Advantatge/Disadvantage mechanic

[bestialwarlust]

So what are thoughts on the disadvantage/advantage rolls? I've been rolling that around in my head too see if it would be worthwhile incorporating as a house rule in my Delving Deeper game. On one hand I like the fact that it can be used to minimize fiddly modifiers when making rolls. But I'm not sure on the overall implications in play. If will add to the fun and keep players immersed in the game world.

[kesher]

Actually, I've been using that mechanic since the very first play test packet. My reaction was along the lines of "I cannot believe that never occurred to me before!"

I love it as a quick way to put some sort of edge on a situation without pausing to figure out a particular modifier.

Actually, I was hoping at first, for the 5e rules, that they'd use A/D in place of bonuses/penalties. However, I think using A/D in conjunction with the single proficiency bonus should work just fine.

[Finarvyn]

Jul 6, 2014 10:27:11 GMT -6 bestialwarlust said:

So what are thoughts on the disadvantage/advantage rolls? I've been rolling that around in my head too see if it would be worthwhile incorporating as a house rule in my Delving Deeper game. On one hand I like the fact that it can be used to minimize fiddly modifiers when making rolls. But I'm not sure on the overall implications in play. If will add to the fun and keep players immersed in the game world.

I think that the advantage/disadvantage system is simple and elegant and I like it. Brings back a certain "roll dice and wing it" feel to the rules, in my opinion. For example, if you are surprised I could have you attack at disadvantage. Clinging to the side of a cliff while trying to attack could be at disadvantage. That kind of thing. Makes it easy to step away from the rulebook and just make decisions about what's happening in the game session.

[Zenopus]

Can you mimic the rolls just by giving an advantaged character a +3 and a disadvantaged character a -3?

[Finarvyn]

Jul 6, 2014 15:02:07 GMT -6 Zenopus said:

Can you mimic the rolls just by giving an advantaged character a +3 and a disadvantaged character a -3?

You could do that, but I think it spoils some of the fun of rolling 2d20. Of course, with advantage it's nearly impossible to have a fumble and with disadvantage nearly impossible to roll a natural 20. Using a +/- 3 system would change those odds a lot.

[talysman]

I have used a bonus version of the advantage/disadvantage system before, and I've written draft versions of games that used a reroll/best of two rolls mechanic, but I never thought of merging thne two. I think it's a good idea.

Jul 6, 2014 11:35:09 GMT -6 kesher said:

Actually, I've been using that mechanic since the very first play test packet. My reaction was along the lines of "I cannot believe that never occurred to me before!"

I love it as a quick way to put some sort of edge on a situation without pausing to figure out a particular modifier.

Actually, I was hoping at first, for the 5e rules, that they'd use A/D in place of bonuses/penalties. However, I think using A/D in conjunction with the single proficiency bonus should work just fine.

Same here. I only glanced briefly at the playtest rules, but the idea of mixing 1,2, and 3-point bonuses and penalties with the advantage/disadvantage system seemed needlessly unwieldy. Keep +/- 1 modifiers for minor advantages/disadvantages, maybe, but everything else should just be the 2d20 drop 1 roll.

I was also disappointed that backgrounds turned out to be bundles of mechanical benefits instead of a brief description of the background for use with advantantage/disadvantage. "Oh, you are a former knight in the king's cavalry? You have advantage when taking care of horses or performing tricky riding maneuvers." Also disappointed that skills aren't treated as part of the same system, too. They had a real opportunity to simplify it a lot with this mechanic, and they dropped the ball.

[waysoftheearth]

Jul 6, 2014 10:27:11 GMT -6 bestialwarlust said:

So what are thoughts on the disadvantage/advantage rolls? I've been rolling that around in my head too see if it would be worthwhile incorporating as a house rule in my Delving Deeper game. On one hand I like the fact that it can be used to minimize fiddly modifiers when making rolls. But I'm not sure on the overall implications in play. If will add to the fun and keep players immersed in the game world.

Statistically, rolling 2d20 and keeping the higher gives an average result of 13.825, as compared to 10.5 on 1d20. So, on average, the advantage is +3.325 pips on a d20.

