Post/Author/DateTime	Post
#1 zombiegleemax Sep 07, 2007 16:35:06	I love randomness, random attributes are great. They make you work round weaknesses, they give you strengths where you didn't expect them, they ensure most of all that everyone is different, and that not everyone is the optimal build. In short they create great opportunities for role-playing because flawed characters are great characters, all the best role-playing happens in that gap between a character being viable and a character being optimal. Sometimes 'real' heroes have to work with what they've got, not what they wanted, that's part of what makes them heroes. But randomness isn't popular, points and arrays are, but why? Well, way back when GMs had all sorts of rolling mechanisms, all designed to give a player the choice they needed to play a particular character and not be completely stuffed by the dice. Gradually over a lot of editions that became the 4D6 drop lowest and arrange method now suggested as the standard. And in doing that I think we killed randomness long before point buys or arrays, just take a look a this table:
#2 zombiegleemax Sep 09, 2007 21:17:36	Interesting system actually, it took me a little bit to figure out what all you had to do with it though. Where as I wouldn't enforce this on players, mainly because we play heroic campaigns where the flaws aren't in the stats but the personality. Or player for that matter. But what I would use it for is exceptional NPC generation, has it lends itself very well to that in my opinion. As a mention, I think that if you list out 1 safe stat and then the other 5 degrees of risk you might get the randomness you are looking for. The highest risk, 5th level risk, would obviously get the lowest stat will the lowest risk, 1st level, would get the second highest stat rolled.
#3 Book5 Sep 11, 2007 13:50:03	It is my learned and experienced opinion that a DM should "give em what they want" and it is our art to know what that is regardless of what the players say they want. Umm.. you asking "what are you talking about Shea?" Basically there is no real problem here. If your game is gonna be a character driven play then.. forcing people to "live with their randomly generated stats" is fine. I have played those games. In fact we played a game that was 6 - 3d6 "in order" and you kept the first run (the DM watched). We played an entire "childhood" game session with reduced stats and no class levels (yeah that was a little weird but we weren't focusing on the rules). After that session we got to slide as many as 4 points from one stat to another, and then choose a class. We were RP focused, and the "play the cards your dealt" method works better that way. Your entirely correct. I know a guy who always plays roll 7 - 3d6 set as you like it (throw one away) because his games are intended to be low powered "we're no supermen" games. My games are usually 7 - 4d6 (drop lowest.. and reroll ones or twos) set as you like it because I like power housing things. I don't want to have to worry about killing people simply because I wanted to have a 15 level psion juvenile blue dragon as the librarian. Ahh... memories... they killed that bastard. He did manage to take one of the players out though. They just "haaad to keep the book" ... couldn't "slow down and read it at the library" nooooo. hehe
#4 Dark_Stryke Sep 12, 2007 20:29:57	I'm sorry, but I disagree with you. A character that has to cope with a strange weakness or strength is not suddenly a better character to RP. I find that players are often upset when their character has an unintended weakness or a strength they could have used better elsewhere. If a player wants to RP someone with a strange weakness or strength, they can do that with the standard 4d6 drop lowest or even point buy, by just assigning scores in a non-optimal way. Forcing people to not optimize for the sake of roleplaying just seems like poor DMing to me. Also, take this for example: I want to make a wizard, so logically I should assign intelligence as "safe", right? It's always going to be at least 11. And I should assign something I don't care about as "risk", like charisma, because the score could be anything, good or bad. Well, let's say my first score over 10 is, well, 11. My very next roll? 18. Now I'm stuck with a virtually unplayable int 11 wizard, but I have an awesome 18 charisma that I don't care about! So, in order to avoid this problem, it's probably better to make something you don't want to be low, but don't care if its high into your "safe" score. Likely Con, for most characters. And your dump stat can be your "risk" score, because, while it wont add anything if it's accidentally high, you wont care if it's low. Maybe Strength for a wizard. So, how does this make for better roleplaying, at all? You'll have a bunch of characters with average to high con (or other scores they only marginally care about), and one score that's completely random. That does not make a better character at all. What it comes down to is that, no matter the type of ability score generation you use, roleplayers will roleplay. I personally have assigned a high charisma score to characters who could have benefited from having it elsewhere, but I've done it simply because I've wanted to RP a sociable person. Furthermore, no matter the type of ability score generation, optimizers will optimize. I've already shown you the most "powergame-like" way that I would assign ability scores using your method, it should be hard for others to do the same. Simply put, forcing someone to play a character that they do not want to play through use of enforced randomness is poor DMing. This style really only works when someone has no concept of what they want to play beforehand, which in and of itself lends to poor roleplaying.
