Class and Level Geekery XIV - We are the Geek Pantheon

[ti'esar]

Honestly, I'm not sure what there is to add. Its ability to control Belkar, but not Tarquin or Roy, implies something about its Ego score, but pinning it down isn't really possible, and we know nothing about its other qualities.

Likewise, if this is in relation to the previous discussion, all we know is that it might have either ki focus or some kind of stunning power.

[Quartz]

Just to extend things further, since it's possible that Tarquin has levels of Monk (I've posted on this before) perhaps his stats need qualification - if he's reached 17th level as a Monk (say Ftr 2 / Mk 17 or Kt 8 / Mk 17) then he no longer suffers penalties due to age.

[Reboot]

It seems to me that this is spiralling in on itself:
1) A strip that shows Tarquin making a strike with his axe (i.e., an armed strike) is cited as evidence for him having Improved Unarmed Strike/Stunning Fist for some reason.
2) Someone pulls up a possible way the axe could be counted as an unarmed strike
3) It's suggested that the axe's entry is edited to take (2) as fact.

When the point is that (1) somehow showing IUS/SF and thus needing to be justified as in (2) has never been proven in the first place! It's not like this is something described in text - like a spell shouted out - that then has to be justified...

[knag]

Hi folks, sorry to interrupt, but Haley is listed as possessing an adamantine dagger, the one she bought in #976, where Bandana also takes an adamantine short sword. The only problem with this is that she loses it in #978. One might assume that she retrieved it later, if it weren't for #988, where Bandana says she got a "snazzy new dagger" (not a short sword) out of the encounter. So it would seem that particular dagger is in Bandana's possession, not Haley's. Maybe they traded and she has the short sword? Is this too nitpicky?

[Ephemera]

Hi folks, sorry to interrupt, but Haley is listed as possessing an adamantine dagger, the one she bought in #976, where Bandana also takes an adamantine short sword. The only problem with this is that she loses it in #978. One might assume that she retrieved it later, if it weren't for #988, where Bandana says she got a "snazzy new dagger" (not a short sword) out of the encounter. So it would seem that particular dagger is in Bandana's possession, not Haley's. Maybe they traded and she has the short sword? Is this too nitpicky?

This is the Geekery thread. Nothing is too nitpicky.

FWIW, I think 988 is evidence that Haley doesn't have the dagger but not that she does have the shortsword (which would be an odd choice for her anyway most likely).

[Timy]

Hello all,

I am not enough of a geek to know that but is 1060 a show of a "parry" bonus to CA due to the sword ?

And if yes, (which I think according to this thread rules) what could be giving him this bonus ?

EDITH : actually, after a little research, I found this :

http://www.dandwiki.com/wiki/Parry

Could it be it ? (arf, it works only with weapons the same size or smaller).

OOOR it could be that being a heirloom, it nullify this prerequisite ??? :D

[zimmerwald1915]

Hello all,

I am not enough of a geek to know that but is 1060 a show of a "parry" bonus to CA due to the sword ?

And if yes, (which I think according to this thread rules) what could be giving him this bonus ?

Showing a character parry is merely one of many, many ways to show another character failing to hit their AC. Other ways include the attacking character swinging wide or missing with a ranged weapon, or the attacked character dodging out of the way. Which way is shown is determined by the impression the character is meant to give of themselves (Roy parries because he's solid and strong; Haley might dodge in the same circumstances) and the circumstances of the fight.

EDITH : actually, after a little research, I found this :

http://www.dandwiki.com/wiki/Parry

Could it be it ? (arf, it works only with weapons the same size or smaller).

OOOR it could be that being a heirloom, it nullify this prerequisite ??? :D

If you look, you'll see that thing you've linked to is a homebrew feat. And all it does is, linking back to the point above, give a bonus to AC. It's not necessary to explain anything (because all you need to explain the giant failing to hit Roy's AC is the giant rolling low), and it's from a bad source besides.

[Quartz]

Are you perhaps thinking rather of a Defender function of the sword?

Anyway, can we give the sword some statistics from that panel, like minimum HP?

[Ephemera]

All these rounds of combat and nothing to add? Maybe we need to loosen our evidentiary standards.

Is it possible that we can draw any meaningful conclusions about the frost giants from their confidence in their ability to survive the fall from Mechane's height? Obviously we're not going to stat the giants for this thread (unless they hang around a lot longer than we expect), but if we can conclude anything it's possible that it will in turn lead to some conclusions about Roy or another character depending on how the fight goes...

