Post/Author/DateTime	Post
#1 oxybe May 30, 2011 16:10:48	i was wondering if anyone's tried using 3d6 instead of a d20 for task resolution? i've been thinking about hacking the game apart and applying "fixes" to the issues that have been irking me as of late. one of said fix is how often PCs seem to miss. 4th ed generally requires 10+ to hit, giving most characters a 55% chance of success in most tasks. 3d6 leans heavily towards 10-11, with a 25% frequency of these numbers. however, it also has a 61.5% chance of rolling between from 10-18 (assuming we allow 16-18 to make critical hits and on attacks 3-5 would be a crit fail, both occurring with a ~5% frequency, same as a d20). for higher crit values (like 19/18-20) you can gain a similar approximate with 15-18 (~9.5%) 14-18(~15%).  has anyone tried this before and what were the results?
#2 slobo777a May 30, 2011 16:33:42	GURPS uses this. My experience is probably more with the old Traveller system which used 2d6 for everything, but I've also homebrewed my own dice systems with variations on simple bell-curve systems. Problems:    * Slower. Finding 3 dice and adding up 4 numbers (3d6 plus your skill bonus) takes longer than tracking one die and adding up 2 numbers.    * Probabilities change rapidly in the central area, which means you have a much narrower meaningful range of values to play with. Your own suggestions for crit ranges bear this out. To get between 30%-70% success range in 1d20, there is a band of 9 different offsets, giving plenty of room for feats or bonuses that give +- 1 or 2 steps. For 3D6 the same range is roughly covered in 9,10,11,12, or 4 different offsets, almost no room before you hit the edges of the distribution. Optimisers will generally prefer 3d6 to 1d20, because it is much easier to lock it down to auto-hits or auto-misses by choosing stat and feat combinations. To address the second issue, if I were going with a bell curve for D&D I would go for something like 3d10, and make everything base off 15 rather than 10. That would give IMO about the right amount of wriggle room for all the stats and feat bonuses to work mathematically OK, but it would mean a lot more mental arithmetic to keep the game moving.
#3 slobo777a May 31, 2011 5:00:24	More thoughts on dice systems and "bell curves" in general: 1) You generally use a bell curve or narrower distributions when you want results to be predictable, and to strongly favour the side with an advantage. IMO some skill rolls would benefit from a bell curve roll instead of the two extremes of 1d20 or take 10 . . . but at least the designers have thought of this by giving options such as take 10. The skill challenge mechanic could also be used in this way, to average results over many rolls, although in practice this makes for boring skill chllanges with "skill spam" and is not IMO a very fun way to use them. 2) You generally use wider flatter systems where you want randomness to be important short term, and have advantages play out over the longer term. When the system is wider, you can dish out lots of advantages in relatively small doses, or allow them to stack, and this stacking becomes part of game play. Combat works fine like this, especially as build and tactical choices combine together to construct advantages, you need a wider distribution to accomodate them and make combat exciting/non-predictable. 3) My earlier post is really about the "random width" versus the "systematic width" of any distribution. Whether or not you use a bell curve is largely irrelevant. If the systematic width is larger than the random width, the players or DM have a very easy time making parts of the game highly predictable - for better or worse. Generally players in a point-buy system are going to be looking for advantages they can have long-term, and that makes narrow distributions highly vulnerable to breakage through Char Op. 4) Open-ended distributions with "special" end points such as fumbles and criticals are similar to increasing the random width. 5) You can also move to "two dimensions", e.g Number of dice v Target number, or Number of dice to roll, number of dice to keep. Typically one dimension is stronger, more game-changing than the other, and only a very few game options change it. I have never seen a balanced game system that used these kind of dice systems and was not vulnerable to abuse. My favourite system for these kind of rolls was 7th Sea, because they cleverly made *all* your stats relevent to combat in different ways. 6) D&D 4E has started with the D20 random distribution, and a lot of game elements are designed around it. You should only consider changing random elements in 4E where you feel random component of the game is "broken" or unfun in some way. For example, I might consider having variable Passive Perception based on a 3D6 roll after every Short Rest, rather than the fixed 10+Perception bonus, because auto-detects (or auto fail-to-detect) are less fun to me than having a random element. You don't explain which part of 4E you consider needing a narrow, predictable bell curve to "fix"? What's not working for you?
