New skill idea

19 posts / 0 new
Last post
All d20 systems have had one major flaw when it comes to skills. This is the variance in a single d20 roll. Even when trained, a character can still have 30-40% chance to fail at every task they attempt. If the bonus from training gets too high though you reach a point where the difference between an expert and a novice is so great the expert can always succeed at a task and a novice will always fail. If D&D 5e is striving for flatter math in general, being trained in a skill should reflect this.

I propose that skill training give advantage to ability checks related to that skill instead of a flat bonus. This can greatly improve a characters chance of success without requiring massive math changes. It also helps to normalize the variance of the d20 roll. Skill focus should still exist for players who want be even more specialized and receive an additional +3 bonus.

Any thoughts?
I agree. 

I think the advantage mechanic adds a lot to the game and see no reason it should be combat only. As you said, skills have specific problems in d20 systems and advantage seems to solve a lot without adding much complexity.
I like your thoughts but not your implementation.  I do appreciate Lawolf you putting out ideas for us to discuss though.

I would allow the player to roll 2d10 or a d20.   So if you are desperate and unskilled you can go for the 5% 20 on a d20.   If you are very skilled you can play it safe with 2d10.  Maybe each skill rank adds another d10 and you keep the top 2.

 
 you reach a point where the difference between an expert and a novice is so great the expert can always succeed at a task and a novice will always fail.

Why is this inappropriate? If you put me in front of a forge, handed me a stack of ingots and told me to make a sword, I'd probably fail. I don't know how to make a sword. There is always that remote outlier chance (represented by the natural 20) that I'll hit the right process by sheer dumb luck, but that's not something that can be relied upon. By contrast, if you put Gabe Newell in front of a forge, handed him a stack of ingots, and told him to make a sword, he'd make a sword. He DOES know how. There is always that remote outlier chance (represented by the natural 1) that he'll screw up something really basic and ruin the blade, but that's not something that will happen often.

However: your suggestion doesn't actually solve the issue. In fact, it makes it worse. J. Random Adventurer has an ability score of 10 and no relevant skills. He fails a DC 11 task half the time--50% accuracy. On the other hand, Random Q. Adventurer (no relation) has an ability score of 18, is trained in a relevant skill, and has Skill Focus. He needs to roll a 4 or better to succeed on a DC 11 skill check, and rolls twice. He succeeds on this task with near 100% accuracy (actually 97.8%).If J. Random were to pick up an appropriate skill training, his success rate jumps to 75% despite having no ability bonus. 

Going by the skill system as presented in the playtest material, J. Random still has a 50% success rate, but Random Q.'s success rate falls to 85%, only a 35% difference. If J. Random were to pick up appropriate training, his success rate becomes 65%--only 20% behind Random Q.




 
The bell curve giveth and it taketh.   For those who are unskilled, the bell curve will make them fail more which is fine.   It will make those with skill succeed more.   
The issue with novice/expert is more to do with party dynamics. In 3e especially, but also 4e, it was common for some players to have a bonus 10+ points above other party members. This caused two problems. One, DCs were hard to create that could present an appropriate challenge for the group. Two, the specialist so fat outclassed the rest of the group that it was hard to include everyone during some phases of task resolution.

Now if DCs were along these lines: 5 easy, 10 average, 15 hard, 20 very hard, 25 legendary, then being trained has a big benefit. As you said someone with focus and a high score can almost auto succeed an average task, while someone with an average score and no training only has a 55% chance. The difference is the person with no training can still attempt the task and contribute to the groups success even at difficulties that the trained person auto succeeds. Because the variance is lower between highest skill and lowest, it takes longer for those who aren't trained to be pushed into the "why bother rolling" category. This tends to happen around level 8 in 4e and as early as 1st level in 3e.
@lawolf
Part of this problem arises from a super limited skill list.  I don't have a problem with some people getting good enough to autosucceed at most appropriate tasks for their level.   They've spent a lot of their level advancement options on getting there though.  On the flip side I want someone to be able to be decently competent fairly easily.

For this reason I think that skills should cost more as they advance.   Rank 2 of a skill should cost more than Rank 1 and Rank 3 even more.   That way specialization has a price.   You can be competent in a whole lot of things or super good in a few.   In the arena of skills, I personally think that broad competency is preferable in most cases.  The DCs just need to reflect the fact that one rank of skill means you are competent at all typical tasks and can often do a hard task.

One way to do this would be to say that all skills cost 1 point but the first time you buy it you get a +5, the second time a +3, and the third time and after a +1.   This would accurately reflect how hard it is to improve into the master ranges.

