There was a comment on one of my answers that stated that all instances of average or dice rolling are treated the same by the rules. Specifically mentioned are players gaining HP, Monster Damage and Passive checks.

For Monster Damage in the DM Basic rules (page 6 v0.1) it reads:

Hit. Any damage dealt or other effects that occur as a result of an attack hitting a target are described after the “Hit” notation. You have the option of taking average damage or rolling the damage; for this reason, both the average damage and the die expression are presented.

For player's gaining hp when they go up a level in the Basic rules (page 10 v.01) it reads:

Each time you gain a level, you gain 1 additional Hit Die. Roll that Hit Die, add your Constitution modifier to the roll, and add the total to your hit point maximum. Alternatively, you can use the fixed value shown in your class entry, which is the average result of the die roll (rounded up).

Passive checks explicitly states that you do not ever involve any die rolls. (Basic Rules V.02 page 59)

A passive check is a special kind of ability check that doesn’t involve any die rolls. Such a check can represent done repeatedly, such as the average result for a task searching for secret doors over and over again, or can be used when the DM wants to secretly determine whether the characters succeed at something without rolling dice, such as noticing a hidden monster.

The one other place the concept of taking an average vs rolling dice comes up is with Rolling Stats for PCs there it makes it very clear (and is arguably not really the average)

You generate your character’s six ability scores randomly. Roll four 6-sided dice and record the total of the highest three dice on a piece of scratch paper. Do this five more times, so that you have six numbers. If you want to save time or don’t like the idea of randomly determining ability scores, you can use the following scores instead: 15, 14, 13, 12, 10, 8.

Is there a general rule that can be derived here, (Are there enough data points in the rules to create a general principle?) or is the different use of language in all the instances make them all unique situations which need to be interpreted independently?

Additionally, if there is a general rule, can it be used anytime you want to "save time" or "reduce randomness"? (And if not, why not?)


4 Answers 4


Yes, the general rule that can be derived is that it's either-or: Either roll, or pick; not both roll and pick and then choose the best one or some other permutation that includes extra decision steps not described.

D&D 5e is written in "natural language", which is disparaged in some edition-warry parts of the Internet right now, but only means that the game is written so that the most obvious reading is generally going to be the right one, so that the game is understandable to the majority of readers who never read errata or participate in online debates about RAW, RAI, LFQW and other jargon.

In a natural language reading of all these rules, you either pick or roll, not both. If you were meant to do both, it would clearly say that, because not clearly saying that you're supposed to do both would be misunderstood by the majority of readers, if that was the intent. Since it doesn't say to pick and roll and take the best, the normal interpretations of "or" and "alternatively" are in force: either-or.

Extending the general principle of "when the game says to pick or roll, you pick or roll but not both" to other circumstances in the rules is a non-starter, because it just doesn't apply to any rules that it doesn't already apply to.

That principle isn't what your last question assumed would be derived though, so for completeness, let's now address the general rule that it assumes could be derived: that any roll could be instead replaced by taking the average, maybe if you want to save time or reduce randomness.

Extending "take the average" to other parts of the game where it's not mentioned as an option could be done as a house rule. It would probably not break most uses, and would likely find some corner cases where the extra predictability would break things and produce something that is off the average, but you'd have to play with the house rule for a while to find those, as is usual houseruling procedure.

  • Off the top of my head, it would definitely break to-hit rolls. If nothing else, this is the clearest counter-evidence to suggest that it could have been meant as a derivable general rule.
  • As an example of taking the average producing subtly skewed, non-average results, a wild magic sorcerer who knew that they could take the average damage of a surge effect might end up risking surges more often, thereby producing them more often than they would on average if all the involved rolls were actually rolled.
  • Similarly, not rolling for healing means you can attempt to plan farther ahead knowing not only your current resources exactly, but also one step ahead into the future of your resources with reasonable accuracy. This would alter decision-making, skewing play in who-knows what different direction.
  • 1
    \$\begingroup\$ @gatherer818 I'm fairly certain, as this came out of a dispute we were having in comments about which of either-or vs both-and-take-best interpretation of the rules for HP. The question might need some clarification if that's what you got from it. :) \$\endgroup\$ Aug 19, 2014 at 9:44
  • \$\begingroup\$ @gatherer818 I've clarified intent. You are both right. \$\endgroup\$
    – GMNoob
    Aug 19, 2014 at 9:46
  • 1
    \$\begingroup\$ @SevenSidedDie lol @ wild magic sorcy taking avg of WMS roll. that'd be funny and...pointless \$\endgroup\$
    – wax eagle
    Aug 19, 2014 at 11:40
  • \$\begingroup\$ In regard to attack rolls, the existence of the “Clockwork Amulet” in Xanathars guide allows a player to take 10, can only be used once per day. The limitation of once per day evinces that the general rule of taking the average would not apply to attack roles or otherwise the apparent magic of this device is nullified. \$\endgroup\$ May 4, 2022 at 13:37

Two of the rules you've quoted are actually a symptom of another rule, one of them is an exception to it and the other is unrelated.

