Class/race choices from 1 year of Baldur's Gate 3 vs tabletop 5e

Larian Studios, makers of Baldurs Gate 3, bless their transparent little hearts, have released another update to their player stats - now after a year of play. Once again we get class and race choices of the players. We got two previous updates - the weekend after launch and an update after four months of play that looked different to the typical tabletop choices despite being a pretty faithful implementation of the 5e ruleset.

So here is the rerun of that number crunching - with the launch weekend and 4 month stats in with this 1 year set and the patterns from a bunch of forum surveys and 5e character building apps.

On class the broad patterns stay the same - since the 4 month update - a completely different pattern to the tabletop game with Paladin and Sorcerer much more popular, wizards, warlocks and clerics way down off what the tabletop surveys show. What is especially interesting is that we now see a steadying of choice between the 4 and 12 month surveys compared to the launch weekend. Paladins and rangers launched well but faded off a bit - though paladin still stayed top of the heap. Clerics and fighters were less preferred on launch weekend but became more popular over time.

This full year stabilisation is interesting in clarifying the very different pattern to tabletop - reasons for which we can only hypothesise. Could be rules tweaks (healing via potion drinking leading to less perceived need for clerics?) or could be something to do with the solo play model - people who might play certain classes to flesh out a party of their friends around the table will instead do something different when they can just pick the composition of the rest of the party since they are NPCs.

Response on Race also shows stability from 4 months in - there was a bit of change from opening weekend at 4 months (elves, tieflings and githyanki getting more popular, dragonborn dropping off significantly, half-orcs by a bit) but from 4 months to the end of the year it is pretty stable. Total characters created over opening weekend were 1.7m then another 29.6m until the 4 month mark, then 49.5m more to the end of the year - in each case enough that a major swing in player choices should be enought to swing the total. All those new characters in the last 8 months were solidly aligned with those created as people got into their groove after opening weekend.

All this shows a huge difference to standard table-top D&D choices - whether the 'standard players' or the forum fans. At first glance, the popular choices seem to align to the 'origin character' choices available - but those are just 7% of the characters created.

Trying to pull that apart based on the 4-month stats it looks like:
Gale - Human Wizard - counts for 9% of human picks
Astarion - Elf Rogue - counts for 8% of elves
Karlach - Tiefling Barbarian - counts for 8% of tieflings
Shadowheart - Half-elf Cleric - counts for 7% of half-elves
Wyll - Human Warlock - another 5% of humans
Lae'zel - Githyanki Fighter - 19% of githyanki

So we should in theory shave those fractions off the various race columns to try and get a sense how free choice might look like but that does not change the relative stacks much - humans and githyanki take the largest trim and half-elves, elves and tieflings all take an equal reduction but that does not shift the stats around enough to get them anywhere near the table-top stats.

Humans, dwarves, gnomes and halflings are way down in BG3, drow, githyanki and half-elves are way up. One option is that aesthetic choices in the portrayal of the shorter races that is coming into play here - another is that game mechanics elements are driving play choices similar to the charisma-class lean in the class choices.

As before, take any reading into this with a big pinch of salt, different medium, almost certainly different target audience, etc. etc. We can have somee confidence in these numbers as they now reflect tens of millions of generated characters over a year of fairly intense play - Baldurs Gate 3 still pulling in 100+ players after a year - so this strong variation to tabletop class choices looks like a real thing.

See Class/Race archetypes in 1.2 million D&DBeyond characters and Player Class Stability from 1M D&D Beyond Sheets for more detail on the tabletop sets and sources.