D&D (2024) WotC On One D&D Playtest Survey Results: Nearly Everything Scored 80%+!

In a 40-minute video, WotC's Jeremy Crawford discussed the survey feedback to the 'Character Origins' playtest document. Over 40,000 engaged with the survey, and 39,000 completed it. I've summarised the content of the video below.

High Scorers

• The highest scoring thing with almost 90% was getting a first level feat in your background. This is an example of an experimental thing -- like advantage and disadvantage in the original 5E playtests.
• Almost everything also scored 80%+.

About The Scoring System

• 70% or higher is their passing grade. In the 70s is a thumbs up but tinkering need. 80% means the community wants exactly that and WotC treads carefully not to change it too much.
• In the 60s it's salvageable but it really needs reworking. Below 60% means that there's a good chance they'll drop it, and in the 40s or below it's gone. Nothing was in the 50s or below.

Low Scorers

Only 3 things dipped into the 60s --

• the d20 Test rule in the Rules Glossary (experimental, no surprise)
• the ardling
• the dragonborn

The next UA had a different version of the d20 Test rule, and they expect a very different score when those survey resuts come in.

It was surprising that the dragonborn scored lower than the ardling. The next UA will include new versions of both. The main complaints were:

• the dragonborn's breath weapon, and confusion between the relationship between that dragonborn and the one in Fizban's Treasury of Dragons.
• the ardling was trying to do too much (aasimar-like and beast-person).

The ardling does not replace the aasimar. The next version will have a clearer identity.

Everything else scored in the 70s or 80s.

Some more scores:

• new human 83%
• dwarf, orc, tiefling, elf tied at 80-81%
• gnome, halfling tied at 78%

Future installments of Unearthed Arcana

• The next one will have new ardling and dragonborn, a surprise 'guest', and a new cleric. It will be a shorter document than the previous ones, and the one after that is bigger again. Various class groups.
• Warrior group digs into something teased in a previous UA sidebar -- new weapon options for certain types of characters. Whole new ways to use weapons.
• New rules on managing your character's home base. A new subsystem. Create bases with NPCs connected with them, implementing downtime rules. They're calling it the "Bastion System".
• There will be a total of 48 subclasses in the playtest process.
• New encounter building rules, monster customization options.
• New versions of things which appear in the playtest after feedback.

Other Notes

• Playtests are a version of something with the assumption that if something isn't in the playtest, it's still in the game (eg eldritch blast has not been removed from the game). The mage Unearthed Arcana will feature that.
• Use an object and other actions are still as defined in the current Player's Handbook. The playtest material is stuff that has changed.
• Thief subclass's cunning action does not interact with use an object; this is intentional. Removed because the original version is a 'Mother may I?" mechanic - something that only works if the DM cooperates with you. In general mechanics which require DM permission are unsatisfying. The use an object action might go away, but that decision will be a made via the playtest process.
• The ranger's 1st-level features also relied too heavily on DM buy-in, also wild magic will be addressed.
• If you have a class feature you should be able to use it in the way you expect.
• If something is removed from the game, they will say so.
• Great Weapon Fighting and Sharpshooter were changed because the penalty to the attack roll was not big enough to justify the damage bonus, plus they want warrior classes to be able to rely on their class features (including new weapon options) for main damage output. They don't want any feats to feel mandatory to deal satisfying damage. Feats which are 'must haves' violate their design goals.
• Light Weapon property amped up by removing the bonus action requirement because requiring light weapon users to use their bonus action meant there were a lot of bad combinations with features and spells which require bonus actions. It felt like a tax on light weapon use.
• Class spell lists are still an open question. Focus on getting used to the three big spell lists. Feedback was that it would be nice to still have a class list to summarize what can be picked from the 'master lists'. For the bard that would be useful, for the cleric and wizard not necessary as they can choose from the whole divine or arcane list.

The playtest process will continue for a year.

 

[EzekielRaiden]

I think they probably just don’t feel the need to deep dive on their statistics, and just talk overall satisfaction, rather than boring 90% of video viewers with “this is how we derive the overall satisfaction percentage.”

It doesn't need a deep dive. "Ratings for dragonborn were 3.76 +/- 0.95 (two SD)." That gives REAMS more information than "70% of responses were positive." Even for a total layman, you now know that almost everyone gave a 3 or a 4, leaning toward 4--some people gave 5s, more than the people who gave 2s, and 1s are quite rare. With that, we can see that the response is favorable but not fervent.

And it is trivial to present that information in table form.

For goodness' sake, this is the kind of thing we demand from political polling, which almost everyone is at least loosely aware of. We know that polls have a listed "margin of error" and that that margin of error is relevant and needs to be accounted for. Giving something similar here cannot possibly be that difficult, if they really want to stick with proportion statistics rather than something more generally useful.

 

[Neonchameleon]

Yup. These surveys haven't had a statistician near them. And they're being used like a drunk uses a lamp post; for support rather than illumination.

They're a basic temperature check more than anything else.

 

[Jaeger]

I think you mean proportionally.

Yes, that thing. If it is better than what I said, then that's what I meant!

