D&D General Data from a million DnDBeyond character sheets?

[ichabod]

Does anyone see the code for getting the characters?

I wanted to check to see if these are explicitly shared.

I dint think characters are available unless explicitly made public which, if correct, means these were. Likely to be used in a shared game.

Or are the free tier characters public?

I have not seen anything from the guy who scraped the data about how he did it except that he used Python requests. I poked around online and I think I found the url he is using for his request. If I'm right, and I'm understanding how he read the json, all the ability scores in the data are wrong. The 'stats' key in the json just shows the initial stats. It doesn't show the racial bonuses and I'm not sure it shows the ASIs. I think this explains why there are so many characters in the data with the starting array and no racial bonuses. I have confirmed that a new character I made does not show the same stats online as the 'stats' key in the json. I have also confirmed that one of the characters in the data with the starting array does not have the starting array when you load it online, but does have it in the stats key in the json.

To your specific question about availability, the character of mine that I pulled had not been shared, but was from a free account. It worked in another browser where I was not logged on the D&D Beyond. If someone could give me the character ID for a character from a paid account that had not been shared, I could test it to see if I can view it.

I can see a lot more data in the json than what is in the dataset. I can see other HP fields, user ID, ability generation method, and others. I haven't had a chance to look through it all, though.

 

[Oofta]

Does anyone see the code for getting the characters?

I wanted to check to see if these are explicitly shared.

I dint think characters are available unless explicitly made public which, if correct, means these were. Likely to be used in a shared game.

Or are the free tier characters public?

They are not shared as far as I can tell. Anyone can see the data of random PCs.

EDIT: When you create a character by default they are created as "Public" I would assume so that other people in your campaign can see them. However, if you're like me I've never scrolled down to the bottom to change that. So a significant percentage of the characters can be viewed by anyone. If I cared, I'm sure I could likely get data on thousands of PCs unless DDB can detect the scraping and stop it.

 

[Lanefan]

It would be interesting if it was possible to narrow the characters down to ones that have been actively played.

As long as there's not a threshold for how long or how many times it was played. A character that enters play, dies in its first session, and is never played again is for these purposes just as statistically valid as a character who lasts for 300 sessions.

 

[TwoSix]

In practice, when a completely new player wants to play a spellcaster they usually ask me or an experienced player for help choosing. If I am helping, I just go for the standards (guidance, fire bolt, sleep, magic missile, bless, healing word, etc.) and encourage them to spend some time reading though the list and seeing what they might want to swap to make the list feel more like their character.

Which is fine, but I think even simpler than that is good. Ideally this class won't need to even look at the spell list because of a class feature.

Besides 1d8 HD, simple weapons, light armor, they should get within the first two levels:

1) Choice of stat mod for attacks between Int, Wis, and Cha. Possibly tie a ribbon feature to the choice, but nothing mechanically important.

2) A ranged attack of 2d6+stat mod, 60' range. Energy type of player choice (not force, possibly not psychic), default fire. Attack deals 1d6 extra damage every odd level (up to 11d6 at 19th).

3) Bonus action at-will utility, strength roughly equivalent to BA help at range. (Give advantage, give one enemy disadvantage, etc.) Small menu of options (3-4) to allow for customization.

4) Small defensive ability, preferably something simple and active, to push the class close to fighter level defensiveness, but through "magic". Maybe a magic ward that lowers damage taken by 1d4 per attack?

 

[Sorcerers Apprentice]

As long as there's not a threshold for how long or how many times it was played. A character that enters play, dies in its first session, and is never played again is for these purposes just as statistically valid as a character who lasts for 300 sessions.

It's quite possible that certain classes are more likely to die at low levels than others, so indeed it might affect the accuracy of the results if such characters were excluded.

 

[ECMO3]

I suspect the default options are so over represented that it's going to be hard to draw any real conclusions about player preferences from much of this data.