That's a pretty significant advantage in OD&D terms; equal to a +3 magic weapon (the best in game), and also equal to +16.625% which is (happily!) just about equal to +1 on a d6 (+16.667%).

FWIW, if you combine the advantage mechanic with a +1 on a both d20s it adds exactly 1 to the average outcome, and your overall advantage becomes +4.325 pips on a d20 (or +21.625%)


I agree that the main boon of the advantage mechanic is that it enables you to ignore a bunch of fiddly modifiers. I don't think there are too many cases where you'd want to combine 2d20 with another plus, because there are not too many situations in OD&D/DD that warrant a +4 (or greater) adjustment on a d20. There are some though.

Advantage could outright replace the halfling's hurled missile adjustment (+3 in DD), and possibly be used with a -1 for all short range missile fire (which is otherwise at +2 to hit on a d20).

Attacking from behind is at +2, or +4 with surprise, so might be another candidate case. There may be others too...

[kesher]

That's pretty fascinating---especially since it was on my mind to ask someone to break down those very possibilities!

What happens if you add another d20 to the mix (i.e., rolling 3d20 as opposed to 2d20)?

[waysoftheearth]

Jul 6, 2014 19:21:05 GMT -6 kesher said:

What happens if you add another d20 to the mix (i.e., rolling 3d20 as opposed to 2d20)?

The average result of "3d20 pick the highest" is 15.4875, which is a 4.9875 pips on a d20 (or +24.9375%) better than the average result on 1d20.

It's very nearly equivalent to +5 pips on a d20 (or +25%).

[kent]

Jul 6, 2014 19:21:05 GMT -6 kesher said:

That's pretty fascinating---especially since it was on my mind to ask someone to break down those very possibilities!

What happens if you add another d20 to the mix (i.e., rolling 3d20 as opposed to 2d20)?

An interesting problem. The average max on 3d20.

The number of ways you can roll all 3 numbers less than k is (k-1)^3

The number of ways you can have at least one k as the max number is k^3 - (k-1)^3

The average value over all possible k is [1/8000] . SUM from k=1 to 20 of ( k^3 - (k-1)^3 ) x k

... which is [1/8000] . SUM from k=1 to 20 of 3.k^3 - 3.k^2 + k

... which is [1/8000] . [ 3.(20.21/2)^2 - 3.[ (20^3)/3 + 20.20/2 + 20/6 ] + 20.21/2]

... which is 15.4875

(Now there may be an calculation error, its late here, but the principle should be sound.)

[oakesspalding]

In his contribution to the article "Git Er Done" in Fight On! Magazine #6 (2009) M. J. Harnish introduced a similar system, though it featured 3 d20's. So depending on the difficulty you either take the highest roll, the middle roll or the lowest roll. I have no idea whether 5e came up with this independently, cribbed off of it, or whether they both cribbed off an idea that had been floating around for a while.

I have four "problems" with such a system. I put "problems" in quotations, because I am not against it, per se, but rather think these should make one cautious:

1. It fusses up the aesthetic of the rules.
2. It adds fuss to the minds of the players and thus potentially detracts from the immersive experience.
3. It adds another layer of potentially annoying complexity for the players and the GM.
4. The existence of the mechanic has the potential to add dice rolls where there shouldn't be any, thus detracting from the flow of the game.

Here are three "fixes" that take care of these "problems":

1. Don't include it in the rules, especially any set of "basic" rules. I would say that one of the principles (or whatever you want to call it) of the OSR is that modularity is a very good thing.
2. Put the emphasis on it being something the GM (not the players) should deal with and understand. Mainly, a fiddly mechanic interferes with the immersive experience only because the players are aware of it. The rule can exist and be used, but still reside in the GM's head.
3. Use the mechanic sparingly and with good sense. Obviously this depends on the GM.

I waffle a bit on this, but I think I prefer using a six-sided die or dice for task resolution, partly for aesthetic reasons and to so to speak, preserve the sacredness of the twenty-sided die for combat and saving throws. And it just seems to fit for me that the range of chances for success at a non-combat task shouldn't be very granular. I also tend to want to de-emphasize the mechanical importance of ability scores except perhaps at the extreme ranges.

