The stat thread apart from degenerating into a bit of a pissing contest also highlighted something quite interesting about the ability range in d20.

People don't thing the range is wide enough, or to be more specific people don't think the extremes of the spectrum are rare enough.

For example although this is a geeky website with over 200 registered users the idea that someone might have 18 Intelligence is ridiculed although statistically that level of Int occurs in 1 in 216 individuals.

Now the site is a self selecting pool so a true statistical correlation is unlikely but you might expect something approximate.

In any case the question that occured to me is that perhaps the d20 ability range is flawed perhaps we need to widen and more tightly condense the range.

So it occured to me that perhaps d30 was a better paradigm. Establish stats with 5d6 and using a d30 to take on the  roll of primary random determiner might satify people more.

Now a stat of 30 is going to be 1:7776 people a stat that far more accurately predicts Olympians and Chess Grand Masters.
5d6 had a standard diviation of 3.82 so in a normative distribution curve c. 80% of all stats would fall into the "average" range of  13 - 22 (average being 17.5) and the top and bottom ends would indeed be far more exceptional

People find intelligence intimidating. It brings out the worste in people and tends to amplify insecurities in the same way that many women react badly to girls with big boobs and guys to, I dunno, other guys with bigger schlongs I guess. Or better hair.

The excessive statistical frequency of extreme stats is only a factor if you're randomly rolling them up - I use stats from 1 to 10, but someone with intelligence 10 would be exceptionally rare, maybe one person in ten thousand.

Or to put it another way, the process used for chargen isn't neccessarily indicative of distributions throughout the general population.

Quote from: The Traveller;688410The excessive statistical frequency of extreme stats is only a factor if you're randomly rolling them up - I use stats from 1 to 10, but someone with intelligence 10 would be exceptionally rare, maybe one person in ten thousand.

Or to put it another way, the process used for chargen isn't neccessarily indicative of distributions throughout the general population.

Yeah, this.

Quote from: The Traveller;688410The excessive statistical frequency of extreme stats is only a factor if you're randomly rolling them up - I use stats from 1 to 10, but someone with intelligence 10 would be exceptionally rare, maybe one person in ten thousand.

Or to put it another way, the process used for chargen isn't neccessarily indicative of distributions throughout the general population.

So you use a logarithmic scale which is fine.

But the d20 model is predicated on a range determined on the 3-18 range randomly rolled.

Quote from: jibbajibba;688415So you use a logarithmic scale which is fine.

But the d20 model is predicated on a range determined on the 3-18 range randomly rolled.

Which doesn't change what I said.

And even there it doesn't make much practical sense - is a person with STR 18 twice as strong as someone with STR 9? Are they 20 times as strong or whatever? Naturally not, so I don't see why it's worth worrying about what is already a representative scale created at chargen spread across the whole population.

Quote from: The Traveller;688417Which doesn't change what I said.

And even there it doesn't make much practical sense - is a person with STR 18 twice as strong as someone with STR 9? Are they 20 times as strong or whatever? Naturally not, so I don't see why it's worth worrying about what is already a representative scale created at chargen spread across the whole population.

No I know you have a different system that works for you how does that change a discussion round the 3d6 model used in d20.

Sorrry am I missing a point? Are you saying that the 3-18 range is fine even though in discussion here it was felt that the top end of the range should be much rarer (thus my suggestion of 5d6).

Well I'm saying a couple of thing here - firstly and most importantly is that you needn't apply chargen maths to the entire population. A lot more than one in ten thousand PCs have a score 10 stat in my games. Conversely a lot less than one in 216 normal people might have an 18 stat in a d20 system.

Secondly even within the d20 system, it's not that internally consistent. Say an average person can lift 40 kilos max at STR 9. Does that mean the maximum anyone can lift is 80 kilos at STR 18? No, that's not right. Going by percentage chances however, someone with STR 18 should be able to lift ~20 times what someone with STR 9 could, or 800 kilos, almost a ton. Also clearly not right. So how does that map? Not very well.

The long and the short of it is the system isn't very internally consistent and should only be used in chargen anyway rather than being used to determine the number of people with 18 in a stat in any given population.

PCs are special, in other words. If a GM wants to change things to make PCs less special, that's up to the GM, but it shouldn't be because of population trends.

Yes, I like the idea of stats as a kind of PC-only subsystem. Monsters and NPCs just have HD with bonuses or penalties to their rolls as appropriate (-2 for a weak old man on an opposed STR check with a PC, +1 to damage for a big orc chief, whatever).

Quote from: The Traveller;688424Well I'm saying a couple of thing here - firstly and most importantly is that you needn't apply chargen maths to the entire population. A lot more than one in ten thousand PCs have a score 10 stat in my games. Conversely a lot less than one in 216 normal people might have an 18 stat in a d20 system.
.

Okay I see fine that makes some sense, although since PCs tend to be generated on a 4d6 drop lowest model so the PCs are already special.

I personally prefer to put PCs into the general population and found that the 3d6 roll for PCs which generates more less able PCs than the 'NPC rules from AD&D (1's count as 3s and 6's count as 4s) was an dd idea.

QuoteSecondly even within the d20 system, it's not that internally consistent. Say an average person can lift 40 kilos max at STR 9. Does that mean the maximum anyone can lift is 80 kilos at STR 18? No, that's not right. Going by percentage chances however, someone with STR 18 should be able to lift ~20 times what someone with STR 9 could, or 800 kilos, almost a ton. Also clearly not right. So how does that map? Not very well.
.

