Mike Mearls is trolling...

[erik]

Have they not heard of sampling bias? The people visiting the site right now are the people playing/preferring 4e almost overwhelmingly. Of course that sample set is going to use the most positive value statement for their preferred selection. If 4e wasn't the most "just right" for any question, then there's probably something fishy going on.

I thought it would be useful to pause this week and talk a little bit about the polls that have been appearing in this column. Some folks think of them as poorly disguised marketing research. In all honesty, they’re simply an attempt to engage in a dialogue. We already have an entire department here at Wizards of the Coast dedicated to collecting data, running official surveys, and so on. Plus, I also took enough statistics in college to understand that a self-selecting audience is by no means a sound foundation for the sort of polling we’ve been running in Legends & Lore.

The biggest problem wasn't the self-selecting. It's that it was self-selecting presumably from a crowd that was already overwhelmingly dominated by 4e fans. That he notes self-selecting without noting that the pool itself was tainted, is telling. If you do phone surveys then you are only surveying people that use phones. If you do a survey on a site that is support for 4e games, then you are largely only going to get input from 4e players.

There are useful bits of information still to be gleaned from these polls as others have wisely noted, but mostly only as indirect things like showing how many players presumably started 4e without any DnD background, how many players still visit their forums despite not playing 4e. It is not a totally useless survey, but I still doubt their competence to make accurate use of such information.

If they had more questions at once then they could have used these questions as a kind of filter to tell them about their audience so they could figure out what future responses meant. They can still loosely of do that, but it is a lot less solid than if they could tie these answers directly to other answers from the same respondents.

[Plebian]

without numbers behind the percents I'm having trouble inferring much of anything from the poll, most especially because every option can be submitted individually.

so 1000 people, total, could've voted on 3e while 5000 voted on 2e and 10000 voted on 4e, but the way the percentiles are given alone just means out of the people who answered that one question.

so, eh, it's a talking point. not sure how he's trolling, though.

[gourdcaptain]

We're talking here about a company that released a monster builder tool today that can't build monsters (it can only rename attacks and change their level). Trolling is probably not out of the question.

[wotmaniac]

We're talking here about a company that released a monster builder tool today that can't build monsters (it can only rename attacks and change their level). Trolling is probably not out of the question.

not that I'm surprised .... that's the whole design philosophy behind the entire game.
i.e., they've now revealed how they build monsters.

[hogarth]

Still at the end of the day, if someone told me non-devil's advocately that he thought 3E was more complicated than 1E, I'm not certain enough of how I'd quantify that to make more than a token counter-argument. But "insufficiently complicated"? Only if one's idea of the ideal level of complexity is Advanced Squad Leader.

Yeah, "insufficiently complicated" is a ridiculous comment. "If only this were written upside down and backwards and in Pig Latin, it would be perfect!"

[shadzar]

Let's make the statistics slightly more meaningful.

You cant.

The poll had all those options, but you could only pick one.

So you only got to vote for ONE edition. There is no information on total votes even to attempt to break out the 5 individual polls that exist within that one.

To make the statistics meaningful, you would have had to be given an option to vote once per edition, not once per all of D&D.

5 polls of 4 choices, would have been a proper poll, not 1 poll of 20 choices trying to answer 5 questions.

Yes there was 5 submit buttons, but all connected to the same radio select option that allowed only a single choice, not 5 choices. They totally fucked up to form creation and HTML for it to actually work. IF they went back and fixed the poll hours after it came out, then they likely didnt discard existing votes, as you could only vote once and by keeping the data from the badly coded form, means the poll, like all others form WotC, is completely worthless as is its data.

[shadzar]

Speaking of which, and I've been wondering this for a while, for a genre depending on a lot of dice rolling and numbers, why are there almost no if any published Tabletop RPGS where the makers understand basic math and probability? (Or at least run it past someone who does, let alone people who can competently use a system).

Because those people are hired for accounting jobs, not game design jobs, because they are best suited at working with real things, not imaginary things.

AD&D 1st edition had charts showing probability of a d20 roll as linear, and the 3d6 as a bell distribution model. Gary could do math.

Also it isnt that important, cause bigger is better now, so simpler sounding math, rather than correct math is all that is important to games of the 21st century. Just ask a player to subtract a negative number now a days and they will put a hemorrhoid just form staring at you.

Quite funny when their mechanics designers, fails to comprehend the math of the game, and not funny in the way Einstein failed his college math.

[Count Arioch the 28th]

Einstein never failed college math. That's a myth.

[shadzar]

Einstein never failed college math. That's a myth.

I said funny, not factual....

The point was laughing at Einstein failing math when he did things beyond what a college would teach or people even know, well that makes for a good little laugh.

Mike Mearls inability to do math, and that is his area of game design...the mechanics...is the kind of funny until you realize its true, you laugh then cry.

[erik]

Einstein never failed college math. That's a myth.

I said funny, not factual....

The point was laughing at Einstein failing math when he did things beyond what a college would teach or people even know, well that makes for a good little laugh.

Mike Mearls inability to do math, and that is his area of game design...the mechanics...is the kind of funny until you realize its true, you laugh then cry.

You can just as easily make the point about Mike Mearls without mentioning a facetious example of a historical figure at all. It is superfluous and in this case also outright wrong.

[shadzar]

You can just as easily make the point about Mike Mearls without mentioning a facetious example of a historical figure at all. It is superfluous and in this case also outright wrong.

woohoo lets go way off-topic to troll and derail a thread....

give me another well known math example that people can easily recognize.

Fibonacci has one attributed to him maybe? Would lurkers of the Den even know who Fibonacci is?

Again for those inept at understanding comparisons....Mike Mearls does fail at math, wherein Einstein didnt. That is the crux of it.

Holy fucking shit...it isnt like Mearls hasnt intimated that he is the Einstein of RPG mechanics either...so the comparison works. Since all gamers have become geniuses with "The Incredible, Expanding Gamer Brain".

[tzor]

Because those people are hired for accounting jobs, not game design jobs, because they are best suited at working with real things, not imaginary things.

It's actually worse than that. Accountants are good at number crunching, (making sure two sets of numbers add up,) but the level of math required for a good role playing game isn't number crunching, but probability. People who are really good at probability work as actuaries in the insurance industry for BIG MONEY.

But even then, you need a lot of math minds to make a game work. Even if you have the probability correct, you need to have the scalability correct. The numbers could easily work for level 1-5 but suddenly fail at higher levels because the scaling system had inheritant flaws.

[gourdcaptain]

I mean, to get the idea of how badly they understand math, look at 4e pre expertise feats (an annoying band-aid), or the infamous 3.0e Drizzt build they put out I looked at and figured, sans his HD and BAB, I could easily take him with a LV 4 Kobold Fighter 2/Rogue 2 I was playing at the time.

[TheFlatline]

Fibonacci has one attributed to him maybe? Would lurkers of the Den even know who Fibonacci is?

1 1 3 5 8 13...

[virgil]

You missed a couple in the beginning there. It's supposed to be: 0, 1, 1, 2, 3, 5, 8, 13, ...

[shadzar]

Because those people are hired for accounting jobs, not game design jobs, because they are best suited at working with real things, not imaginary things.

It's actually worse than that. Accountants are good at number crunching, (making sure two sets of numbers add up,) but the level of math required for a good role playing game isn't number crunching, but probability. People who are really good at probability work as actuaries in the insurance industry for BIG MONEY.

But even then, you need a lot of math minds to make a game work. Even if you have the probability correct, you need to have the scalability correct. The numbers could easily work for level 1-5 but suddenly fail at higher levels because the scaling system had inheritant flaws.

well WotC doesn't hire actuaries, the giant hedge funds people hire then and WotC hires the hedge fund owners for that...was the closest number related person i could compare to that WotC would hire cause their marketing people don't know how to use numbers properly any better than Mearls does...

You missed a couple in the beginning there. It's supposed to be: 0, 1, 1, 2, 3, 5, 8, 13, ...

Thank you both....but he doesnt still have an example of him being cited at being a good person with math and failing, even if people do know who he is.

I would call it a golden number if ONE person at WotC actually cared to check the math on things they did.

Wasnt WotC the one claiming that a d20 offered the same probability of 3d6 for creating stats?

No it was a d20 had an average of 10 or 11... in regards to skill checks.

[name_here]

Um, the author of that post was in too much of a hurry to actually bother to see if what he was talking about made the slightest damn sense. D20 is a roll-above system, and slightly over half of all possible rolls on a d20 equal or exceed 10. It's really swingy, but you do get a roughly 50% chance of succeeding when you need a natural 10.

[shadzar]

Um, the author of that post was in too much of a hurry to actually bother to see if what he was talking about made the slightest damn sense. D20 is a roll-above system, and slightly over half of all possible rolls on a d20 equal or exceed 10. It's really swingy, but you do get a roughly 50% chance of succeeding when you need a natural 10.

He was basically saying that the people at WotC didnt know a thing about probabilities wherein they didnt understand a linear distribution model.

