Humans, Fighters, and Life Domain Most Popular On D&D Beyond

[FrogReaver]

While nothing is perfectly representative, I would say their 9 million characters might be more representative than your three gaming groups. I expect DDB takes your stars with a grain of salt.

I find it bizarre that in 2019 we still have to say “anecdotes are not data”. Their data isn’t perfect, of course. But it’s not an anecdote.

At least we know all the characters in the games he is stalking about are actually being played

[ad_hoc]

While nothing is perfectly representative, I would say their 9 million characters might be more representative than your three gaming groups. I expect DDB takes your stars with a grain of salt.

I find it bizarre that in 2019 we still have to say “anecdotes are not data”. Their data isn’t perfect, of course. But it’s not an anecdote.

Yeah.

The only people who likely have better data are WotC.

I'd love to hear this sort of stuff from Mearls. Here and there he drops statistics if you pay attention.

[Jay Verkuilen]

This data is presented as a high-level look at distribution of race, class, and subclass selection for active characters on D&D Beyond. It serves that purpose just fine.

It could certainly "confuse" or "amuse" those who want to read too much into it or think it is trying to achieve a different purpose than it is.

When you have a fairly numerate fan base, expect issues like the fact that the percentage of Life Clerics exceeds the percentage of Clerics to generate a WTF?!? reaction. This happened because comparing the one-way margin of subclasses is a strange thing to do. Not actually having your data, I'd look at the breakdown within class, for instance, of Clerics, which archetypes are chosen? Comparing Life Clerics to Champion Fighters is... odd. I also think that giving the breakdown as "Given what Class is chosen, which Archetype is then chosen?" really helps answer which is the most popular within a particular class. It's less bothered by multiclassing, too.

While you're right that back of the envelope calculations aren't as good as having the real data, they are often quite useful. For example, it turned out that assumptions about binning made threw off the mortality calculations that were part of what won Angus Deaton the Nobel Prize in economics a few years back. Later on, some statisticians, including the (in)famous Andrew Gelman, figured out there were mistakes made that made the story much less dramatic. These started from back of the envelope calculations.

As you said elsewhere I doubt the dominant trends of "humans, half elves, and elves are the most popular, along with fairly straightforward classes" would be undermined by any re-analysis.

Actual decisions could be (and are) made from this data. For instance, when looking at future subclass design, maybe the bar would be set at Life Domain and not at the least selected domain. It's a safe bet to say that draconic sorcerers "do well" in the community and other design should target that, or that perhaps people would want to know more about half-elf culture since they are so often chosen. The data never pretends to dig any deeper than that.

Oh I'm sure there's a lot of useful data in there. As I said, process data is indeed very valuable, though still limited.

[cbwjm]

Maybe I should rephrase. I didn't think you could use homebrew classes and subclasses.

You can use homebrew subclasses. Homebrew classes aren't available yet.

[Dausuul]

Less than half of gaming groups use feats.

No. Less than half of characters (on DDB) use feats. We have no idea how many gaming groups use them.

In a group which allows feats, not all of the characters will actually have feats; many players will choose the stat bump instead. By definition, in a group which does not allow feats, none of the characters will have feats. So the number of characters who have feats is almost certainly understating the number of groups that allow them.

[ad_hoc]

No. Less than half of characters (on DDB) use feats. We have no idea how many gaming groups use them.

In a group which allows feats, not all of the characters will actually have feats; many players will choose the stat bump instead. By definition, in a group which does not allow feats, none of the characters will have feats. So the number of characters who have feats is almost certainly understating the number of groups that allow them.

Mearls stated that less than half of gaming groups use feats.

I trust the data he has.

[FrogReaver]

This is my point that no one here (except me) knows the actual dataset and many assumptions are being made. Multiclassing and homebrew subclasses are going to throw off any napkin math you all can do.

If we entirely removed multiclass characters, for instance, does this actually still give us the most accurate look at class popularity? Is a class still popular if it is chosen, even if for only a level or two?

We can absolutely remove multiclass characters (and we have before), but I can tell you the distribution doesn't actually change that much.

I'll have to take your word for it. I'm surprised that's the case, probably more so than I should be lol. That's very interesting.

I'm also really curious about whether there is anything you can tell us about how active characters are determined?

[Dausuul]

Mearls stated that less than half of gaming groups use feats.

I trust the data he has.

Where did Mike Mearls say this? The only place I have seen this assertion made is in a thread title here on ENWorld, and the data cited was showing that less than half of characters use feats - in other words, the person creating the thread misinterpreted the data.

[ad_hoc]

Where did Mike Mearls say this?

It's in one of his Happy Fun Hours.

He tends to drop info like that here and there.

[Kevin Smith]

I think what "useless" meant was "miscalculated." As in "how is it that the subclass percentage exceeds that of the base class given that all life clerics are, necessarily, clerics?" That's really, really... weird.

