days where rolling stats meant you didnt choose where they went. You rolled 3d6 and they went in order down the list. From those stats you decide race/class... and create a background and story around all that.
From this thread (http://www.giantitp.com/forums/showthread.php?399959-To-start-with-18-or-not&p=18871459#post18871459).

Now, looking at that, it seems fun for fluff, but also potentially bad for fun gameplay. (All 3s? Sucks to be you. All 18s? Boring to be you. Chances of either happening? <1/1E14) So my thoughts are making random rolls happen within point buy bounds.

Here's how it works:

Roll 1d10-1. This is how many points you spend on a stat, rounding up if needed for higher numbers. (E.G., roll an 8, round up to 9.) Then, roll 1d6 to determine which stat gets this many points-1 Strength, 2 Dexterity, etc.

Repeat until you have determined five stats. The 1d6 becomes a 1d5, then a 1d4, all the way to 1d2.

Any leftover points get added on an appropiate die (E.G. 7 points left, roll 1d8-1) and then a 1d6 for which stat.

Example stat line-up:

1d10-1 1d6
1d10-1 1d5
1d10-1 1d4
1d10-1 1d3
1d10-1 1d2
1d10-1

Edit: Right, forum dice roller is being finicky, so I'll be rolling and posting from home.

9 in (5) Wisdom
8 (round up to 9) in (5) Charisma
1 in (2) Dexterity
9 (only 8 points left, round down to 7) in (2) Constitution
2 (only 1 point left, round down to 1) in (1) Strength
0 (no roll needed) in Intelligence

Giving us the following line-up:

Strength 9
Dexterity 9
Constitution 14
Intelligence 8
Wisdom 15
Charisma 15

A wise and charming (wo)man, with the endurance of an ox, but unfortunately possessing a brain bearing similarity to an ox as well, and clumsy and weak besides.

Mechanically, a good fit for a dwarven heavy armor Cleric/charisma caster multiclass, with excellent casting stats and a great constitution, but in sore need of something to make up for the other subpar physical traits.

[Knaight]

This seems needlessly convoluted. It's basically just a way to roll stats that involves checking a bunch of tables. If you want a particular bound, just come up with a die mechanic that covers it. If you also want a total, just do something like the 21-23-25 method where you roll 3 times, subtract the three things rolled from those numbers, and use that for the other three stats. If you want to keep the randomness of which stats get which rolls (which I would recommend against in systems not explicitly built for it), just have you roll for a particular stat, with a particular other stat assigned automatically. 21-23-25 does let you introduce a skew with that, but that's not necessarily an issue.

[JNAProductions]

What's the 21-23-25 system? This is the first I'm hearing of it.

Edit: It's not that bad, rolling it out. It's also not particularly fast compared to regular stat-gen, so take that as you will. Easier if you memorize stat positions and the point cost table (which I, the massive nerd I am, have).

[Knaight]

What's the 21-23-25 system? This is the first I'm hearing of it.

Basically, you roll 3 times (either 3d6 or 4d6, best 3). Subtract one of these from 21, another from 23, another from 25. Those are your rolls. So, for an example - I roll 12, 12, and 5*. I can't assign the 5 to the 25 since that goes out of bounds, so there are two different sets possible.

Assign the 5 to the 21: 5, 12, 12, 16, 11, 13 (ordered: 16, 13, 12, 12, 11, 5)
Assign the 5 to the 23: 5, 12, 12, 18, 13, 9 (ordered: 18, 13, 12, 12, 9, 5)

*Actual rolls.

So there's some randomness. The character has a 5 to assign and will suck at something, the two twelves guarantee some pretty average rolls. However, the stats will always total 69, so it prevents characters from being either too amazing or completely screwed.

[JNAProductions]

I do believe I'll be using this from now on. Thanks for explaining, Knaight.

[Pex]

21-23-25 was originally 23-25-27, which I prefer, but I think lowered due to 5E paradigm. Let's compare.

Rolls: 15, 12, 10

23-25-27 vs 21-23-25

27-10=17 *** 25-10=15
25-15=10 *** 23-15=8
23-12=9 *** 21-12=9

Add 2

9 + 2 = 11 *** 8 + 2 = 10

17, 15, 12, 11, 10, 9 *** 15, 15, 12, 10, 10, 9

If 17s were allowed, *** 27 5E Point Buy value on the dot
29-30 5E Point Buy value

Play a variant human

18, 16, 12, 11, 10, 9 *** 16, 16, 12, 10, 10, 9

Personally I still prefer 23-25-27, but 21-23-25 is closer to 5E Point Buy. Since I loathe 5E Point Buy, my preference doesn't change.

Trying different rolls, just 21-23-25

Rolls: 11, 18, 14

25 - 18 = 7
23 - 11 = 12
21 - 14 = 7

Add 2

18, 14, 12, 11, 9, 7
If 5E Point Buy allowed 18 and 7, 18 cost 15, 7 cost -1, this is 29 Point Buy cost

23-25-27

27 - 18 = 9
25 - 11 = 14
23 - 14 = 9

Add 2

18, 16, 14, 11, 9, 9 this is 5E 38 Point Buy

Conclusion: Both methods provide a variance in total value 5E Point Buy cost, but 21-23-25 standard deviation is closer to 5E's Point Buy value.

I prefer 23-25-27, but I could get over it with 21-23-25 just because I dislike 5E Point Buy so much. It's not the "bigger numbers" with 23-25-27 just more what I'm used to. I would have to agree 21-23-25 fits 5E more where 23-25-27 fits quite well in 3E/Pathfinder. It doesn't forbid an 18 at 1st level which is one reason for me to dislike 5E Point Buy, and the inherent three odd scores you get are nice incentives for feats that give +1 to an ability score and general ability score increase option of +1 to two scores.

