D&D 5E The Philosophy Behind Randomized and Standardized Ability Scores

[Snarf Zagyg]

I just read a comment by @FireLance -
One possibility that I've been toying with, although I admit I've never actually used in any of the campaigns I've run, is to start with the standard array (15, 14, 13, 12, 10, 8), assign as desired to the various ability scores, and then roll 3d6 in order. If the number rolled for an ability score is higher than the assigned number, use the rolled number instead.

It's an interesting proposal! The thing is, from the very beginning (3d6 in order) to the new random (4d6k1) to the variants introduced in the 1e DMG and UA to modern methods of point buy and the standard array, the one constant I have seen is the desire to explore different methods of ability score generation.

That said, I think that there has been a singular focus on one aspect of the ability scores- the effect on the singular PC. I will use the continuation of Firelance's quote-

The standard array guarantees the baseline level of competence that the game expects, and the 3d6 rolled in order gives a chance that one or more ability scores may be improved - and not necessarily the ability scores that the player would have chosen.
Of course, the downside to this approach is that the PCs will have better ability scores, on average, than the game expects and the standard challenges will be that much easier for them.

From this perspective, the perspective of the individual PC, the issue of ability score generation seems to boil down to a few basic decision points:

A. Randomized vs. standard scores overall. For this, the choice is usually between a default level of competence, or the chance for amazing ability scores, with the concomitant risk of terrible ability scores.

B. Randomized order vs. choose your own order. Again, we have a choice between selection of scores within certain fields (I wanna play a Wizard, so I will put my highest score in intelligence) as opposed to the risk/serendipity of having to make a character based on whatever order of rolls you happen to have (I was going to make a Wizard, but my two highest rolls are in Strength and Constitution ...).

Now, there are various tweaks you can perform (for example, in the proposal above, you get all advantages of the standard array, with the possible upside of better randomized rolls, and you get the choice of setting your character's rolls, with the possible upside of some serendipity making you re-choose your selection; the downside, of course, is that it's all reward, no risk). But at it's core, almost all ability score generation methods involve choices among these axes; it's rare for one not too - such as the infamous 1e Unearthed Arcana method of "select your class, roll outrageous amounts of dice, and also get the minimum score."

However, this focus on the individual PC overlooks one of the biggest issues when it comes to ability scores that, IMO, has driven the adoption of the point-buy/standard array method.

Fairness.

Fundamentally, D&D is a social game that you play with other people. Different people have different methods of handling the game, of competition, and of certain innate ideas of fairness- this is why, for example, the concept of "sharing the spotlight" has become more important in RPGs. However, this is an aspect of the social compact that often gets overlooked when it comes to PC ability score generation.

From the beginning, there has always been an issue with fairness. Some tables handled it better than others- sure, maybe Paul always managed to get a 17 in Charisma for his latest Paladin (Percival IV), and maybe Bob always got that 18 in strength for that sweet, sweet percentile strength ... but, you know, maybe not. Some people, and some tables, are much better at handling that inherent power dynamic (and imbalance) that occurs in D&D when characters have disparate ability scores.

I mention this because I think that this is often given short shrift in discussions of ability score generation- not the effect on a singular PC, but on the effect on the table. If someone gets really really low scores, and someone gets really really high scores, what will happen? Is this going to be seen as good fun- an opportunity for memorable characters? Or is this going to be viewed by the players as unfun and unfair? The type of players and the reaction you get will most likely determine the best methods of character generation.

I would go so far as to say that this sense of fairness is what has likely driven the adoption of standardized ability scores over time (point buy, array). Even some randomized methods allow for this sense of fairness (such as allowing players to bid on scores at an auction, or to chose among a number of random scores).

Anyway, I wanted to put this out for general discussion- the social and table aspect of ability score generation. What do you think? Do you consider this, and if so (or if not), why?

 

[payn]

I do consider it when starting up a new game. The need for fairness largely matters on the system. I dont worry about it in a game like DCC where adventure comes at you fast and life is cheap. 3E system mastery design is a world of endless options, and many of them bad. While I love this edition the most, I've had randomly rolled tables and you can fit the effectiveness gap into the grand canyon. It's difficult to GM a game like that, and folks dont feel good playing BMX bandit next angel summoner. On the flip side, some modern games are designing around standard arrays, where if you roll randomly at all, you can easily end up with an ineffective character. The math of the system simply doesn't consider it.