I don't think so, at least not if your question is what are player preferences overall.

The default SRD options are available to a lot more players, more importantly though they are the only options available to some players, other players only have PHB options, others have PHB and XGE, others have PHB and Spelljammer only, others have PHB, TCE and MOT, and others have the SRD and individual options they have purchased a la carte......

If you are asking the question - What are the preferences for the few players who have access to all the available WOTC books (or worse all the WOTC and 3rd party material)? Then yes if that is the question then the default is necessarily overepresented, but that is a different question and one which is representative of a very small portion of the actual player base.

On the other hand if your question is what are the player preferences overall, considering all players regardless of the amount of material they own, then I do not think they are biased by the default options.

 

[Thomas Shey]

Observation from numerous groups I've played with both P&P and online, both as a DM and a player.

In 5E I am talking about my personal observations from around 30 different campaigns with a different mix of players in each one and another 20 or so one-shots, including players from at least 9 different countries.

While, fair, as people point out to me when I make similar comments in other contexts, its still a drop in the bucket, and you can't count on selection bias not being a factor.

 

[Yaarel]

My takeaway from this graph is, ENWorld (2019) forumers are representative of DnDBeyond (2023) players.

 

[EzekielRaiden]

I don't think so, at least not if your question is what are player preferences overall.

The default SRD options are available to a lot more players, more importantly though they are the only options available to some players, other players only have PHB options, others have PHB and XGE, others have PHB and Spelljammer only, others have PHB, TCE and MOT, and others have the SRD and individual options they have purchased a la carte......

If you are asking the question - What are the preferences for the few players who have access to all the available WOTC books (or worse all the WOTC and 3rd party material)? Then yes if that is the question then the default is necessarily overepresented, but that is a different question and one which is representative of a very small portion of the actual player base.

On the other hand if your question is what are the player preferences overall, considering all players regardless of the amount of material they own, then I do not think they are biased by the default options.

Okay. The problem is, people are not making claims of the form: Fighters are popular relative to their accessibility.

Instead, people are making claims of the form: The Fighter is universally popular, therefore this specific implementation is why it is popular.

Both the premise and the conclusion are suspect. The former depends on questions we can't answer, assuming that frequency of use on DDB is equivalent to being popular, which doesn't hold (after all, many of these characters are NPCs!) Hence, we don't know Fighter is universally popular, and why I've repeatedly said people make far too strident claims for the data. The only thing the data unequivocally says is that many people who use DDB create Fighters, and relatively fewer create Wizards. It does not tell us that Fighters are universally popular.

Moreover, even if it DID tell us that, it would NOT follow that these mechanics are why. I've said, repeatedly, the Fighter is always popular. AFAIK, in every edition, it's been either #1 or at least top 3. Why? We know the quality is uneven--consensus is the 3e Fighter was really quite bad--yet even when it's been bad, it has remained "popular," that is, chosen and played frequently. How to explain that? Well, per Occam's razor, the Fighter's popularity is orthogonal to whether it is good or not! (Same argument applies to humans. Humans are ALWAYS popular. Yet the 3e human kinda sucks! But it was still much more popular than better alternatives.)

D&D players pick based on theme. It's not that they do not at all care about mechanics, though they may not strictly be able to identify what they like mechanically or why. Instead, it's that their picks usually occur before they even look at the mechanics of something. Hence, whether it is good or bad is irrelevant to what people choose to play--unless it is so egregiously, unconscionably bad they can't excuse it. That's why I quoted the Declaration of Independence, because it so neatly summarizes this situation.

 

[Lanefan]

My takeaway from this graph is, ENWorld (2019) forumers are representative of DnDBeyond (2023) players.

Am I the only one who finds that format of chart/graph nearly impossible to decipher?

Still, I appreciate the work that went into making it.

One request, if this is ever re-done: make the line colours a time-based rainbow going from blue (oldest) to red (newest) so we can see if there's any pattern changes over time.