[Theodoxus]

I'm still going with 1d8+7. Sure, there's no bell curve, which kinda sucks, but it's exactly the correct range of 8-15 for 5th Ed

[NotVeryBatman]

I'm still not following. Can the rule for adding two be explained differently? And can it be explained how three rolls gets to six stats?

[Knaight]

On 21-23-25, the actual numbers were something I just pulled from memory. I didn't remember what the original case was exactly, just that it had an odd number greater than 20 (which I'll call N), N+2, and N+4. Maybe 21-23-25 should be used for 5e, maybe 23-25-27 should. Both have effectively the same advantages.


I'm still not following. Can the rule for adding two be explained differently? And can it be explained how three rolls gets to six stats?

Each roll generates two stats, x and C-x, where C is a constant and x is the roll. You have C1, C2, and C3. You also have x1, x2, and x3. The other three stats are generated via Cn-xn.

[Balor777]

On 21-23-25, the actual numbers were something I just pulled from memory. I didn't remember what the original case was exactly, just that it had an odd number greater than 20 (which I'll call N), N+2, and N+4. Maybe 21-23-25 should be used for 5e, maybe 23-25-27 should. Both have effectively the same advantages.



Each roll generates two stats, x and C-x, where C is a constant and x is the roll. You have C1, C2, and C3. You also have x1, x2, and x3. The other three stats are generated via Cn-xn.


So lets say i roll 8 11 13
I keep these for me 3 stats
and the other 3 stats are
2X-8
2X-11
2X-13
For lets say
21 - 8 = 13
23 - 11 = 12
25 - 13 = 12


Final stats 3 blue and 3 green?

Can we use 22 - 24 - 26 too?

[Myzz]

the only part I think is too complex is the double randomness...

random roll to determine stat value
random roll to determine where stat goes...

simplify and just go down the list Str, then Dex, then Con, then Int, then Wis, then Cha...

still has a random value in each place.

still has the possibility that you have all 15's... but thats the nature of random rolling

[Knaight]

Can we use 22 - 24 - 26 too?

You can, but it's not advised. With odd numbers you always get 3 odds and 3 evens. With even numbers, if you roll an even you get two evens, if you roll an odd you get two odds - which essentially gives you a -1 penalty to one stat moved down, and a +0 bonus to one stat moved up once you get to actual rolls.

[Chronos]

The one problem I have with this method is that some 5e classes need only three ability scores, while some need four. It feels like this method of stat generation would be much kinder to the former than to the latter. What's the best-case scenario for, say, a barbarian who needs Str, Dex, Con, and Wis?

[JNAProductions]

the only part I think is too complex is the double randomness...

random roll to determine stat value
random roll to determine where stat goes...

simplify and just go down the list Str, then Dex, then Con, then Int, then Wis, then Cha...

still has a random value in each place.

still has the possibility that you have all 15's... but thats the nature of random rolling

The problem with that is that it front-loads stat points. High rolls early on exclude high rolls later, so if you roll all 9s, you only get high physical stats every single time.

It's not needed, and the 21-23-25 is probably better.

[Solusek]

Grid Method is still my favorite way to roll stats due to it's mix of randomness along with some player determination of where the numbers go.

http://invisiblecastle.com/stats/help/grid/

Yes, in his example the numbers end up being super high, but thats's because he is re-rolling 1's. If you want more reasonable looking PC stats just don't re-roll the 1's.

[Pex]

I'm still not following. Can the rule for adding two be explained differently? And can it be explained how three rolls gets to six stats?

Sorry, I thought it was understood. You just get a free +2 to put somewhere after all the math. I'll number the steps, using 23-25-37.

1) Roll 4d6, drop lowest, three times. These are your first three scores. Optionally/Ideally have a minimum roll of 7 such that rolling 2, 1, 1, 1 on four dice will still give you 7, not 4.

2) Take any roll and subtract from 27 for your 4th score. Normally there would be a limit of 18 so you can't do 27 - 7 = 20 for a score, but for 5E this can be removed. Even a 1st level character using normal rolling can have a 20 by rolling the 18 and put the racial +2 modifier in it, such as dragonborn strength warrior characters. The rules prevent the score going higher than 20 anyway.

3) Take a second roll and subtract it from 25 for your 5th score. Note that 25 - 7 = 18.

4) The final roll is subtracted from 23 for your 6th score.

5) Add 2 to any one score. Max 18 probably not needed for 5E.

6) Arrange scores as desired and apply racial modifiers. Here allow the 20 since you can get one by normal rolling because of racial modifiers anyway.

Step 5) is important. Some people would use it to get rid of a less than 10 score, especially if they have two of them. Others would bump a 14 to 16 or 15 to 17 for a needed second prime ability score. A 12 to 14 is attractive just to put in Constitution. Step 5) is needed for 3E/Pathfinder. It may not be as necessary for 5E especially if you allow the system to start you with a 19 or 20 at 1st level, but with bias I prefer it remain.

Here's another example.

1) 13, 13, 12

2) 27 - 12 = 15

3) 25 - 13 = 12

4) 23 - 13 = 10

5) 15 + 2 = 17

6) Final array of 17, 13, 13, 12, 12, 10 before racial modifiers.




Can we use 22 - 24 - 26 too?

No, three odd scores are better to guarantee three odd scores in results. If I roll all evens, using 22-24-26 leaves me with all evens. In 3E/Pathfinder having odd scores matter because feats that use an ability score as a prerequisite use an odd number. It also provides interesting choices at ability score increase levels. In 5E it becomes a bigger deal since +1s to two scores becomes a lot more attractive as well as feats that provide +1 to a score. Still nothing wrong with taking +2 to one score, but that you have interesting equivalent choices of juiciness is a feature that makes the fun.