 

[Oofta]

This ... comes up on a somewhat regular basis. Short version?

• I prefer point buy or array and home brewed options long before there were official ones.
• I've seen tables where one person had crap scores and one person had absolutely incredibly awesome scores. Neither person was happy with the result.
• I think people underestimate the difference that you will see if you use the 4d6 drop lowest; in a table of 6 in most cases 1 PC significantly better scores.
• You can still get random by using point buy to generate a couple dozen arrays, randomly place the numbers.

Much longer version ... it's a lot because this has come up before so I'll just copy/paste. Of course this is all personal preference, this is just why I won't use it and would hesitate to join a group that does. IMHO the only thing it achieves is that some characters will be better from a numbers perspective than others.

Spoiler: Long, long semi-rant

The only advantage to rolling for ability scores in my opinion is that you get random results. Sometimes they will be higher than what you get with point buy, sometimes they will be lower. On average the person with the highest ability scores at a table of 6 will be 2 points higher in every ability score, and can be much more significant. A +2 to every ability score may not sound like a lot but it can have a significant impact.

I wrote a program to try to get concrete numbers instead of relying only on experience (partly because some of mine were ... bad). Using standard 4d6 drop lowest, I generated 2 dwarves at level 4 and threw each on a semi-suicidal fight against individual Hell Hounds. The result was that the PC with lowest stats for the group survived 20% of the time, the PC with high stats for the group survived 42% of the time (details are below).Yes, I know it was probably a colossal waste of time. But it was an interesting challenge.

Some people believe you get better numbers on average. If it does it's within a point or two, and depends on how you do your analysis since point buy is a constricted range. When rolling the average low ability for a character will be 8, the high will be 15. If you stay in the constricted range 8-15, a 27 point buy is slightly better. According to an independent analysis of the method, point buy will give you numbers pretty close to average: 4d6 Drop Lowest You can buid the 3.5 elite array with a 27 point buy that is mentioned in the article.

Many groups don't use straight 4d6 drop lowest roll once. They use Fantasy Updated Dice Generation Enhancement, or F.U.D.G.E. when rolling stats for characters. For example if you roll enough characters you will eventually have a good one, something I took advantage of in the old D&D video games like Baldur's Gate.

There's nothing wrong with F.U.D.G.E.ing your results as long as the entire group (DM included) is on board. The resulting number of course will typically be higher than what you get with point buy.

The equivalent point buy to the F.U.D.G.E. method with is to simply give people more points to spend, or use a heroic array from previous editions. For example, give 32 points and consider letting people buy higher numbers. Allow people to buy a 16 for 11, 17 for 14, 18 for 17 points.

In other words, if your goal is to start out with higher (or lower) numbers than standard point buy allows, you can always adjust the point buy system or use arrays that suit your taste to achieve that goal while still ensuring that PCs are balanced relative to one another.

Want randomness? Come up with a list of 20 or so valid arrays that represent a decent spread of numbers and roll for which array to use and assign the numbers in the array to stats randomly.

What does rolling achieve?
The only goal rolling for stats achieves over some variation of point buy/array is that some characters will be statistically more effective at their chosen profession than others based on a one time roll of the dice.

Some people think this is more realistic (I don't), I don't think it adds anything to the game. I look at adventuring companies something akin professional sports teams. They're some of the best at what they do. If you have average or below average stats, you probably stayed on the farm. But even beyond that, I don't think it would be fun to play the water boy while someone else plays a superstar athlete. I could see how a game with all commoners could be fun, I just wouldn't want to do it long term. In addition, does it really make sense that 1 in 200 people is at the top of human potential in any give ability score? That would mean that 1 in 200 commoners could competitively weight lift which I think is silly. YMMV of course.

Why do I prefer point buy?

Story Time
The last time I rolled for ability scores we used straight roll 4d6 drop lowest. My wife and I wanted to use a point buy system, but were told that everybody had to roll for stats. I rolled a decent character, my wife rolled incredibly poorly (a single 14, a single 10 and everything else below) while another gal (Sue) at the table rolled a couple of 18s and a low roll of 14.

