Chess is not an RPG: The Illusion of Game Balance

[Celebrim]

So your solution is give them a bunch of point buy arrays and they should be happy. I've tried giving arrays of pregen stats to my groups, they do not like it. And it doesn't matter if the stats are good or bad.

Of course not, if everyone is special no one is. But if everyone got lucky, well that's different, right?

 

[Celebrim]

Let's look at it my way. [MENTION=4937]Celebrim[/MENTION], who do you think you are to tell me that I don't actually like the Pepperoni Pizza that I've been enjoying eating for the last 20 years? Do you have any idea of how offensive it is to hear that I should eat a Mushroom Pizza instead because you say it is far superior when I have an extreme irrational dislike of Musrooms to the point that if I had a time machine, I'd go back in time and substitute Mushrooms for Jews and have Hitler slaughter millions of Mushrooms instead?

Godwin's Law. Don't be telling me about offensive making that analogy.

 

[Bedrockgames]

Of course not, if everyone is special no one is. But if everyone got lucky, well that's different, right?

Except everyone is not getting lucky. They are getting random stats derived from the 4d6 or 3d6 method. Some are employing two sets because they want the average slightly higher than normal. Some eliminate completely hopeless characters (which can't even really be made using the point buy system because stats begin at 8 in that). These are very simply two different preferences that serve differing tastes. They don't reflect moral or gaming superiority. There is nothing to be gained by lying here. If people really wanted to be assured higher than normal stats, they could achieve that much more easily by doing a 32 point buy method.

 

[Celebrim]

There is a big difference between chucking completely hopeless characters and rolling sets of 4d6 until you get overall higher stats. In the former case you are rolling and accepting what you get, but you might roll a new set if you got like three 6s and nothing over 12 (definitions of hopeless vary though).

In actual practice, I'm not sure there is. Keep in mind that definitions of 'overall higher stats' also vary, and definitions of hopeless will depend on those definitions. A functional definition of hopeful in 1e might be: "At least 1 16 and doesn't have a 5 or less in a conflicting category (for example 16 wisdom and if a 5 or less, then in dexterity) OR at least 2 15's in base class prime requisites (not charisma, in other words) and no 5's or less, at least 8 intelligence and at least 7 dexterity, OR qualifies for Ranger (2 14's and 2 13's and no 5's or less)." So hopeless might be deemed everything else. But another group may have different standards.

The important point is that everyone actually wants to be above the minimum standard, and preferably above them by a good deal. So take the case of a group that is happy with 4d6 drop 3, but also agrees that truly hopeless characters can be rerolled. And, look at the list of example ability scores generated by 4d6 drop 3. Even if the group doesn't do point buy, lets evaluate them as point buy with the idea that average stats are like 28 point buy. The first thing you note is that most sets end up being above 28 point, and therefore satisfy the players desires and expectations to be above average. A few are really above 28 point buy by a wide margin. However, there are a smattering of results where the system generated 9 point buy, 15 point buy, 12 point buy, 18 point buy and so forth. So imagine that happens. Well, SURELY everyone at the table will concede that's just a fluke, a hopeless character, and should be rerolled.

As soon as that happens, you've thrown randomness basically out of the equation. Imagine the similar situation in game where you throw a dice, it's a remarkably low rare result, and you say, "Well gee, that's not supposed happen. I'll just reroll the dice." Once the game starts that is called 'fudging' or 'cheating' depending on the demeanor of the table. What it really means is, "I had a result in mind. This wasn't it. But instead of actually admitting to myself that I'm choosing the results, I'm going to just reroll the results... until I get the result I was going for all a long." For some reason emotionally, for irrational creatures like humans, this lets them mentally believe that they aren't actually choosing the result. But that doesn't mean that this emotional conviction is in any way rational.

Once you grant that the player can reroll until he gets a non-hopeless character, look at that table of results again. By and large that first reroll is going to produce a 'correct' result. A few players may get a disappointing result just below 28 point buy, but there are lots of oppurtunities to 'win' once we throw out all the losers. And observe also what that is doing to the average result. If 4d6 take the best three is on average 28 points in point buy terms, 4d6 take the best three and keep rolling until you get a 'non-hopeless' result is in practice something like 36 point buy. Because the standard deviation is huge, so once you throw out the bottom 20% or so of scores, the remaining scores are really good indeed.