I think that's intentional. 5E is very much a player-focused game. Almost every single book, even the modules, includes PC options...because WotC knows there's 10 players to every 1 DM. Their fanbase is the players. Not the DMs. But, they need the DMs to run their games.

That's the thing.

WotC said that their average "campaign" lasts something like six sessions. I just don't see how that is enough time to engage enough players to become the hobbyist GM's that will support the game years down the road.

My son who is 13 doesn't care about optimization, and would rather get on with the story and creative descriptions of the action (i.e. the rules get in the way for him). He is a voracious reader, but finds the rules as presented in the PHB to be complicated and convoluted for no reason that he can discern. He did DM for one group, but he decided to run a non-DnD RPG game, which the group is enjoying much more. They are having fun, which is the whole point.

In my opinion; A lot of people of this sensibility would be far more satisfied with something like Dungeon World as a rules set. That being said, DW cannot compete with D&D for it's player network, and brand IP cachet...

So D&D it is.

Darn - I really do not like feats. Love that they are optional in 5E 2014 since it makes it easier as a DM to not allow them in my games. I guess I can just ignore them in 5E 2024. Since everyone gets them it should balance if I ignore them completely. Or I can just keep running 5E 2014 - I have all the books I need physically and digitally and it will essentially be the same game.

If WotC really wanted to make them optional, then they should have not put them in the PHB...

Of course, your table your rules. But in the wider D&D sphere, it seems that if it is in the PHB; players have the expectation that it's on the table.

WotC seems to be reacting to that, hence a feat at level one for D&Done.

Get the elf variants back under a half-dozen and we can talk.

Don't get me started!

There are only two types of Elves:

The good Elves that live on the surface, and may choose to live in a forest, or cities. Sometimes they will bless the other races with their presence, then shortly leave because of the smell.

Then you have the Evil, demon worshipping, boogeyman elves known as the Drow. Where even in a D&D world they are considered more of a legend used to scare children at night rather than anything real...

And all Drow info is in the MM and DMG.

Gnomes and Halfings thing isn't just "Being small". Gnomes are small fey-ish folk with the talking to animals, alongside the tinkerer, and are honestly closer to elves than they are halflings, wheras Halflings are their hobbits with the serial numbers filed off selves

Still no need for both.

Take some hobbit traits, give them to the Gnomes, done.

One small race only please...

 

[Blue]

It doesn't need a deep dive. "Ratings for dragonborn were 3.76 +/- 0.95 (two SD)." That gives REAMS more information than "70% of responses were positive." Even for a total layman, you now know that almost everyone gave a 3 or a 4, leaning toward 4--some people gave 5s, more than the people who gave 2s, and 1s are quite rare. With that, we can see that the response is favorable but not fervent.

I work in a highly technical field. I routinely think about my audience when speaking and writing and make sure that I target their skill level. If I was doing a spoken presentation for a large, non-professional audience I would absolutely make sure that I couched items in easy to digest bits aimed at the lowest common denominator being able to assimilate before I moved onto my next sentence. Especially if it's a supporting item. Your average person does not have a grasp of what two standard deviations means, so that's just technogabble getting in the way of understanding. They could understand 3.76 +/- 0.95, but it will likely take longer than my next sentence to work out the details of that so it's either wasted (and therefore confusion-making) detail or it means they will lose the thread and miss stuff as they work it out.

Sorry, what you suggest would be a mistake to present in a video to a large and varied audience. Regardless if they have it in that format. They're going to take the simplest to understand at a spoken rate and say 75%. Not even 75.2% that the 3.76 would have given out of 5.

The format used and the widely varying audience pretty much ensure what you want would not be presented regardless if they have well done statistics or have only what you worry they have.

 

[Whizbang Dustyboots]

It doesn't need a deep dive. "Ratings for dragonborn were 3.76 +/- 0.95 (two SD)." That gives REAMS more information than "70% of responses were positive." Even for a total layman, you now know that almost everyone gave a 3 or a 4, leaning toward 4--some people gave 5s, more than the people who gave 2s, and 1s are quite rare. With that, we can see that the response is favorable but not fervent.

I think you are giving laypeople way too much credit, given how laypeople think that both weather forecasters and political polling are wildly inaccurate. (They're not, and weather prediction in particular has gotten incredibly accurate in the last decade.)

The level of innumeracy in the general public means we'd have even more bad takes on this data, just with added bad math. Admittedly, it'd be funny to see people arguing about how statistics work in YouTube comments, but I think the end result would be about the same as it is now.

 

[Whizbang Dustyboots]

Still no need for both.

Take some hobbit traits, give them to the Gnomes, done.

One small race only please...

Again, we have many PHB species -- independent of elves -- with fewer differences than gnomes and halflings currently have.

Yes, WotC could merge them together (and given what they did to each of them in 4E, I would have preferred they do that then), but they're still plenty distinct, even if they're not the choice for every player. (I mean, I've never played a cleric, but that doesn't mean I want them taken out of the PHB.)

 

[Whizbang Dustyboots]

I remember the strength percentile roll from before 3e. And 18 01% was very different from an 18 100%. That role as a very young person felt like it could make or break a fighter

It was remarkable how many times the percentile dice (which have really dropped in popularity in the WotC era, now that I think about it) would come up with a 100 when the DM and other players went to the bathroom while the fighter was rolling up their character.

