A Radical new D&D Next mechanic idea...

---Here's a radical idea for D&D Next.  It could at the very least be included in the final rules as an Optional Rules Module for DMs that want to give it a try.  ---Our beloved game in all of it's iterations had in many ways relied very heavily on Twenty Shades of Randomness, namely the d20.   The twenty-sided die is Iconic to the game.   I think most of us are cool with that or at the very least used to it...  but it also causes problems; for one, if your dice are cold, you do poorly.  Here's my proposition:    

---Continue to use D20's but read them as if they were numbered 1-10 twice to shrink the scale of randomness down to a more manageable amount.  Call it a "Half-Twenty" (Half-20) or "Short Twenty"  Either use a d20 that is numbered 0-9 twice and have the die inked different colors (green is low, red is high, like my 1984 red translucent Game Science d20 is) or just drop the first digit for 11-20 on a normal d20 die.  A natural 20 is still the only roll that's a Critical Hit.  For normal attacks, spells cast or skills used, you'd have a random number from 1-10 to deal with.  If you have Advantage, extra d20 but use the full scale on both dice for this roll, meaning you can be super successful.  If both Advantage dice come up as Natural 20's it should be some sort of critical success (like TRIPLE damage or a DM's choice event of a hinderance to the PC's foe).  Alternately, roll both dice again, if either one is a Natural 20 again, grant some boon "Your God smiles upon you and lends his strength as you swing your mace into the troll's midsection. There is a sudden flash of light & the troll doubles over in pain. He's stunned for one round.  Well done! who's up next?"  --And if you happen to roll Natural 20's on both dice when you reroll them here, let's call that an instant kill... it would be ultra, ULTRA rare.

Likewise, for Disadvantage, you roll the extra d20 but use the "Half-20" 1-10 scale on both dice and take the lower result.  If you end up getting Natural 1's on both dice, your PC would be affected by some sort of a Fumble effect, dropping a sword, slip, trip, fall, etc.  Would be rare but also fun & realistic.  Even a LV 20 fighter could lose grip on their axe in rare cases (and thus would want to avoid incurring a Disadvantage).   Saving Throws should probably use the full spectrum of 1-20 dictating a higher LV or Randomness. 

--DM's can also use this "Half-20" mechanic to exercise more flexibility & control over challenges.  "OK, you have a minor Advantage as you've climed these walls before. Take two Half-20's, tell me the best result"...  Meaning two dice, you get the best result of the two on the 1-10 scale.  Not FULL Advantage, just a MINOR or PARTIAL One.  Or you could just say the Rogue succeeds, but rolling is fun & I'm of the mind that a player gets more satisfaction out of the game when they have to work to overcome obstacles and less satisfaction when they are given too much.  Rolling 10's on both dice in this case could mean a near flawless effort, "OK, nice rolls, you easily make it to the top of the wall in record time & manage to seclude yourself in a shadowy area to the side where the guards can't see you."

The game math has to be reworked quite a bit but almost all of it can be done so my neatly dividing it in half or dropping 10 fromt he total.   Armor Classes are adjusted to start at Five rather than Ten perhaps.  The Rule we've seen from time to time where a Rogue can "Take 10" can be redesigned to "Take 5" (no pun intended, or maybe 6 if you have a Feat).  Likewise, bonuses given out have to be cut in half as evenly as possible to be adjusted to this scale.  Once a character reaches a certain level, maybe LV 10, they can roll with a minor advantge more often, say Twice per Encounter or all the time for a Rogue and their 4 mains skills...