All four of these are special rules and unrelated. I don't think anything can be derived from looking at them together. The monster damage and passive checks rule is subject to D&D's "always round down" rule. Hit dice at level up explicitly breaks that rule.

The ability score array...well, it doesn't have anything to do directly with dice averages as far as I can tell (I mean it probably does, but it's a lot more technical than simply taking an average and rounding up or down).

I'm not sure there is anything to be gained by trying to define a general rule here. As SSD points out, this is a natural language text and trying to form general rules from something like that can actually be dangerous (you can impart a much more complicated interpretation on something that a plain reading won't turn up).

Taking the three relevant rules we'd get "when you roll dice, you may take the average and sometimes round up, sometimes round down", not only is this not a very useful rule, it's also incorrect a lot of the time. d20 rolls for attacks and active skill checks should use the average (all the randomness is gone from combat). We don't want to use this for PC damage (you totally could, but it's not permitted anywhere in the rules, and it would make combat less interesting). So on and so forth.

Basically, take caution when seeking general rules, this is a good example of a place where we should just simply take the rules at their face value and move on instead of looking to derive a larger rules meaning from the combination (it could well be that the DMG will present an always use averages option, but we don't have one codified in the PHB text in the way the question suggests there might be).


When the rules indicate you can choose between the average and roll, you may do so. The DM especially may use average monster damage to both speed up play (especially with lots of monsters) and minimize the chance of randomly killing a PC (randomness never favors the players).

When making stats, that is an acceptable array and not necessarily indicative of any "average".

In other situations, you must roll. Players don't calculate their average damage on a fireball and then apply it every time they cast. PCs are dynamic beings that shine or fail spectacularly, not apply the same numbers by rote.

However, letting them take average on hit points is just a means of not allowing the dice to completely cripple a character because ONE important roll failed a few times in a row (again, randomness and players).

  • \$\begingroup\$ Ah, when you said "You're both right", I thought you meant you were asking that in addition to the question I'd already answered. Answer reverted. \$\endgroup\$ Aug 19, 2014 at 10:07

These are all special cases. In general, outcomes are decided by rolling. This is part of the thrill of the game. If there is a general rule here, it's this: Taking the average favors the PCs, so there are some particular places where it's a good idea, but it shouldn't be overdone.

Case by case:

Point-buy or array for ability scores

This is primarily about inter-player balance. If someone gets really unfortunately rolls, their character could have a lifetime handicap, or in the other case, be a superhero compared to the others at the table.

Imagine starting the game by picking someone at the table and telling them that they add 3 to all of their d20 rolls for the next year of campaigning, and someone else and telling them to subtract 3.

In my opinion, rolling works best if your group changes characters frequently — either you do a lot of one shots, or you have a high expected mortality. (In the former case, variety can be fun; in the latter, the stronger characters will tend to survive and the weaker ones — well, better luck next time.)

There's also another reason here: many gamers are a litttttle cheaty when rolling when no one is really looking. Not in a malicious way, but tending a little towards "okay, that was a test roll" or "that guy didn't come out so well; I'll just roll up five or six characters and pick my favorite" — and the favorite will happen to have a pair of 18s. Asking players to use one of the fixed methods lets them create characters at home without this temptation.

Hit Points

Exactly the same.

Monster Damage

This has two reasons. First, the DM may be running a whole horde of monsters, so reducing math speeds up the game. I haven't used 5e much yet, but in 3e with higher level monsters with lots of dice to roll, this can be significant. I often used averages for the "average" enemies, and rolled for the more unique ones — bosses, NPC rivals, legendary opponents, and so on.

Second, an unfortunately high roll can kill a player outright, when maybe you didn't mean to. Using averages prevents this, which is especially important if you don't like to fudge things.

Passive ability checks

Here, the answer is in the name. These checks are actually "always on", not individual occurrences. Most crucially, the DM cam easily tell if you might notice or succeed at something without raising player suspicions by asking for a check for no apparent reason. It's also to the players' advantage because they can be basically aware of their chance of success or failure in a given situation — time doesn't need to be wasted on worrying about a low roll and rechecking multiple times just in case.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .