Basic mechanics first: Roll and take highest. Meaning you roll a bunch of dice, take only the highest result. 2d6 could roll 3 and 4, so the result is 4.



Let's use dice to represent a character's abilities. A skill could have five levels for example:

• 1d4, 1d6, 1d8, 1d10, 1d12

So our possible results are 1-12, a reasonable range for determining success or failure. To think more practically, we could say that 3 is "easy", 7 could be "hard" and 12 would be "impossible".

Now let's say we use a similar scale for depicting attributes and items. So a character with a skill of d8 (good) in Swordfighting and d6 (decent) Strength armed with a Fantastic Sword (d8) would roll 1d6+2d8 when attacking, taking only the highest result. If we think a bit further, we could use this single roll to represent both hitting and damage in a combat system. You can view the probabilities of this roll here http://anydice.com/program/1872


So basically, a variable-dice-take-highest system. What we have

• Fast rolls
• Simple results
• No calculating
• Higher is always better
• Probabilistic control
• Temporary penalties are easy to control (restrict amount of dice or subtract a number)

However

• Lack of unlikely "critical" rolls, some results are impossible with lower dice
• Temporary bonuses are a bit more difficult to control

I have a few additional ideas in mind, but this is the basic premise. I'm pretty much a total newbie in game design and doing this for fun (the goal is to make a simple system maybe), so it would be really interesting to hear what sort of thoughts you guys have. If you know of any system that uses a mechanic like this, I'd like to hear about it and see what I can learn from it

The basic idea looks a bit reminiscent of Marvel Heroic Roleplaying (MHR) - it uses a similar set of different size dice, although the exact sources of the dice are different and it adds the 2 best (with 1s generating further complications and one dice assigned to damage or other effect).

I'm not sure what you mean by "probabilistic control". It looks a bit difficult to determine what exact probabilities are.

In most take-highest systems, the highest result is going to be  the most probable result (i.e. if you roll 2d6, "6" is the most likely single result) which is perhaps a bit anti-climatic. In your game, different sized dice perhaps moderate that a little bit, though (if you only have 1d8 among your d6s, a 7 or 8 is still uncommon).

You might want to look at Ironclaw/Jadeclaw series of games, as that's pretty much the dice system in a nutshell.

Savage Worlds works like that.

Would you consider any auto fails?  Or cancellations?

Say you get 1d6 and 2d8 as stated, you get a 4 and 8 on your d8, but a 1 on your d6.  The auto fail would say the 1 is just a fail, no matter what the other dice read.  Cancellation would say the 1 knocks out the 8, leaving an effective 4.  

Just a thought on balance.  The always take high rolls work well, but sometimes make everything a success.  Variable dice help as well.

Quote from: Bloody Stupid Johnson;594101The basic idea looks a bit reminiscent of Marvel Heroic Roleplaying (MHR) - it uses a similar set of different size dice, although the exact sources of the dice are different and it adds the 2 best (with 1s generating further complications and one dice assigned to damage or other effect).

I'm not sure what you mean by "probabilistic control". It looks a bit difficult to determine what exact probabilities are.

In most take-highest systems, the highest result is going to be the most probable result (i.e. if you roll 2d6, "6" is the most likely single result) which is perhaps a bit anti-climatic. In your game, different sized dice perhaps moderate that a little bit, though (if you only have 1d8 among your d6s, a 7 or 8 is still uncommon).

The probabilities are practically impossible to calculate by hand when you have more than two dice. We can summarize though: the more dice you roll and the higher dice you roll the better your result will probably be.

One side-effect, for example, is that even with just 3 dice the probability of getting 1 as a result is less than 1%. So characters with good skill, good stats, good gear - they can still fail but it gets more and more unlikely. For the 1d6+2d8 example, the probability of getting less than 4 is 16%, 5-8 are the most common results. This is kinda what I mean with probabilistic control - a character with a certain skill level, let's say "superb", should get mostly "superb" results.