[Zenopus]

Elsewhere (DF, I believe) it was pointed out that rolling two d20s for a bunch of monsters can be onerous for the DM. For example, normally if 5 orcs are attacking the DM could roll 5 d20s at once. But with this system you can't roll ten d20s unless each set is color coded.

[oakesspalding]

Jul 6, 2014 21:29:07 GMT -6 kent said:

An interesting problem. The average max on 3d20.

...15.4875

Well, running it 5x1000 times using the Excel RandBetween function yields 15.66, 15.51, 15.31, 15.45 and 15.66. So that sounds right.

[waysoftheearth]

Jul 6, 2014 21:59:48 GMT -6 Zenopus said:

Elsewhere (DF, I believe) it was pointed out that rolling two d20s for a bunch of monsters can be onerous for the DM. For example, normally if 5 orcs are attacking the DM could roll 5 d20s at once. But with this system you can't roll ten d20s unless each set is color coded.

A crude solution for this is to always roll 1d20 but with advantage 1=20, 2=19, 3=18, and 4=17. With disadvantage 20=1, 19=2, 18=3, and 17=4. It's not perfect, but it's probably near enough.

Another crude solution is to get some d24s and similarly re-assign the faces over 20; 21=17 or 4, 22=18 or 3, 23=19 or 2, and 24=20 or 1.

[kesher]

You guys are all awesome!

I assume the statistics work the same in reverse (i.e., Disadvantage)?

Elsewhere (DF, I believe) it was pointed out that rolling two d20s for a bunch of monsters can be onerous for the DM. For example, normally if 5 orcs are attacking the DM could roll 5 d20s at once. But with this system you can't roll ten d20s unless each set is color coded.

Wouldn't a simple solution, in a case like this, be to give the characters disadvantage, and leave the orcs' rolls as normal?

[cooper]

I actually don't know many groups who mass roll dice for enemies. I can't remember the last time I saw a DM roll 5 goblins attacking at once.

[Porphyre]

Personally, I really like the mechanism, but I am afraid that its implementation could become uselessly fiddly.

In my opinion, this could be ideal for "fixed" advantages like missile fire for hobbits, secret door detection for elves, Saving throws for dwarves , weapon mastery for a fighting man or background skills.

But it could become like the worst parts of the 3E rules if it is applied for circumstantial bonuses: "If I attack under circumstance X, and you have feat Y, then you have advantage" is assured headache for a DM.

[Finarvyn]

Jul 7, 2014 8:09:42 GMT -6 kesher said:

Wouldn't a simple solution, in a case like this, be to give the characters disadvantage, and leave the orcs' rolls as normal?

That's what I do. I hardly ever have monsters roll advantage or disadvantage, but put the focus on the players instead.

Jul 7, 2014 14:07:40 GMT -6 Porphyre said:

Personally, I really like the mechanism, but I am afraid that its implementation could become uselessly fiddly.

<snip>

But it could become like the worst parts of the 3E rules if it is applied for circumstantial bonuses: "If I attack under circumstance X, and you have feat Y, then you have advantage" is assured headache for a DM.

I suppose this could be a problem, but I do it so that I get to decide who gets advantage and when.

* Surprised? Attack at disadvantage.
* Roll a natural 1 on an attack? Next attack at disadvantage.
* Thief's first attack from behind? Make it an advantage.

That kind of thing.

I try not to get to hung up over the mechanic of the thing, but instead look for cases where one side clearly has an edge over the other and award it as needed.

[Finarvyn]

Jul 7, 2014 10:14:27 GMT -6 cooper said:

I actually don't know many groups who mass roll dice for enemies. I can't remember the last time I saw a DM roll 5 goblins attacking at once.

I don't, either, but maybe some folks do.

I point to a mini and say "now this guy" or, if I'm not using minis I might say "the guy attacking Alan..." and roll for him.

[talysman]

Jul 7, 2014 8:09:42 GMT -6 kesher said:

Elsewhere (DF, I believe) it was pointed out that rolling two d20s for a bunch of monsters can be onerous for the DM. For example, normally if 5 orcs are attacking the DM could roll 5 d20s at once. But with this system you can't roll ten d20s unless each set is color coded.