That doesnt ean your using 4d6 to somehow game a point buy you were never employing in the first place.

No, of course. I'm not claiming that the players had some knowledge of point buy to compare it too. I'm just using that as a means of measuring just how good, or not good, the various rolls are. I'm just showing just how wide the range of characters real randomness would produce if it was actually employed in earnest - which I'm asserting it almost never actually is.

Worse for me though was the fact that this illusionism around randomness meant that the bar on what was hopeless was being continually raised, particularly as people began to figure out what they actually needed to have the best shot of a highly successful career with a character. When we were using 3d6 straight up, that was a pretty low standard. At least not mostly scores lower than 11 was enough, which should have clued us in right away that our definition of hopeless was already 'anything below the average expectation of the method'. However, those characters tended to have short lives compared to the few lucky ones, and most people were - if not exactly cheating - working around the rules any way. So we went to 4d6 and the standards of what was playable went up, and conversely what was hopeless went down. In practice, it became 'not mostly scores lower than 13'. 12, 12, 11, 11, 9, 7 which might have been considered playable previously, gradually became understood to be hopeless. And I discovered that, I really couldn't make people play what they didn't want to play. So scores got better, but the percentage of 'do overs' - via some methodology - didn't decrease. And as I understood the math better, my ability to see why 13, 13, 13, 13, 13, 13 wasn't actually a good result, but a bland character that would always be inferior to his peers lagging in XP and with no reliable abilities, my understanding of what a good character for the system looked like evolved. I began to realize that guy with a 7 and two 8's wasn't actually paying much of a penalty if he had a 17 and a 15, and making him play it wasn't really a hardship nor was choosing to play it being really hard core.

 

[Bedrockgames]

Again i am not going to spend all day taking this point by point. Point buys do not appeal to some folks who like 4d6. One reason for this is randomness. You keep creating false choices or say randomness no longer matters once players remove hopeless characters (and your definition of hopeless doesn't match what I have ever seen in actual play, where it usually means something like three really bad stats below 9 and nothing over 12 or 13---but this will vary from group to group for sure). The fact is with point but you select the numbers you want or you start out with presets someone devised before hand. You can almost achieve the same numbers with point but that you can with rolling 4d6 drop the lowest---point but starts at 8 so unless there is a way to reduce that I am forgetting the big difference in value is point but won't see numbers like 3,4,5,6, or 7. And because you can set point but to any value really the only difference between Lonny but and 4d6 is the randomness and the fact you roll. So I think those are the only real 2 explanations for why people like 4d6---the randomness or the act of rolling. If they really just wanted to exceed a 25 point buy they would be doing like a 28 or 32 point but method. I suppose the only other possible explanation is the range of values: yes you can get 1 or more high stats with 4d6, but you can also get 1 or more very low stats with 4d6. So they might also like it because it gives you the possibility of having an 18 AND a 3. Either way probably best to take people at their word rather than essentially accuse them of lying or being deluded.

 

[Bedrockgames]

Of course not, if everyone is special no one is. But if everyone got lucky, well that's different, right?

Again looking at the character sheets from some of my own campaigns, the values vary. People seem to like having a mix with some folks getting above the 25 point value and some below. I don't see how you explain this as them secretly wanting to bust the 25 point buy cap when A) the GM is free to set a higher cap and B) both the players who have under 25 and over seem happy with the method, their characters and the campaign.

 

[prosfilaes]

not all people who do 4d6 drop the lowest do two sets though. Lots of people do one set and only eliminate characters they consider hopeless. We have been covering a lot of ground here dealing with many different takes on the method.

And I have different responses to the different takes. I'll let Celebrim handle the 3d6 crowd; I'm skeptical, but whatever. But 4d6 drop the lowest, two sets, is a method that produces 3/5ths broken characters and 1/5 weak characters. If you drop any hopeless characters out of that, you've got a method for building straight-up overpowered characters. If the players and especially the DM don't know that, it skews how they view the game and interferes with the DM properly running the game. To properly handle powerful characters, one must know how and why they are powerful.

Yet it doesn't bother me that people want 4d6 and say they want it because it gives them that spark of random, a bit of excitement and a little cushion against weaker characters. Seems reasonable to me and I will happily adopt this method if that is what the group as a whole wants.

Things are more likely to seem reasonable to you if you get really annoyed at anyone arguing about them or even discussing them and refuse to look at how they actually work out.

