D&D 5E Array v 4d6: Punishment? Or overlooked data

pemerton · Oct 22, 2014

Sacrosanct said:
Let's try to break this down, because this is your first core mistake right here. "Depend heavily upon stat modifiers."

Looking at the mechanics of D&D we can see the following:

To Hit = d20+proficiency+stat mod+weapon mod vs. AC value in most cases. If you have one PC with a 15 in a stat and another who rolled a 16, all else being equal for say a 1st level PC against an AC opponent, you'd have:

d20+2+2 vs AC 15. On a 1-10 he misses, on an 11-20 is a hit
d20+2+3 vs AC 15. On a 1-9 he misses, on a 10-20 is a hit

Seems to me that the odds on hitting the target do not "depend heavily" on the stat difference

I don't see how you reach that conclusion. The character who hits on 10 rather than 11 has a 10% higher success rate.

And suppose one character's stat in the comparison is 15 (+2) and another's is 18 (+4) then the success rate for the stronger character is 20% higher (succeed 12 in 20 rather than 10 in 20).

Sacrosanct said:
And that's even assuming that the person who rolls random stats will always have higher stats than the point buy or array, and that's already been demonstrated to not be true.

It's been shown that the expected result from rolling is higher than that from point buy, hasn't it?

In any event, I'm not talking about expected values. I'm talking about actual values. If two players roll their stats, and one gets a 15 as the highest and another an 18, the expected results are irrelevant. One of those players has a greater capacity to impact the fiction, via action resolution, than does the other.

Sacrosanct said:
I imagine a typical response you (or someone else) might have to this is, "No, not every stat should be the exact same, but the pool of stat bonuses should be, and the players can choose where to put them." yes, that's true. But the players also choose whether to use array, point buy, or random rolls.

I don't see how this is relevant to my point. Choosing whether to use array or take a gamble on starting conditions - ie the distribution of stat modifiers - doesn't make the outcome of those starting conditions fair. As I said upthread, if the point of the game is to play the same PC for a long time and thereby impact the fiction, the fact that players A, B and C all had the same opportunity to roll a powerful or an underpowerd PC is irrelevant to the fact that the person with the powerful PC has more capacity to impact the fiction, and the one with the underpowered PC has a lesser capacity. And it is those differening capacity which, given the point of the game as I play it, that are unfair.

Sacrosanct said:
This sentence above seems to imply that a player who demands that every PC at the table has the exact same score for each of the abilities as every other PC at the table, that is not an immature behavior, and isn't fair*.

I have neer said that a player who demands that every PC at the table has the same exact score isn't fair. Nor have I said that such a player is fair. I haven't said anything about players demanding anything. That's your word.

I've said that, in a game which (i) is expected to run a long time, and (ii) is expected to be driven by the player's engaging in action resolution, it is not fair if one player has a greater cpacity to impact the fiction via action resolution. And the fact that, at the start of the game, each player had the same chance to roll those better stats is irrelevant to this unfairness.

Why is it unfair? Because it is an inequality in capacity to play the game (by impacting the fiction via action resolution) and that inequality is arbitrary. It is nothing but the result of random rolls made at the start of the game.

Sacrosanct said:
every PC makes every type of stat roll at some point in the game, whether it be an attack roll, a saving throw, or an ability check. You're talking about how an ability modifier "endures over the course of a game", and how every PC should have the same equality in that measurement. That means, and is dependent on, every PC having the same ability modifier (equal chance) for every one of their stats since every stat is used "over the course of the game."

I don't really follow this.

First, a small thing: it is players who make rolls, not PCs.

Second, different players use their PCs to impact the fiction in different ways. Classes are an important part of this. If player A is playing a rogue, his/her PC's DEX is more important than his/her PC's CON as far as impacting the fiction goes; if player B is playing a fighter, his/her PC's STR is similarly more important than INT; etc.

So it's simply not true that unless all PCs have identical stats their players can't have equal capabilities to impact the fiction.

Sacrosanct said:
If everyone was given the same choices, complaining about a player who chose to roll and got a higher stat as unfair is literally the same thing as complaining about a player who chose to put his +2 modifier in dexterity when you chose to put a +0 modifier there (both using the same arrays) when you both have to make a dex saving throw.