#5 the_apologist Sep 15, 2007 12:27:28	First of all, ten sets of rolls on each table doesn't really prove anything. You'd have to go at least 50, more likely 100, for that to be a respectable sampling of data. For example, flipping a coin goes heads/tails 50/50. However, over ten flips I might have 7 tails and 3 heads. Only over very large samplings do statistical points work out. That said, I agree with your premise, that heroes should have to roll with some bad stats sometimes, and not just be able to dump their low rolls off onto a stat that's not going to matter. Flawed heroes really are more interesting, and if your players are good roleplayers it enhances the character as they try to explain why a clumsy oaf is a warrior, etc. However, in my experiences, the reason for arrays and tables is less a case of "I don't want to get screwed" and hinges more on two things: party balance and character variability. The first is a big one when you're playing with a big group. I have a group of nine people, and the Law of Big Numbers states that, over several campaigns or one-shots, someone usually rolls ridiculously high, and someone usually rolls ridiculously suck. In my longest-running campaign to date, we had a gnome that was eight levels bard and the rest barbarian. Her initial stat roll (and I watched this) was 16/16/14/15/18/18. On the other hand, we had a straight druid with rolls of 11/10/11/14/13/15. Just by the virtue of her rolls, the gnome was not only stronger and tougher, but quicker, wiser and smarter than the druid. The bard-barbarian, despite being a newer player without much of a head for tactics, dominated the game, and by the end she had more HP than anyone else, including the paladin and the fighter, even though over half of her HD were d6. This is just one (extreme but representative) example of a recurring problem: when you play with a lot of people, statistics says that you'll have a spread. Someone is going to have a worthless character, someone is going to have a character that blows everyone else away. Character variability could be considered a lesser concern, but it's important none the less: in D&D, you're designing a personal avatar to represent you on the field of play, and the game is supposed to allow you to pick the avatar you decide upon, within reason. You should be able to make your character concept pretty easily. However, and this happens in almost every game we play, the pure randomness of die rolls make your decided-upon character concept inviable. Time for another real example, this time with me as the slighted. I was building a rouge who was going to be a suave, swashbuckling noble's son. I wanted his dexterity and charisma to be high, and I wanted a severely low wisdom and slightly low constitution to make him interesting to play. However, my stat roll was 13/14/14/12/10/11. Meanwhile, the Wizard rolls 18/17/13/10/11/13. The wizard puts his 18 into Intelligence and, because he has low HP and wants to be able to dodge, his 17 into Dexterity. Just by the sheer randomness of the dice, my character concept is foiled. I wanted someone who was slick and quick and thick as a brick, someone who could steal the rings off of your fingers and get away before you noticed, but if you hit him with a crossbow bolt you'd probably finish him. But the dice said otherwise. Of course, I was free to -lower- my stats to whatever I wanted, thus achieving my -low- stat ideas. But I was stuck with a max of 14 in any given stat. The spellslinger was faster than me just for existing, and because of the gear-advancement based nature of D&D I would have to be level 16 before my base dexterity could overcome his, no matter how hard I trained or how much I focused my character in that direction. So arrays and point buys are popular not because the players despise having even one low stat. Heck, most of the people in my games drop a point or two off a stat to make their character more interesting. The bigger problem is that the randomness invites all manner of unintended consequences. And as much as getting something you didn't ask for and playing with the cards you're dealt is a part of real life, D&D isn't a real life simulator. It's a game that employs verisimilitude only where it enhances the fun of the experience. Players should be able to create the concept they see in their head, and there should be some reasonable assurance to the DM that no matter what their concepts are, the players will be somewhat equal in power unless they choose not to be for roleplay reasons. The randomness of rolling stats often accomplishes neither, especially in large groups where statistics can really get played out.