And I can see other usefulness for this as well; a Figher or Rogue can take a full d20 roll once per Encounter, as if they had Advantage but with only 1d20... Maybe let fighters also add a Half-20 roll to damage once per Encounter.  Maybe let Rogues & Rangers get One Full-20 roll per Encounter and one Minor advantage (as stated above, 2d20 using the 1-10 scale, keeping the better result).  Clerics might be able to do the same either on a Healing spell or on a Turn Undead attempt (or maybe an attack).  Wizards could be granted an extra Half-20 to spell damage oncer per Encounter or add a Half-20 to number of creatures affected by Slow, Grease, Web, etc once per Encounter.  The utility of this idea is only limited by our imaginations.  Seeing the slower progression of Base Attack Bonus & such, the smaller scale of the Half-20 seems to make sense on many levels.  Why not just use d10's you ask??  Because this is D&D, we roll a TWENTY-SIDED DIE!!!  It might seem overly complicated to some reading this but I've tried it and it really is easy, my players loved it, were less plagued by cold dice, less DM beat-downs when my dice were hot, etc.  After using a few times, you get used to it quickly.  

Anyway, that's it in a nutshell.  Not quite so "Simple yet Elegant" as the BRILLIANT idea of the Advantage/Disadvantage but interesting nonetheless.  Thoughs, Comments, Criticisms, High Praise for the idea??  Please reply below.  You won't hurt my feelings if you don't like; after nearly 30 years playing this game, I'll be OK, heheh!  Hope some of you like...
I've pondered this sorta idea myself. To the point where one week I replaced the d20 with 3d6 to average things towards the middle a bit. It only lasted one week for us though. I don't think the math really has to be reworked so much if you just want to narrow the range while allowing the player to freely make up the range. Which it somewhat depends on your intent. Anyway her's an idea:

For your example, any time a d20 is asked for, the result is really 10+d10. I'd specialize crits in a different way. Which is, if you roll 0 on the 1d10, then roll another d10. If that d10 is 6 or higher then let's call it a crit. If you want to narrow the crit range to 1%, then they have to roll two 10s in a row. So interesting result. Of course this gives a super strong level 1 fighter +18 to hit.1d10+18 is going to hit most ACs in the game without the roll. So maybe ACs would need to go up or maybe the free +10 is really just too much and needs to be +5. 1d10+13 instead?


As far as four 20s being insta kill or even three 20s being very painful or putting in crit misses. I'll never insert that into my camp (and 4 20s has never happened in our games anyway). The odds are 1 in 160,000 for 4 20s to occur and 1 in 8000 for 3 20s to occur (which we have had our table from time to time - I always make people reroll until they stop rolling 20s). When it does occur, it is more likely to hinder the PCs than the monsters (since the PCs suffer more attacks overall). I've played in games with crit hit tables and crit miss tables though. I just don't like them as a player very much and so I haven't tortured my players with it. To each their own. It is a rare circumstance to be sure. Just don't want to punish my players when my dice are hot. I guess I wouldn't care so much if 'they' could do it to the monsters.

Currently running a playtest, weekly, online D&D Next Session using a virtual table system called roll20.
Interesting look at the math.  The way I ran it, a fighter would only start with +2 or +3 to hit, so could not hit an armor class above a 13 on a "Half-20" without Advantage... but as I stated in the OP, I dropped the starting AC down so Chain Mail (or whatever armor worn) is added to 5... so a value of 9.  On further reflection, It would probably work better if the armor values were cut in half.  In this case the Chain Mail would be an AC of 8, Scale Mail an AC of 7 (plus up to 2 Dex points).  It works & is comparable to the latest armor table from the 10/29 playtest. 
ARMOR   AC VALUE
Leather   4+Dex
Displacer  5+Dex
Mithril      6+Dex

Studded    6+Dex(up to 2)
Scale         7+Dex(up to 2)
Drag Scl     8+Dex(up to 2)

Ring         7
Chain        8
Banded     9
Splint        9
Plate        10
Shield still adds +1 to AC

I totally hear you on the critical chances.  Rolling two Natural 20's would be exceptionally rare.  After that, rolling any more 20's would be exceedingly rare.  And my thoughts on this were not to grant the monsters this, only the PCs.  You might note that I didn't mention any examples of the Bad Guys getting it, right?  --The monsters & bad guys have enough abilities & odds stacked in their favor; giving them more is not a good idea.  On occasion, giving that to a recurring enemy or "boss" the critical chance would be interesting.  --Possible fumbles would apply to all... no chart, just a DM judgement call:  "Your fighter's axe gets lodged in the bugbear's shield.  Drop the weapon & draw another or your next attack will be at a Disadvantage due to you having to dislodge it."  or "You are off balance; make a Balance check DC8 or fall prone" (this is assuming my new scale).  Was looking to have Optional Rules for this type of stuff, anyhow...


