Skills over 100%, the mega charts edition thread

[Lordabdul]

There are a few rules discussions that invariably generate endless debates around these forums, and the rules for over-100% skills is definitely one of them... I saw a lot of complaining, a few attempts at making house rules to "fix" it (assuming you're in the camp who think the rules need fixing... all camps are valid!), but nobody really tried to use Science!(tm) to address the problem. And by "Science!(tm)", I of course mean every game designer's favourite tool: spreadsheets!

Disclaimers

I started a spreadsheet on this topic a while ago but never got around to finish it... thanks to the quarantine, I now have it in a more-or-less usable state, and I thought I would share a sneak peek here. But before this degenerates into an unproductive debate between RQ3/Mythras/HERO/GURPS fanboys or something, I want to make it clear that the goal here is not to try and find a rule that absolutely everybody agrees on (although that would be nice). My goal here is just to offer a tool that people can use to see how this or that rule scales, and therefore what are the pros/cons of each rule. I have included a number of rules, but I definitely didn't include all known rules... feel free to point me to some rules you'd like to see graphed!

I didn't include the RQ3 rules because (unless I'm mistaken?) they are different between normal contest rolls (subtract one skill from the other) and combat (nothing? just... roll your score? I'm not sure).

The most important disclaimer however is this: those graphs assume that we are striving for a linear skill curve!  This means that, just as 80% skills is twice better than 40% skill (obviously), we want the rules to scale up and stay linear, i.e. 300% is twice better than 150%, and 600% is twice better than 300%. This might seem like a given, but not necessarily: some people might want to have the scale bias a bit "upwards", to help represent super-powered beings without having to go with ridiculous skill scores of 2000%. My spreadsheet can probably model this with a bit of extra work and extra tweakable parameters, but for now that's not the goal.

Rules

The rules I have here are:

• The "ideal" rules: these are the rules where you bring the highest skill score down to 100, and take out your calculator to compute the exact value of the other skill score ("lowest*100/highest"). It's not meant to be used at the table unless you're fast with a calculator But it gives us a baseline.
• RAW: the RQG rules. Lower the highest score to 100% if it's higher, and reduce the lower score by the same amount.
• JJPG1 and JJPG2: these were house-rules briefly mentioned by @jajagappa in this thread.
• GURPS: rules inspired by my "go to system" when it comes to great ideas for crunch vs playability. In this rule:
  • In most cases, you just roll as-is. So with 300% vs 75%, you just roll that 300% as-is, special'ing and crit'ing like crazy and wiping the floor with the other character.
  • If both skills are above 80%, to prevent endless rolling until someone crits or fumbles, lower the lowest skill to 50%, and lower the highest by the same amount.
  • If both skills are above 150%, it's not enough to "flatten the curve" (hey, I heard that somewhere else already), and that where GURPS, in typical GURPS fashion, makes you take the calculator out: you lower the lowest skill to 50%, and lower the highest skill by "50/lowest".
  • Note that my spreadsheet allows to tweak those numbers. So if you want to only take out the calculator above 300% instead of 150%, I can do that. If you want the rule to kick in above 100% instead of 80%, I can do that also.
• SimpG: same as the previous rule, but calculator-free! Just lower the lowest skill above 80% to 50%, and lower the highest one by the same amount.

Graphs

And now, for the piece de resistance: the glorious graphs!  These graphs here show the ratio between the "ideal" (linear) differences in scores, and the effective difference in scores (as expressed by the adjusted skill scores). The graphs take one fixed skill score (like, say, 120%), and pits that score against a range of skills going from 20% to 600% in 20% increments. Because those graphs go to 600%, don't forget to only look at the range you're interested in for your particular style of play. I tend to run more low-powered/gritty/"street-level" games, so I'm more interested in the left half of the graph (<300%).

For example, if a data point on the "150% graph" indicates that, in the "150% vs 220%" case (the value on the horizontal axis is 220), the difference ratio is "2.28" (the value in the vertical axis is 2.28), then it means that the difference of effective skill scores is roughly two and half times higher than it should be when rolling 150% vs 220%. This is for instance the case with RAW: the ratio between 220 and 150 is 1.46 (220 is 1.46 times better than 150), but after applying RQG rules, you get 100% vs 30%, and 100 is 3.33 times better than 30... 3.33 is 2.28 times 1.46 so the RQG rules make this particular skill contest 2.28 times harder/easier than what it "should" be... makes sense? Am I making any math errors here?  (maybe! please tell me and I'll adjust the spreadsheet formulas... which... are a bit convoluted to account for the minimum/maximum 5% and 95% who are always a success or fumble...).