Average to the non-mathematically inclined means most often rolled, and 10 and 11 on a d20 is NOT most often rolled, as each number has the same probable chance of coming up as the next, but the average of 2d6 DOES make 10 and 11 average as with the ability score rolls.

Whoever wrote the passage didn't check to make sure their language agreed with their math. Another example also of where WotC writers dont understand the meanings of the words they are using. There is no such thing as a 10 or 11 being the statistical average on a d20 as each number has a 5% chance, so all numbers 1 through 20 are the average number that would be rolled.

If someone understood math enough, they could have written the part of 3.0 PHB pg. 58 in order to express it properly.

On average, Devis will roll a 10 or 11 on the d20,

...does not mean the same as...

On average, Devis needs to roll a 10 or 11 on the d20,

[tzor]

OK, time for a Fibonacci joke. Why can you attribute the following sequence to him ...

0 1 2 3 4 5 6 7 8 9 ...

Fibonacci is best known to the modern world for the spreading of the Hindu-Arabic numeral system in Europe, primarily through the publication in the early 13th century of his Book of Calculation, the Liber Abaci.

Leonardo Pisano Bigollo did not invent the sequence named after him; he just used it in an example in his book I mentioned in the spoiler.

[name_here]

On average, Devis will roll a 10 or 11 on the d20,

...does not mean the same as...

On average, Devis needs to roll a 10 or 11 on the d20,

That second one is also not what they were going for. I'm pretty sure they meant:

On average, Devis will roll at least a 10 or 11 on the d20,

Since that's a true statement and fits with how they configure their DCs, they probably just screwed up their phrasing.

[TheFlatline]

You missed a couple in the beginning there. It's supposed to be: 0, 1, 1, 2, 3, 5, 8, 13, ...

*palmface* I'm full of fail today.

[TOZ]

Always fun to see the different ways the same thread goes when posted to multiple forums.

[shadzar]

On average, Devis will roll a 10 or 11 on the d20,

...does not mean the same as...

On average, Devis needs to roll a 10 or 11 on the d20,

That second one is also not what they were going for. I'm pretty sure they meant:

On average, Devis will roll at least a 10 or 11 on the d20,

Since that's a true statement and fits with how they configure their DCs, they probably just screwed up their phrasing.

No, your way doesnt work either.

The reason is simple, on a single die, with the same surface area for each face, and each face having a number unique to only one face with no numbers repeated, then that dice will yield each of the faces with the same frequency.

The only way to use a d20 with any sort of "on average", would be to say...

On average, Devis will roll at least a 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 on the d20,

But it would be very dumb to say that, as each has a 5% chance of appearing in a linear distribution model.

On a d20 a 10 is as likely to come up as often as an 11, as much as they both are likely to come up as any of the other 18 numbers.

You are likely to roll a "10 or 11" more often than rolling a 1 because, and ONLY because you have grouped two chances together to give a 10% chance out the outcome as opposed to the other 18 numbers, which still have a 5% chance each to be the result of the roll.

[duo31]

not a joke about the man per-say, but an amusing Fibonacci joke just the same.

-duo

[fectin]

...does not mean the same as...

That second one is also not what they were going for. I'm pretty sure they meant:

On average, Devis will roll at least a 10 or 11 on the d20,

Since that's a true statement and fits with how they configure their DCs, they probably just screwed up their phrasing.

No, your way doesnt work either.

The reason is simple, on a single die, with the same surface area for each face, and each face having a number unique to only one face with no numbers repeated, then that dice will yield each of the faces with the same frequency.

The only way to use a d20 with any sort of "on average", would be to say...

On average, Devis will roll at least a 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 on the d20,

But it would be very dumb to say that, as each has a 5% chance of appearing in a linear distribution model.

On a d20 a 10 is as likely to come up as often as an 11, as much as they both are likely to come up as any of the other 18 numbers.

You are likely to roll a "10 or 11" more often than rolling a 1 because, and ONLY because you have grouped two chances together to give a 10% chance out the outcome as opposed to the other 18 numbers, which still have a 5% chance each to be the result of the roll.

You're both wrong. Average means either a mean, a median, or a mode. A mode is pretty rare, and doesn't really parse as used in the original sentance. The arithmetic median is 10.5 (top of the third quartile + bottom of the second quartile/2). The arithmetic mean is 10.5 (sum from i=1:20 of (i*.05)). The geometric mean and median are rediculously inappropriate, so we'll ignore them.
Expressing the result as "On average, Devis will roll at least a 10 or 11 on the d20" is wierd, and slightly wrong, but "On average, Devis will roll at least a 10 on the d20" is factually correct, assuming sufficient trials.