Quote from: rway218;594150Would you consider any auto fails?  Or cancellations?

Say you get 1d6 and 2d8 as stated, you get a 4 and 8 on your d8, but a 1 on your d6.  The auto fail would say the 1 is just a fail, no matter what the other dice read.  Cancellation would say the 1 knocks out the 8, leaving an effective 4.  

Just a thought on balance.  The always take high rolls work well, but sometimes make everything a success.  Variable dice help as well.

They could work well, but I'm not sure if they'd work with these mechanics. When you get more dice, your probability of rolling 1 with at least one die also increases, so your chance of forced failure increases. Though then again, rolling 1 on a 1d10 is less likely than rolling 1 on a 1d4, so this is worth investigating.



One addition I have in mind is explosions. Whenever you roll the maximum of a die, you get to roll an extra die, one that's also one step higher. So rolling 4 on d4 gives you a d6 to roll. This way you could theoretically roll 12 even with a single d4. This mechanic would add a low-probability chance of rolling very high results. Does looking for maxes in the rolls seem like it would distract from the play too much?

http://anydice.com/program/1876

The probabilities are very attractive, but is it worth the extra trouble? I like the idea of very rarely getting totally ridiculous results even with newbie characters. It's realistic too, I guess.

With exploding dice you have a higher chance of rolling an 8 on a d6 than you have of rolling an 8+ on a d8.

A solution is to say that the dice implode on a 1, and you roll the next dice up instead. A 2 is an auto-fail.

Quote from: Omnifray;594177With exploding dice you have a higher chance of rolling an 8 on a d6 than you have of rolling an 8+ on a d8.

A solution is to say that the dice implode on a 1, and you roll the next dice up instead. A 2 is an auto-fail.

Here's how the probabilities would work with a d6 and a d8 and the explosions. No matter how many dice we roll, we only take the one highest roll.

Maybe this explains it better than I did
http://anydice.com/program/1878

So a d6 has a 1.67% chance of rolling 8, while d8 has a 10% chance of rolling 8. Explosions do get rarer with higher dice, though, so d8 rolling [max+2] is 1% less likely than d6 rolling [max +2].

Quote from: davidov;594167The probabilities are practically impossible to calculate by hand when you have more than two dice. We can summarize though: the more dice you roll and the higher dice you roll the better your result will probably be.
 
One side-effect, for example, is that even with just 3 dice the probability of getting 1 as a result is less than 1%. So characters with good skill, good stats, good gear - they can still fail but it gets more and more unlikely. For the 1d6+2d8 example, the probability of getting less than 4 is 16%, 5-8 are the most common results. This is kinda what I mean with probabilistic control - a character with a certain skill level, let's say "superb", should get mostly "superb" results.

Well...if skill is just one dice and if a higher skill = a larger dice, then the results a character generates are getting more unpredictable the greater the skill is. Its only when all the factors line up (high stat + high skill + equipment and other bonuses) that the characters' result numbers become more reliable.
 
 
I can imagine attributes being a problem; having a stat of d12 won't win contests with an any degree of certainty, but will give a character a chance of very high results (e.g. wins on opposed rolls) on any skill based off that stat. Though perhaps you can get around it by having stat dice be generally smaller than skill dice, or cost much more.

Quote from: Bloody Stupid Johnson;594187Well...if skill is just one dice and if a higher skill = a larger dice, then the results a character generates are getting more unpredictable the greater the skill is. Its only when all the factors line up (high stat + high skill + equipment and other bonuses) that the characters' result numbers become more reliable.
 
 
I can imagine attributes being a problem; having a stat of d12 won't win contests with an any degree of certainty, but will give a character a chance of very high results (e.g. wins on opposed rolls) on any skill based off that stat. Though perhaps you can get around it by having stat dice be generally smaller than skill dice, or cost much more.

Definitely, stat dice will have to be much harder to get than skill dice. Having d12 in a stat would pretty much be reserved for godlike beings.