[Peelee]

All these rounds of combat and nothing to add? Maybe we need to loosen our evidentiary standards.

Is it possible that we can draw any meaningful conclusions about the frost giants from their confidence in their ability to survive the fall from Mechane's height? Obviously we're not going to stat the giants for this thread (unless they hang around a lot longer than we expect), but if we can conclude anything it's possible that it will in turn lead to some conclusions about Roy or another character depending on how the fight goes...

Assume the standards were loosened. What would you propose? I can't think of anything.

Is it possible that we can draw any meaningful conclusions about the frost giants from their confidence in their ability to survive the fall from Mechane's height?

How high up is the Mechane from the ground?

Obviously we're not going to stat the giants for this thread (unless they hang around a lot longer than we expect), but if we can conclude anything it's possible that it will in turn lead to some conclusions about Roy or another character depending on how the fight goes...

I agree, but there have been no meaningful clues yet that I am aware of.

[Ephemera]

Assume the standards were loosened. What would you propose? I can't think of anything.

I was pretty much kidding. I'd thought about proposing that we keep dual stats, one set min-maxed, and one set of statistical probabilities. If I were going to seriously propose that, I'd suggest something like that any one action can be excused as chance, but that if we see someone do something more than once (without anyone commenting on the unlikelihood) we assume that is, say, a more than 1% probability. This, however, sounds like an awful lot of work to even agree on rules, let alone agree on conclusions (although it would possibly result in having somewhat more information in the thread).

How high up is the Mechane from the ground?

We don't have a totally conclusive picture, but given the size of the Mechane (as well as the variable terrain), I think we could safely conclude that the giants are planning on taking max falling damage. This is actually kind of interesting, since if we assume they're competent, they have to be confident that they can at least take average (if not something closer to max) falling damage.

And just when I thought I was getting somewhere, it turns out that average frost giants have 133 hp, so they can all survive max falling damage of 120. Oh well, it was worth a shot (and I apologize for not doing my own homework before I opened my mouth).

[littlebum2002]

I was pretty much kidding. I'd thought about proposing that we keep dual stats, one set min-maxed, and one set of statistical probabilities. If I were going to seriously propose that, I'd suggest something like that any one action can be excused as chance, but that if we see someone do something more than once (without anyone commenting on the unlikelihood) we assume that is, say, a more than 1% probability. This, however, sounds like an awful lot of work to even agree on rules, let alone agree on conclusions (although it would possibly result in having somewhat more information in the thread).

This seems like a good second thread. Having so much information in one thread would be cumbersome and confusing. However, if you started a second thread with "most likely stats" instead of "stats that we know they have based on observable evidence" I could definitely see that being popular as well since it is proposed so often.

[Peelee]

Honestly I am eagerly awaiting the rekindling the fires of Tarquin's arrow-catching ability now that we've seen him in new art, and know that dude wears gloves.

I was pretty much kidding.

Ahhh. Sorry!

[Chei]

Honestly I am eagerly awaiting the rekindling the fires of Tarquin's arrow-catching ability now that we've seen him in new art, and know that dude wears gloves.

I still don't understand how that argument went on so long before everyone understood that "not explainable using core 3.5" should be a valid enough answer.

[Emanick]

This seems like a good second thread. Having so much information in one thread would be cumbersome and confusing. However, if you started a second thread with "most likely stats" instead of "stats that we know they have based on observable evidence" I could definitely see that being popular as well since it is proposed so often.

I was actually thinking of starting something like this - well, sort of like this, anyway.

In terms of format, it would be kind of like the Stick Awards, in that it would have a "central" thread with stats for all the characters, and other threads, each of which lasted for a week or so, in which people had a chance to vote on how to fill in the unknown gaps in the characters' stats - a crowdsourced version of what the community thinks the characters' stat blocks most likely look like. For instance, in Week 1 a thread would be created for Roy, in which people would vote on his HP, level, ability scores, feats, and probably other things, subject to a minimum and maximum and maybe with some other provisions in place to prevent outliers from skewing the mean.

I'd have to give the design of those threads some careful thought (I've done some of this already, but it's by no means complete), and in order to get community input on those designs, I'd probably use the "central" thread to put up a proposed template for how to vote on a character's stat block a week before said character's thread was launched. That way I can hopefully create a voting format that balances the need for structure with the need for - well, a system that most people like, or can at least accept as the best this forum can do at creating "crowdsourced" full stat blocks for OOTS characters.

