If we all rolled the normal way for stats, how come he has three 18's?

[FreeTheSlaves]

First up just to be clear Philotomy, I’m enjoying what you’re saying. I’m just exploring this topic again from a new viewpoint.

Of course, "unviable" is qualitative. In any case, "unviable" can be addressed by throwing out a "hopeless" set of rolls, or by skewing the rolls so you get numbers within an acceptable range (e.g. 2d6+6 or whatever creates the range you're willing to consider). Or by playing the hopeless PC -- if he's truly hopeless, he'll die. Actually, I've seen some "hopeless" PCs survive and do fine; personally, I find that good play is more important than good stats, but that probably varies with the group and with the system.

I don't think unviable is wholely qualitatitve, poor stat placements can lead to poorly performing characters. An extreme example could be the high int coupled with low dex and con, while the overall stats do not qualify for reroll. (Also, a reroll system imo is a bit of a cop out – shouldn’t a system be self-contained and require as little looping as possible?)

I have played 3d6 in order and had fun. However I think we had fun despite the system, rather than because of it. Making lemonade out of lemons and all that. The 4d6 in order (organic) method proved to be better, but even so we felt the need to skew it by making it 2d6+6.

Skewing the rolling system is another approach I'm perfectly fine with. So long as the worst result can co-exist with the best result without disruption, such a system can capture the best aspects of the rolling methods. I'd imagine a more perfect system to have relatively small variations in results and probably have something like a different method for every 'set' of stats. A set would be a sub-grouping of the 6 stats that would have their value range predetermined.

I doubt anyone really benefits from going through the motions with the suicidal character. It’s more work for the poor dm and disjointed play for the other players. It’s not a solution I’d like to contemplate.

Could “good play” be independent of all methods? (I’m also not sure what you mean by good play, is that fun play or tactical play?) Yep, I’ve seen the player hurl his 17,16,17,16,16,16 fighter into certain death and been bummed out by the result. I guess that’s not good play however you define it. But I think that player was prone to act like that whatever the method.

However, low stats needn't be an automatic deal-breaker for player satisfaction; it depends on your assumptions going in. As I mentioned up-thread, a good deal of the fun in rolled stats is the challenge of "playing the hand you're dealt." You see what fate gives you. You exercise your creativity to make an interesting PC from the rolls you came up with. You look at low stats as a challenge -- can you hack it?

The short answer for myself is “no, I cannot hack it.”

I don’t want poor stats but I'm creative. Give me good stats and I’ll remain just as creative, and I'll be happier and have more fun.

It would be nice if whatever the method used, I would get nice high stats – with a bit of variety because high stats need context to be regarded as high. I’d also like that variety to exist on my character sheet, not by comparing my sheet with the sheets of my friends.

 

[Felnar]

someone check me on this:
the odds of getting an 18 on 4d6 drop lowest are 21:1296?

 

[SavageRobby]

I get 1.6% - so yeah, thats about right.

 

[Zustiur]

OK, when you say "six sets" do you mean roll, for example, 3d6 a total of 6 times, or do you mean roll 3d6 a total of 36 times, in groups of 6? I am not familiar with using the term "set" in this way. It would help with the calculations.

I meant sets as in, you roll one set for each ability.

That is, you roll 3d6 for strength, this is one 'set'
3d6 for dexterity is the next set, and so on.

Zustiur

 

[Dr. Harry]

Here's a little experiment you can do: You'll need two friends. Give them a coin. Have one of them flip the coin 50 times & record the results. Have the other one make up a list of results on their own. Tell them to decide who does what after you leave the room. When they're done, come back & look at the lists. The one that you think looks random will be one that was made up.

The thing about dice is that they are random. While even straight 18s may be rare, it can to happen. If you're using dice to generate ability scores, then you have to accept that that is a possibility.

Of course, I prefer to either have everyone roll their scores at an actual session or go point-buy if that isn't an option. Because, personally, I enjoy getting good rolls infinitely more when there are others there to experience them with me. Engendering trust is just icing.

I kind of picked this post to respond to randomly, so no offense is meant anywhere ...

The odds of ending up with a score of x is y (for an individual score).

x y (%)

3 0.08
4 0.31
5 0.77
6 1.62
7 2.93
8 4.78
9 7.02
10 9.41
11 11.42
12 12.89
13 13.27
14 12.35
15 10.11
16 7.25
17 4.17
18 1.62

If anyone is curious about my methodology, I use brute force, as well as copy, paste, and fill down to set up a table of all 1,296 possible results in Excel.

The odds of rolling six 18's in this way is (0.0162)^6, as I make it, or 1.81 x 10^(-11). We should expect to see all 18's come up 1.8 times for every hundred billion sets of rolls.

