Prima System - Core Mechanic discussion.

Michael Morris · Feb 21, 2014

I'm starting a new thread since I really don't know what page the old one has fell to, and there are major changes since the last time I wrote on this. I am going to review how I reached this point.

I started with a blend of d20 and Savage Worlds. You rolled

d20 + ability die + skill die

The ability and skill dice fell in the d4 to d12 range. Early on I hit upon allowing multiple dice to be used for extreme skill or ability, but to keep the target numbers sane (a major criticism I have of d20) you only keep two of the dice. So a dragon with 4d12 for its strength and 2d12 for fighting skill rolled d20 + the best 2 of 6d12.

Playtests where conducted. It seemed natural to put in equipment and magic as items that added dice. I soon hit a problem of having lots of low dice, especially if there were d12's in the mix.

Also, the limit 2 rule had made it a bit less fun to be high level. While running Pathfinder, towards the end of my game, it hit me that it would be nice if my big bads could do more on their turns. High level characters getting more actions? Well, fighters do get more attacks. So I decided that if someone was rolling more than 2 dice, then could divide the dice among targets, but had to assign at least 2 dice per target.

That worked, but it still had problems. My latest idea is to decouple the dice from the various effects somewhat as follows.. This will be hard to follow cause I'm still working out terminology.

Abilities, skills, traits, equipment, spells, all these have a rank of 0 or higher. When you perform an action, you add the ranks of whatever you have that is relevant. At the very least, an ability will be relevant. The rank determines the dice you use with the d20 as follows:

1 d4
2 d6
3 d8
4 d10
5 d12
6 d12+d4
7 d12+d6
8 d12+d8
9 d12+d10
10 2d12
11 2d12+d4
...

And so on, infinitely. If you have more than 2 dice to roll with the d20, you can break the action into multiple targets. If you want to take more than one action in a turn you may do so, but the dice pool used will be that of the action you have the least skill in from among the ones you are attempting, and it must have at least 3 dice (be rank 11).

Consider a fighter with a combat trait rank of 2, longsword weapon skill rank of 2, and a Strength rank of 5. He rolls d12+d10 when he attacks with that sword. These sorts of numbers aren't unknown for a low level character.

I've also worked out the rudimentary advancement scheme, which lies somewhere between class/level and character point systems. Instead of receiving experience points that can be spent on anything, you receive points of different types.

Skill points are the baseline, received at the end of the session. A count of skill points is the same as your level in other games. Skills are boosters to actions the character specializes in. Specific weapon proficiency is a skill. The one restriction on skills is you can't boost the same skill two sessions in a row. You have to alternate.

Edges are obtained once every even level. They are analogous to minor feats.

Traits are obtained on levels divisible by 3. They are analogous to Base attack bonus, saving throw bonuses, or for those not coming at this from d20 traits are the skills that everyone wants. A character's combat trait for example affects *all* attack rolls. Magic use will be a trait.

Feats are obtained on levels divisible by 5. They are analogous to major feats (the ones high up on feat trees in Pathfinder) or to class abilities (uncanny dodge, et al) They include raises to attributes.

This approach insures that something happens at the end of each session (a skill goes up) but major character feature advancement is more spread out. DM's can advance more than one tick of course.

"Classes" in this system will be precrafted advancement maps that sit on top of the above.

Thoughts?

GMMichael · Feb 23, 2014

First two thoughts:

Michael Morris said:
1 d4
2 d6
3 d8
4 d10
5 d12
6 d12+d4
7 d12+d6
8 d12+d8
9 d12+d10
10 2d12
11 2d12+d4
...

And so on, infinitely.

This looks like a horrific table to wade through. When I thought it might not be so bad, I saw the "infinitely" part.

Michael Morris said:
I've also worked out the rudimentary advancement scheme, which lies somewhere between class/level and character point systems. Instead of receiving experience points that can be spent on anything, you receive points of different types.