Wouldn't a simple solution, in a case like this, be to give the characters disadvantage, and leave the orcs' rolls as normal?

Either that, or just roll one extra d20 for the orcs as a whole. 5 orcs? Roll 6d20. If they have the advantage, drop the lowest roll. If they have disadvantage, drop the highest.

[waysoftheearth]

Jul 7, 2014 10:14:27 GMT -6 cooper said:

I actually don't know many groups who mass roll dice for enemies. I can't remember the last time I saw a DM roll 5 goblins attacking at once.

I used to roll individually for each monster versus each PC.
In the last few years I have taken to rolling monsters in bulk.

I have found that rolling in bulk:
a) is a huge time saver,
b) really promotes the notion of "abstract" combat.

Regardless of whether or not you roll monster attacks in bulk, the same issue exists for a single PC/monster with multiple attacks per round. Do you want to roll your fighter's four attacks all at once? Or individually?

[krusader74]

Jul 6, 2014 18:24:55 GMT -6 waysoftheearth said:

Statistically, rolling 2d20 and keeping the higher gives an average result of 13.825, as compared to 10.5 on 1d20.

Just wanted to put in a plug for my new dice calculator, Palamedes, discussed here. In the latest version* I uploaded today, you can easily do this calculation with the succinct code...

stats(max(2#d20))

to get the result...

stats (max (2 # d20)) → {"min":1,"max":20,"mean":13.825,"median":15,"mode":[20],"sd":4.711090638058235,"count":20}

which confirms waysoftheearth's calculation.

The hashmark # between the 2 and the d20 means take a sample of two d20 rolls, whereas 2d20 would mean roll two d20s and sum, which isn't what's desired here.

If you wanted to see a table of probabilities for each outcome, use the code...

table (prob (max (2 # d20)))
to get the result...

table (prob (max (2 # d20))) →
                key     value
                  1     0.0025000000000000044
                  2     0.007499999999999991
                  3     0.012500000000000011
                  4     0.017499999999999974
                  5     0.02250000000000002
                  6     0.027500000000000024
                  7     0.03249999999999996
                  8     0.037499999999999936
                  9     0.04250000000000004
                 10     0.04750000000000004
                 11     0.052499999999999936
                 12     0.05750000000000005
                 13     0.06250000000000006
                 14     0.06750000000000006
                 15     0.07250000000000012
                 16     0.07750000000000001
                 17     0.08250000000000013
                 18     0.08750000000000002
                 19     0.09250000000000014
                 20     0.09749999999999948


Jul 6, 2014 21:29:07 GMT -6 kent said:

The average max on 3d20... which is 15.4875

Likewise, Palamedes can confirm kent's calculation (up to JavaScript rounding errors) with the code...

table(stats(max(3#d20)))
table(prob(max(3#d20)))


to produce the output...

table (stats (max (3 # d20))) →
                key     value
              count     20
                max     20
               mean     15.487499999999997
             median     16
                min     1
               mode     20
                 sd     3.8676018086147406


table (prob (max (3 # d20))) →
                key     value
                  1     0.00012500000000000033
                  2     0.000874999999999999
                  3     0.002375000000000002
                  4     0.004624999999999994
                  5     0.007625000000000005
                  6     0.01137500000000001
                  7     0.01587499999999998
                  8     0.02112499999999997
                  9     0.02712500000000001
                 10     0.03387500000000003
                 11     0.04137499999999994
                 12     0.04962500000000006
                 13     0.058625000000000066
                 14     0.06837500000000002
                 15     0.07887500000000014
                 16     0.09012500000000001
                 17     0.10212500000000024
                 18     0.11487500000000006
                 19     0.12837500000000024
                 20     0.14262499999999922

You can easily do even more complicated order statistics such as this: Roll four d20s and keep only the second highest. Use the code...


table(stats(orderstat d20, 4,3))
table(prob(orderstat d20, 4,3))

to get the results...

