The Gaming Den

So, basically, in my experience, there's about two:

• d[x]+mods v target number
  • d20
  • d% systems
• Pool d[x]s+mods, look for hits against a TN (ST, AS, ORE)
  • Storyteller
  • After Sundown
  • One Role Engine (from a game called Godlike, you roll a pool of d10s, count the number of matching numbers, and the number of dice above a TN to determine whether you succeed and how well)

Um... that seems really broad, the fact there's only two makes me really wonder if I missed something.

But also, what models of character advancement are there? You have class-based (D&D and variations), Point Buy (GURPS, M&M, BESM, Godlike), and... point buy with set powers? (After Sundown, WoD).

Again, am I missing anything? I figure it'd be handy to have some kind of list of this shit, what with all our varied game design projects.

what about:
d[x]+mods vs d[x]+mods

seems like a GW favorite in the form of warhammer and HeroQuest style games.

where the opponent actually gets some kind of roll to block incoming damage as opposed to a set "armor", something like that. could be done with pools on both sides as well.

You missed the WEG d6 system, and the chaosium advancement system.

Edit: Oh, and FATE - style variable d resolutions, and the after sundown card advancement.

Your characterization of rolling dice and looking for a target number (like After Sundown) and rolling dice and checking for arbitrary features of die result relatedness (ORE) as being "the same", or even similar is batshit. Those aren't even remotely the same. That's like saying Champions (3d6, roll under) is the same as Craps (2d6, repeated, looking for matches). It's not.

Basically you got:

• Single Result, Flat RNG. In this system you roll a dX, and you apply modifiers to either your roll or the target number, and you try to roll over or below the target number. The dX is commonly a d20 or d%, but it could just as easily be a d10 or any other random number generator you wanted. The primary draw of such a system is that every +1 changes your odds by 1/X.
• Single Result, Curved RNG. In this system you roll two or more dXs and add the results togethter. Then you add modifiers to either your roll or the target number and try to roll above or below the (modified) target number. Common choices for this are 2d6, 3d6 or 2d10, but really the sky is the limit. Some games try to subtract numbers from the result (Fudge uses 4d3-8) or have one of the dice subtract from the other - but these sorts of shenanigans do not affect the odds at all. The primary draw of such a system is that bonuses mean more in absolute terms the closer to the center of the curve you get. The secondary draw is that the most extreme results are comparatively extremely rare, which can make crits/fumbles similarly uncommon.
• Variable Dicepool, Fixed TN. In this system, you roll a pile of dice and count how many land above a specific number. Usually these are d6s or d10s, but in principle, any die could be rolled. The primary draw is that average results are easy to calculate (dicepool * Chance per die), but variance rises as dicepools rise. This means that your chances of total failure never go to zero even with very high dicepools, meaning that threats never completely expire.
• Variable Dicepool, Variable TN. In this system, you roll a pile of dice and count how many land above a variable number. This is theoretically able to generate virtually any conceivable chance of success, but in reality it is crazy nightmare realm shenanigans. Back in 1989, people thought this was a good idea. It is not.
• Fixed Dicepool, Variable TN. In this system, you add modifiers to the number rolls on each die, and roll a series of dice. It's basically the same as the flat RNG, single result option, but it matters best two out of three, or whatever. This makes more likely results more likely and less likely results less likely. You'll note that is very similar to simply having a curved result on a fixed RNG - and it's almost close enough for me to label it a cosmetic change like shifting the average point or inverting one or more of the dice.
• Various Poker Bullshit. Lots of games roll dice and then check of fives or matches or straights or something. Yahtzee gaming systems exist, but are generally shitty enough that I can't even suggest an advantage they might have.

-Username17

There's also variants of 'die pool totaled against a TN' like Over the Edge, Roll and Keep, Weapons of the Gods, etc.

But there's a lot of weird shit out there: Everway compares static numbers and uses the MC's interpretations of a variant tarot deck to modify the outcomes. Dread uses Jenga.

Godlike's ORE is "roll pool, compare to TN to determine success, look fro matches to determine degree of success." So it is similar to ST et al, in that it's the same, plus an extra thing.

Edit: Ok, I mis-remembered, Godlike's ORE doesn't really have a TN, you just roll Stat+Skill d10, and look for the number of matches (width) and the highest number you rolled (height), written [width]x[height]. Width determines speed, height determines degree of success, and pools cap out at 10, because if you got 11, you'd automatically get a width of two. So if you've got 5 dex and 3 shooting, you roll 8d10, getting 1, 3, 4, 4, 6, 7, 9, 10; for a 2x4 width by height, which means that you did shitty, and it took a while. Powers can give auto-10s, called hard dice, and wildcards (pick number after rolling), called wiggle dice.