As such, the "ideal" rules are a more-or-less flat line (give or take the rounding to the next integral number) that stays close to "1" all the way through (that is: the "adjusted" scores preserve the difference between the original scores). So the goal here (at least for me! you might have different gameplay goals!) is to find a rule that is a good compromise between keeping as close as possible to "1", while also being easy and fast to implement in play. BUT (yet another important disclaimer), this is all impersonal math! These graphs don't tell you if the rules are going to make things fun and exciting! For instance, you could just use no rules at all! Just roll the scores as they are, but (1) although this would be mathematically "correct", it doesn't take into account the fact that a d100 only rolls up to one hundred, and how you're not actually using the whole range of the skills and (2) some people don't think this is fun because when both skills are above 100%, the characters just succeed all the time (well, 95%) and it just boils down to rolling under your special/critical threshold, making the scene boring, and the results needlessly more brutal than a "normal" combat. But hey, I'm not here to tell you which is better -- I'm just here to make pretty graphs!

Ok, so now, here's the graph for a 75% skill:

 

Next, the graph for a 150% skill:

 

And last, the graph for a 300% skill:

 

Preliminary Commentary (before people find any mistakes!)

It's pretty clear, given those graphs, why people are unhappy with RAW. It has this sudden "wall" where opponents that are twice as better than you appear in practice to be 10 times better.

Jajapagga's house rules delay this effect a bit, which might be good enough for certain ranges of scores (and therefore certain "styles" of play, i.e. low-powered/gritty vs. "medium" heroic vs "we are play Argrath").

The GURPS rules are pretty good, but that's not a surprise since they are pretty close mathematically to "the right answer".

I'm mostly surprised by how the SimpG rules hold up -- they are pretty much as simple and straightforward as RAW (just a tiny bit more complex since subtracting 50 is a bit less intuitive than subtracting 100, but really not that much), and they yield a curve that is frankly very good in my opinion. I think I'm going to use those rules in my games.... (and hey Jason/Jeff, if you want to add that rule to the upcoming GM book, feel free to steal it ).

Note the small little artifacts around certain combinations of skills. Those are what adds a little spice to life! (but also they're weird).

That's it! I hope this is as fun to look at as it was for me to make. Cheers!

Edited May 15, 2020 by lordabdul

[Ultor]

My PCs have several skills well above 100% and have been taking every advantage of RAW. I'll be using the Simple Rule in my game tonight to see how it goes.

[HreshtIronBorne]

I really like the way that Simple version of the Gurps rule looks in the graphs. I think I will propose it to my group and see if we want to play around with it as skills increase above 100%. Does the graph change drastically for the Simple G rules if you only use them above 100% instead of 80%? I am thinking of 100% being easy to remember, maybe 90% as we all played a lot of RQ3 and Stormbringer so that is where we situate weapon Mastery intuitively. 

[Lordabdul]

54 minutes ago, HreshtIronBorne said:

Does the graph change drastically for the Simple G rules if you only use them above 100% instead of 80%?

Nope, it doesn't change the graph much, if at all. If anything, moving the threshold to 100% actually reduces a couple of very small artifacts around the 120-180 range (where it dips a bit below "1" on the 2nd and 3rd graphs... it keeps 2 more data points on "1").

GURPS is 3d6-based, but kicks in with its rule if both skills are above 14 (not 18), because, I think the designers are mostly concerned with making contests resolve faster, not necessarily with handling "always succeeds" rolls (and a 3d6 roll of 14 or less is around 80% chance of success, which is why I chose that threshold here). But yeah I agree that using a threshold of 100% feels a bit more BRP-ish, and is easier to remember. I would probably use that threshold too.

Edited May 15, 2020 by lordabdul

[Rodney Dangerduck]

I disagree that 80% is "twice as good as" 40%.  Is a .400 baseball hitter "twice as good as" a .200 hitter?  Absolutely not.  The .400 hitter is a priceless top tier Hall of Famer, the .200 is crap at the infamous "Mendoza Line". 