Does this seem like a good idea? I think it could be fun, and might also relieve some of the frustration some people tend to have over the very high standards in this thread by creating an alternate channel for expressing opinions about character stat blocks. I was worried that it might be perceived as too similar to C&LG, but if it's not - and I don't think it is - it could be a fun experiment.

[Ephemera]

I was actually thinking of starting something like this - well, sort of like this, anyway.

I'm not totally sure the mods would let us have a second thread like that but you could try. FWIW, I had in mind a slightly different project, setting some kind of probabilistic threshold for measured feats (as opposed to voting on most likely stats), but yours is probably a lot more practical--my method would involve a lot of math and would probably only give the bare appearance of rigor while actually still just coming down to peoples' opinions in the end.

It does seem like it would be awfully hard to vote on this, however--are you going to take a survey of all possibilities for each stat of each character (and then what, average/take the result with highest votes)? Or propose some possible stat blocks (how many?) and then someone come to consensus on the most likely one?

I'll be interested to see what you can come up with if you decide to do it, anyway...for now, we'll just have to all live with the fact that we've decided to require a fair amount of rigor, which makes things somewhat clearer (not that it avoids arguments!) but results in pretty modest conclusions.

This seems like a good second thread. Having so much information in one thread would be cumbersome and confusing.

I don't expect to get any enthusiasm for this project here (nor am I going to push), but I don't think it would be that hard. It would just result in something like:

Roy:
Str 18+, 20+ (90% confidence)

and we'd FAQ what the confidence intervals mean--just a single statistical one based on whatever criteria we were using, and it would only apply to stats where we'd seen enough feats to have some confidence.

Since a lot of the calcs are pretty similar to ones we already do and would involve the same people discussing the same feats, I actually think this would be clearer and less redundant than keeping two threads. As I say, however, I'm not actually proposing that we do it at this time...

[Jasdoif]

It does seem like it would be awfully hard to vote on this, however--are you going to take a survey of all possibilities for each stat of each character (and then what, average/take the result with highest votes)? Or propose some possible stat blocks (how many?) and then someone come to consensus on the most likely one?

Off the top of my head...If you held votes for the distinct values in a range, the way to choose a single one would be: For every value in the range, sum the squares of the differences between the value and each vote; and then choose the value with the smallest sum...in other words, choose the value with the smallest variance (and therefore smallest standard deviation).

Multi-choice methods like the Borda Count wouldn't work too well for ability scores I don't think, because of the paired nature of them (the only difference between 12 and 13 for an ability score is feat prerequisites, so the only real distinguishing feature will be if there's a feat a prereq's needed for...and in theory, an accepted feat is going to set a lower bound on the possible range and exclude one of the possibilities anyway). They'd work fine for resolving conflicting scenarios, though.

[El Dorado]

Honestly I am eagerly awaiting the rekindling the fires of Tarquin's arrow-catching ability now that we've seen him in new art, and know that dude wears gloves.

They are *clearly* 2nd edition Gloves of Missile Snaring.

muhahahaha

[Ephemera]

Off the top of my head...If you held votes for the distinct values in a range, the way to choose a single one would be: For every value in the range, sum the squares of the differences between the value and each vote; and then choose the value with the smallest sum...in other words, choose the value with the smallest variance (and therefore smallest standard deviation).

Multi-choice methods like the Borda Count wouldn't work too well for ability scores I don't think, because of the paired nature of them (the only difference between 12 and 13 for an ability score is feat prerequisites, so the only real distinguishing feature will be if there's a feat a prereq's needed for...and in theory, an accepted feat is going to set a lower bound on the possible range and exclude one of the possibilities anyway). They'd work fine for resolving conflicting scenarios, though.

Wait, did you say "Borda Count"? Are we going there?

The major problem I see your proposed method is that it's going to give undue weight to outlandish (or just outlier) guesses. That said, I guess we should wait until someone actually sets up this hypothetical thread first...and then derail that one discussing the merit of different methods of voting.

[Jasdoif]

The major problem I see your proposed method is that it's going to give undue weight to outlandish (or just outlier) guesses.

You're liable to have that problem if the range of options includes "outlandish" options, regardless of the method used. Constraining the range (acceptance that a character has Power Attack puts the lower bound at 13, that kind of thing) would be effective there.