When you disparage Variable Dicepool, Variable TN, you neglect to mention that Single Die, Flat RNG is that when rolling more than one attack/save at a time.

And Single Result, Curved RNG can be made with dicepool-shaped dice (fixed dicepool, fixed TN). It does some interesting things to the way probabilities get almost exponentially more likely as your bonus increases, before becoming automatic.

Don't forget the weird non-dice shit:

• Maybe I'm Psychic? - The GM draws Zener Cards dependent on your skill and you have to read his mind to determine which ones he drew. Potentially rewards players who are actually psychic I guess? As seen in the psychic system for Conspiracy X 1st edition.
• Actual poker hands - Actually drawing cards from a deck and attempting to make poker hands to do stuff. Differs from dice poker because each card drawn affects the probability of the other draws, and is therefore even more of a clusterfuck to work out your chances of doing anything. As seen in Deadlands.
• Resource management - You have some amount of a resource and you can use it to do stuff either without a roll or to give bonuses to rolls. This allows the player to effectively decide which activities are important to them and can introduce a decision-based minigame if you're into that kind of thing.
• I'm bored, let's play another game! - Using one game as the resolution method for tasks within another game. Is an easy way to add a fully developed minigame into your system without all that pesky work. See Dread using Jenga for task resoution or D&D using Chainmail for resolving combats. LARPS often use this by making violent actions into rock, paper, scissors.
• Winds of Fate - Draw cards that provide or supplement your available options. Magic is the archetypal example, Frank has expounded on possible RPG applications in various threads around here.

The initial two categories are very broad, too broad to be useful.

A couple of obvious things you've missed here are

*Step Dice: you roll a dice with a variable number of sides depending on how awesome you are a la Savage Worlds, Cortex, or Earthdawn.

*dice pool systems where you roll a set of dice and take the single highest result, like the '20+' thing Hicks was doing (although that also has uses plusses) or the Heavy Gear system.

*many 80s games ran off table-based systems - James Bond or Marvel Super Heroes for instance. You roll a d100 and have to check a table to generate a result level.

*There are also roll-under systems where you're trying to roll under a stat with a varying number of dice (the opposite of Star Wars D6...) a la Summerland or The Fantasy Trip.

There's a thread on therpgsite where I tried to list all the variants I could think of of these sorts of things, among others.

http://www.therpgsite.com/showthread.php?t=21479

I have to apologize for the complete lack of organization, but people may find it interesting - I'd always intended to reorganize it before linking it to here but would require too much effort. Core Mechanics are from posts 11-16 (plus 24-26 when I thought of some more things).

Other methods include:

Plot Tokens: You start with X plot tokens. You can spend them to dictate the narrative, and the rules dictate how to gain/lose them. Descent uses something like this with the DM's "make a dungeon" points.

Roll and Add and Repeat:
1. You roll a die.
2. If this beats the TN, you succeed. Else, go to three.
3. Roll another die and add your previous result. Go to two.
This means that you basically can't fail at any task where this is the mechanic, but it might take you a shit-ton of rolls to do it.

Roll X, Keep Y: You roll a bunch of dice and then keep a smaller number of them based on an algorithm like "3 highest" or "2 lowest" or something like that. Then you add them together and compare to the TN. This is basically the same as the Single Result Curved RNG, except it gets more and more skewed toward high numbers the more dice you add.

Roll and Multiply: Roll some dice, multiply the results together and compare to a TN. 1d6 times 1d6 vs TN 10. Is wildly fucking swingy with and kind of modifier added to the rolls, but works somewhat reasonably if you add/subtract from the TN.

Physical Challenge: Arm Wrestle Mr Cavern. If you win, you get to say what happens. If you lose, Mr Cavern gets to say what happens. Alternatively, Rock-Paper-Scissors for it.

Grek wrote:Roll and Multiply: Roll some dice, multiply the results together and compare to a TN. 1d6 times 1d6 vs TN 10. Is wildly fucking swingy with and kind of modifier added to the rolls, but works somewhat reasonably if you add/subtract from the TN.

For anyone who plans to use this system, here's a comparison of the result distribution:



"Required TN" means the number you need to roll on the dice themselves to succeed. (e.g., if you have a +1 bonus and the DC is 11, the required TN is 10, and you have a 6/36 chance of passing on add, and 19/36 chance on multiply)

Foxwarrior wrote:When you disparage Variable Dicepool, Variable TN, you neglect to mention that Single Die, Flat RNG is that when rolling more than one attack/save at a time.