[Lordabdul]

1 hour ago, Rodney Dangerduck said:

I disagree that 80% is "twice as good as" 40%.

A d100 is going to succeed twice a much at rolling under 80 than under 40... that's... err... basic math. I'm not sure what I'm missing here about your argument? I'm really just talking about rules, not what the rules are modeling.

Edited May 16, 2020 by lordabdul

[GAZZA]

So just to clarify, the proposed Simp G is:

• If at least one score is 79 or less, do nothing.
• If both scores are 80 or higher, find X such that X = Lower Skill - 50, and then reduce both skills by X.

Just checking the math here.

[Julian Lord]

6 hours ago, Rodney Dangerduck said:

I disagree that 80% is "twice as good as" 40%.

You're right -- and having number crunched the heck out of this for my mass combat house rules, I worked out that BASIC's system of skill progression versus d100 dice rolling of both skill use and opposition follows a log scale ; which surprisingly enough turned out to be mathematically coherent with classic wargame rules, except that these rules assume 100% skill efficiency whereas roleplaying games assume 50%.

In that RPG model, 80% is less than 1.5 times better than 40% ; 90% is 1.5 times better than 40, it goes in 50s, although in a wargame-based system 80% v 40% would indeed be double.

You can actually check this for yourself ; but against a basic 50% opposition roll, a skill at 80% will be successful roughly one and a third times as often as a 40% skill, off the top of my head and IIRC (results do wobble a bit with very low skill % numbers, and with % numbers in the low 100s)

The reason why subtracting the difference between the lowest skill and 50 from both or all opposing skill numbers works is that it preserves that log progression a lot more closely than the official RQG system does -- by never allowing a skill to fall lower than the basic "passive" 50% opposition that the core game system expects.

But I still found in my own game that allowing a Heroic success chance of 1% for each full 100% skill, equal to a crit success but with an immediate second skill attempt at the same % ; and then otherwise just use the basic % numbers worked perfectly well to get rid of the lengthy dice rolling sequences.

Edited May 16, 2020 by Julian Lord

[Lordabdul]

1 hour ago, GAZZA said:

So just to clarify, the proposed Simp G

I'm not proposing anything, but yes I think you have the right math. It's basically like the RQG rules, but you lower to 50, not 100, and you lower the lowest skill, not the highest one.

54 minutes ago, Julian Lord said:

You're right -- and having number crunched the heck out of this for my mass combat house rules, I worked out that BASIC's system of skill progression versus d100 dice rolling of both skill use and opposition follows a log scale ; which surprisingly enough turned out to be mathematically coherent with classic wargame rules, except that these rules assume 100% skill efficiency whereas roleplaying games assume 50%.

Good points, but you're not quite comparing the same things. I was not saying that 80% vs 50% will have twice as many successes as 40% vs 50%. What I was saying is that 80% vs 50% is roughly equivalent to 40% vs 25% (dividing both by the same denominator). But that's not true either, and I was wrong about that... if I get my math right, 80% vs 50% has a win chance of ~47% (when taking into account specials and criticals and fumbles), but 40% vs 25% has a win chance of ~32%. So thanks for that, I'll have to re-adjust my spreadsheet actually. If anything, I think it will flatten the curves at higher points, making the GURPS and SimpG lines even closer to the "ideal" line... we'll see.... stay tuned for updated graphs. Thanks!

Edited May 16, 2020 by lordabdul

[AndreJarosch]

I am not a mathematican, and our group has no problems with these skill levens, since noone has skills over 100% (yet), except when using augments.

Maybe we should adapt the rules about masteries from HeroQuest?

Instead of 
95%   100%   105%

It would be 
95%   100%  5M%

For instance: 
a 150% attack Humakti is fighting a 90% attack Yelmalian

In RQG RAW it would become
100% Humakti vs 40% Yelmalian

If we adapt the HQ rule it would be
50M% vs 90%

The M means: No fumbles possible, a fail becomes a success, a success becomes a crit, a crit is still a crit

 

[GAZZA]

27 minutes ago, AndreJarosch said:

I am not a mathematican, and our group has no problems with these skill levens, since noone has skills over 100% (yet), except when using augments.

