D&D 5E Rolled character stats higher than point buy?

[FrogReaver]

Why is it that rolled character stats are normally always higher than a character created through point buy? I pretty much never see someone post a rolled character with stats worse than point buy. Does anyone know why?

 

[lowkey13]

*Deleted by user*

 

[FrogReaver]

That explains many of the higher arrays with point buy. However it fails to answer the question of why you virtually never see poorly rolled characters.

An average rolled (4d6, drop) provides an average advantage of two points, and, more importantly, the advantage occurs with one of the high rolls (you average getting a 16 instead of a 15).

In addition, while there is downside (~15% chance of getting below an 8) there is tremendous upside (~56% chance of getting a 16 or higher). This variance makes it much more likely you're see some extreme (and good) characters.

Finally? People cheat.

 

[Mistwell]

Oh I had a 5e character with craptacular stats...I just never posted him because, well, in what context would I post the stats of a character with low stats?

 

[lowkey13]

*Deleted by user*

 

[Umbran]

However it fails to answer the question of why you virtually never see poorly rolled characters.

I would imagine that few people are enforcing that one play such a character. If you roll 3 stats of 8, your GM is probably going to let you reroll.

 

[Hriston]

If you're using the 4d6 drop lowest method (default for 5e), the average roll (rounded) is 16, 14, 13, 12, 10, 9, with an average score of 12.24. The average score on the standard array is only 12, so 4d6 drop lowest is a little better. With point buy, you can get an average score of between 11.5 (15, 15, 15, 8, 8, 8) and 12.5 (13, 13, 13, 12, 12, 12), so it can actually be a little better than 4d6 drop lowest, but I think what you're seeing might be that players don't want to have close to the same score in every ability. They want to have at least one 15, for which they end up sacrificing more points.

 

[Oofta]

Most people cheat, usually with the approval of their DM. I did some analysis way back and a significant percentage of randomly rolled characters will have a truly bad ability score (4 or less). The average is fairly close to the point buy system.

For example if you have a 3 intelligence you can:
- Ignore the fact that your character has the intelligence of a monkey.
- Play them as if they had a 7 or 8 intelligence - really slow on the uptake but basically functional.
- Reroll the stat.

Most people reroll until they get something they like, or the house rule is that you reroll until you have higher than an 8 or similar.

On a personal note, I've never liked randomly rolling for stats so I may be a little biased. I know some people will praise it up and down but I've never seen the appeal of having some members of the group being Olympic super stars and others only being qualified for the Special Olympics. YMMV.

 

[lowkey13]

*Deleted by user*

 

[Bayonet]

Rolled stats = "I made up these stats and will pout if you call me on it."

 

[Umbran]

On a personal note, I've never liked randomly rolling for stats so I may be a little biased. I know some people will praise it up and down but I've never seen the appeal of having some members of the group being Olympic super stars and others only being qualified for the Special Olympics. YMMV.

Yes, well, that's stated as, "I don't see the appeal of the worst possible result of this, if the GM exercises no oversight on characters," so, yeah, you wouldn't see the appeal.

For me, the appeal is in designing a character with exterior restrictions. Point buy asks the question, "What stats to I want to have, and how close can I get to them?" Random rolls ask the question, "What can I do with this particular mess of stats?" I find answering the latter to often be more fun, and often leads to me creating more interesting characters than the former.

A well-designed point-buy system should guarantee that the generated character will be reasonably viable (just like a well-designed class does not allow one to shoot oneself in the foot too badly, such that characters meet a certain base effectiveness unless one either really doesn't understand the system, or chooses to hamstring themselves. And, by extension, a point buy system helps keep a party from being too out-of-whack, power-wise.

Random rolling generally can't guarantee either that a particular character will be viable, or that the party will be balanced among themselves. That guarantee must come from player and GM oversight.

 

[Hriston]

Why is it that rolled character stats are normally always higher than a character created through point buy? I pretty much never see someone post a rolled character with stats worse than point buy. Does anyone know why?

Upon rereading your post, I realize your premise is flawed. Rolled stats are not "normally always higher" than point buy. They are, however, higher on average.

 

[FormerlyHemlock]

That explains many of the higher arrays with point buy. However it fails to answer the question of why you virtually never see poorly rolled characters.

Darwin. Both on the Internet and at the table. There's normally no reason to post your characters on the Internet, so the only time I ever see rolled character stats on the Internet is either in the context of a "let's make a bunch of characters" thread (which often has variant rolling techniques like 3d6-in-order and results in a ton of poor rolls and interesting characters), or a "I rolled two 18s, how should I optimize this?" thread. So the Internet has a bias towards showing only the highest rolls.

Then at the table there's a little bit of Darwin too. For good or for bad, players invest more time in characters that excite them, which often means that a character with a statline like 10, 9, 12, 12, 7, 10 gets less attention than one with a 17, 15, 14, 11, 13, 16 or even than one with 6, 8, 9, 7, 4, 10, 7. Unusual stuff is more interesting, and so is awesome stuff, so there's an additional Darwinian bias towards characters that are simultaneously powerful and unusual.

In short: in what context would you expect to see poorly-rolled characters? There is no global database of PCs for you to look through.