[Ephemera]

You're liable to have that problem if the range of options includes "outlandish" options, regardless of the method used. Constraining the range (acceptance that a character has Power Attack puts the lower bound at 13, that kind of thing) would be effective there.

Not necessarily...if we used, say, ordered choice with a virtual runoff or even something like approval voting then anyone who had a theory outside the general consensus range wouldn't exert any influence on the outcome. Approval voting would actually be interesting here--for those whose geekery doesn't extend to voting methods, this would mean that everyone votes for as many options as they want to and the one with the greatest number of votes win. You'd get a different answer than a lot of other methods, but (leaving aside anyone's attempts to game the system) you'd wind up with the answer that's palatable to the most people.

[Jasdoif]

Not necessarily...if we used, say, ordered choice with a virtual runoff or even something like approval voting then anyone who had a theory outside the general consensus range wouldn't exert any influence on the outcome. Approval voting would actually be interesting here--for those whose geekery doesn't extend to voting methods, this would mean that everyone votes for as many options as they want to and the one with the greatest number of votes win. You'd get a different answer than a lot of other methods, but (leaving aside anyone's attempts to game the system) you'd wind up with the answer that's palatable to the most people.

That depends on what constitutes an outlier. Under a virtual runoff (or whatever instant runoff voting is called these days), a push for a score at either extreme can have an advantage over a theory that allows multiple primary choices (say, a Con score--anything based on the save or HP modifier being less than maximum is going to be equally valid with 16 or 17, whereas if the upper limit is even or the lower limit is odd there's no equivalent alternative to split the votes). Worst case, an outlier single-score theory with just over half as many supporters as a more common double-score theory could still come out ahead...resulting in the process favoring the more extreme theories to a large degree.

It's a general oddity with voting for points on a single spectrum: The differences between adjacent points are rarely significant enough to matter on the individual level, and rarely consistently so if they are significant, so you end up with oddities when trying to operate on the most common first listed choice. Methods like approval voting (where every independent choice has the same weight) and Borda Count (where every ranked choice has a corresponding weight towards the final outcome the whole time) are less volatile, but can still have issues in close matchups (approval voting where not everyone votes for all alternatives, or Borda Count where not everyone chooses the same alternative as their first choice, could result in a single-option theory outperforming a slightly more popular theory).

Which is why I think a numeric aggregate is most appropriate. Outliers here are not equipment errors, they're the opinions of people with unusual views; and the significance of what number they choose to stand behind on a spectrum extends beyond only that number. And personally, I think the idea that every vote will have some influence on the result, rather than discarded outright if the view behind it is unpopular, is a feature. The number with the least variance across all votes should be the most palatable on the whole, after all.


...all that said, my personal preference would be to avoid voting on points on a range, if at all possible, precisely because it's prone to weirdness. Sadly, excising ability scores from the system isn't an option Probably bundle ability scores along with feats/choices/equipment that would rely on them, so you'd have the score hashed out before trying to vote on them ever comes up. If a vote for an ability score turned out to be absolutely necessary, I'd do a vote for the modifier instead of score (so the almost-equivalent scores are grouped together) to mitigate oddities, and choose the corresponding score to support the three-odd-and-three-even arrangement used for presenting monsters if necessary.

---

Hopefully it's apparent why complex processes should be entered into willingly, and not forced on curators (And that Emanick has sounding boards before/if his hypothetical thread gets started)

[Peelee]

Under a virtual runoff (or whatever instant runoff voting is called these days)

I've heard Instant Runoff or Ranked Choice.

[Emanick]

I was actually considering using the middle 95% of votes (in practice, disregarding the highest and lowest guesses no matter what, since getting more than 40 votes on any given variable would be unlikely), or discarding outliers that fall a certain number (3, perhaps) standard deviations from the mean. I suspect that this might result in a more accurate representation of the overall community's view - but both of these voting systems could be unpopular and/or galling for people with outlier views, so I'm not sure I will go with either idea.

I'm thinking of setting up the initial thread around the beginning of the New Year, but it might take a couple weeks, because I'm on tour through January 15th. We'll see. And of course I still don't know for sure whether the mods will be okay with the whole idea.