No it's not.

In the case of the DEX mods, the player who put a 10 into DEX will have put a better score into some other stat, and will - during the course of play - be able to get the benefits of that better score.

Whereas the player whose stats are just superior, due to lucky rolling, has achieved the better prospects with respect to one stat without trading off against other stats. S/he is just better equippped to successfully engage the action resoution mechanics. And hence has an unfair (because arbitrary and unwarranted) advantage in playing the game.

Sacrosanct said:
where do you draw the line?

<snip>

Basically, in order for your argument to have any sort of weight, you have to get rid of random rolling altogether in the game, and every PC would have to have the exact same stats.

Nonsense. For instance, if every PC has 18 STR and 6 INT, then the players playing fighters would have a better chance to impact the fiction via their PCs than would the players playing wizards.

Even within the same class, too, different approaches to build can prioritise different stats.

That doesn't mean that there is no meaningful scope for comparing capacity to impact the fiction.

As far as random rolls are concerned, if every player is getting to roll the dice with every action declaration, then no player is going to be systematically disadvantaged relative to the others (assuming that all are rolling fair dice in an honest fashion).

Of course, if you play a different sort of game, in which action resolution mechanics are not very important, then fairness and unfairness will be different again. But that's not a refutation of my point - it's a reiteration of it!

Sacrosanct said:
the success of a to hit roll is a lot more heavily dependent on a random d20 roll than an ability modifier, and the only way to achieve "equity over the course of the game" is to eliminate randomness and make everyone the same.

I find it interesting that you keep talking about "to hit rolls" when I have talked consistently about action resolution. That is more evidence that we approach the game in quite different ways.

Putting that to one side, your statement is obviously wrong. To achieve equality of capacity to impact the fiction over the course of the game you don't need to eliminate randomenss. (Of course, you might, but you don't have to.) You just need to make sure that every player has roughly the same range of opportunities to declare actions and roll dice to resolve them, and to make sure that there are enough such opportunities that every player will roll lots of dice, and hence won't be saddled with the results of just a small number of unlikely outcomes.

Sacrosanct · Oct 22, 2014

I'm not going to reply with a giant wall of text because it would probably fall on deaf ears, and you have a critical error right from the beginning that impacts everything else you say.

It seems your problem is that you're basing your argument on the extremes, assuming the highest values will be rolled. Either you don't understand how math works, or you're being disingenuous in your analysis. For one, needing a 10 on a d20 (50%) as opposed to an 11 on a d20 (55%) is not a 10% difference, so I highly suspect it's the first of those two reasons. And a 5% difference is not something that a result would be "heavily dependent on".

When you're using analysis, you need to look at the average likelyhood of the result, and not cherry pick the highest possible result. Especially when you're ignoring the lowest possible result. That's just plain disingenuous methodology. For example, if you include a PC that has the likelyhood of getting an 18, you also have the likelyhood of getting something like a 5 or 6--lower than you'd ever have with an array.

The rest of your post is also filled with similar errors, but there's really no point in creating a huge block of text, is there?

Majoru Oakheart · Oct 22, 2014

Hussar said:
Die rolling has very little to do with "fairness" and everything to do with powergaming.

I've gone back and forth on the issue of rolling stats and have been on both sides of this arguments at various points. However, I've never equated rolling with powergaming. I got fed up of seeing everyone having maxed stats in their primary stat in 3e, and 4e when using point buy. It was like stats weren't even on the character sheet at all. Everyone had the exact same set of stats. Anyone who didn't min-max their stats were berated by the rest of the group for being an idiot.

The difference between someone who rolls a 15 and an 18 in their primary stat isn't overwhelming any longer in 5e. Both are viable characters. And it's difficult to berate another player for rolling too low in their strength. So it means people stop paying as much attention to their stats and therefore concentrate more on the game and less on powergaming.