 
All d20 systems have had one major flaw when it comes to skills. This is the variance in a single d20 roll. Even when trained, a character can still have 30-40% chance to fail at every task they attempt. If the bonus from training gets too high though you reach a point where the difference between an expert and a novice is so great the expert can always succeed at a task and a novice will always fail. If D&D 5e is striving for flatter math in general, being trained in a skill should reflect this. I propose that skill training give advantage to ability checks related to that skill instead of a flat bonus. This can greatly improve a characters chance of success without requiring massive math changes. It also helps to normalize the variance of the d20 roll. Skill focus should still exist for players who want be even more specialized and receive an additional +3 bonus. Any thoughts?



I like it, except that I would have training give +3 AND advantage, with skill focus an additional +3.  Training should offer more than just a better chance of rolling higher than someone who is untrained, while still allowing someone untrained to have a shot at it (providing it isn't a trained only skill).  Skill focus would be for the experts/savants in the field.
A downside to granting advantage with training: then there's no more benefit to be had from a situation or other effect that also grants advantage.

Another possibility: the rules already recommend that players autosucceed on trivial checks. You could say that in addition to the +3 bonus, a character with training can autosucceed on moderate checks.
I think the problem is not the difference between the skilled and expert. The problem is the misunderstanding of the concept of DCs.

A DC of 21 is supposed to be impossible for a normal with +0. Only experts are supposed to even succeed at 21+ DCs unaided. Same with opposed checks. A trained thief sneaks by average joes at a high percentage. In online multiplayer games, the experts beat the new players at a high rate of success.

The problem is the misunderstanding of DCs. DCs are not relative to the attempter. An easy Search check is easy for a normal and automatic for a expert. A hard spot check is difficult for a normal and easy for an expert. An impossible search check is impossible for a normal and difficult for the expert.

DMs just have to understand DCs better.

Orzel, Halfelven son of Zel, Mystic Ranger, Bane to Dragons, Death to Undeath, Killer of Abyssals, King of the Wilds. Constitution Based Class for Next!

I think the problem is not the difference between the skilled and expert. The problem is the misunderstanding of the concept of DCs. A DC of 21 is supposed to be impossible for a normal with +0. Only experts are supposed to even succeed at 21+ DCs unaided. Same with opposed checks. A trained thief sneaks by average joes at a high percentage. In online multiplayer games, the experts beat the new players at a high rate of success. The problem is the misunderstanding of DCs. DCs are not relative to the attempter. An easy Search check is easy for a normal and automatic for a expert. A hard spot check is difficult for a normal and easy for an expert. An impossible search check is impossible for a normal and difficult for the expert. DMs just have to understand DCs better.


True, and I think the DMG in Next needs to be very explicit about this. They neglected to include this detail in 4e on the assumption that it was obvious and... well, we see how well that worked out for them.

True, and I think the DMG in Next needs to be very explicit about this. They neglected to include this detail in 4e on the assumption that it was obvious and... well, we see how well that worked out for them.



I agree.  4e was misleading (not intentionally I'm sure) for many people with their adjustable DCs by level.  DCs for a particular task should remain fixed no matter anyone's level.   I'm sure their intent was provide challenging tasks that equate out to these DCs when you want to provide a challenge.  I'm also sure they didn't mean for the same wood door to get a higher DC over time.   But some DMs interpreted it that way and I personally think they brought it on themselves with their wording.

 
I was thinking of something simple that allows someone who is good enough at a skill but not able to take 10, due to being in a stressful situation, to be able to mitigate a skill failure. The rule would be something like this:

If your Passive Score (just 10+skill check bonuses) equals or exceeds the DC of a check (so you would succeed if you could take 10), and you fail that check, you may instead make the check a success but suffer Disadvantage on your next d20 roll (this represents your loss of focus while you attempted to recover from your skill check so as not to fail) or grant Advantage to the next d20 roll used against you, whichever comes first.

If you want to make this more of an ahead of time thing, give all PCs the ablity to take 10 even if they cannot, but if they choose to take 10, they suffer Disadvantage on their next d20 roll or grant Advantage to the next d20 roll against them, whichever is first. If you use this option, they cannot use this (as the above rule) if they fail, they would need to make that choice before the roll, and if they risk it because they do not want the Disadvantage, they suffer the failure if they roll low.

Allows you to focus on a skill at the expense of something else. 
One of the reasons I came up with this idea is due to how a skill challenge ended up with the most trained player rolling poorly.