Rather than changing everything though I think it is only going to work if the conversion is easy and automatic. Like: All attacks are now 1d10+5 and all ACs remain the same. This changes numbers in one place. I have a big feeling that in some of the upcoming playtests, the math is going to change a lot too. So maybe that will give us some ideas on how to make things a little less swingy. Which I think they'll attempt by narrowing down the accuracy or damage ranges somehow? Speculation, for sure, but I'll be interested in seeing what comes out.

Also, why does your fighter only have +2 or +3 to hit (other than you just planned it this way)? They have +3 to hit at level 1 atm with a 10 in that primary stat.
Currently running a playtest, weekly, online D&D Next Session using a virtual table system called roll20.
Yep, planned it that way (Fighter Attack Bonus: using the previous iteration of the classes, not this most recent one).  I figured that number would represent a middle of the road bonus for average (point buy) generic LV1 characters .  Survival of characters beyond LV1 would be better, of course.  One of my hopes with the "Half-20" idea was to speed things up while also offering some added utility with the dice.... AND some chances for some phenomenal successes of the PCs.  --One of the things I enjoyed from 4E was the action point; most of us know it but for those who don't, you could cash it in for a reroll, an additional action or basic attack, etc.  

Anyhow, the narrower dice range means less wild variation... eventually, at higher levels, such as LV 15 or higher PCs, the die roll becomes less and less of a factor (best to keep the rule as it stands, a Natural 1 always misses).  And FluxPoint, thanks for the feedback.  Good insight!  I'm interested in seeing if your prediction comes true and we do see some "math changes" such as what you and I have suggested (or something else entirely).
The problem with the d20 is high swing at low levels, low swing at high levels.

i.e. If you have +0 mod and I have a +2 mod, I'm only at a slight advantage when we engage in a contest of power.
If I have a +80 bonus and you have a +60 bonus, I'm going to beat you every single time.

While your fix helps the low level swing, it exacerbates the high level problems.
The 3d6 method is the best alterative I've seen.

BUT, I think that Advantage/Disadvantage clears alot of this up, actually.

3d20 take the middle would give us a distribution.
Advantage and Disadvantage give us distributions that make it alittle better too.
In both environments the 20 provides the same swing. The problem with high levels in the past has been that the bonuses are simply not close enough together. With 4e, I'm somehwat hoping the bounded accuracy makes this difference much less than 20. If so, the d20 still makes sense as something you can 'win' on. If it bounds it enough, the d10 may make the same thing.

With 0 v 2 , you're at a 10% advantage with d20s. With the d10 mechanic, you're at a 20% advantage. It certainly exacerbates the bonuses to die rolls. The biggest issue really is that in actual role/game play, how swingy do you want to be?

d20 just seems to leave such a big range up to luck. 
Currently running a playtest, weekly, online D&D Next Session using a virtual table system called roll20.
I think 3d6 as an optional rule would work just fine. Also I beive they are trying to avoid ever letting things get to +60 vs +80 in this ediition.
I think 3d6 as an optional rule would work just fine. Also I beive they are trying to avoid ever letting things get to +60 vs +80 in this ediition.



This is EXACTLY what Gygax wanted to avoid: Bell Curve distribution.  He wanted that wildness whereas 3d6 generates mildness.  The more dice you add the harder to the center it pulls.  Even 2d6 pulls pretty strongly to 7.

3d6 is better for grittier, less pulpy action.  If you want very pulpy, swashbuckling extremes, trade the d20 for a d12.  Crits and fumbles will abound.