I don't really think some players should roll and others use point buy, however. They do produce different results.

ingeloak · Oct 22, 2014

Piratecat said:
I have a weird personal divide: I love rolling my character stats, but the game generally plays better when it's point-buy, so it's what I encourage as a GM.

same here. it took my group having some inflated stats and crushing everything in their path for me to realize that 5E is largely intended for standard array or point buy system. there are some clever custom methods floating around out there as well. they offer a touch of uniqueness without making the characters too powerful.

Umbran · Oct 22, 2014

Folks, some of you are being unacceptably rude.

Several people have told you why they like random dice rolls, giving reasons that are not about character power. You reject those, and tell them that no, really, they like the power - their protestations notwithstanding. You are claiming to know why another person plays better than they do themselves - generally with folks you've never so much as met in person, much less ever played a game with.

That, folks, is jerkish behavior.

If I now said that you were engaging in this behavior because you were all egoists whose main joy is to rhetorically beat people into submission until you "win" on the internet, you'd be upset with me for ascribing motives to you.

So, why do you do much the same thing to your fellow posters and hobbyists? The Golden Rule applies. Please, respect, and try to learn from, your fellow posters.

Thanks for your time.

billd91 · Oct 22, 2014

pemerton said:
I don't see how you reach that conclusion. The character who hits on 10 rather than 11 has a 10% higher success rate.

And suppose one character's stat in the comparison is 15 (+2) and another's is 18 (+4) then the success rate for the stronger character is 20% higher (succeed 12 in 20 rather than 10 in 20).

I can see what you're doing with that math. You're looking at it from a subjective viewpoint - the viewpoint of how much more successful the higher stat is from the lower stat's point of view. That skews the perception of the difference high for some ranges of success. It's a 100% increase if you need a 20 to hit and the +1 allows you to also hit on 19. It's a 10% increase if you hit on 11 and a +1 allows you to hit on a 10. The rate of increase isn't the same which could allow you to bias the argument by cherry-picking your target numbers.

If you want to argue the significance of the difference from the point of view of the game as a whole, you have to take a more objective viewpoint and look at the difference from the whole range of results. A +1, from that point of view, is a 5% increase in hitting the target number, any target number.

Ridley's Cohort · Oct 22, 2014

dd.stevenson said:
My point being, the only things stat increases and level increases have in common is that they are both objective increases in character power, and they accomplish this partially by increasing some of the same numbers. I don't think that's sufficient grounds to claim that rolling for one makes as much sense as rolling for the other. Or else, if it is sufficient grounds, then I suppose we're free to claim that rolling for any quantifiable element of the game makes as much sense as rolling for any other quantifiable element. Right?

Exactly.

Sitting in a box in my garage is a copy of Chivalry & Sorcery (1st edition), published circa 1977, which predates the publication of AD&D.

In that game, you roll stats, you roll social status (yes, you may be a peasant who needs to grovel while near their "betters" in civilized lands), you roll your race, you roll your alignment(!).

At least at some superficial level, rolling for this stuff is fair, because everyone rolls on the same table. A lot of modern RPGers would say "WTF!" at being forced to roll all those things, and be expected to play the PC that results because it would not feel like their own character (among other things). There is no right or wrong here; it is a question of how expectations are set and managed.

The foundational assumption of an RPG played at the table is that the rules + DM + houserules + campaign rules + group agreements + etc. collectively put certain constraints on which kinds of "fair" are included in this particular game, which kinds of fair are "in play" in front of the screen, and also which things are excluded because they are either "unfair" or "less fun".

Furthermore, if we stop and think about what a DM does, why would a DM ever reject a suggestion or request from a player? There are many possible reasons. In no particular order, the more common ones would be "that is too good (unfair compared to options available to other PCs)", "that is too good (trivializes a kind of challenge I want to put against the party, so the game will be less fun)", "does not fit in this campaign (adds work for me, so I have less fun)", "does not fit in this campaign (straying from my vision will make it less fun for me/DM, and maybe the players as well)", "I do not like that (you are adding work by asking me to think too hard right now about something you do not need to have fun)". (I am sure there are others, by I believe I covered the common ones.)

How many on that list (or your list) are tied to "fairness"? A minority. In fact, DMs (and/or players) veto all kinds of "fair" options all the time, because they have other reasons that take precedence to fairness.