 

[Bedrockgames]

I am happy to discuss the topic but I don't have enough time in my day to spend on posting responses if folks are going to squeeze insults into their rebuttals. All I can say is if you like point but, the. Go for it. But I am not going to question the motives of folks saying they like random rolls when that seems quite a reasonable statement in light of the reasons I provided earlier a. What does not seem reasonable is assuming people are lying or have a hidden agenda because they say they like the random spark 4d6 provides.

 

[Celebrim]

Again i am not going to spend all day taking this point by point.

I'm getting a bit tired of the topic myself, and finding it to be a longer and longer digression from the thread topic. To sum up, most methods of random chargen produce highly unbalanced characters. To resolve this, players that use chargen systems based on randomness (for example 1e D&D), develop elaborate ways of controlling for the randomness to produce less unbalanced results.

or say randomness no longer matters once players remove hopeless characters

Because by definition, if you steer the results, it's no longer random.

(and your definition of hopeless doesn't match what I have ever seen in actual play, where it usually means something like three really bad stats below 9 and nothing over 12 or 13

Which is pretty much exactly the examples that I gave, for example, "At least not mostly scores lower than 11...", or "At least not mostly scores lower than 13" And obviously, from my explanation its pretty clear that I considered just about anything with at least 1 16+ playable. So, actually, I'd guess we are pretty much exactly on the same page - exactly as you'd expect if my theory that what was 'hopeless' was derived from expectations about the mathematical averages.

As for the rest, it's not worth responding to at length. It would be fairly easy to devise point buy to allow for scores of any range, and in a well balanced system this would be fairly easy. I'm pretty sure the only reason that D20 didn't do this by default was an admission that not all abilities were created equal in the default game. Likewise, the choice to use point buy to create cookie cutter characters is just that - a choice. And if when presented with the opportunity to create cookie cutter characters you do, then chances are diversity isn't the reason you like randomness. Exceeding 25 point buy implies that there is a choice between using point buy and random - which there often isn't. I never encountered point buy as a choice for D&D until I played 3e. So if the goal was to exceed the equivalent of 25 point buy, only by manipulating 'random' chargen could you consistently do that. However, even where point buy a choice, there would still be people who'd prefer random to it precisely because you can manipulate it and do so without really reflecting on the fact that you are doing so. There is a certain boost to ones self-esteem that comes from feeling you earned a 40+ point buy character when you 'randomly' roll it up, for example, that you don't get when the DM says, "Sure, make 45 point buy characters." People would rather think of themselves as "hard core" and "old school" and whatever.

yes you can get 1 or more high stats with 4d6, but you can also get 1 or more very low stats with 4d6. So they might also like it because it gives you the possibility of having an 18 AND a 3.

Except no one actually plays characters with multiple 3s nearly as often as they play characters with multiple 18s - or at all. For one thing, if you actually did randomly generate a 1e character with 2 or more 5's or less, it's literally unplayable (by the rules, you qualify for no class). We've already established that the possibility of playing 1 or more very low stats doesn't really exist (except in parallel to also having mitigating good stats). Everyone in this conversation admits that do overs were done when the character was "hopeless". So in fact the real results aren't sometimes high and sometimes low, but high and high. And that isn't actually random - though some of us aren't admitting it. You can pretend its random all you want, but after you've admitted in practice the methodology isn't random, it's no longer unreasonably to note that it isn't random regardless of what you claim about it being random.

 

[billd91]

And I have different responses to the different takes. I'll let Celebrim handle the 3d6 crowd; I'm skeptical, but whatever. But 4d6 drop the lowest, two sets, is a method that produces 3/5ths broken characters and 1/5 weak characters. If you drop any hopeless characters out of that, you've got a method for building straight-up overpowered characters. If the players and especially the DM don't know that, it skews how they view the game and interferes with the DM properly running the game. To properly handle powerful characters, one must know how and why they are powerful.

Considering 4d6 is the default method of character generation, if it tends toward generating overly powerful characters (in your opinion), then maybe the problem is the point-buy value low-balls the amount PCs should be given. In any event, rolling 4d6, even rolling two sets and taking the preferred result, certainly doesn't create overpowered characters nor are they broken. I'd venture to say that they are less game breaking than point-buy characters considering randomly rolled characters tend to ameliorate differences between single-attribute and multi-attribute classes - a problem fed by point-buying stats.