Part of it might be due to the fact that archetype is only chosen at particular levels that vary for the class in question. Clerics choose that at 1st level while many others don't choose until 3rd level. Also the class breakdowns are pretty much guaranteed to have multiple counting due to multiclass characters.

So, yeah, these data are misleading as presented, or at least confusing.

"Maybe misrepresented," or at least, "not given proper context" might be the culprit too. If 62.8% of active PCs are levels 1-4, and 8.4% of active characters are life clerics, could the 10.2% of active subclasses be due to the fact that Clerics start with a subclass? Fighters and Rogues are the two most popular choices representing 24.4% of active classes (in total) and they don't receive a subclass option until 3rd level. So, it seems it takes quite a bit of interpretation to make meaningful sense of the data without context provided by D&DBeyond.

But I think we tacitly agree the presentation of the data could be improved.

[Yaarel]

Cleric and Ranger − people play these two classes, but eventually reject them.

They are among the top four classes at tier 1, but by the time they reach tier 4, Cleric and Ranger disappoint among the bottom four classes.

[Doc_Klueless]

They are among the top four classes at tier 1, but by the time they reach tier 4, Cleric and Ranger disappoint among the bottom four classes.

That surprises me in that, in my group, the two classes that have lasted the longest have been Cleric and Ranger.

And, yes, I know Anecdote =/= Data. I just thought it was interesting.

[5ekyu]

Cleric and Ranger − people play these two classes, but eventually reject them.

They are among the top four classes at tier 1, but by the time they reach tier 4, Cleric and Ranger disappoint among the bottom four classes.

A slice of data I would love to see for higher tiers is "played from lower" vs "started at tier".

It should be an easy cut given past days is stored for things like rests.

I can see a pretty big skew between generating a character for long term play that gets to high level and generating a character for a specific high level event. Are the top tiers rare enough to be influenced by the number of "let's go level 20 arena tonight" subset?

This is also a place where MC treatment would matter.

[FrogReaver]

"Maybe misrepresented," or at least, "not given proper context" might be the culprit too. If 62.8% of active PCs are levels 1-4, and 8.4% of active characters are life clerics, could the 10.2% of active subclasses be due to the fact that Clerics start with a subclass? Fighters and Rogues are the two most popular choices representing 24.4% of active classes (in total) and they don't receive a subclass option until 3rd level. So, it seems it takes quite a bit of interpretation to make meaningful sense of the data without context provided by D&DBeyond.

But I think we tacitly agree the presentation of the data could be improved.

The subclass graph title says “active characters”. Normally that means the the whole graph represents all active characters and the percentage breakdowns are then shown relative to that whole.

In this case the graph actually should be labeled “active characters with a subclass”. That would be accurate but then it obviously forces the thought “what kind of characters don’t have a subclass” and then maybe even a thought about “fairness”. Such as fair is it to include level 1 clerics in that breakdown but not include level 1 fighters. If we do that isn’t the chart possibly just showing us that there are a lot more active level 1 clerics than active level 3+ fighters? Etc?

[Whizbang Dustyboots]

Weird that forest gnomes come in behind the other subraces. I think those are clearly the best for gnome illusionists. (Bias: I've played a gnome illusionist for more than a decade.)

[doctorbadwolf]

I must admit this information supports my belief that the Swashbuckler subclass was awful. I don’t know how they ever settled on that version but it really missed the mark.

Strongly disagree. It perfectly hit the mark for me. If I want a more technical play experi now with a swashbuckled, I’ll take 3-5 levels of Fighter and/or take the Martial Adept feat for some maneuvers, but that’s a mechanical preference thing.

Well I should start off by saying I certainly appreciate other points of view. Not trying to throw shade or anything.

I can see how this could be a fun subclass just not a swashbuckler per say. When I think of the classic swashbucklers such as Zorro or the Musketeers they are more than panache and feet’s of dexterity. They are also incredible swordsmen bar none. Technical masters of light fighting. That is just not represented in the current swashbuckler.

I think part part of it is that the swashbuckler does not fit well in any class. To me it is really part fighter, part rogue and even a little bard minus the spells of course.

i don’t mind the subclass it just doesn’t ring true as a swashbuckler. Swashbucklers don’t backstab they defeat you in a duel with wit, acrobatics and superior swordsmanship.

The SB rogue doesn’t backstab. They go toe to toe with enemies, besting them with superior swordsmanship, wits, and mobility. A level 3 SB rogue will strike first most fights, is lethally accurate with strikes as long as they can create an advantage or are fighting one on one with no other combatants immediately involved, and can whip around enemies with impunity, allowing fluid stunts without reprisal.

Combine that with Expertise in the right skills, and you’ve got Zoro, or the Musketeers. Although, for max fun, I’d play each musketeer as a different class to approach swashbuckling from different angles. It’s dnd, so I’d make Aramis a Dex Paladin with a couple levels of Bard.

I have been using homebrew a lot for my current game and so far DDB makes it pretty easy. Classes are not there, as was stated, but it's well put together so far.

It’s a great system for using homebrew.

