A Radical (Not to Mention Controversial) Change to Characteristic Determination

[pbd]

I have to aggree with genshou about the probability issue. Using that many dice to roll stats you are going to hit average just about every time. Getting a 90 strength (equivalent to 18 for a human) would be nigh on impossible.

Basically evryone is going to be very vanilla. Too many dice = not enough statistical variation.

Rolling up 3 sets of stats using an applet (mikes helpers) rolling 7 times

1: 51, 55, 53, 48, 53, 60, 53

2: 42, 50, 59, 67, 52, 52, 47

3: 60, 46, 60, 46, 58, 50, 53

This would also lead to the difficulties that Jack Smith pointed out with gaining higher level spellcasting abilities.

I suposue you could do roll 20d6 drop the 5 lowest, but then it would take a week to roll a character...

 

[Man in the Funny Hat]

The problem with stat generation is that when you use multiples of dice you have a CURVE of probability, but the modifiers that are then derived from the stats are a LINEAR progression. What you want is either a more curved progression of modifiers to better complement a curved progression of stat probability, or a linear method of generating stats to better complement the linear modifiers. I think it would be much more beneficial to adjust the linear side of the equation than to torture the curve.

 

[Raigon]

The problem with stat generation is that when you use multiples of dice you have a CURVE of probability, but the modifiers that are then derived from the stats are a LINEAR progression. What you want is either a more curved progression of modifiers to better complement a curved progression of stat probability, or a linear method of generating stats to better complement the linear modifiers. I think it would be much more beneficial to adjust the linear side of the equation than to torture the curve.

I think that this hits the nail square on the head. As it is the system works fairly well becasue the range of the scores is so small that it does not make much of a difference. I feel that if the bugs are worked out it could develop into a realistic and playable system

My opinion on the matter is that these large numbers make the game to complex. In general i find that the more realistic the rules become the more cumbersome they are. Which bogs down the game.

 

[SBMC]

Ever have a house cat fight a peasant?

I could give it a try but most likely the cat would run away when a tall creature attacks it. Recall the rules are designed to be interpreted by the DM and work ALONGSIDE the DM. Attack a house cat in real life and see if it runs away - do it a thousand times to a thousand different cats then tell us how many actually stood and fought it out. Of course if they are all cornered in a cage or in a sealed box they will fight - but so would a peasant against a Demon in that case as well.

One major point of mechanics in DnD is simplicity (as simple as possible). To 10+ year players we can think up all this fancy stuff, changes, better ways to do things but WoTC (and TSR previously) is not selling just to us.

Anyone can look at any class and figure it out. Your system makes that a lot harder not to mention the mathmatical issues raised by others.

For example: Hit Points
HP are not just damage; they represent the staying power in harsh conditions (not just combat - think about forced marches and extreme weather), your ability to stay concious and alert etc. Barbarians are "tougher" than the rest so they get d12's. Fighters and the like are "tougher" than other classes so they get d10's. Wizards and the like get d4's because they do not focus on "toughness". Toughness, here, as I define it for this purpose, is, among other things, prowess under bad conditions (comba included but not exclusive), keeping in solid physical shape, becoming mentally tough (as in "heart" - in a physical way that is not to confuse it with Will saves), etc.

It is an intangible conept however a more "tangible" alternative would revolve around identifying a hit - where on the body it hit, how much damage to that part of the body was done, how it effects that body part (and effects on other things) and how Armor protects you in what areas - if ya get hit in the leg while wearing Breast Plate Armor; what happens? Exactly what part of the leg? The part with or without the geave?. In full plate armor; exactly how do you get hit? Only with peircing weapons and the hits always have to be in the joints unles you roll a natural 20? What happens when I bludgeon your helmet with a heavy mace - does your helmet visor get partially closed; reducing your vision? What If I bludgeon the knee joint of the armor - is your leg stuck in place because the armor is bent up? How exactly could I target a specific place? Would that leave me more open to attacks from my target? What if the target knew what I was doing and went to defend that part? what then?

That all is not an easy thing to do (recall Battletech anyone? But with machines a lost leg is a bit different)

I compare your Ability score alternaive to my example above here regarding complexity. Would you go after this alternative as well?

 

[genshou]

The problem with stat generation is that when you use multiples of dice you have a CURVE of probability, but the modifiers that are then derived from the stats are a LINEAR progression. What you want is either a more curved progression of modifiers to better complement a curved progression of stat probability, or a linear method of generating stats to better complement the linear modifiers. I think it would be much more beneficial to adjust the linear side of the equation than to torture the curve.

Which is why heroes need to use a method that keeps the same probability curve as the standard method, as the suggestion I made above does. If you only roll 4d6 and drop the lowest, then multiply the result by 10, it'll give the same ability score variance as if you never multiplied in the first place. If you're bothered by all heroic ability scores being a multiple of 10, simply add 1d10-1 to each ability score.

Of course, if we want to keep with the original idea suggested by mythusmage, a character with 20d6 (drop 5 lowest) is still going to be closer to the middle than a normal 3E/3.5 PC would be. Rolling even more dice than that and dropping yet more would alleviate the problem somewhat, but it'll be a real hassle for those who don't use dice roller programs. I've been trying to figure out how I could program a Java app to predict the average result on an "XdY, drop Z lowest" roll, as well as outputting a bell curve image to show the probability of each individual number being rolled. That would help with figuring out how to best calculate heroic abilities in this system. If anyone can provide me with any help in this task, let me know *wink*

 

[SBMC]

Which is why heroes need to use a method that keeps the same probability curve as the standard method, as the suggestion I made above does. If you only roll 4d6 and drop the lowest, then multiply the result by 10, it'll give the same ability score variance as if you never multiplied in the first place. If you're bothered by all heroic ability scores being a multiple of 10, simply add 1d10-1 to each ability score.

Of course, if we want to keep with the original idea suggested by mythusmage, a character with 20d6 (drop 5 lowest) is still going to be closer to the middle than a normal 3E/3.5 PC would be. Rolling even more dice than that and dropping yet more would alleviate the problem somewhat, but it'll be a real hassle for those who don't use dice roller programs. I've been trying to figure out how I could program a Java app to predict the average result on an "XdY, drop Z lowest" roll, as well as outputting a bell curve image to show the probability of each individual number being rolled. That would help with figuring out how to best calculate heroic abilities in this system. If anyone can provide me with any help in this task, let me know *wink*

I don't see the issue ehre with the cuurve versus linear - the ability score roll has equal proability per dice of 1 to 6. If I follow your mathmatic logic then that means that the ability scores are meaningless as rolling 3 ones equals rolling 3 sixes.

The two are not related except that the higher ability score (which has equal probability as any other of coming up) gets a higher modifier; why? Well because of that same probabiluty; just like gambling.

We could talk in circles about this and bore all the non mathmeticians here but think about it; the modifier relates to the number not the role. If you take dice off (4d6 rolled) then the prbability of gaining a higher number increases (as you just yanked off the lowest 25% of your average).

 

[genshou]

I don't see the issue ehre with the cuurve versus linear - the ability score roll has equal proability per dice of 1 to 6. If I follow your mathmatic logic then that means that the ability scores are meaningless as rolling 3 ones equals rolling 3 sixes.

The two are not related except that the higher ability score (which has equal probability as any other of coming up) gets a higher modifier; why? Well because of that same probabiluty; just like gambling.

We could talk in circles about this and bore all the non mathmeticians here but think about it; the modifier relates to the number not the role. If you take dice off (4d6 rolled) then the prbability of gaining a higher number increases (as you just yanked off the lowest 25% of your average).

Actually, 3 sixes rolled nets an 18, a +4 bonus. 3 ones rolled nets a 3, a -4 penalty. The statistics of getting either is the same, and the mechanic works to the same level in the opposite direction.