Neither my wife nor Sue were happy with the characters. Sue felt guilty, my wife had a character with stats that she felt could not represent the heroic character she had envisioned when we were discussing what we wanted in the campaign.

When my wife asked if she could reroll or use the point buy system from the living campaign, the DM just laughed, and said something along the lines of "that's too bad you rolled poorly but it's fair because everybody rolled".

Sue eventually committed suicide-by-goblin because she felt guilty. Since you can't be forced to testify against your spouse, I can neither confirm nor deny that my wife secretly adjusted her numbers to something reasonable.

I mostly relate this story as a cautionary tale. Don't assume that everyone wants to use random rolls, or believes that the results of random rolls are fair. The DM in this story was generally a good guy, he just had a huge blind spot and didn't realize that 2 of his players were unhappy with the results they were forced to use.

Random is not "Fair"
Analogy time: you apply for a job somewhere that advertises average pay per hour of $10.50. When you get there, you roll dice to see what your actual salary is. Will someone making $3.00 per hour working next to someone making $18.00 per hour doing exactly the same job with exactly the same qualifications consider this "fair"? I would not.

Personally I'll take PCs that start out on roughly equal footing.

Heroic Characters
I admit it. When I play the game I want to play a character that's just a little better than the average Joe. I want my character to be a good at their job, perhaps one day rising to the ranks of the best of the best.

Some people may enjoy the challenge of playing a character with below average stats. I don't. For me the game is about escapism, pure and simple. I want to play George Clooney's character from Ocean's 11 or Vin Diesel's character from The Fast and Furious (or any of the other characters from the movies, they're all exceptional in their area of expertise). I don't want to play Ralph "it tastes like burning" Wiggum from The Simpsons.

For that matter, I don't want to play a superhero either. I like a sense of growth, and don't really care for ability scores of 20 at first level. The math seems to work for 5E a little better without super high stats at first level as well.

Whether or not any particular character will "shine" from an RP or achievement perspective is a completely separate, unrelated issue. Whether or not you personally enjoy playing characters with potentially vastly different capabilities out of the box is a preference.

Character Vision
I start thinking who I want my character to be when we first start discussing a new campaign. I develop a background story, basic description, goals and so on long before I determine numbers. I will frequently post a "prequel" story before the campaign starts as an introduction for everyone else for my character.

If I roll for stats, I may or may not be able to build the character I had envisioned. I've always been able to do so with point buy. Of course this means that my vision is "stronger than average" not "stronger than any mortal that has walked the face of the earth". I do have to restrict my vision to what the game allows.


Party Ability Score Variance
It's almost inevitable that there will be winners and losers in the random result lottery. Sue and my wife's experience were extreme, but for every character with an 18, there will be a character with a 5 or less.

As mentioned above, a while back I wrote some code to compare "groups" of 6 randomly generated characters using 4d6 drop lowest (I was bored). I assigned point buy values to 3-18 (above 15 used +2 for every number) and then compared point buys. It's not a perfect comparison but gave me a general feel for different power levels.

The majority of "tables" fell into the 25-40 point difference range. What does that mean for real numbers?

Well compare 25 point buy diff for a couple of randomly selected tables.
Diff 25
Cost: 13: 12, 12, 8, 8, 9, 12 - hit the snooze alarm boring
Cost: 38: 16, 10, 7, 15, 12, 16 - pretty good, only 1 low stat

Diff 35
Cost: 15: 13, 12, 11, 11, 9, 7 - not completely horrible, at least he has a 13.
Cost: 50: 17, 15, 15, 14, 14, 10 - no weaknesses, scores too high for my personal taste.



To me, those are significant difference, the low end represents stats of someone who may be decent an their company softball team. The high numbers represent people that play in the major leagues. The last character is probably a superstar on their team.

How big of an impact does this difference have on character effectiveness?
It is my belief that it should be obvious that ability scores matter. There can be a fairly significant difference between the most powerful and least powerful member of a party if you use straight 4d6 drop lowest.