Weird that forest gnomes come in behind the other subraces. I think those are clearly the best for gnome illusionists. (Bias: I've played a gnome illusionist for more than a decade.)

They make incomparable mage hunters, too. Adv on mental saves vs magic, Minor Illusion, and wholly mundane “speak with animals”? Absolutely boss tier.

[ClaytonCross]

This data is presented as a high-level look at distribution of race, class, and subclass selection for active characters on D&D Beyond. It serves that purpose just fine.

Just curious, I assume their is a window for "active characters", is it something like 30 days or 120 day? etc. I ask because I play 2 monthly games that shifts due to player availability. If its a 30 day window of activity I play the first two weekends one month, play the last two weekends the next month, and the D&D Beyond polling is done on in the middle of the next month, its possible that despite being active characters they would not have been active in the time frame of the active pole. This would of course effect more than my group and impact the numbers. I understand you can't get everyone and it maybe that your interested in 30 day range changes. I am just curious what the data polling window is, it that's something you allowed to answer. If you can't I do understand it there is a need to keep some things behind closed doors.

I am also curious about how you feel about players having access to character tags to aid D&D Beyond in poling from a personal view, though I understand, you being you its hard to answer personal views without it being taken official unless you add a clear disclaimer. Maybe even then.

Clarifying what I mean, I would be more than happy to Mark my characters:
"Untagged" Default?
"Theory crafting only - currently not in use"
"Alternate/Backup character - currently not in use"
"GM - NPC", "Player Sidekick - currently not in use"
"Player Sidekick - In use"
"In use Player Character" understanding this tag does not mean active unless it also meets active character criteria in X window of days.

Would this be useful, in your opinion, for improving data for D&D Beyond data poling or do you think lake of participation and status changes would just muddy the water? Also, I am 100% clear that even if you would like these feature it does not reflect a desire by the community or prioritization of such a request that it would ever be implemented or implemented in the near future.

If you can answer this I apricate it. If you can't I understand it. Thanks for reading if you get to it... I know I am not concise, sorry about that.

[Jay Verkuilen]

J
Clarifying what I mean, I would be more than happy to Mark my characters:
"Untagged" Default?
"Theory crafting only - currently not in use"
"Alternate/Backup character - currently not in use"
"GM - NPC", "Player Sidekick - currently not in use"
"Player Sidekick - In use"
"In use Player Character" understanding this tag does not mean active unless it also meets active character criteria in X window of days.

Would this be useful, in your opinion, for improving data for D&D Beyond data poling or do you think lake of participation and status changes would just muddy the water? Also, I am 100% clear that even if you would like these feature it does not reflect a desire by the community or prioritization of such a request that it would ever be implemented or implemented in the near future.

I don't know their data specifically, or their general structure, but a mechanism that more clearly reveals the intent of the user would very much improve the nature of the data, so long as it was cheap to gather and consistently filled out. Otherwise there's uncontrolled heterogeneity in the data. Upthread I offered myself as an example: Of the five characters in my DDB account, only one is an active PC. The others are either NPCs or, more commonly, test builds of some sort, which I find DDB to be particularly useful for. This is pretty typical for me. Is it typical for others? I have no idea, but without a method for the user to signal this, it's a guess.

(I speak as someone with a graduate degree in statistics who teaches the subject at a graduate school. That doesn't make me infallible, but does mean I have a better idea than Joe Average.)

[ClaytonCross]

I don't know their data specifically, or their general structure, but a mechanism that more clearly reveals the intent of the user would very much improve the nature of the data, so long as it was cheap to gather and consistently filled out. Otherwise there's uncontrolled heterogeneity in the data. Upthread I offered myself as an example: Of the five characters in my DDB account, only one is an active PC. The others are either NPCs or, more commonly, test builds of some sort, which I find DDB to be particularly useful for. This is pretty typical for me. Is it typical for others? I have no idea, but without a method for the user to signal this, it's a guess.

(I speak as someone with a graduate degree in statistics who teaches the subject at a graduate school. That doesn't make me infallible, but does mean I have a better idea than Joe Average.)

I have 19 characters. 2 are active . 17 are Theory Crafting. At one point I had 2 NPC mules as second characters which represented actual mule NPCs and held separate inventory (They died, so deleted for now). I had 1 for a bag of holding simply to separate inventory so it would not impact encumbrance, however I moved it to a Google Sheet use for the campaign so we can share it and it auto calculates weight changes on the fly without going through the painful notes and inventory functions of D&D Beyond. (Feature Requests are in for notes (with GM sharing) and containerized inventory that I hope will remove the need for these.)

So no your not the only one for sure.lol

[Mistwell]

Weird that forest gnomes come in behind the other subraces. I think those are clearly the best for gnome illusionists. (Bias: I've played a gnome illusionist for more than a decade.)

Their "Speak with Small Beasts" also makes them good Druids. Conjure Woodland Creatures becomes a better out of combat spell with that ability. Of course, as you point out a free Minor Illusion cantrip and bonus to Dex makes them excellent illusionists.