One option would be to forget stats and attributes, and instead have more focus on skills, for example ranks being

• 2d4, 1d4+1d6, 2d6, 1d6+1d8, 2d8.. etc.

http://anydice.com/program/1879 (click "Graph" and "At Least")

Quote from: Bloody Stupid Johnson;594187Its only when all the factors line up (high stat + high skill + equipment and other bonuses) that the characters' result numbers become more reliable.

Then again, this might be a desirable effect. Having a high skill with no 'talent' or 'gear' to support it, you won't get as reliable results.

Quote from: davidov;594186Here's how the probabilities would work with a d6 and a d8 and the explosions. No matter how many dice we roll, we only take the one highest roll.

Maybe this explains it better than I did
http://anydice.com/program/1878

So a d6 has a 1.67% chance of rolling 8, while d8 has a 10% chance of rolling 8. Explosions do get rarer with higher dice, though, so d8 rolling [max+2] is 1% less likely than d6 rolling [max +2].

Ah! yes, I get you now. You roll a d6, it explodes on a 6, that means you've got at least a 6, but you're rolling a d8 as well, and if you get a 7 or 8, your score improves to a 7 or 8. [I think I misunderstood the dice-rolling you had in mind because I assumed it was more like Savage Worlds / Deadlands.]

Have you worked out what odds a d4 would have of beating a d8 or a d12 on an opposed roll? I suspect they're quite close together... cf. my other posts in a nearby thread on dice-pools.

Quote from: Omnifray;594205Have you worked out what odds a d4 would have of beating a d8 or a d12 on an opposed roll? I suspect they're quite close together... cf. my other posts in a nearby thread on dice-pools.

No I haven't.. and calculating that might not be a trivial matter especially if we want to include the explodes and more complicated systems with different amounts of different dice.

For ordinary dice:
d4 "wins" d6 with a 25% probability (6/24, 6 situations where d4 wins out of  6*4=24 possible situations. Should be right..)
d4 wins d8 - 18.75% (6/32)
d4 wins d12 - 12.5%

For more complicated systems, it might almost be easiest to solve the probabilities through simulating.

Here we see one complicated system, 2d4 vs. 2d10 (with explodes)
http://anydice.com/program/187d

A quick simulation of 1000 rolls would indicate that 2d4 wins 2d10 roughly 9% of the time. If we turn off explosions, that probability goes down to 5.5%. For the sake of accuracy, the probability I got for "d4 wins d6" by simulating 1000 rolls was 26.7%.


To summarize.. more and higher is better

To test and explore these mechanics a bit further, here's a way of using them in Fudge/Fate instead of Fudge dice. Remove skill numbers and the fixed dice mechanic, and replace those with this.



How-to:
Roll the dice for the relevant skill, take highest result.

Results above 10 exist too of course, and they're better than "Legendary".

Players should start from Mediocre (2d4) as in Fate. Poor (1d4) probably shouldn't even be there. These levels exist mainly for the players, the GM can use whatever dice he wants to.


Pros:

• No calculating at all
• Less dice to roll
• No negative results
• Faster

Cons:

• The odds work somewhat differently. The basic idea remains the same though, you get mostly results that are to be expected for that skill level.
• Although you can't get certain high results with a low skill as in Fudge, you can get all lower results with higher skills. It's unlikely, but still.
• .. less dice, but more different dice.

Data:

• Basic probabilities for each rank, you'll want to click "Graph" and "At Least" to make any sense of this data - http://anydice.com/program/18b3
  
• Each rank is equivalent to a +1 bonus.
• Interestingly opposed rolls on various levels are very similar to those using Fudge-dice mechanics - http://anydice.com/program/18af
  
• In fact, we can compare the rolls of this mechanic and Fudge directly with some success as long as we adjust away the negative results from Fudge results - http://anydice.com/program/18b1
  

The explosion rule described earlier in this topic could be implemented as well, and it would have beneficial effects, but at the cost of rerolling.