Logically speaking, fairness is a necessary but not sufficient reason for including an element in the game.

(And, of course, there are those rare cases we very purposefully include outright unfair options, but that not the run of the mill scenario.)

So, when someone says "This is fair, and I do not care about nuances about fun" they are fundamentally misunderstanding the very nature of how roleplaying games are usually played. Sure, everyone is entitled to their opinion, but it is very inauspicious line of persuasion.

pemerton · Oct 23, 2014

Sacrosanct said:
Either you don't understand how math works, or you're being disingenuous in your analysis. For one, needing a 10 on a d20 (50%) as opposed to an 11 on a d20 (55%) is not a 10% difference, so I highly suspect it's the first of those two reasons. And a 5% difference is not something that a result would be "heavily dependent on".

With all due respect, my maths is fine. The 5% you refer to is a measure of an increment. On it's own, it tells you nothing about effectivenss. If my chance of hitting is 1%, and yours is 6%, then you enjoy an incremental increase, compared to me, of 5%, but your effectiveness is 500% greater than mine (ie you are 6x as effective).

Whereas if my chance of success is 50%, and yours is 55%, then the incremental step-up of 5% confers an increase in effectiveness of only 10%.

billd91 said:
The rate of increase isn't the same which could allow you to bias the argument by cherry-picking your target numbers.

I didn't cherry pick any numbers. I used the target numbers that [MENTION=15700]Sacrosanct[/MENTION] stated in the post to which I replied.

billd91 said:
f you want to argue the significance of the difference from the point of view of the game as a whole, you have to take a more objective viewpoint and look at the difference from the whole range of results. A +1, from that point of view, is a 5% increase in hitting the target number, any target number.

This is nothing to do with subjectivity or objectivity.

There is not a 5% incease in effectiveness. In fact, when success is determined on a d20 it is impossible to have an increase in effectivenss as small as 5%, because the smallest possible increase in effectiveness - from 19 in 20 to 20 in 20 - is a 1 in 19 increase in effectiveness, which is greater than a 5% increase.

But it is true that the likelihood of succeeding increases by a 5% increment.

At 1st level, prof bonus is +2. What is a typical target number? Let's say that it ranges from 10 to 15.

So, with a stat bonus of +2 the roll needed for success ranges from 6 to 11. Hence the chance of success ranges from 15 in 20 to 10 in 20.

With a stat bonus of +3, the roll needed ranges from 5 to 10. Hence the chance of success ranges from 16 in 20 to 11 in 20. That is an increase in effectiveness of 1 in 15 (more than 6%) to 1 in 10 (exactly 10%). Is that neglibible? Over the course of play I don't think that it is. After all, profiency bonuses of +1 are regarded as non-neglible. Magic swords +1 are regarded as rewards. Rings of Protection that confer +1 to AC and to saving throws are valued items.

If the stat bonus is +4, the roll needed ranges from 4 (17 in 20) to 9 (12 in 20). That is an increase in chance of success, compared to +2, of over 13%, and 20%, respectively. I think that is obviously non-neglible.

Sacrosanct said:
It seems your problem is that you're basing your argument on the extremes, assuming the highest values will be rolled.

When you're using analysis, you need to look at the average likelyhood of the result, and not cherry pick the highest possible result.

No.

As I explained in the post to which you replied, I am not concerned with expected values. Expected values are useful for predicting outcomes for repeated trials. And they can help us anticiate outcomes in circumstances of uncertainty. But they tell us nothing about the actual stat distribution in this particular game, here and now.

My whole point is that rolling permits higher values, or lower values, depending on what is rolled, and in certain RPGing contexts that is an unfair consequence, because leading to arbitrary but significant differences in character effectiveness.

If you want random generation of stats but want to avoid the sort of unfairness I am talking about, the solution is fairly simple: use the point buy rules to generate 6 or 10 or 12 or 20 or however many arrays, and then have players roll a die of the appropriate size to determine which array they use for their stats. This way there is randomness rather than the uniformity that [MENTION=2]Piratecat[/MENTION] and others have expressed concerns about, but there is not the arbitrary difference in character effectiveness, which for my preferred approach to RPGing is an unfairness.