Ability score generation has never been "linear." Even in point-buy, higher ability scores cost more than lower ones. It gets harder and harder to get to the extremes of the curve the more dice you use to roll, so if every human ability score is calculated via 15d6 as the first post in this thread suggests, ability scores in the "middle ground" (a range of modifiers called the Doldrums or Mundanes by my players) are going to be more frequent.

I hate to say this, but if you're bored by the mechanics aspects of the game and our mathematical discussion of this house rule, why are you reading this thread? If you don't want this level of math in a thread you participate in, there are plenty of other threads on this forum, and you can even start one of your own. So please leave us to our complex tinkering if you must, but don't criticize us for "boring" anyone with our "talking in circles"–there are many types of gaming discussions, and as this is a public forum we are free to get as boring as we want.

 

[Bront]

I don't think this helps at all. All it does is change the numbers. Current system means that 1 ability point results in a 2.5% difference in ability (A + or - 1 to a D20 roll is a 5% influince, and it takes 2 ability points to make a + or - 1). This seems like a perfectly fine range for me, especialy with size and other modifiers being used to adjust things like con, damage, carrying capacity, and such.

Now, the flaws of the D20 system show up in the extremes, such as 1st level where a lucky hit can kill instantly and at 20th level, falling several hundred feet is unlikely to. Skill have similar problems, from too easy, to too hard. Changing the stat numbers won't fix the issue, and honestly I don't know quite what will. Since most adventures end up running 5-10th level, which is where the system is near it's best, it's usually not a big issue.

 

[SBMC]

Recall my point to your statement:

Which is why heroes need to use a method that keeps the same probability curve as the standard method, as the suggestion I made above does. If you only roll 4d6 and drop the lowest, then multiply the result by 10, it'll give the same ability score variance as if you never multiplied in the first place. If you're bothered by all heroic ability scores being a multiple of 10, simply add 1d10-1 to each ability score.

Of course, if we want to keep with the original idea suggested by mythusmage, a character with 20d6 (drop 5 lowest) is still going to be closer to the middle than a normal 3E/3.5 PC would be. Rolling even more dice than that and dropping yet more would alleviate the problem somewhat, but it'll be a real hassle for those who don't use dice roller programs. I've been trying to figure out how I could program a Java app to predict the average result on an "XdY, drop Z lowest" roll, as well as outputting a bell curve image to show the probability of each individual number being rolled. That would help with figuring out how to best calculate heroic abilities in this system. If anyone can provide me with any help in this task, let me know *wink*

And my response to that:

I don't see the issue here with the cuurve versus linear - the ability score roll has equal proability per dice of 1 to 6. If I follow your mathmatic logic then that means that the ability scores are meaningless as rolling 3 ones equals rolling 3 sixes.

The two are not related except that the higher ability score (which has equal probability as any other of coming up) gets a higher modifier; why? Well because of that same probabiluty; just like gambling.


We could talk in circles about this and bore all the non mathmeticians here but think about it; the modifier relates to the number not the role. If you take dice off (4d6 rolled) then the prbability of gaining a higher number increases (as you just yanked off the lowest 25% of your average).

----------------------------------------------------------------------

Actually, 3 sixes rolled nets an 18, a +4 bonus. 3 ones rolled nets a 3, a -4 penalty. The statistics of getting either is the same, and the mechanic works to the same level in the opposite direction.

Really? Please tell me how the probability of rolling dice has to do with a -4 or a +4. That was the point I had made in response to the "linear" and "curve" statements made by several others as well here.


The CONSEQUENCE is the modifier; that does not drive the probability of 3 1's or 3 6's. The chances are equal for either one. Feel free to look that one up - you'll find it under "Game Theory". It dates back to the 1960's. In their it identifies the proof why casinos will always end up ahead.


As I had said "gambling". When you roll dice to generate ability scores you are hoping that you end up on the for right of the curve.

Ability score generation has never been "linear." Even in point-buy, higher ability scores cost more than lower ones. It gets harder and harder to get to the extremes of the curve the more dice you use to roll, so if every human ability score is calculated via 15d6 as the first post in this thread suggests, ability scores in the "middle ground" (a range of modifiers called the Doldrums or Mundanes by my players) are going to be more frequent.