To try to add some proof to this other than anecdotal evidence and opinion, I added to my code. The code generates groups of 6 characters and calculates how much it would "cost" to buy the ability scores of the group using point buy. From that I selected the result sets with the highest and lowest cost and then simulated thousands of fights using those 2 PCs in a fight with a Hell Hound. I picked a hell hound because at CR 3, it's a creature they could be expected to fight (admittedly normally not solo, but stuff happens).

Tactics, environment don't really matter for the basic hypothesis. This doesn't speak to whether one player (Tom) might have better strategy than another player (Joe). If the claim is that "Tom may do better than Joe even with a lower score character" then Joe will have even worse if he rolls up the character with lower scores. Individual player ability does not affect the basic math. It also doesn't have anything to do with whether someone may have fun playing the guy with the low stats. This is simply a method to get an indication of how much ability scores can matter.

So, after way too much time coding, and somewhere around a million simulated fights for a hundred different randomly generated groups ...

Assumptions

• Dwarven fighters because it's one of the better options for a low(er) stat character
• Strength gets highest roll, Con gets next best
• Dex is assigned the lowest roll available
• AC 20 (plate + shield)
• Dueling fighting style (+2 damage)
• Champion (crit on 19-20)
• Second wind when at half HP
• Level 4, add to Strength if it's less than 20 then add to Con if its less than 16 (splitting the ASI on odd numbers). Otherwise take the Shield Master feat.
• No other limitations, if the low roll was a 3 it was assigned to dex, if highest was an 18 the dwarf had a 20.

Results
It's been a tough adventure. Minnie and Maxie, dwarven comrade in arms are separated from their party when they see two pairs of red glowing eyes in the darkness. With no chance to run, each faces off against their foe. Who will win?

Minnie, who was low man on the ability score results for the group wins the fight 22% of the time.
Maxie, who was high man on the ability score results for the group wins the fight 45% of the time.

If you make them 5th level PCs the results are similar, and there's a big jump in survivability (the extra attack helps a lot)
Minnie wins 80% of the time
Maxie wins 95% of the time

On a side note I was curious if the breath weapon was skewing things so I added a flag to turn it off. I thought that not relying on a Dexterity save would narrow the gap. Turns out without it, MInnie wins 51% of the time, Maxie wins 80%. It appears that winning initiative (ties always go to the PC) is more important than I thought.

Analysis Conclusion
Twice as likely to survive a deadly fight is a significant difference IMHO. Ability scores can make a significant difference in the effectiveness of a character to fulfill their roll in the party.

Conclusion
I think it comes down to player empowerment and agency. Different people play for different reasons, but I see no value to forcing someone to play a randomly generated character if they do not want to. I don't think playing someone with below average ability scores makes you a better role player or person.

This is not about "competition" between characters or that I believe stats dictate whether or not I can roleplay a character. It's about feeling like I can contribute to the team on even footing with the rest of the characters.

I want to play the vision of a character that I have, not a character that is forced on me by random luck.

I have no problem accepting that some people prefer random characters. Just don't tell me that it makes you somehow superior, or that rolling for ability scores is "the one true way" to play D&D.

Standard Disclaimer
There is no intent in this posting to be insulting, I'm not trying to demean or put anyone down if they like rolling for stats, if some word or phrase bothered you I apologize. The opinions in this post merely reflect my opinion, which hopefully I've backed up with logic and analysis.


Like I said, we've been down this road many times. So a disclaimer: If you respond to this complaining about:

1. You find the phrase "F.U.D.G.E." personally offensive for some reason. I'm just acknowledging that many groups use alternative dice roll methods and roll but don't really use 4d6k1.
2. Somehow I'm implying you only want to play characters with high ability scores, or that even if I was that it's a put down. Play characters with all 20's if it's fun for you. I won't think less of you.
3. That random results are "fair" and that until I change my opinion you will continuously troll my responses. Whether rolling is "fair" is an opinion. You are entitled to yours, I've given mine.
4. That comparing two characters of the same class and race with different ability scores in a "white room" fight is somehow meaningless. My analysis is showing how effective a PC could be in a particular role. I picked a role I could simulate. It clearly shows that ability scores matter from a game perspective. Whether they matter to you as a player is a preference.
5. All of the other people stating that their opinions are "facts" because they have really, really strong opinions.