 

[FrogReaver]

Upon rereading your post, I realize your premise is flawed. Rolled stats are not "normally always higher" than point buy. They are, however, higher on average.

Care to offer observational proof that the phenomenon I described doesn't occur?

 

[lowkey13]

*Deleted by user*

 

[FormerlyHemlock]

Care to offer observational proof that the phenomenon I described doesn't occur?

Sure. Here's my observation: I've seen plenty of 5E characters with mediocre stats. The probability distribution I've observed matches what you'd expect from the math: maybe 20% or 25% of the characters have no bonus over +2 even after you account for racial stat boosts. These characters become Sharpshooter archers, Sorlocks, or Moon Druids. Many of these characters never see actual play, but the same is true of PCs with all 18s--when creating a character takes five minutes of spare time, and playing a character takes hours a day over the course of months, there's a lot of interesting characters that never actually get to see play.

I've also played (briefly) with one table where everyone but me had suspiciously-high stats, multiple 17s and 18s. I didn't play at that table for very long; maybe seven sessions over the course of five months.

 

[Oofta]

To the OP, rolling for stats will give you an average roll that is only a couple of points higher than a point buy. In addition, it will guarantee that your character can't be too gimped.

When I wrote "cheat," it may have had some bad connotations (go figure), and sometimes it is out-and-out cheating; but that's rare. More often, it's house rule/ DM acquiescence (as pointed out in this comment and Umbran's).

This is something that has been baked into the system since (IMO) 1e, which required certain classes to have extremely high minimum scores for entry, and made it extremely difficult to increase you ability scores (ioun stones, other certain magic items, occasional weird events). Moreover, players wanted to play "special" characters (you know, heroes!) and gameplay advantages didn't start accruing for abilities until 15 (16 for strength!).

So, again IMO, there has long been a ... relaxed culture regarding die rolling tracing back to 1e.

Many people do simply cheat/lie/have the dog eat their character sheet if they roll bad. If you have the DMs approval then I guess it's not technically cheating. Call it generous house ruling if you wish.

All I can say is that I've been in games where one person rolled super-high stats (a couple of 18s, nothing lower than a 14) while my wife rolled a character with a high stat of 14 (most were below 10). When my wife asked if we could use point buy or reroll, the DM laughed and said basically "Wow that sucks. Too bad you have to play that character." I didn't think it was fun or fair back then, I still don't.

I wrote a little program a while back that cranked out thousands of randomly generated groups and found that in most cases there is a significant power difference between characters if you use random rolls*. I find that unfair, others don't.

I've gone down this road of "rolling for stats is the schizzle because I always roll awesome character or just ignore the fact that my PC has an intelligence lower than a baboon". I don't see the point of having it again ...

*I used the point buy system to determine power level differences. It's an imperfect yardstick but the best we have. If you want the numbers I can find them, but most people don't care because most people roll until they have a character with stats they like.

 

[FrogReaver]

maybr suboptimal is the wrong word. Mediocre I expect and see. It's the below average I don't.

Sure. Here's my observation: I've seen plenty of 5E characters with mediocre stats.

I've also played (briefly) with one table where everyone but me had suspiciously-high stats, multiple 17s and 18s. I didn't play at that table for very long; maybe seven sessions over the course of five months.

 

[FormerlyHemlock]

maybr suboptimal is the wrong word. Mediocre I expect and see. It's the below average I don't.

I edited the above post for more detail, but by "mediocre" I mean something like 12, 11, 6, 9, 12, 10. I see an array like that and I'm like, "Wow, I actually can't think of anything awesome to do with that at first level." Even a single 13 is enough to make a Dex 14 Sharpshooter, but with all those 12s... that's what I mean by "mediocre". Mathematically you expect those kinds of characters only a small fraction of the time (10%? less? not sure off the top of my head) but they do happen occasionally.

For my current campaign, my rule for PCs is:

(1) You can have more than one PC. (Of course. I always use character trees.)
(2) If you roll up a PC so bad that you don't want him, you can roll up five more characters using 3d6 (with names and classes and everything), and then donate all six characters to me as DM for use as NPCs. Then roll a new PC.

#2 is designed to underscore the point that a PC with a 14 is actually quite good at what he does (e.g. Str 14 is bouncer-level Str, and 16 is Arnold-level Str) so that players don't get sucked into the power creep thinking you need all 18s to be something worthwhile.

So far no one's taken advantage of option #2, which is kind of too bad because I like player-created NPCs. But it's still doing its job just by existing as an option.

 

[FrogReaver]

Sorry but talking about characters that don't actually see play that you rolled up on your spare time don't really count.

Sure. Here's my observation: I've seen plenty of 5E characters with mediocre stats. The probability distribution I've observed matches what you'd expect from the math: maybe 20% or 25% of the characters have no bonus over +2 even after you account for racial stat boosts. These characters become Sharpshooter archers, Sorlocks, or Moon Druids. Many of these characters never see actual play, but the same is true of PCs with all 18s--when creating a character takes five minutes of spare time, and playing a character takes hours a day over the course of months, there's a lot of interesting characters that never actually get to see play.

I've also played (briefly) with one table where everyone but me had suspiciously-high stats, multiple 17s and 18s. I didn't play at that table for very long; maybe seven sessions over the course of five months.