#4 slobo777a May 31, 2011 6:36:26	Finally, I re-read your OP, and noted you are concerned about how often PCs miss their attacks is the problem. They are unhappy to be missing so much. D&D 4E has an element of gambling in it, esepcially when you roll out a heavy Daily power - it is frustrating to spend that resource and see it fail due to a bad roll. A 3D6 bell curve works for this provided the players would currently hit on a 10 or better. However, it would equally apply in other important ways: 1) Monsters that hit on 10,9,8 or better will hit more often as well. E.g. Soldiers, or monsters that are specified as accurate-but-low-damage. The players might find it equally frustrating to be hit so reliably by the monsters. 2) Anything that needs an 11, 12, or 13 to hit will be worse off with the new system. E.g. Brute monsters, anyone suffering a penalty, builds that fall behind the to-hit curve (e.g. most Defenders) 3) Trade-offs that reduce accuracy for increased damage would be shunned. 4) Tactical issues such as cover/concealmenst and combat advantage will dominate battleboard play. 5) Differences between defense types become a Scissor/Paper/Stone affair - once a monster or player can make use of a vulnerability (or their own invulnerability) they can push that button. None of these prevent you going 3D6 instead of D20 . . . but personally I would look for an alternative solution. Find balanced ways that the players could gamble less on hit or miss, but with lesser returns. There are plenty of PC builds that can already offer this without altering the dice system - reliable daily powers for example, or powers that are based more around the effect line. Unfortunately the players might find that less satisfying too - what they probably want is a magic die that always rolls 15-20's when it's required so they can hit with massive damages and effect that they are hoping for.  And if they actually had this, then they would discover that actually this is quite boring. Ultimately, I prefer to have the gambling element still in the game. I want part of the game to involve taking risks, and there's no risk without the chance of failure. Disappointment from a miss, and enjoyment from a hit are part of the game, and you don't get the latter without the former.
#5 oxybe May 31, 2011 12:37:43	grr... power outage at work ate my post... i'll try to be a bit more succinct in case of a repeat power blink, so i apologize if i might come off as a bit rushed. first of all, thanks for the replies slobo. lots of great stuff here to consider and mull over. second, and i should have mentionned this in the OP, i wasn't planning on using this in a current campaign immediately but rather give it a test in a one-shot/adventure module rather then throw it onto pre-existing characters. i'm going to reply post by post rather then simply throw ideas at the wall i'll also preface this by stating my stance on randomization: unless i see the need for it, i don't see it's use. if the only penalty for failure is "i try again" then i simply let the players dictate how it happens. basically, unless you are in a stressful situation, one where failure has heavy consequences or there is a big unknown element involved i won't have someone roll.  but in a game like D&D, i want the characters to feel heroic, where while they have a chance of failure, they should be expected to win, for the most part. should all the numbers fall into place, 3d6 seems to give a more appropriate expectation of what i would want.  i think this would require a bit of testing "in the wild" to get a good idea of if it actually works or not. either way, nice brainstorming! 1st post i can understand how in d20 each little +1 is a bit more important, but my group isn't high optimization. they do the basics, like having a high main stat, keeping your weapon+'s at a good level, etc... but nothing really OP'd. like all houserules, i'm making this one for my my group in general rather then a "this is how i think D&D should work" . 2nd post 1) this is actually what i'm going for. i want a more predictable result from player actions. as for skill challenges, i rarely announce we're starting one, but rather spread the rolls/important junctions along the adventure and adjudicate the adventure based on how well they do. 2) eeeehh... i don't think random = exciting in combat, especially in a game where you're supposed to be the Big Darned Heroes. for me it's more a mix of the creatures in the encounter, their tactics and the overall design of the battlefield then any given roll of the dice. i don't mind players hitting more frequently. if anything it might help speed things up a bit. 3-4) like i said before, the players aren't big optimizers going for every possible +1. not worrying much there, but i will keep an eye open in the test run. as for crits, i dunno. i always figured it was simply to let the players occasionally output big damage. 5) never played 7th sea, so i'll try to look it up to get a better idea where you're coming from here. is there a free/demo version online? this is the one i found with a quick google search. i'm not working tomorrow so i'll have some time to digest it then. 6) like i said, the main issue i have is the more or less 50-55% chance of success on any given check is a wee bit too low and overall way too random. i'm looking for a bit more consistency and a higher chance of success when dice are rolled. post 3 1) this is expected and a wanted change. no real comment here. 2) this i can understand the issue. of the rolls fall into the 10-14 bracket, the success rate is: 10+ 62.7% / 55% 11+ 50.2% / 50% 12+ 37.7% / 45% 13+ 26.1% / 40% 14+ 16.4% / 35% i think this is the key issue, and until appropriate tweaks are done will require a bit more tactical play on everyone's part, or at least a minor retooling of a system, be it the monster attk/def  or player's attk/def.  3) i never see anyone trade-off accuracy for damage as is, at least no more then a single point of accuracy. i doubt i would see a change in the current playstyle but i could be wrong 4) i don't mind more tactical play. 5) isn't it already like that? test the waters & spam applicable attacks? like i said some might sound a bit rushed, but we had a power blink at work so i'm trying to be brief. either way thanks a whole lot for all the feedback, i really appreciate it. this will definitely give me something to think about on my day off.