Um that makes no sense.... the fact that the population sits along a normalised bell curve in no way is indicative of a linear progression (or a logarithmic progression or geometric one)  of weight lifted or lanuages learnable etc. So nothign int eh system of stats lying in a 3-18 range implies that someone with 18 can lift double that of someone with 9. the d20 Strength tabel for exampel makes no such statement.

QuoteThe long and the short of it is the system isn't very internally consistent and should only be used in chargen anyway rather than being used to determine the number of people with 18 in a stat in any given population.

PCs are special, in other words. If a GM wants to change things to make PCs less special, that's up to the GM, but it shouldn't be because of population trends.

I can't really agree with that. You use a generation methodology that best maps to the level of realism or emulation you want in your game.
If we find out that the top 5% of humans can lift 100kg then our char gen curve should try to emulate that. Obviously if you are emutlating super heroes or giant robots then YMMV. In most games though I would say you want the ability ranges to map to the real world as well as you can becuase it makes working shit out easier (how far can I jump with 18 Strength.... um an average elite long jumper can hit 8m so 5.5+1d3m if you make a dex check if you fail the dex check then 2.5 +1d6m etc).
I would have thought with your position on Realism in falling damage for example you would be more inclined to try to follow a model that mapped to real statistics where possible.

Having stats 1-30 and rolling d30 seems to have an assumption that Human Maximum Ability = 100% chance of doing stuff. If anything I imagine this is doing human maximum ability (compared to the average) a disservice.

My preference is probably 1-30 stats, with average still about 10 and a long-tailed distribution. Maybe keep using the d20 and add extra penalties for difficult tasks. Avoids some of the usual 'the 18 Str fighter rolled a 19 and failed to break down the door, then the 6 Str wizard rolled a 4 and it burst open'.

Quote from: jibbajibba;688429So nothign int eh system of stats lying in a 3-18 range implies that someone with 18 can lift double that of someone with 9. the d20 Strength tabel for exampel makes no such statement.

And I'm likewise making no such statement, merely pointing out two ways that stats might be interpreted which don't map well to the average weight liftable by a normal man versus the maximum weight liftable by a weightlifter.

If the bare stats don't map well to actual activities in real life, why are you trying to lay the percentage chances of getting those stats onto the general population?

Quote from: jibbajibba;688429If we find out that the top 5% of humans can lift 100kg then our char gen curve should try to emulate that.

You could indeed, but I'd find it both a pain in the balls and limiting in that PCs would then have a very very minimal chance of getting very good stats, when as PCs they're meant to be the ones who survive.

Quote from: jibbajibba;688429I would have thought with your position on Realism in falling damage for example you would be more inclined to try to follow a model that mapped to real statistics where possible.

I've always said I favour playability over realism. These unique games provide a bridge to the imagination, and let us interact almost directly with what comes across. Realism strengthens the bridge and helps us to believe in what we're doing, playability lets us enjoy what we're doing, which is why I don't use death spirals or hit locations in combat, despite these being very realistic.

What I don't like is when people start claiming that realism doesn't and can't possibly exist in RPGs, or if it did it wouldn't have any effect on play, a demonstrably incorrect statement, before flying into a handbag hurricane when proven wrong.

Quote from: Bloody Stupid Johnson;688432Having stats 1-30 and rolling d30 seems to have an assumption that Human Maximum Ability = 100% chance of doing stuff. If anything I imagine this is doing human maximum ability (compared to the average) a disservice.

My preference is probably 1-30 stats, with average still about 10 and a long-tailed distribution. Maybe keep using the d20 and add extra penalties for difficult tasks. Avoids some of the usual 'the 18 Str fighter rolled a 19 and failed to break down the door, then the 6 Str wizard rolled a 4 and it burst open'.

quite happy to add modifiers to a roll once the maximum a human can do is set at a 30

So if we are looking for someon to life a 200KG portcllis Make a Str -4 check (or whatever )

Quote from: The Traveller;688435And I'm likewise making no such statement, merely pointing out two ways that stats might be interpreted which don't map well to the average weight liftable by a normal man versus the maximum weight liftable by a weightlifter.

If the bare stats don't map well to actual activities in real life, why are you trying to lay the percentage chances of getting those stats onto the general population?


You could indeed, but I'd find it both a pain in the balls and limiting in that PCs would then have a very very minimal chance of getting very good stats, when as PCs they're meant to be the ones who survive.


I've always said I favour playability over realism. These unique games provide a bridge to the imagination, and let us interact almost directly with what comes across. Realism strengthens the bridge and helps us to believe in what we're doing, playability lets us enjoy what we're doing, which is why I don't use death spirals or hit locations in combat, despite these being very realistic.

What I don't like is when people start claiming that realism doesn't and can't possibly exist in RPGs, or if it did it wouldn't have any effect on play, a demonstrably incorrect statement, before flying into a handbag hurricane when proven wrong.

so realism can exist in rpgs but not in stats and PCs just have to have high stats?

Quote from: jibbajibba;688442so realism can exist in rpgs but not in stats

If stats were realistic then each player would have a one in ten thousand chance of getting the highest in each stat. It can exist but it would be crap, so I don't know why you'd want to do that, unless you wanted to play average folks or something.