Here is an example of how I see the difference between this working vs 4e.

4e: the party has to cross a river of average difficulty (15). Mr athletic has a +8 and joe has a +0. Mr athletic succeeds 70% of the time while joe only success 30% of the time. Mr athletic has a decent chance to fail even though he is a trained athlete, and joe has a pretty good chance to start drowning.

Advantage: the river now only has a DC 10 because numbers are lower. Mr athletic has +3 but has advantage. Joe is still +0. Mr athletic has a 91% chance to cross the river and joe now has a 55% chance to. This means mr athletic has a great chance to succeed at the tasks he is trained for, but the rest of the party does not have to fear for their lives at the thought of attempting the same task as someone who is trained.

Now, I am open to the possibility of additional levels of training to increase the top end of skills, but bonuses to skills for a feat tend neglected for more interesting options. Having to spend 3 feats just to have a +8 (as is the current proposed model for next) seems excessive.
Having to spend 3 feats just to have a +8 (as is the current proposed model for next) seems excessive.

Where are you getting this from? I see nothing of the sort in the available material.

Remember also that skill DCs are going to scale much differently in D&DN than they have in previous editions, if what we see in the playtest packet is any indicator. You're not going to have anymore DC 40 checks because we're using ability mods rather than skill training + 1/2 level + mod 1 + mod 2 etc. Also, the vaguery involved with skill usage is most likely going to help rather than hinder in D&DN, provided your DM is reasonable.

I don't know if advantage/disadvantage would be a good idea for replacing skill bonuses, simply because it is too easy to cancel someone's advantage, whereas a bonus to the d20 roll is always there (presumably).

Eh. My 2 cents, anyway.
The issue with novice/expert is more to do with party dynamics. In 3e especially, but also 4e, it was common for some players to have a bonus 10+ points above other party members. This caused two problems. One, DCs were hard to create that could present an appropriate challenge for the group. Two, the specialist so fat outclassed the rest of the group that it was hard to include everyone during some phases of task resolution. Now if DCs were along these lines: 5 easy, 10 average, 15 hard, 20 very hard, 25 legendary, then being trained has a big benefit. As you said someone with focus and a high score can almost auto succeed an average task, while someone with an average score and no training only has a 55% chance. The difference is the person with no training can still attempt the task and contribute to the groups success even at difficulties that the trained person auto succeeds. Because the variance is lower between highest skill and lowest, it takes longer for those who aren't trained to be pushed into the "why bother rolling" category. This tends to happen around level 8 in 4e and as early as 1st level in 3e.



I tend to play in Living campaigns, and it's not uncommon for me to have an entirely different group of players from adventure to adventure.  This gives a different perspective on skills.  A real problem with many "pick up" groups in Living Forgotten Realms is that players will attempt and fail skill checks despite the knowledge that there is an expert in the group.  It's a symptom of being selfish in my opinion (based on personal experience) in most cases. 

When you reduce the DC into the manageable range, you can run into the problem of certain selfish characters taking over skill checks.  It doesn't solve the lack of participation; it just shifts it to where a selfish player does not have to invest into a skill to successfully take away participation from others.  

Now, an experienced DM can usually handle this situation (most of the time; I did see one situation that ended badly - I mean real badly - in the last ten years); however, an inexperienced DM is at risk.  The goal of the game should be to provide a fun game "out of the box."  A flatten system, while it does promote participation, does give less than ideal players tools they shouldn't have.  The more traditional system, while still abused to a degree, gives the other players more of a right to exclaim "WTH!" and regulate themselves.  This along with DM mediaton tends to reduce overparticipation.  
        
I tend to be alittle cynical on how successful this part of the system will end up due to those experiences... 
One way to do this would be to say that all skills cost 1 point but the first time you buy it you get a +5, the second time a +3, and the third time and after a +1.   This would accurately reflect how hard it is to improve into the master ranges. 

+1

I suggested something similar in another post (only with a +3, +2, +1 progression).

Hmm, if taking 10 doesn't retun, then I think you should still be able to take 10 if you are trained in a skill.  Though, I guess if we went with auto-success, then if you are trained you should auto-succeed if the DC is equal to your ability score or less, rather than the DC+5 being equal to your ability score or less (remember +3 bonus is like +6 to the ability score).  That could be done instead of taking 10.  (I suspect the auto-success shall be tweaked a great deal, and this should be tweaked accordingly).

I also think that if you are trained, then you ignore any penalty an ability modifier might give you to your check.