Skill points are the baseline, received at the end of the session. A count of skill points is the same as your level in other games. Skills are boosters to actions the character specializes in. Specific weapon proficiency is a skill. The one restriction on skills is you can't boost the same skill two sessions in a row. You have to alternate.

This advancement scheme looks similar to mine - each session you advance a bit in something useful. So, +1!

Michael Morris · Feb 23, 2014

DMMike said:
This looks like a horrific table to wade through. When I thought it might not be so bad, I saw the "infinitely" part.

It's actually not that bad if you break it down into two steps. First, divide the rank by 5. That's the number of d12's. The remainder is the other die 1 - d4, 2 - d6, 3 - d8, 4 - d10. If the rank is a multiple of 5 then it's all d12's.

No matter how many dice get cast, only the highest 2 count. Hence the while the odds of performing a difficult task go down, the target numbers never change. They are:

12 - Very Easy
15 - Easy
18 - Medium
21 - Hard
24 - Very Hard

Very easy is made just over 50% of the time by the weakest of characters - rank 1 - someone rolling d20+d4 (1 average roll). Easy is mad just over 50% of the time by someone at rank 4 (d20+d10). Medium is made half the time at rank 6 (d20 + d12 + d4, 19.5 avg), hard is made just over half the time at rank 8 (d20+d12+d8, 21.5 avg) and very hard is made at rank 10 just under half the time (d20+2d12, 23.5 avg). Target number 24 is also significant in that it's the highest target number a rank 1 character can hit, though the odds of rolling 4 and 20 on d20+d4 are 1 in 80, so it's a long shot.

Unless a character wants to have a lot of actions, the issue of diminishing returns comes up. Once you have 2d12 to roll on an action, the next five points you spend are probably more useful spent on other things. The next five after that are almost certainly better off elsewhere.

The neat part of this is the game therefore doesn't need the complexity of changing the cost to purchase a power as it goes up. For example, in a lot of games the point cost of a skill is often the skill's current value. So, rank 1 costs 1, rank 2 costs 3, rank 3 costs 6. Eventually it becomes very expensive to increase a skill, encouraging the player to branch out. Prima is sneakier in this regard - the cost doesn't change, but the odds benefit (aside from extra actions) per point spent drops over time, even from the beginning.

GMMichael · Feb 23, 2014

Michael Morris said:
the target numbers never change. They are:

12 - Very Easy
15 - Easy
18 - Medium
21 - Hard
24 - Very Hard

This looks like a Modos RPG table too, but then, I've also seen it in the D&D 3.0 books. So now I'm wondering what happened to the Savage Worlds influence in your game. Is that the use of dice instead of points for skills? I'm afraid you might have wandered a bit far from those roots, because Savage Worlds eliminates a lot of math by leaving target numbers at 4. (Correct me if I'm wrong, I'm not a SW expert). You've reintroduced the math by adding the d20 in. Further, one could say that you haven't eliminated skill points, since each die can be considered a fixed average: d4 is a 2.5 bonus, d6 is a 3.5 bonus, etc. Granted, the dice are a lot more interesting than adding 4.5 to a d20 roll.

Michael Morris said:
The neat part of this is the game therefore doesn't need the complexity of changing the cost to purchase a power as it goes up. For example, in a lot of games the point cost of a skill is often the skill's current value. So, rank 1 costs 1, rank 2 costs 3, rank 3 costs 6. Eventually it becomes very expensive to increase a skill, encouraging the player to branch out. Prima is sneakier in this regard - the cost doesn't change, but the odds benefit (aside from extra actions) per point spent drops over time, even from the beginning.

Imma throw Modos RPG up for contrast, since we have some good ideas in common.

Skills are purchased in Modos for one point each (and each is effectively a permanent bonus to one type of skill roll). Sounds cheap, like D&D, right? Well, you get one skill point per level, or two if you sacrifice a perk. So make it count. Further, players are encouraged to diversify by virtue of fighting and defenses being skills as well.