The problem with variable pool / variable TN is that people are too lazy / rushed / math-inept to calculate whether 5 dice vs TN 7 is better than 6 dice vs TN 8, so they don't and the system gets all stupid. Many of these people are game designers. I kind of see your point, because the poster child for variable-TN terribleness (oWoD) also severely undervalued multiple attacks, but a lot more games ask the GM to make on-the-spot judgments about variable TNs than encourage the GM to hand out discretionary extra actions.

Also it's important to remember that the information you're interested in calculating in multiple attacks and multiple saving throws in a D&D environment is actually much simpler than the information you're being asked to calculate in a variable TN, variable dicepool scenario. That is, the question for multiple saving throws is "What are my chances of passing every save?", which is relatively easy (chance of success on one, raised to the power of the number of attempts). And the question for multiple attacks is "What is my average damage output?", which is even easier (chance of success per attack, times the number of attacks).

For the multiple dice, variable TNs scenario, you're being asked to calculate something much shittier, which is "What are my chances of rolling more successes in my dicepool than that guy over there rolls on his?", and that is fucking terrible (in that it is a ghastly set of permutation expressions multiplied by each other and then added together). That is fucked. People don't do it. Gamers don't do it. MCs don't do it. The game designers don't do it. And then the rules get designed and played in an unexamined way and they produce stupid results constantly and no one knows whether they are looking at swans that are black or white.

-Username17

Then there is The Dark Eye's slightly weird

Roll and compensate:
1. Roll a set number of dice, perhaps each die against a different difficulty.
2. Spend skill points to boost failures into successes.
3. If you end up with negative skill points, you fail; otherwise the remaining skill points show the amount of your success.

crasskris wrote:Then there is The Dark Eye's slightly weird

Roll and compensate:
1. Roll a set number of dice, perhaps each die against a different difficulty.
2. Spend skill points to boost failures into successes.
3. If you end up with negative skill points, you fail; otherwise the remaining skill points show the amount of your success.

That is in no way different from d20. You roll the dice, you add your modifiers, you compare the result to the TN, and your margin of success or failure is the amount above or below the TN you get. I mean, it's Das Schwarze Auge, so they explain it in the most obtuse and confusing possible way. But it's not actually any different from the Single Result systems we'd been talking about earlier.

-Username17

ModelCitizen wrote:The problem with variable pool / variable TN is that people are too lazy / rushed / math-inept to calculate whether 5 dice vs TN 7 is better than 6 dice vs TN 8, so they don't and the system gets all stupid.

Given that you're using d10's and that you succeed with at least one success and that there are no exploding dice, those give roughly the same odds of success, right?

If you only need 1 success, then you succeed 92.2% with 5 dice v TN 7, versus 88.2% with 6 dice v TN 8. You're also more likely to get multiple successes with the 5 dice, which is a result that I would imagine strikes some people as unintuitive.

Wrathzog wrote:

ModelCitizen wrote:The problem with variable pool / variable TN is that people are too lazy / rushed / math-inept to calculate whether 5 dice vs TN 7 is better than 6 dice vs TN 8, so they don't and the system gets all stupid.

Given that you're using d10's and that you succeed with at least one success and that there are no exploding dice, those give roughly the same odds of success, right?

No.

5 Dice, TN 7 = 92.2% chance of 1 or more hits, 2.0 average hits, 1.02% chance of 5+ hits.
6 Dice, TN 8 = 88.2% chance of 1 or more hits, 1.8 average hits, 1.09% chance of 5+ hits.

And that is why the concept is shitty. The second is slightly but noticeably worse but has a very slightly but tangibly higher chance of getting a critical success. There just isn't any genre emulation that succeeds at. Also, it's mathematically intractable for most people.

-Username17

Ah for some reason I had the chance of success for TN 8 at 35%. I knew my maths was wrong.

Wrathzog wrote:

ModelCitizen wrote:The problem with variable pool / variable TN is that people are too lazy / rushed / math-inept to calculate whether 5 dice vs TN 7 is better than 6 dice vs TN 8, so they don't and the system gets all stupid.

Given that you're using d10's and that you succeed with at least one success and that there are no exploding dice, those give roughly the same odds of success, right?

5d10 TN 7 is better unless you need 5 or 6 successes.

This thread needs more graphs.

Here, have the d20 vs similar RNGs:




(In this context) the CDF is the probability of getting a given number or higher, and the PDF is the probability of getting exactly a given number.

This being a discrete RNG, the PDF is the change in the CDF between steps, but in a continuous RNG, the PDF is the derivative of the CDF.

Normally, CDF starts at zero, but I'm assuming roll-over, not roll-under.

...You Lost Me wrote:This thread needs more graphs.

Graphs that are relevant to the topic, or graphs in general?

Everything in that chart is necessary for it to look like Pac-man. Therefore, it is incorrectly labeled.