--

2d10 is not like 1d20.  As stated above, more dice increases pull to the average. Crits would be very (1% instead or 5%) rare unless bonuses counted as part of the Natural Roll.

--

There's no such thing as hot or cold dice.  Luck is imaginary.  Poorly made or loaded dice will skew results.  Get some Game Science dice.


This is EXACTLY what Gygax wanted to avoid: Bell Curve distribution. 



I am somewhat of a research nerd, so out of sheer curiosity I was wondering if you have any source or link I can see where gygax says he wanted to avoid bell curves? Not that I am saying your wrong, just that, I am very intrested in any insight of his intent. knowing my luck it's in the DMG. 

To work out how 3d6+modifier works in play, don't pay attention to the chance to hit a particular DC.  Pay attention to the chance to roll at or over a given DC.

Graph it beside the chance that 1d20+modifier has to roll at or over a given DC.

When you look at the chance to hit a DC, these two curves look way different.  When you look at the chance to roll at or over a DC, they start looking really similar after a simple rescaling.

Take 3d6+modifier vs 10+modifier, and compare it to 1d20+modifier*1.7 vs 10+modifier*1.7.  The difference between these two is nearly impossible to detect in actual play.  Or 2d10+modifier vs 10+modifier against 1d20+modifier*1.4 vs 10+modifier*1.4.  Or 1d10+5+modifier vs 10+modifier compared to 1d20+modifier*2 vs 10+modifier*2.

Almost the only thing that changing the dice you add up does is change the average and the standard deviation, plus maybe change how the crit miss/hit mechanics work, and they scale how "important" each +1 to DC and +1 to roll is.

You can continue playing 1d20 D&D and just double every modifier to d20 rolls, and double every static bonus to DCs away from 10, and you'll get a game that is less swingy (and more predictable) than standard D&D.  The bell curve of 3d6 is a red herring.

But you don't have to trust me -- just graph the chance that 3d6 rolls at-or-over a given value, then scale it horizontally by sqrt(3) (with 10.5 fixed), and compare it to the bog-boring d20 rolling at-or-under a given value.
Why is this thread in the Dungeon Master Playtest Forum?
because it's from a long time DM looking for thoughts from other DMs and there was no place in the D&D Next Forums for "Proposed Rules"... but if one of the Moderators feel it belongs under General discussion or somewhere else, so be it.  And it's an Optional idea to cut down the wild variation of a straight d20 while also creating a few other options for resolving actions quickly in the game to keep the action moving.

 -UrielApeiron wrote: "There's no such thing as hot or cold dice.  Luck is imaginary.  Poorly made or loaded dice will skew results.  Get some Game Science dice."  I've played this game for 30 years and have seen some really bad rolling streaks even with Game Science dice (see OP in which I specifically name them).  Luck may be imaginary but sometimes really bad rolling streaks are a reality, sadly.
 I've played this game for 30 years and have seen some really bad rolling streaks even with Game Science dice (see OP in which I specifically name them).  Luck may be imaginary but sometimes really bad rolling streaks are a reality, sadly.



I'd love to see a dice system that gives a bonus/penalty to the success/fail roll and a balancing bonus/penalty to the degree of success and failure based on streaks.  The more successes you get in a row, the more confident you are and the more amazing your successes become... but also the more likely that hubris will strike you down with huge failures.  The more failures you get in a row, the more humble you become and so your degree of failure (and first, tentative successes) decreases, but your caution makes it all the more probable that you will be successful.

Of course, this could produce its own sort of disparity.  Rather than one player having streaks of bad luck and another player having streaks of good luck, you could get one player having streaks of mildness while another player has streaks of cinematics.  But I think even this could be adjusted by applying the bonuses and penalties to the succes/fail roll in a nonlinear fashion.  And maybe you could remove a bonus or penalty more quickly than you added it.  For instance, maybe the first instance of a humble success or hubristic failure takes you right back to normal immediately.  There are a lot of variables here that could be manipulated to produce the desired flow of play.