Maybe we should adapt the rules about masteries from HeroQuest?

Instead of 
95%   100%   105%

It would be 
95%   100%  5M%

For instance: 
a 150% attack Humakti is fighting a 90% attack Yelmalian

In RQG RAW it would become
100% Humakti vs 40% Yelmalian

If we adapt the HQ rule it would be
50M% vs 90%

The M means: No fumbles possible, a fail becomes a success, a success becomes a crit, a crit is still a crit

Mathematically, that makes it a big difference between 99% and 101%. At 99%, you have a 1% chance of fumbling, a 4% failure chance, a 70% chance of success, a 20% special chance, and a 5% critical chance. At 101% (or 1M) you'd have 0% fumbling, 1% failure, 4% success, 70% special, and 25% critical (assuming you meant that a success becomes a special, otherwise it's even worse). That means that 101% will utterly dominate a 99% opponent even though they are only 2% higher.

[Stephen L]

1 hour ago, GAZZA said:

At 101% (or 1M) you'd have 0% fumbling, 1% failure, 4% success, 70% special, and 25% critical (assuming you meant that a success becomes a special, otherwise it's even worse)

I've never read the HQ rules, much less played them, though I do own a shrink wrapped copy!, but my reading of Andre's post is:

101% become 1% with 1 level of mastery, so its: 0% fumble, 5% fail, 90% success, 4% special, 1% critical.

If it is, then I'd stick with the 99%.

[Stephen L]

13 hours ago, lordabdul said:

SimpG

This is very interesting, thanks for the post!

I always though that RQ works best with skills in the 50%-80% range.  If everyone makes their parries (which was often proof against a special), the fights used to go on forever.  But with lower level skills, you were basically waiting for you opponent to fail their parry.

(as an aside, combats are better now, where weapons/shield seem to suffer much more damage, and that's when you get your chance for the killer blow)

The Gurps seems to push things into what I was thinking was the sweet spot...

For my campaign, I'm encouraging the players to split attacks when their above 100%.  

[Stephen L]

(damn this new keyboard - hitting submit again in error!)

For non combat skills above 100%, I'm tending to Rules as Written (if that's the acronym RAW?), but ruled on a case by case basis, by what feels right for the moment (my players really don't care about rules, so they, fortunately, don't worry about consistency)

[AndreJarosch]

26 minutes ago, GAZZA said:

Mathematically, that makes it a big difference between 99% and 101%. At 99%, you have a 1% chance of fumbling, a 4% failure chance, a 70% chance of success, a 20% special chance, and a 5% critical chance. At 101% (or 1M) you'd have 0% fumbling, 1% failure, 4% success, 70% special, and 25% critical (assuming you meant that a success becomes a special, otherwise it's even worse). That means that 101% will utterly dominate a 99% opponent even though they are only 2% higher.

My assumption was: 

 

99%
1-5         Crit
6-20       Special
21-95     Success
96-99     Fail
100         Fumble

1M%
1          Success/Special/Crit = Crit     
2-99     Fail = Success
100      Fumble = Fail

20M%
1          Crit = Crit
2-4       Special = Crit
5-20     Success = Special
21-99   Fail = Success
100      Fumble = Fail
 

[GAZZA]

Sorry, you've lost me. You're saying that 1M% can't special any more? But 20M can?

[AndreJarosch]

43 minutes ago, GAZZA said:

Sorry, you've lost me. You're saying that 1M% can't special any more? But 20M can?

A crit is also a special, always!

Each crit is also a special, but not all specials are crits. 

Edited May 16, 2020 by AndreJarosch

[Stephen L]

58 minutes ago, AndreJarosch said:

1M%
1          Success/Special/Crit = Crit     
2-99     Fail = Success
100      Fumble = Fail

Unless you use the convention that 01-05 always succeeds, in which case:

1          Success/Special/Crit = Crit      
2-5       Success = Special
6-95     Fail = Success
96-00   Fumble = Fail

Which is the values I assumed.

2 hours ago, Stephen L said:

101% become 1% with 1 level of mastery, so its: 0% fumble, 5% fail, 90% success, 4% special, 1% critical.

I admit, it's largely prejudice, but I've never felt the urge to explore HQ as a system!