[Doug Lampert]

I was actually considering using the middle 95% of votes (in practice, disregarding the highest and lowest guesses no matter what, since getting more than 40 votes on any given variable would be unlikely), or discarding outliers that fall a certain number (3, perhaps) standard deviations from the mean. I suspect that this might result in a more accurate representation of the overall community's view - but both of these voting systems could be unpopular and/or galling for people with outlier views, so I'm not sure I will go with either idea.

I'm thinking of setting up the initial thread around the beginning of the New Year, but it might take a couple weeks, because I'm on tour through January 15th. We'll see. And of course I still don't know for sure whether the mods will be okay with the whole idea.

If you're trying to eliminate outliers, go all the way, take the median.

This avoids giving outlandish votes excess weight, avoids ignoring anyone's vote, and gives the central value.

[Ephemera]

If you're trying to eliminate outliers, go all the way, take the median.

This avoids giving outlandish votes excess weight, avoids ignoring anyone's vote, and gives the central value.

I suspect that in a non-secret-ballot system like we'd have to adopt, that's a little too subject to gaming (or at least people casting votes primarily intended to game the system as opposed to reflecting their honest opinion. Assuming, of course, that we wouldn't have a different problem, which is too few people actually voting for any of these systems to work.

My best guess is that our actual problem will be that the mods will say "nah, you already have a thread for discussing character stats." We shall see, however.

[littlebum2002]

It's worth a shot. I'd certainly participate.

[Doug Lampert]

I suspect that in a non-secret-ballot system like we'd have to adopt, that's a little too subject to gaming (or at least people casting votes primarily intended to game the system as opposed to reflecting their honest opinion. Assuming, of course, that we wouldn't have a different problem, which is too few people actually voting for any of these systems to work.

My best guess is that our actual problem will be that the mods will say "nah, you already have a thread for discussing character stats." We shall see, however.

How do you game "median"? Your best vote is effectively always your honest opinion, more extreme doesn't help at all.

The situation where gaming is POSSIBLE requires ALL of the following:
1) There are an odd number of other voters.
2) The result you want has zero actual votes or other supporters.
3) The result you want has no votes between it and the median.
5) A vote for the result you want would result in a tie for the median value involving the result you want and the median of the other votes.

In that specific situation, you can vote more extremely and get a small extra movement toward the result you want, in pretty much any other circumstance, open voting or not, you're best off voting for the result you want.

[Crusher]

I suspect that in a non-secret-ballot system like we'd have to adopt, that's a little too subject to gaming (or at least people casting votes primarily intended to game the system as opposed to reflecting their honest opinion. Assuming, of course, that we wouldn't have a different problem, which is too few people actually voting for any of these systems to work.

Having run hundreds of these sorts of things (at another site on another topic), I'm always surprised at how seldom people try to game the system. Certainly people do sometimes, and I think they did it more in the beginning, but by now they're long since tired of those shennanigans and just plop in their answers and move on. In a way its peculiar because while I collect the aggregated information (its event prediction), I also recognize who did the best on each particular one. There's no prize other than bragging rights, but this is the internet so that's kinda like the gold standard. Still, I'm always surprised the last few people to put in their numbers don't always come up with something like 10.001 to try and play the system.

The bigger issue is invariably the second point. If its something big and punchy (What level is Xykon?) then you'll get tons of interest. If you're asking people what Durkon's STR score is... well, at least its easy to figure out the median and stndev when you've only got 5 votes.

[Tobtor]

In such a thread using voting: as you are voting one individual subjects, how are you going to handle conflicting voting results?

Lets say it is voted that character X is lvl 20, and character Y is voted to be lvl 18, but we then find out that the lvl gap between them is more than 2? This would invalidate the votes. As it will never be the same people voting, you have to decide who is wrong (with a new vote, with new people).

Even if we already know the lvl gap, then using the average or median or whatever results on individual characters, it would likely mean runitely giving obvious "wrong" results. Again a lvl gap of exactly 2 between two characters is known. Two votes on their levels are made, Character A gets a median on 13, and average of 13,8 (rounded to 14) while for character B we get a median of 16, and average of 15,4 (rounded to 15). Results like this is very likely, especially since different people might vote on the different characters.

Then you could make the votes on both characters simultaneously, but then you might get problems comparing character A with character X, Y and Z etc.

What if new people joins the discussion? Should you re-run every vote so and so often? Or have it a continues vote? Can you chance your vote due to good arguments (or simply that you think a character should have levelled since last time)?

That would lead to extremely many votes all the time...