#6 zombiegleemax Sep 17, 2007 5:54:05	Thanks for the replies, there are a lot of good points here. One mistake I made in the original article was to wrap up the idea of risk/safe stat with the attribute system. Dark Stryke's complaints are mostly about that and I would agree with him. The purpose of the dice rolling mechanic was to avoid exactly what The Apologist was getting at. The big three problems of random dice generation: ---------------------------------------------------------------- "I have a group of nine people, and the Law of Big Numbers states that, over several campaigns or one-shots, someone usually rolls ridiculously high, and someone usually rolls ridiculously suck." This is problem number one. Points values and arrays solve it by making everyone equal, which is often undesirable. However it is not as The Apologist believes a product of dice but rather of the particular dice rolling mechanism used. The chart below shows total attribute bonuses for the 4D6 method and for a variant on the mechanism above The red line shows the 4d6 (with its +8.2 typical total mod), against the blue method VII (which has an average of +5.7), from a sample of 498 characters. You can see how much the variation amongst characters is reduced. The reason that the problem with one character being more powerful than another is there is because the 4D6 method encourages it - it would not be unusual to have characters with 10 or 11 points between their total bonus in a party. It is this element of variation which needs reducing (but for those who enjoy randomness not completely eliminating). ---------------------------------------------------------------- "I wanted someone who was slick and quick and thick as a brick, someone who could steal the rings off of your fingers and get away before you noticed, but if you hit him with a crossbow bolt you'd probably finish him. But the dice said otherwise. Of course, I was free to -lower- my stats to whatever I wanted, thus achieving my -low- stat ideas. But I was stuck with a max of 14 in any given stat." This is problem number two, and I think the biggest one. Characters that don't have bad stats but are simply very average. Again, I think this is something that has been encouraged - and one of the points I was getting at - to avoid characters being completely nerfed by poor stats, more and more dice are rolled, which means characters tend more and more towards an average. This is essentially a matter of the highest attribute, which is shown below: The chart compares the characters highest attributes for the 4D6 method and method VII. Remember method VII has an average total bonus 2.5 points lower than the 4D6 method, yet it still avoids a lot of the 11, 12, 13 highest results. ------------------------------------------------------------- "Heck, most of the people in my games drop a point or two off a stat to make their character more interesting." The problem is they have to. Frankly you have to be bloody unlucky to get a -2 attribute under the 4d6 system. The system never really presents you with an opportunity for an interesting flaw. The table below shows attribute distributions: Notice how vanishingly small the chance of getting a low attribute is under the standard system. It is at exactly at this point in the system that you really want randomness, and it is this point in the character generation system that the various mechanisms kill randomness. ------------------------------------------------------------- "So arrays and point buys are popular not because the players despise having even one low stat." And I wasn't trying to say they were. I've played not just D&D but hundreds of games over several decades and the point I was making was that the 4D6 type dice rolling systems were a response to players getting poor attributes and that the attributes and points were a response to the inadequacies of those heavily 'averaged' dice rolling mechanisms. However, they are a product of a particular sort of dice rolling, not dice rolling in general. "The bigger problem is that the randomness invites all manner of unintended consequences. " Yes, but some unintended consequences are bad and some are good. Anyone who has played the game Golden Heros will know that it produces characters that would never be built using the points build systems in other super-hero games - it doesn't matter who your group are. And those characters and games are more interesting for it (but harder work for GMs). Randomness will often give you the nudge you need to really use your imagination (and your character will ultimately live or die, succeed or fail by dice rolls so you have to deal with it at some point - or go and play Falkenstein or Amber). The Method VII is a variant on the system I suggested above (I ran a number of different variants in a spreadsheet after I had written the original article). The core elements of all the systems were: You have three types of roll, a high roll, a medium roll, a low roll. The high roll is used when you have had two low attributes (whatever you define low as), the low roll when you have had one previous high attribute. Method VII which I put in the chart had a high roll of 2d6+6, a medium of 3d6, and a low roll of 2d6+1. You made your first roll as a high roll and the trigger for a low was the last attribute was 15+, and for a high roll was the last two attributes were 14-.