I will ignore them. Thank you.


P.S. Yes, it's sad I have to put a disclaimer on this post. This topic has come up multiple times and rather than discussing the issue things devolve into personal attacks on me and my opinions pretty much every time.

 

[Charlaquin]

In the interest of fairness, I always provide the option to take the standard array or point buy if you want it. You can also choose to roll 4d6 (drop lowest) if you want to, and whether you roll in order or re-arrange is up to you. That way players who like the risk/reward of rolled stats or the challenge of making a character that suits the abilities chance gave them have that option, and players who would rather be assured of a baseline level of competence have that option as well. I also allow players who decide to roll the option to re-roll if they’re really unsatisfied with the scores they rolled, but I encourage them to keep their stats if they think they can live with them. If you choose to roll, get “bad” stats, and want to re-roll, you’re missing the point of rolling IMO.

 

[BookTenTiger]

I love randomized ability score generation.

However, I've noticed a shift in how my D&D group has generated character ideas. Way back in the day a player would say something like "I want to be a barbarian!", roll up stats, and then build their character.

Now the players in my group say something like, "I want to play a soulful aasimar barbarian whose god fills him with devine rage, but he feels guilty after battle." The ability scores then serve to tell this story, and pregenerated scores tend to be easier to fit around a pregenerated character concept.

I remember one of my players rolling stats and saying "I can't make the character in my head with these scores."

I think randomized ability scores work better for shorter campaigns. If you roll "poorly" and you're stuck with that character for three years, it feels a little bad. If you roll "poorly" and you know this story will end in six weeks, it's a little easier to deal with.

 

[Blue]

Snarf, that was positively concise for you. (Love ya posts, that was a gentle tweak.)

The point you bring up about fairness is generally an important one, and in 5e perhaps even moreso as the "haves" then get to play with feats much earlier while the "have nots" more often get stuck with taking ASIs to catch up on some math first.

I've seen several times people post systems where the players work out rolls for the entire party. In many different ways. It doesn't seem to be hugely popular as a variant, but especially for 5e I could really see that as a fair but organic method of developing ability scores.

As a side note, while you mention fairness there is another aspect about point buy I like in 5e. It leaves you with scores that both an be improved by ASIs but also are good enough you could take a feat. I really like the Faustian bargain of having to take one or the other when both are attractive. It makes it a meaningful choice that helps define your character mechanically vs. others of the same class/etc.

 

[BookTenTiger]

I like the idea of a group using Point Buy to make six different arrangements of stats, then each player rolling randomly for which numbers they will use.

For example, the first person may design a 15, 15, 14, 8, 8, 8. And if you rolled a 1, you would use that array.

 

[Blue]

Just a quick anecdote. Reacently on a FB 5e forum, someone was gribing abut their rolled ability scores, adn saying that they were so low that they had to go for a SAD character like an archer rogue. The scores were all better than the standard array arranged highest to lowest, and I mentioned that plenty of people play point buy and/or standard array and make characters that need more than one ability score.

What it eventually came down to was that they rolled the lowest at their table, and wanted to minimize that by leveraging their one good score.

There will always be someone who rolls lowest when determining randomly. It seems something that could bring discontent for the months or years that a campaign lasts.

While I love unexpected and organic characters, rolling doesn't really give that - it mostly just changes the exact numebr while leaving the order you'd place them. You won't end up with a single-classed fighter with a 14 STR, 10 DEX and 17 INT, because the person that would assign like is the exception in the D&D community. So going for fair for everyone beats out what individual variations you would get rolling in terms of how it shapes your character.

 

[payn]

I played with some old schoolers that loved random rolling. Took what the dice gave them and I was pretty impressed with that. If it was bad though, they would just suicide the character at the earliest opportunity.

 

[Krachek]

I played with some old schoolers that loved random rolling. Took what the dice gave them and I was pretty impressed with that. If it was bad though, they would just suicide the character at the earliest opportunity.

Suicide or simply the death of unfit character make great story too. The last Deadpool and Suicide squad just made amazing funny scene with that!