Table of 5 is the consensus group size

tl:dr; 4-5 players appears as the common group size even where more players available in friend circle or gaming club

Older surveys asked how many people were in the broader circle of friends or club members which suggested a larger number of players at a table.

Multiple surveys across forums, from twitter and as academic research indicate table sizes of 4-5 players.

Comparing this to the older figures for gaming clubs and friend circles it implies there was a single table running at a time amongst most groups with the suggestion that second tables of typical size for most groups would share players as ~80% of groups were 8px or smaller and would not have fitted a second concurrent table or more than 4 players.

Another interesting note is that 4-5 players breaks across the 'weak/average' threshold for most Adventurers League games - 5 players of equivalent APL to the adventure will find the challenge average, 3-4 players will be weak. This implies that many DMs are having to tune this adventures from 'as written' before being able to run them.

An additional point could be that the 'ideal group size' identified from the Kind DM twitter survey is smaller than most players actual tables (as measured directly by the subclasses survey and Trygstads thesis) implying many players are at ~5 person tables and would prefer to be at 4 person tables.

Preferred classes correlate to favourite aspects of gameplay

tl:dr - class preferences correlate to preferred aspects of play, as seen in Reddit 2014 survey

Cross checking responses to D&D Survey (N = 1734, 2014) which asked “What are your favourite class/gameplay types?” and “What aspect of gameplay do you prefer?” shows:

- Everyone is here to role play >40% of preferred aspect of play for everyone
- Problem solvers prefer defensive combat spellcasters
- Combat oriented players prefer fighter and sneaky types
- No-preference lower in jack-of-all-trades and ranged fighter types suggesting these are play-styles more chosen on purpose than others

Broadly coherent with ‘spell-casters for complex play’ and ‘martial classes for getting stuck in’ straight-forward combat oriented play that was hypothesised from the different preferences seen in different player populations broadly defined as 'general players' and 'enthusiasts'.

Comparing range of race choices across surveys of D&D players

tl:dr; humans are most popular but not under all circumstances nor by huge margins

Update: clearer graphics and more recent data can be found here.

Following an update by the developers of Baldurs Gate 3 where they encouraged players to try more exotic race builds, it is interesting to look at the spreads coming out of other surveys.

The variation with race options has a wider spread compared to classes and more variation between surveys - where 2 archetypes were clear for forums and general users in class choices, this pattern is weaker here.

We can see that across all these surveys - 3/4 of general players prefer 'not-human' and among forum respondents from 3/4 to 5/6 of players choose races other than humans with indication that more people choose something other than humans over time.

This all suggests the efforts of the Baldurs Gate 3 developers will not go to waste.

Rate of new joiners to D&D has increased significantly

TL;DR the number of D&D players is expanding rapidly with 50% of players joining since 2015.

In the previous post we looked at ages, in this post we look at length of time people have played

The fraction of new joiners increases with the more recent surveys. Surveys of players in 2019-2020 show ~ 80% joined since 2010, ~ 50% joined since 2015.

The 'old school' profile highlighted by the OSR 2019 survey shows flatter growth over the years - 80% of respondents played at least 30 years, 50% played at least 14 years.

This suggest the significant new interest in D&D is driven by 5e which is coherent with media statements about 'explosive growth' by Hasbro.

Comparison of Hasbro player age estimate against surveys of new generation and OSR gamers

Comparing the age profiles from many surveys scraped from the web throws out a couple of different profiles:
- the contemporary forum user
- the old school gamer
- Hasbros estimate of the general populace

Two high-detail surveys return a year by year curve we can use to classify populations:
- a twitter demographics survey 2020
- the OSR survey by Questing Beast et al in 2016

Comparing these to the WotC 2000 and Hasbro 2020 estimates of player populations shows that the publishers are counting a population of younger gamers who are not appearing in forum responses.

This gives relatively distinct player profiles:
- OSR survey, show a population with half younger than 31, 80% younger than 42.
- the new generation player population appears to be half younger than 24-25 and 80% younger than 29-34.
- Hasbro estimates 50% of players younger than 27 and 80% younger than 36; closer to the new generation.

The split between the Hasbro curve and the general population response is an indicator that the players of D&D and those talking about it online are not perfectly overlapping.
- it suggests there are players too young to participate in online discussions (~12% under 12 years old)
- the Hasbro numbers also suggest there is a larger tail of older gamers undercounted in online discourse: survey responses would suggest only ~23% of players are over 31 years old but Hasbro estimates they are closer to 40% of the population.

Ranking 5e classes by preference within player populations

Splitting sources into Apps (Apps and reports from DnD Beyond) and Forums (surveys from Reddit, Giant in the Playground, etc) and aggregating the results weighted by number of participants lets us see the average differences between the two player populations.

Sorting the classes by greatest difference in preference (most preferred by app uses to left, most preferred by forum users to right) we see martial classes are preferred by app users, spellcasters are preferred by forum users.

Differences in 5e class preferences by source

tl:dr - preferred D&D 5e player class shows distinct difference depending on whether sourced from apps/online tools or from forum response.

From digging about on the internet I pulled together a number of data sets on preferred race and class combindations of PCs over the past ~4 years (2017-2020). I wanted to see if there was a change over time but instead I found a distinct difference depending on where the data was sourced from.

First - setting everything up according to date does not show much in the way of a distinct pattern.

However once this is sorted by source type we see a difference between the relatively smooth curve of the online tools and apps (orange) and the forum survey responses.




Initial hypothesis is that there is a large block of players who just want to sit down and play that tend towards melee types - "give me my sword and board and lets roll dice!" - compared to those who answer forum surveys that show more interest in spell casters both divine and arcane.

Aggregating these into a pair of archetype finger prints weighted by the number of contributors to each source highlights this difference. A single forum response deviated significantly from the forum archetype: a survey on RPG.net in 2017 requesting only those characters that responders had played for at least 1 session.

Data sources include 2 sets of raw data (OGANM app, Reddit Subclasses 2019) and 1 estimate based on sub-forum pageviews (RPGBOT) - the others are processed outputs from the creators of those sources (e.g. the report-out by a survey creator, DnDBeyond Dev webcast, etc.)