The contests use a d20 as a base. And they add modifiers for ability and skill. However, you can take half (of the max die roll) at any time. So an easy-ish die roll at 50% becomes automatic. This is how you can reduce the chaotic sway of a d20 from the player's side. And all target numbers are actually target bonuses: the GM rolls a d20 in opposition, or takes half. So if a player's result on a contest is 5, the GM can still roll and get below that.

One big difference here, since we have d20s and modifiers in common, is that Modos RPG players can basically add their ranks to die rolls, instead of occasionally checking a chart to see which dice they should be using.

doghead · Feb 24, 2014

Michael Morris said:
Thoughts?

It looks like an interesting system. Once you get your head around the "divide by 5 and use the remainder to determine the last dice" then it is also quite simple to step from total ranks to dice pool.

I like the idea of having the choice between loading the dice one one task to ensure the best result i.e. Total Rank 11 (best 2 of 1d20+2d12+1d4), or splitting the dice to try and achieve multiple actions. Based on the above it looks like you only get one 1d20, with subsequent actions rolled without the d20 (first action: 1d20+1d12 and second action: 1d12+1d4). Is it an option to go (first action 1d20+1d4, second action: 2d12)?

Alternatively, you could say that in order to attempt iterative actions, you have to burn a dice. Thus making Total Rank 16 the threshold for 2 actions, 26 the threshold for 3 actions etc.

2 actions: 1d20+1d12, 2d12 and (burn 1d4) - Min Total Rank 16 (3d12+1d4)
3 actions: 1d20+1d12, 2d12, 2d12 and (burn 1d4) - Min Total Rank 26 (5d12+1d4)

If you required the character to burn a dice for each iteration, the Thresholds become Total Rank 16 for two actions, Total Rank of 31 for three actions etc

2 actions: 1d20+1d12 (free), 2d12 (burn 1d4) - Min Total Rank 16 (3d12+1d4)
3 actions: 1d20+1d12 (free), 2d12 (burn 1d12), 2d12 (burn 1d4) - Min Total Rank 31 (6d12+1d4)

The former is a little more elegant in progression (16/26/36/46/etc), the latter definitely pushes the multiple iterations into the domain of the high level character.

thotd

Michael Morris · Feb 24, 2014

DMMike said:
You've reintroduced the math by adding the d20 in. Further, one could say that you haven't eliminated skill points, since each die can be considered a fixed average: d4 is a 2.5 bonus, d6 is a 3.5 bonus, etc. Granted, the dice are a lot more interesting than adding 4.5 to a d20 roll.

That's only true to a point. Once three or more dice are tossed, the fact that only 2 are kept throws this analogy out the window. It's also the reason why the target numbers can stay the same at all levels at play, whereas with fixed bonuses you eventually will reach a point where the target numbers that where average risks in early play become all but impossible to miss at higher level play. D&D 3.0 and its descendants suffer from this problem a great deal - once you hit +23 for skill checks then target numbers have entered a different scale. It gets worse when you have one character with a +4 and another with a +34 in the same party for the same skill. Essentially any target number the weaker might be able to hit is a give away for the stronger. Skills stop becoming a guide at what a character is good at and become a straight jacket for what they can and cannot do. 5e addresses this by capping bonuses, and that will work but I think this plan works better.

Imma throw Modos RPG up for contrast, since we have some good ideas in common.

Skills are purchased in Modos for one point each (and each is effectively a permanent bonus to one type of skill roll). Sounds cheap, like D&D, right? Well, you get one skill point per level, or two if you sacrifice a perk. So make it count. Further, players are encouraged to diversify by virtue of fighting and defenses being skills as well.

That's not what I'm doing. Yes, players get a skill point at each session end, but skill points are just that, skill points. Other things come at other rates and I'm still testing what should be at which rate. At the moment I'm testing traits on even levels, edges on levels divisible by 3 and feats on levels divisible by 5. So at the 5th level, or 5th session, you get a skill point and a feat. At 6th you get a skill point, a trait and an edge. The points are not interchangeable - you can't spend two skill points to get a trait. Part of this is to make it difficult (though not impossible) to make a completely gimped character.

doghead said:
It looks like an interesting system. Once you get your head around the "divide by 5 and use the remainder to determine the last dice" then it is also quite simple to step from total ranks to dice pool.