This does not account for other influences, such as:

1) This roll does not count because the dice is touching that piece of paper/book/bit of cheeto.

or

2) Rolling the dice consecutively and aiming the throw of a later die to hit and effectively reroll a previous, less desired result.

or others. Maybe it would be a fun thread to collect as many dice rationalizations as possible. That might even overtake the Hive Mind thread

There is a 1.34% chance that someone could "roll without rationalizations" and get no scores lower than 13. If the player rolled ten characters and only kept the best one, there is a 12.7% chance that at least one of them has no scores lower than 13. For twenty characters, the odds of "success", (if you want to call it that) go up to 23.7%


The odds of a character getting at least one 18 are 9.30 out of 100, or 9.30%

If I have done the math correctly (just flying off the top of my head) the odds of a character getting at least two 18's would be this 9.30% for one score to be 18 out of six times the chance that one of the five remaining scores is an 18, or 0.73 %. The odds of three 18's would be 0.046 %.

(Looking at it like this gives me a chance of getting six 18's as a more common one in one hundred million, as opposed to one in one hundred billion, so I think that the numbers in the previous paragraph trend high. Why? I haven't got time to sort it out right now.

 

[Elf Witch]

No, it's true you can get lucky. I call even one 18 very lucky. An 18 and 17, also very lucky, but very possible.

But 13, 14, 15, 18, 18, 18 rolled in that order isn't lucky. It's impossible.

It is not impossible. We roll for stats and we roll in front of each other. One of the players rolled 14 15 11 18 18 18 in that order.

It is highly unlucky but not impossible.

 

[Dr. Harry]

It is not impossible. We roll for stats and we roll in front of each other. One of the players rolled 14 15 11 18 18 18 in that order.

It is highly unlucky but not impossible.

Three 18's should happen in about one of every 2,000 characters using 4d6 six times and drop the lowest.

Other methods such as 4d6 seven times, drop the lowest, disregarding rolls that "don't count", and especially rolling character after character will dramatically increase those odds. Not that I'm accusing Elf Witch's compatriots of doing that (if they only made one pass through the 4d6 and rolled that, I'd give 'em a pat on the back and tell 'em to live it up), but the situation that led to this thread is why I either give players a point-buy system or give them one roll that they make in front of me.

I used to have the opinion that the point buy took something, some "let the dice fall where they may" out of the game, but I've seen a *lot* of players cheat here.

Part of this may harken back to when characters who wanted to play a certain class had to have high numbers in one or more stats, like the paladin.

 

[Jolly Giant]

One of the guys in my group is always making a mockery of the laws of probability when rolling his stats. You know how the core rules allow you to reroll your stats if the sum of your modifiers are no higher than +1, or if you doesn't have a single stat over 13? Everytime he rolls up a new character he has to roll several times before he manages to roll a set that doesn't qualify for a reroll.

The last time we started a new campaign, he had to roll 14 times before he got a set that was useable. He had several sets of stats where the sum of modifiers where from -4 to -6, and in two of his stat sets his highest stat was an 11!

 

[RFisher]

Three 18's should happen in about one of every 2,000 characters using 4d6 six times and drop the lowest.

People get into trouble, however, when they don't realize that this is still talking averages. It's entirely possible that you get three 18s in a row for three characters in a row. (...& then--perhaps--not again for around the next 6,000.) It's entirely possible to roll 6,000 characters without ever getting three 18s.

This is the thing that can really be hard to wrap your head around with probabilities. When it comes down to it, it doesn't really matter how improbable something is. If it is possible, it can happen.

 

[Dr. Harry]

People get into trouble, however, when they don't realize that this is still talking averages. It's entirely possible that you get three 18s in a row for three characters in a row. (...& then--perhaps--not again for around the next 6,000.) It's entirely possible to roll 6,000 characters without ever getting three 18s.

This is the thing that can really be hard to wrap your head around with probabilities. When it comes down to it, it doesn't really matter how improbable something is. If it is possible, it can happen.

True enough. On the large scale, if you gather several thousand people's experience together, someone on there will more than likely have the three 18's. It might be that when people consider characters, they don't put it in context with the number of rejected characters (like the computer rolling fellow whose character sheet noted that his cool, lucky stats came on his 27th try) to realize just how many charaters have been rolled.

Where this becomes problematic is the "it could happen so it can happen" not when used by RFisher, but when used by the player walking in the door with his fourth character with 3+ 18's in gaming group campaigns.

One guy in a group I was in who I know just out and out changed his stats (I would consider this cheating) had a character in a Marvel Super Heroes game. In MSH, the character generation requires rolling percentile dice - this guy had to have rolled 9/18 above 90, with three 00.

This was not a "let the dice fall where they may"