#7 icefractal Sep 19, 2007 13:51:53	What's wrong with this table?
#8 zombiegleemax Sep 21, 2007 22:53:22	Your post is based on the flawed assumption that being weak makes you better at role-playing. Your personality is completely unaffected by your ability scores, and you role-play with your personality. Prove to me in any way you can devise that having low ability scores makes you a better role-player and I will take back this statement. Anyway, as others have said, randomness breaks the game. My first campaign involved our favourite level 1 orc barbarian with 22 Strength called Ug. First boss fight? He criticals and deals 41 damage. Splat. Later he was weakened when the level got to the point at which casters take control, but you get the picture. Rolling 18s can kill balance when other people get mediocre rolls. And this makes things worse by forcing you to put certain scores in certain abilities. You can balance a game where everyone is a level 1 commoner with 3s in everything (and he would be the perfect character to role-play because he sucks, I'm sure), and a game where everyone has 18s in everything and gets 50 LA worth of free templates. But how do you balance a game in which the wizard has less Intelligence than the barbarian, who can rip every enemy to shreds in one hit? Unless all your enemies know not to pick on the wizard and stab themselves for critical damage every time the level 20 fires his only spell, Magic Missile, at them, it can't work.
#9 uberwoot Sep 22, 2007 10:16:59	I would just like to point out that your system does not -always- guarantee your "safe" attribute will be at least 11. While statistically unlikely (bordering on impossible, since you would have to roll at best [1, 1, 1, 2] on four successive rolls- a chance close to one in four and a half billion, not even counting the probability of not rolling better than a 10 on your first two rolls), it is possible to never achieve an 11. In fact, it is possible that your best score is an 8 if your dice are really, really off. Ok, I'm done playing devil's advocate for now. /woot
#10 Book5 Sep 30, 2007 23:59:15	While I must smile and nod saying that being forced to deal with stats does not make you any more capable of RPing ... what I am saying is that only hard core RP players would actually want to. In a combat sim point build and array assignment is a lot more "fair" than rolling any number of dice, dropping or rerolling etc. If you going to roll you dice first .. you should pick your class second. The character and background and everything is generated from the array of stats not vice versa. Thats Old school, when you were leveling what? To three or six or what not. Kinda like that ol' literary exercise "write a poem from three words" ya know? Some people are real big on it - usually 2nd ed heads.
#11 bekeleven Dec 23, 2007 23:16:38	Best solution I've found so far: Campaign Arrays. My gamers looooove the power, so they will want good rolls. I give them what they want. Make 6 rolls (make them better than elite array!) and all the character use them, assigned how they see fit. I used the following stipulations: One 18. Most players I have will actually commit suicide with their characters if they don't have it, but that's a problem for another post. At least one other great stat (16-17). I encourage them to play multi-stat characters like Pallies, so I try not to screw 'em. About 3 10-15s. Once again, this is so the players can be versatile, and to encourage MADs. At least one negative. If you want to role-play, this is your chance to shine! If not, suck it up. You're not perfect. This usually ends me up with a lot of unattractive party members, but that's hard to avoid. With this I ended up with: 18 17 14 12 11 8. Once again, a little bit powerful, but that's how I run 'em. ...And, of course, the Stormwind fallacy figures pretty heavily in this topic. It's been mentioned but not by name.