I like the idea of having the choice between loading the dice one one task to ensure the best result i.e. Total Rank 11 (best 2 of 1d20+2d12+1d4), or splitting the dice to try and achieve multiple actions. Based on the above it looks like you only get one 1d20, with subsequent actions rolled without the d20 (first action: 1d20+1d12 and second action: 1d12+1d4). Is it an option to go (first action 1d20+1d4, second action: 2d12)?

When doing multiple actions the d20 is cast once. It is added to each group. So if you split your dice d12, d12+d4 against two targets, the d20 roll becomes critical for both attacks. On the contrary, if you split 2d12, d4 then it's possible to hit even with a low d20 roll with the 2d12, but the d4 is almost an afterthought to that same d20 roll.

Alternatively, you could say that in order to attempt iterative actions, you have to burn a dice. Thus making Total Rank 16 the threshold for 2 actions, 26 the threshold for 3 actions etc.

2 actions: 1d20+1d12, 2d12 and (burn 1d4) - Min Total Rank 16 (3d12+1d4)
3 actions: 1d20+1d12, 2d12, 2d12 and (burn 1d4) - Min Total Rank 26 (5d12+1d4)

If you required the character to burn a dice for each iteration, the Thresholds become Total Rank 16 for two actions, Total Rank of 31 for three actions etc

2 actions: 1d20+1d12 (free), 2d12 (burn 1d4) - Min Total Rank 16 (3d12+1d4)
3 actions: 1d20+1d12 (free), 2d12 (burn 1d12), 2d12 (burn 1d4) - Min Total Rank 31 (6d12+1d4)

The former is a little more elegant in progression (16/26/36/46/etc), the latter definitely pushes the multiple iterations into the domain of the high level character.

thotd

I don't follow any of this. I also don't expect ranks to get up into the 40's either, at least not for awhile. I'll probably test things out to around 100 ranks

GMMichael · Feb 24, 2014

Michael Morris said:
Once three or more dice are tossed, the fact that only 2 are kept throws this analogy out the window. It's also the reason why the target numbers can stay the same at all levels at play, whereas with fixed bonuses you eventually will reach a point where the target numbers that where average risks in early play become all but impossible to miss at higher level play. D&D 3.0 and its descendants suffer from this problem a great deal - once you hit +23 for skill checks then target numbers have entered a different scale. It gets worse when you have one character with a +4 and another with a +34 in the same party for the same skill.

Let me get this straight - at rank 5, a player rolls a d20 and a d12 for his outcome, and his range of outcomes is between 2 and 32? A little swingy. Is that what you're going for? Swing by anydice.com and check out some of the outcome distributions.

I see what you're saying about the D&D skill-monkey versus skill-n00bs. Now, a discussion about D&D rules doesn't fit this thread...UNLESS D&D rules are helping to inform and shape the creation of your own rules. So...
You're right: a character with +23 in Swimming is going to make his wizard buddy with -2 in Swimming look like a fool. But I don't think this is a system problem; I think it's a GM problem. The GM should plant non-competitive DCs firmly in the 15-25 point range. This gives characters with really low modifiers a chance, while it allows characters with high modifiers to shine. Obviously, it's unbalanced if the mer-man is directly competing in a swimming contest against the wizard. But this overlooks an important question:

Isn't +23 already a little unbalanced? And the companion question: shouldn't the wizard, hopefully at equally high level, be able to well outshine the swimmer in a different category?

What this means for Prima: narrowing the range is fine. But do you want a system that brings the novices closer to competing with the demi-gods?