table (stats (orderstat d20, 4, 3)) →
                key     value
              count     20
                max     20
               mean     12.4999875
             median     13
                min     1
               mode     14
                 sd     4.010367813535776

table (prob (orderstat d20, 4, 3)) →
                key     value
                  1     0.00048125000000000126
                  2     0.0032187499999999964
                  3     0.008281250000000007
                  4     0.015218749999999977
                  5     0.02358125000000002
                  6     0.03291875000000004
                  7     0.042781249999999924
                  8     0.05271874999999995
                  9     0.06228125000000001
                 10     0.07101875000000007
                 11     0.07848124999999989
                 12     0.08421875000000012
                 13     0.08778125000000009
                 14     0.08871875000000007
                 15     0.08658125000000017
                 16     0.08091874999999982
                 17     0.07128125000000018
                 18     0.05721874999999976
                 19     0.038281250000000044
                 20     0.014018749999999858

Or this: Roll six d24+1 and keep the fourth highest...


table(stats(orderstat (d24+1), 6,4))
table(prob(orderstat (d24+1), 6,4))
Note the parentheses around the dice expression in this case. This gives you...


table (stats (orderstat (d24 + 1), 6, 4)) →
                key     value
              count     24
                max     25
               mean     15.214297740711272
             median     15
                min     2
               mode     16
                 sd     4.209065650668775

table (prob (orderstat (d24 + 1), 6, 4)) →
                key     value
                  2     0.00004224947287058456
                  3     0.0005880285192418994
                  4     0.002337556480543766
                  5     0.005734154553406843
                  6     0.010953644175588338
                  7     0.01794202304834857
                  8     0.02645314115882734
                  9     0.03608637680242087
                 10     0.04632431260515804
                 11     0.05657041154607667
                 12     0.06618669297960067
                 13     0.07453140865791641
                 14     0.0809967187533489
                 15     0.08504636788073872
                 16     0.08625336111981852
                 17     0.0843376400375887
                 18     0.07920375871069618
                 19     0.07097855974780853
                 20     0.06004885031199103
                 21     0.04709907814308467
                 22     0.03314900758008088
                 23     0.019591395583499538
                 24     0.008229667757764347
                 25     0.0013155943735799758
Anyway, I created this tool to experiment with just these kinds of computations with minimal effort.  Hopefully, someone here might find it useful. 

---

* Worked these calculations in Palamedes version 1.0-a25 (20140708). Unfortunately these examples would fail in the prior version 1.0-a24, due to a regression that treated numbers as strings in certain functions. It would have worked OK in versions 1.0-a23 and prior. Note this software is still alpha quality and contains bugs! If you have issues with it, please PM me with a description of the problem and the version. Thanks!

[kesher]

Jul 7, 2014 17:40:54 GMT -6 waysoftheearth said:

Jul 7, 2014 10:14:27 GMT -6 cooper said:

I actually don't know many groups who mass roll dice for enemies. I can't remember the last time I saw a DM roll 5 goblins attacking at once.

I used to roll individually for each monster versus each PC.
In the last few years I have taken to rolling monsters in bulk.

I have found that rolling in bulk:
a) is a huge time saver,
b) really promotes the notion of "abstract" combat.

Regardless of whether or not you roll monster attacks in bulk, the same issue exists for a single PC/monster with multiple attacks per round. Do you want to roll your fighter's four attacks all at once? Or individually?

waysoftheearth, when you say "in bulk", do you mean five dice for five orcs, or one die for five orcs (all treated as having the same result)?

[jeffb]

FWIW, I eliminated the normal Prime Not Prime system in C&C and simply used advantage for characters with a prime. Target number was flat 15 +/- challenge level. Primes roll twice. It worked well enough, better than the standard system (which drives me nuts).

[kent]

Jul 8, 2014 5:54:03 GMT -6 krusader74 said:

Just wanted to put in a plug for my new dice calculator, Palamedes, discussed here

Some excellent work there. I lost interest in stats and computing a while ago but I would use the statistics language R for such things, have you heard of it? What you are doing is user friendly for those with no coding experience but still tricky for them. It would be interesting to see if they could use it.

Your EXAMPLES link is very large and perhaps should be more prominent.

[krusader74]

Jul 8, 2014 11:37:52 GMT -6 kent said:

Jul 8, 2014 5:54:03 GMT -6 krusader74 said:

Just wanted to put in a plug for my new dice calculator, Palamedes, discussed here

Some excellent work there. I lost interest in stats and computing a while ago but I would use the statistics language R for such things, have you heard of it?