JamesonCourage · Oct 23, 2014

pemerton said:
Nonsense. For instance, if every PC has 18 STR and 6 INT, then the players playing fighters would have a better chance to impact the fiction via their PCs than would the players playing wizards.

I actually think the ability to impact the fiction is much, much more reliant on factors other than stats.

For example, proactive players have a much higher chance to impact the fiction because they engage with it more.

Players that have more robust mechanical options for changing the scope of play (say, the Teleport spell in 3.5 D&D) will have a better chance of impacting the fiction. Of course, I mean when it comes up (Teleport is much more useful in a large sandbox setting than urban campaigns located in a single city).

Players that have abilities that match the campaign theme have a much larger chance of impacting the setting than those that don't. Another example: in the past five sessions at my table (about six hours each), there have been three combats. However, there has been ample discussion, investigation, knowledge checks, and the like. Players have a much higher chance to impact the fiction based on skill choice and character build than based on stats. (A player with 6 Int that invested in investigation, negotiation, and knowledge would have a higher chance of impacting the fiction than a 18 Str / 6 Int Warrior with none of the skills listed.)

I'm not saying that stats don't matter. But they're only one piece of a huge puzzle that makes up the list of factors that come into play when seeing which player gets to impact the fiction the most. And sure, you can say "all other things being equal, stat generation and distribution matters greatly." While that's theoretically true, that's not how things generally work in a practical sense; all other things aren't equal. Some players enjoy the spotlight and are proactive, others don't mind following along and making less rolls. Some players pick mechanical options that fit their concept regardless of the campaign pitch, even if it's less useful, while others take the campaign pitch to heart and come up with characters with very "relevant" mechanical choices. Some players want something simple ("I hit it in the neck") while others choose something with many mechanical options (Teleport, Plane Shift, etc.).

Again, I'm not saying that stats don't matter. I'm just disputing what you've said above.

pemerton · Oct 23, 2014

dd.stevenson said:
the only things stat increases and level increases have in common is that they are both objective increases in character power, and they accomplish this partially by increasing some of the same numbers. I don't think that's sufficient grounds to claim that rolling for one makes as much sense as rolling for the other. Or else, if it is sufficient grounds, then I suppose we're free to claim that rolling for any quantifiable element of the game makes as much sense as rolling for any other quantifiable element. Right?

Ridley's Cohort said:
Exactly.

Sitting in a box in my garage is a copy of Chivalry & Sorcery (1st edition), published circa 1977, which predates the publication of AD&D.

In that game, you roll stats, you roll social status (yes, you may be a peasant who needs to grovel while near their "betters" in civilized lands), you roll your race, you roll your alignment(!).

At least at some superficial level, rolling for this stuff is fair, because everyone rolls on the same table.

Just adding to what Ridley's Cohort said:

In Classic Traveller PC generation involves rolling for stats, admission to a desired profession, the possibility of conscription, number of terms served, skills acquired, and benefits received upon mustering out.

In Stormbringer, stats, nationaity (= race), class, some skills, and starting money are all rolled randomly. Character progression is also determined by random rolls. Runequest is similar.

In D&D, starting money used to be rolled for on 3d6; in AD&D there were still rolls, but the amount varied by class; in 4e starting money was fixed at a uniform amount for all classes; in 5e starting gear can either be fixed (by class and background) or rolled for (by class, and on mutiple d4 tending thereby to produce tight clustering around the mean).

In D&D, when stats are rolled for, practices differ widely over whether players get to allocate the stats, or adjust them (as in Moldvay Basic) or choose from multiple rolled sets (as per some of the options in Gygax's DMG) or have to take the result of a single set of rolls. The latter can become tantamount to rolling for class, especially in AD&D where many classes have rather strict ability score requirements.

So it seems to me that rolling for any quantifiable element does make as much sense, a priori, as rolling for any other quantifiable element.

Of coures, no actual game determines these things purely a priori! We're looking for approaches to PC build that are fair and fun relative to our RPGing goals. Because those goals differ, so will the appropriate approaches.