Michael Morris said:
That's not what I'm doing. Yes, players get a skill point at each session end, but skill points are just that, skill points. Other things come at other rates and I'm still testing what should be at which rate. At the moment I'm testing traits on even levels, edges on levels divisible by 3 and feats on levels divisible by 5. So at the 5th level, or 5th session, you get a skill point and a feat. At 6th you get a skill point, a trait and an edge. The points are not interchangeable - you can't spend two skill points to get a trait. Part of this is to make it difficult (though not impossible) to make a completely gimped character.

Oh, I'm with you. In the Modos RPG game I'm running, PCs get one award after each session: a skill point, ability point, or perk. Gain one of each, gain a level. Can PCs make a completely gimped character? Yes. But there are two safety mechanisms:

1) If you roleplay your gimped character well, this provides you with more hero points (which can be used as either special abilities or luck).
2) The GM has Rule Zero: the Guide of Modos is always right. So even if the rules say you can spend all your points in the Metaphysical ability, the skill Underwater Basket Weaving, and perks converted into more Metaphysical points . . . Rule Zero. The GM says the gods take pity on you and grant you some extra Parry skill points...just to make you less gimped.

doghead · Feb 24, 2014

Michael Morris said:
When doing multiple actions the d20 is cast once. It is added to each group.

Ah. I thought that it was used only once.

Michael Morris said:
I don't follow any of this. I also don't expect ranks to get up into the 40's either, at least not for awhile. I'll probably test things out to around 100 ranks

That's OK. Its not really relevant as it was based on using the d20 only once. You mentioned originally that multiple actions was something you thought would be good for higher level characters. Not knowing what 'higher level characters entailed in this system, I was just fooling around with numbers. I wanted to see what Rank would be required in order to be able to do multiple actions, and how this changed by requiring characters to spend/sacrifice/burn a dice:
1. to allow any number of additional actions, or
2. for each additional action attempted.

thotd

Michael Morris · Feb 26, 2014

DMMike said:
Let me get this straight - at rank 5, a player rolls a d20 and a d12 for his outcome, and his range of outcomes is between 2 and 32? A little swingy. Is that what you're going for? Swing by anydice.com and check out some of the outcome distributions.

I'm quite aware of that and ok with it. I also know what bell curves are and how they work.

What this means for Prima: narrowing the range is fine. But do you want a system that brings the novices closer to competing with the demi-gods?

I want a system where the target number for "very hard" doesn't have to be moved up to something unreasonably high at high level. Because of how bell curves work, high level characters become frighteningly consistent at very hard tasks (24) and can even try harder than hard (27, 30, 33 ...) - but eventually the curve peters out. Each point raised beyond thirty lowers the odds, even for high level characters with lots of dice, by an order of magnitude or more. 44 is the max target number with d20+xd12 keep 2. And after awhile the gain for putting a point on a skill becomes minuscule, how minuscule depends on the extra actions rules which are being worked on.

2) The GM has Rule Zero:

which is NOT an excuse to write a game system so shoddy the GM has to routinely invoke it. In my mind, any time the GM invokes this "rule" the game system has failed. No system can cover all situations, but if the GM is constantly overruling the system rather than using it then something is wrong.

Michael Morris · Feb 26, 2014

Bennies

Prima, like Savage Worlds has bennies. They work differently though because they aren't a re-roll mechanic. Also, they are dice based.

A player starts with 3 d20's in the benny pool (or hero pool or whatever it ends up being called). When the player does something particularly noteworthy the GM can give him additional dice for the pool - they can always be d20's, or the GM can come up with a graduated system and put other dice in - the base rules assume they are always d20's.

These dice can be used on any roll. If called before the roll they go right in as d20's. So, if you really need to make that check you can roll 2d20 plus the dice you make on the check as determined by your rank. These d20's can replace any of the dice in the throw - the normal d20, or if it did roll high it can displace one of the rank dice. If your rank is 5 or less the die can augment the roll outright.

If you fail at a roll and want to nudge it up you can do that, but the die tossed drops a step. d20's become d12's when used this way. The newly rolled die can only replace one of the dice in the failed roll.