Point-Buy is not really measurable by any sort of probability as there is little "chance". It is a decision by an individual.


Even when rolling 15d6 the probability for any of the 6 numbers coming up on any dice is indeed equal. However given the mathematics involved and 15d6 the likelihood of the numbers being evenly distributed over a wider numeric spectrum is higher. However note that the probability of each die is still the same per die if you roll 1 or a million of them. It is the combination of those curves that push the numbers towards the middle.


HOWEVER if you do what most folks do and DROP some lower dice then the probability of having a higher average increases exponentially based on the number of die rolled. Tink about it. If I have


6
5
4
3
2
1

My average is 3.5

I drop the lowest 2
6
5
4
3


My average is 4.5

Choose any numbers you want (unless they are all the same of course - which has an equal probability of occurring as anything else per die; less probable for the whole lot simultaneously) and the same thing happens. Thus your statement is no entirely correct.

I hate to say this, but if you're bored by the mechanics aspects of the game and our mathematical discussion of this house rule, why are you reading this thread? If you don't want this level of math in a thread you participate in, there are plenty of other threads on this forum, and you can even start one of your own. So please leave us to our complex tinkering if you must, but don't criticize us for "boring" anyone with our "talking in circles"–there are many types of gaming discussions, and as this is a public forum we are free to get as boring as we want.

Did you actually read the statement I made? Here it is for reference:

We could talk in circles about this and bore all the non mathmeticians here but think about it; the modifier relates to the number not the role. If you take dice off (4d6 rolled) then the prbability of gaining a higher number increases (as you just yanked off the lowest 25% of your average).

The point I was making is that I, or others here most likely, could start throwing mathematics, diagrams, mapping of curves, etc. up on the actual board and others may not like it. Perhaps I am wrong. I also, in making my statement, did not want to start getting into a discussion about the actual calculation of probabilities - but it appears perhaps you do want that.

You don't hate to say it - you wanted to say it otherwise you would not have. I would venture to guess it is because you ran out of anything substantial to say given you spent half your post criticizing a portion of mine and virtually ignoring the meat of it.

 

[SBMC]

I don't think this helps at all. All it does is change the numbers. Current system means that 1 ability point results in a 2.5% difference in ability (A + or - 1 to a D20 roll is a 5% influince, and it takes 2 ability points to make a + or - 1). This seems like a perfectly fine range for me, especialy with size and other modifiers being used to adjust things like con, damage, carrying capacity, and such.

All very true.

Now, the flaws of the D20 system show up in the extremes, such as 1st level where a lucky hit can kill instantly and at 20th level, falling several hundred feet is unlikely to. Skill have similar problems, from too easy, to too hard. Changing the stat numbers won't fix the issue, and honestly I don't know quite what will. Since most adventures end up running 5-10th level, which is where the system is near it's best, it's usually not a big issue.

The idea of the "lucky hit" killing the 20th level character I agree with. However that could also happen at any level to anyone. I saw a 3rd levl expert (smithy) kill a 11th level rouge once; a lucky hit.

However the characters, in DnD, when the hit higher levels are supposed to be, in my own words "super human". Like the heroes in movies swinging their sword in the midst of scores of combatants and leaving unscathed. The aveage person in the world is first level; if your 10th that means that you are 10 times as skilled as they are. Adventurers are supposed to be excessivly rare as are people with the ability scores the would have.

The aveage human ability score is a 10; wizards normally start off with an Intelligencet score as close to an 18 as they can get. The figher the same with Strength (or maybe Dexterity). That is "8" "levels" above average. Modifier wise an 18 is 4 times better than the average (+0 vs. +4) (modifiers is what I think WoTC looks a when they came up with their Abiliy score ideas). Not to mention that, as an example off the top of my ehad, a heavy warhorse has a strength of 18; I can have a character literally a strong as a horse! At first level!

So as far as the ability to do things; I think that the characters should be able to do those "super human" things as indeed they are just that.

Just my opinion is all...