Back in the summer of 2010 I started work on a similar project in R. It piggybacked off the DiscreteDistribution class defined in the "distr" library. I've put some of the R functions I wrote in a text file---see the attached file below. It includes the function to compute order statistics that we're using in this discussion. I like R, but I wanted to create something that could run in a web browser without installing any special software, so I chose to re-implement in JavaScript. JavaScript is non-typed, which has been a blessing and a curse. It makes many things simpler to implement, but also allows bugs to creep in... like the bug in the previous build which prevented the order statistics function from working because comparisons were being done on strings instead of numbers.

Jul 8, 2014 11:37:52 GMT -6 kent said:

What you are doing is user friendly for those with no coding experience but still tricky for them. It would be interesting to see if they could use it.

I also wanted to make something more user friendly than my set of R functions, so that it could be used by non-statisticians. I don't think I've achieved that goal yet, but it's still a work in progress, and I hope to add more features, improve the documentation, make the web interface friendlier, and exterminate the bugs.

As noted, it's still an alpha preview. My goal is to get a stable version done by the end of this summer. If anyone has more suggestions for improvement, feature requests, or bug reports, please PM me.

Jul 8, 2014 11:37:52 GMT -6 kent said:

Your EXAMPLES link is very large and perhaps should be more prominent.

In the "instructions" frameset, I'll add a link at the bottom, something like, "click here to see more EXAMPLES." The EXAMPLES file itself needs to be better organized, perhaps with a table of contents that lists particular topics of interest. As I add more functionality, the file may need to be split up into a "user guide" divided into chapters.

Attachments:

Dice-in-R.txt (18.23 KB)

[vladtolenkov]

I think this mechanic is particularly useful for things like charging--you gain advantage on you attack (or maybe just the damage roll), but you also grant advantage to your opponents. Or something like that. . .

[waysoftheearth]

Jul 8, 2014 9:36:02 GMT -6 kesher said:

waysoftheearth, when you say "in bulk", do you mean five dice for five orcs, or one die for five orcs (all treated as having the same result)?

I meant ten dice for ten orcs.

Specifically, for normal combat (attacks against normal-types), I've taken inspiration from the combat examples in EPT.

Damage dealt to normal monsters is dealt to the whole group (well, the whole group that is engaged in the melee), eliminating the weakest first, with damage rolling over from one slain enemy to the next.

Attacks made by normal monsters are allocated randomly to PCs engaged in the melee.

A nice side effect of this is that "the melee" can be treated as an abstract area where fighting is happening. You don't need to know precisely who is where, or who is fighting who. All you need to know is who is in the melee, and who is not. Those "in" the melee are potential targets, those outside it are not.

This method seems quite efficient for handling large groups of normals, and I like that it makes melee a crazy, unpredictable mess. It seems a particularly good fit for one minute combat turns when an individual be really be expected to be in "one place" for a whole minute. It is also good to know precisely when a spell caster is "in melee" or not, and whether a potential shooting target is in melee or not.


There are usually far fewer heroic/fantastic opponents, so for these I don't mind setting up duels where we know precisely who is attacking who.

[jeff]

I'm wondering if it could be used for enhancing any mechanic that involves rolling...


For example:

 Elves gain advantage when searching for secret doors.

 Halflings gain advantage when hiding outdoors (thinking of Moldvay D&D).

 Dwarves gain advantage on reaction rolls with other dwarves and gnomes (3d6 for reaction roll, take the 2 highest).



It could also seamlessly integrate with Backgrounds suggested by Zenopus.
 Berserkers could gain advantage on combat rolls when they are enraged
 Gemcutters gain advantage on any rolls used for negotiations.

It could also replace some rules, like surprise. Instead of losing a whole round when you are surprised, you attack at a disadvantage and your opponents gain advantage.

I'd like to try it out in practice to see how well it works.

[jeffb]

 

Berserkers could gain advantage on combat rolls when they are enraged

13th Age does this for the Barbarian's Rage mechanic.

A/D has alot of potential,, IMO. I have thought of using it in the same manner for surprise, racial abilities, etc.