 

[EzekielRaiden]

(They're not, and weather prediction in particular has gotten incredibly accurate in the last decade.)

The problem isn't that weather forecasting is inaccurate; it is, on the macro scale, indeed incredibly accurate. Sometimes accurate out to weeks in advance, rarely off by more than a few degrees, etc.

The problem is that the very specific singular thing people REALLY REALLY want to know--"will the weather be inclement during this specific time frame in this specific urban location?"--is an extremely difficult question to answer. It depends on a host of variables which are very difficult to directly observe, especially in real time, conditions which change rapidly and often without warning, and which frequently covers (from a meteorological perspective) extremely slim tracts of land, perhaps even as small as a single street or park. That kind of prediction is still spotty at best, but of course that's the thing people want to know the most and always have. "Will I need my umbrella if I take a walk now?" You can tell a whole city if they should expect rain everywhere, but telling a single neighborhood to expect intermittent light showers for a couple hours is extraordinarily difficult.

This causes laypeople to think weather prediction is garbage, because for their lived everyday experience, yes, it is an extremely rough measure. (The fact that the percentage chance of precipitation you hear on weather reports may actually be any of the following (with made-up numbers):
A) Under conditions resembling the (predicted) conditions of the day in question, 40% experienced some level of precipitation, thus there is an 40% chance that precipitation will occur
B) Under conditions resembling the (predicted) conditions of the day in question, the forecaster is 80% certain that precipitation will occur, but only cover 50% of the area under analysis, thus there is a 40% chance that any given randomly-chosen area will experience precipitation.
C) There is a precipitation source (e.g. cloud front) moving toward the area, which will apply 100% coverage of precipitation, but the forecaster is only 40% certain that the source will reach the destination in the prediction period, thus there is a 40% chance that precipitation will occur.

ALL of these mean something different from what most people think "chance of rain" means, which they interpret in the casual, ordinary-people sense of "this is the chance that, if I were to go outside today, rain would fall on me." The second case, for example, could be as high as 100% "chance of rain" in the casual sense if the person wanting the report is expecting to travel around the area in question quite a bit.

 

[doctorbadwolf]

It doesn't need a deep dive. "Ratings for dragonborn were 3.76 +/- 0.95 (two SD)." That gives REAMS more information than "70% of responses were positive." Even for a total layman, you now know that almost everyone gave a 3 or a 4, leaning toward 4--some people gave 5s, more than the people who gave 2s, and 1s are quite rare. With that, we can see that the response is favorable but not fervent.

And it is trivial to present that information in table form.

For goodness' sake, this is the kind of thing we demand from political polling, which almost everyone is at least loosely aware of. We know that polls have a listed "margin of error" and that that margin of error is relevant and needs to be accounted for. Giving something similar here cannot possibly be that difficult, if they really want to stick with proportion statistics rather than something more generally useful.

We demand more detail from political polling because it matters.

Beyond that, others have replied quite well to this, so I’ll it at that for now.

 

[doctorbadwolf]

The problem isn't that weather forecasting is inaccurate; it is, on the macro scale, indeed incredibly accurate. Sometimes accurate out to weeks in advance, rarely off by more than a few degrees, etc.

The problem is that the very specific singular thing people REALLY REALLY want to know--"will the weather be inclement during this specific time frame in this specific urban location?"--is an extremely difficult question to answer. It depends on a host of variables which are very difficult to directly observe, especially in real time, conditions which change rapidly and often without warning, and which frequently covers (from a meteorological perspective) extremely slim tracts of land, perhaps even as small as a single street or park. That kind of prediction is still spotty at best, but of course that's the thing people want to know the most and always have. "Will I need my umbrella if I take a walk now?" You can tell a whole city if they should expect rain everywhere, but telling a single neighborhood to expect intermittent light showers for a couple hours is extraordinarily difficult.

This causes laypeople to think weather prediction is garbage, because for their lived everyday experience, yes, it is an extremely rough measure. (The fact that the percentage chance of precipitation you hear on weather reports may actually be any of the following (with made-up numbers):
A) Under conditions resembling the (predicted) conditions of the day in question, 40% experienced some level of precipitation, thus there is an 40% chance that precipitation will occur
B) Under conditions resembling the (predicted) conditions of the day in question, the forecaster is 80% certain that precipitation will occur, but only cover 50% of the area under analysis, thus there is a 40% chance that any given randomly-chosen area will experience precipitation.
C) There is a precipitation source (e.g. cloud front) moving toward the area, which will apply 100% coverage of precipitation, but the forecaster is only 40% certain that the source will reach the destination in the prediction period, thus there is a 40% chance that precipitation will occur.

ALL of these mean something different from what most people think "chance of rain" means, which they interpret in the casual, ordinary-people sense of "this is the chance that, if I were to go outside today, rain would fall on me." The second case, for example, could be as high as 100% "chance of rain" in the casual sense if the person wanting the report is expecting to travel around the area in question quite a bit.

Is there a context in which you feel this adds to the actual discussion?