#6 lofgren May 31, 2011 13:33:14	I would think a big concern would be any penalty to hit status effects could become crazy powerful or only slightly useful depending entirely on the target's attack bonus.  If you hit on an 9 with 1d20, that's goo enough, and if you take -2 you can still hit on a 11 which sucks but isn't the end of the world.  If you go from 9 to 11 on 3d6 your chances of success just went from 75% to 50%.  That's pretty extreme.  If your target gets cover or somesuch, your chances of success just went down to 25%. And forget ever attacking anything invisible or with concealment.  Even if you normally hit on an 8 you'd have to be desperate to attack anything with -5 to hit. I can see a lot of benefits to this approach in a system that was built to handle it from the ground up.  It would make actions besides attacking way more valuable in combat, such as aiding another, buffing defenses or attacks, negating penalties, etc., which is very cool and helps eliminate one of the annoyances with 4e which is that optimizing around damage output is often more valuable than other aspects.  But I don't think you could just tack it on to a system that is designed around 1d20 without major unintended side effects.
#7 slobo777a May 31, 2011 16:44:13	Glad you found time to digest all my comments. I cannot tell from the site whether 7th Sea is still using the dice system I liked. I seem to recall that a few by that publisher went D20. Not sure if it's what you are looking for anyhow - would be far too complicated to retrofit it to D&D, I was just saying there are other options, depending on what you are looking for. I wouldn't say that random=exciting, just that in combat you make many rolls over an encounter. That in itself is already an averaging mechanism which will over time pay back players who build or find advantages. There is no need for players to get hung up on individual hits/misses if they have solid tactics. As for the scissor/paper/stone issue, then yes it is already a sensible style of play. It will just become more polarised, and more worth doing when the prize is finding the defense where you need 8 to hit (as opposed to 11), or facing the monster that attacks your best defense (making the monster need a 14 to hit you). You may find you like the style of play that 3d6 encourages, and I may not be accurate in all my predictions, they are based on my understanding of the math you are playing with, I obviously cannot  predict what your players will actually do, just spot patterns in the math.
#8 bone_naga May 31, 2011 16:49:57	I've thought about using something similar, but to be honest I kind of like the swinginess of a d20. Besides, everyone lives for the moments when a natural 20 or 1 is rolled, and those moments come along a lot less often with a bell curve. If I did decide to use it, however, I would go with a 2d10 instead of a 3d6. I think the 2d10 gives you a little more variation and 2-20 feels closer to a d20 roll than 3d6.
#9 slobo777a Jun 01, 2011 5:28:00	Just a thought: You don't appear to have a problem with a specific distribution, the only stated problem is that players want to hit more often. A bell curve doesn't actually solve this at all - it just so happens that 3d6 has a better chance of rolling 10+ than 1d20. A change to 3d6 changes loads more, some of which you might not want. A simpler, and possibly better solution: Grant all attacks an inherent bonus +2 to hit. With that change, all players will hit roughly the same amount better as if they were in the "8-10 needed" category that you are targetting. Also, all 4E's D20 adjustments will work pretty much the same as before. This also focuses you on your core problem. Would a +2 to hit really solve your stated problems with the system (your math example focuses heavily on this as the main advantage you want to achieve) - if not, why not - what are the still unstated needs?
#10 Rustmonster Jun 05, 2011 15:56:18	Personally, I dislike bell-curves because I like being able to know, at a glance, what an increase or decrease in numbers means. I appreciate that in a d20 system, a plus or minus 1 is plus or minus 5%, always, all the time, forever. In a bell-curve system, I have to take into account what the number already is before I add/subtract, and I have to deal with the fact that sometimes a +2, for example, means something different than it did before. I find this, personally, unappealing. Why is a trained swordsman effected more from being blinded than one who is an amatuer? If your main concern is players not hitting enough, I'd go with what Slobo said above and just give players an attack bonus. 3d6 might address your problem under perfect circumstance, but as soon as things start going away from the norm, you have to deal with all the little hickups of the system.
#11 oxybe Jun 08, 2011 12:06:32	i've given it some thought since the last post and i'll keep playing it as is for now. after a bit of deliberation i'm leaning towards giving the expertise feats for free, so that might help alleviate the issue. thanks for everyone's contribution. also total side note: that 7th sea system is interesting. not something i'd run but the X keep Y dice roll is different. i kinda like it.