[David Scott]

3 hours ago, AndreJarosch said:

I am not a mathematican, and our group has no problems with these skill levens, since noone has skills over 100% (yet), except when using augments.

Maybe we should adapt the rules about masteries from HeroQuest?

Instead of 
95%   100%   105%

It would be 
95%   100%  5M%
 

While I thought this looks easy, the added complexity is that if used for combat, there is an additional level of special, which doesn't have much effect on normal skills. Keeping the special is the problem.

Ignore the special and it's fine...

[AndreJarosch]

3 minutes ago, David Scott said:

While I thought this looks easy, the added complexity is that if used for combat, there is an additional level of special, which doesn't have much effect on normal skills. Keeping the special is the problem.

Ignore the special and it's fine...

And as i saied A Critical is always also a Special. 

 

[Stephen L]

1 hour ago, AndreJarosch said:

A Critical is always also a Special

I think the problem isn't with Criticals (which as you say are Specials), it's with Specials that aren't Criticals.

[Joerg]

I suppose that specials in  @AndreJarosch 's proposal would still be calculated from the percentage above 100. Given the way RQG rounds always in favor of the player heroes, a skill of 101% or M1 1% would yield a special chance of 21, a skill of 120% or M1 20% a special of 24%.

What the mastery system doesn't deal with is ties within the same success bracket, unless you inherit the "high roll is good" rule from HeroQuest for resolving ties. High roll within the success bracket, that is. That's where the system becomes a bit un-intuitive, although it is effectively the same as in Pendragon, only applied to three different tiers of success level.

I am not convinced by the complete absence of fumbles. On a roll of 00, roll for the skill again, without the mastery. If that roll is above your skill (above the mastery), it is still a fumble.

Mathematically, this creates a continuous system. The question is whether dealing with all these marginal victories or parried attacks is what you want. What is the benefit of a marginal victory?

RQ3 had the crutch of fatigue points that hardly anybody ever played as written. Applying damage from marginal victories to fatigue and letting fatigue slowly draw down effective skill levels would be a simulationist's answer (house-ruling RQ3, not RQ3 RAW), and may be tolerable in a game where a computer gives you your adjusted stats at the start of every melee action. Seeing it applied in a pen&paper environment seems unlikely to me.

(Disclaimer: On occasion I do play a German language rpg by the name of Midgard that has a system similar to this, but that system uses a D20 and doesn't have the special success level to take into account.)

[Julian Lord]

RQ3 fatigue was great on paper, basically unusable around the table.

A RuneQuest game isn't an aptitude test for chartered accountancy ...

[Lordabdul]

The idea of re-using the HQ mastery system is interesting! But I'm not sure it goes well with the RQ philosophy where anybody, even a small trollkin, has a chance to kill you... the mastery/bumping system is more designed to make heroes be able to dominate non-heroes AFAICT. But I'd love to hear the feedback of people playtesting it.... and if you need something graphed, I can do that

As far as the baseline in my graphs, though, I'm not sure what to do with it... we need some kind of a baseline to be able to judge if adjusted scores are preserving the "difference of power" represented by the unadjusted skills. The adjusted skills behave the way normal d100 rolls (and opposed rolls) behave. Unadjusted skills behave... whichever way we want. What does 650% vs 380% "represent"? How stronger is one supposed to be compared to the other? Well that's up to us to decide, right? And these scores will behave in practice in different ways depending on what house rules we use (i.e. one house rule will make it into an effective 3:1 difference of power, while another house rule will make it into an effective 5:1 difference of power). What we can at least strive to achieve is to make sure the curves don't have weird inversions of power, roller coaster up-and-down ratio changes, etc... (i.e. I'm less interested in "how stronger is this compared to that" than in "make the progression of difficulty predictable"). Apart from that I'm not sure what else we can do? I'm open to suggestions!

Edited May 16, 2020 by lordabdul

[Runeblogger]

13 hours ago, AndreJarosch said:

Maybe we should adapt the rules about masteries from HeroQuest?

[...]

The M means: No fumbles possible, a fail becomes a success, a success becomes a crit, a crit is still a crit

Maybe a critical when you have a skill over 100% could reduce the level of success of he attempt to parry of your adversary. So if you get a result of 5 when you have 115%, not only you get a critical, but you reduce your adversary's successful parry to a failure.