The Pitfalls of D&D Beyond Data

FrogReaver

As long as i get to be the frog
Huh? I don't care.

I'm just saying that the manner in which you assigned blame is revealing, and contradicts the assertion that this thread is somehow "scientific". Certainly not more scientific than the work it's criticizing.

And I’m saying that’s an inaccurate and misleading statement given the full context of the comment. I’ve pointed that out but you keep repeating yourself.

It’s also inaccurate and misleading to ignore the stance that the overall thrust of a thread can be scientific without every comment in the thread being so.

Yea, it’s easy to tell who is the actual biased one here...
 

log in or register to remove this ad

Mistwell

Crusty Old Meatwad (he/him)
#1. The subclass chart was titled “subclasses (active characters)”.

No, it was not. Here it is for everyone to see:

YhhU0Pk.jpg


That designation means it’s a breakdown of the subclasses of all active characters. That isn’t what it showed. Instead it was a breakdown of all active characters with a subclass.

I read it as the later (there is no distribution of something which does not exist so it obviously meant the distribution of subclasses which are active and not of all characters even if they had no subclass), many others read it as the later, it's one reasonable interpretation of that title to read it as the later (and I'd argue it's the more reasonable interpretation in fact) and most importantly YOU KNEW TWO DAYS AGO THAT'S WHAT IT MEANT AND NEVER SAID A WORD HERE.

That you're claiming there is only one way to read that is false. There are two ways to read it, you read it one way, some others read it another way, you got a clarification on how to read it, and you're stubbornly still acting like you had to be right and there is no other option.

#2. Before he even claimed that’s what that particular chart meant I had already deduced that was what it meant.

So you knew all along what it meant. This whole thread is bizzare! What the heck are you even arguing anymore if many knew it right away from the title (even if you did not) and even you knew it after thinking about it for a bit?

He didn’t correct my understanding on it, I called out the chart for what it was and he validated my understanding.

No dude, that is not what happened. But if that's what you need to tell yourself to sleep better at night, fair enough. The rest of us all saw what happened, it's copied for everyone to see to this thread.

I think we're done now. This conversation has been "invaluable".
 
Last edited:

FrogReaver

As long as i get to be the frog
So I think the multiclassing problem has been solved. There’s at least 2 solid solutions for dealing with it.

Any ideas on the best way to handle the subclass problem?
 

G

Guest 6801328

Guest
And I’m saying that’s an inaccurate and misleading statement given the full context of the comment. I’ve pointed that out but you keep repeating yourself.

What? That you were trying to assign blame, and trying to give the "technical people" a free pass? I don't see how that changes anything. Or am I missing something?

It’s also inaccurate and misleading to ignore the stance that the overall thrust of a thread can be scientific without every comment in the thread being so.

True, but blatant biases, and premature conclusions, call both motives, objectivity, and methodology into question. (You yourself use your belief that business needs are driving the methodology here...although it's not clear to me why the business side would have a vested interest in misrepresenting the data. Surely they could care less if the 0.01% of the market that whines and complains about the fighter class continues to do so.)

Yea, it’s easy to tell who is the actual biased one here...

I'm not the one claiming to be doing science in this thread.
 
Last edited by a moderator:


billd91

Not your screen monkey (he/him)
Note the numbers involved with this data are more than political polling.

The data set may be larger, but size has relatively little effect on sampling error and bias. Whether you're consulting 1000 or 10,000 pieces of data, you're still polling only a fraction of the entire population and therefore you're still contending with sampling error. While a larger sample size does decrease the error, it does so with diminishing returns. This is one reason political polls can contact only a few thousand voters out of over 100 million and come up with reasonably accurate results.

The greatest weakness of D&D Beyond's data and ability to infer the campaign length or class preferences of the general 5e-playing population lie in its systematic bias as a self-selected sample rather than being a true random sampling of D&D 5e players and DMs.
 


FrogReaver

As long as i get to be the frog
I want to apologize to anyone here that is for a legitimate discussion. The thread should be a lot better now though it’s against the rules for me to comment on why.

Hopefully the rest of the thread can avoid biased personal attacks against those that dare criticize the methodology used to create some of the graphs that D&D Beyond has shared with us.

Who knows, maybe we can even hammer out some solutions for how the data should summarized in the future to avoid these pitfalls.
 

ClaytonCross

Kinder reader Inflection wanted
Without double checking I believe the total number of characters on D&D Beyond was cited at 8.8 million. Of course there is also a bit of debate about whether those charts include all those character or some subset of them called active characters (whose criteria thus far is undefined to us).

I feel like the lack of defining what an Active charcter is removes context and could be easily resolved. If they put "Active in the last 120 days" and defined active as characters assigned into a campaign that have done 3 of the following taken damage, taken short/long rests, leveled up, updated notes, changes in inventory, use off spell slots, and/or use of special abilities. I feel like that would be a greats start to making the data useful outside of their internal functions. Also, creator tags on characters as I mentioned before.

I do realize writing this that there may actually be an active attempt by D&D Beyond to show slides in a vague manner in order to prevent competitors from using their data for product improvement.
 

Mistwell

Crusty Old Meatwad (he/him)
I want to apologize to anyone here that is for a legitimate discussion. The thread should be a lot better now though it’s against the rules for me to comment on why.

Hopefully the rest of the thread can avoid biased personal attacks against those that dare criticize the methodology used to create some of the graphs that D&D Beyond has shared with us.

Who knows, maybe we can even hammer out some solutions for how the data should summarized in the future to avoid these pitfalls.
Apology accepted.
 

Remove ads

Top