Quote from: Omega;1113687That was because intitially a d6 was all they had to work with, jackass.

That's wrong. Chits have always been available and were in use for some games.

Even without chits it's still wrong. Multiple six sided dice were available And various modifications of a simple 1d6 were already in use before D&D was published, e.g. 1d6+1, 1d6-1, or an averaging die. (Chainmail actually used an averaging die for morale for mercenaries, results: 2, 3, 3, 4, 4, 5).

Quote from: Omega;1113685This has pervaded board gaming as well slowly to the point now several designers and players are very against math in games at all. Apparetly 1+2+3 is just too complex for "modern" gamers. You have to dumb it down to the point a preschooler is mocking you.

You realize that a good chunk of the fanbase is drunk while playing?

Quote from: Bren;1113725"Fortunately" is not the first word that comes to mind for someone coming up with that system. You certainly did provide the requested example, though.

The key is that while calculus was behind the game, as a player or GM there was no need to understand the calculus. You could see the bell curve in the table, and I had been long familiar with bell curves from various things including seeing the 3d6 curve in the Dragon or something.

QuoteIf the problem is the missing number 19, you could roll 2D10 instead of 1D20 and reroll on doubles. Upen ended at the top with no holes in the middle.

It's not 19 that's missing, it's 20, but yes, you can fix it. On a 20, you can roll again and add to 19. But it's still a mostly long flat probability curve with some long flat lower steps at the extreme(s).

With 3d6 there's a nice open ended curve by taking an 18 and re-rolling the 3 dice, and each 6 adds one and you can pick that die up and roll again looking for more 6s (or the reverse on a 3 reroll 1s). It still will be bumpy but much more elegant than a d20 (or d100) rolll again and add mechanic. L5R's mechanic of re-rolling 10s and adding is also more reasonable, though with Xkx you will (almost) never produce exactly 10 * X (you can score a 30 on 3k3 by rolling a 10, a 6, a 5, and on the re-roll a 9, it's only Nk1 that you can never produce a 10), on the other hand that actually doesn't matter much since you only need the open ended roll to meet high TN, degree of success doesn't matter.

QuoteCreated by engineers. I'm shocked. Shocked I tell you.

:-) There are other things in the system that make it feel very much like something only an engineer or mathematician (or computer scientist - I got my family friend who became a professor at my college hooked) could appreciate.

Quote from: ffilz;1113764You could see the bell curve in the table, and I had been long familiar with bell curves from various things including seeing the 3d6 curve in the Dragon or something.

Yes, bell curves do have different properties compared to the linear distribution of a d20 or d100 which make them nice for modeling natural processes. I know that's one reason some people prefer GURPS since you get results closer to a normal distribution.

QuoteIt's not 19 that's missing, it's 20

Doh! 20 is what I meant, but not what I said.

QuoteOn a 20, you can roll again and add to 19. But it's still a mostly long flat probability curve with some long flat lower steps at the extreme(s).

Rolling dice gives integer results so any choice will give you a step function. What changes are how far apart the steps are and whether the steps are of uniform size or not. D20 gives you 20 steps, d100 gives you 100 steps and each step is the same size (i.e. equally likely). 3D6 only gives you 16 steps and each step is not the same size (i.e. not equally likely). The graph of the results is symmetric about the mean and probabilities increase from 3 to 10 and decrease from 11 to 18 with P(10) = P(11) and more generally P(N) = P(21-N) where N is a natural number between 3 and 10 inclusive.

It seems that your friend wanted to keep the same sort of symmetric step structure with probabilities increasing towards and decreasing away from the mean as 3D6, but with open tails on either end. And he was willing to put up with up with both a table lookup table and the possibility of sequential dice rolls to get that result. I have two issues with the method.

I'm not especially fond of table look up tables. There's three reasons for that. The first is clutter. As the GM a lookup table is one more bit of paper for me to keep track of or place on a GM screen. Ditto if the players are going to have to access the table. The second is related to clutter. Needing to use a lookup table increases the handling time for resolving an action. The third is aesthetic. I prefer simple formulas and symmetry. (I note that your friend's system does yield symmetry and it is formulaic, but the formula isn't what I would describe as simple.)

• Simple attack tables are OK. Attack tables used to be common in board game versions of wargames. They needed to be small tables printed on a single card and the game usually didn't require the players to have many (or any) other bits of paper to keep track of or shuffle through. Everything else was out on the map board.
• Tables in OD&D are OK. There were only two look up tables (To-Hit and Saving Throw). The former was a simple linear formula once you knew what it took for a character to hit AC 2 (or AC 9) so you needed a single lookup at the start of a combat and the latter only applied to spells, poison, and special monster attacks so it wasn't needed for most of the rolls required in a combat round and sometimes it wasn't needed at all.
• Lookup tables are no problem in Runequest 1-3. Since the Resistance table was based on a simple linear formula that compared Attack Value vs Defense Value I never needed a look up table. Probability of success was 50% if AV = DV, with a 5% increase/decrease for every point of difference. So 17 vs 15 was a 60% chance of success for the 17 and a 40% chance for the 15. Similarly P(critical hit) and P(special hit) were based on simple linear formulas so again no need to look things up if you can do the arithmetic in your head.

Sequential dice rolling increases handling time. For me this is a pretty big issue.

• For known die roll sequences, this can be mitigated if one can get players to roll multiple differently marked dice at the same time e.g. in D&D one could roll the attack and damage rolls at the same time similarly in Runequest one could roll the attack roll, the hit location roll, the damage roll, and even the parry roll all at the same time. In practice I've had minimal success getting players to do this.

• Part of that is psychological, in part based on our common understanding of cause and effect. Unless I hit, I won't do any damage, so why would I roll damage before knowing whether or not I hit.
• Part of it is players don't always have enough differently colored dice to do this.
• Part of it is for that to work the player has to leave the dice on the table while the various outcomes are figured. Players often seem to want to pick up the dice shortly after rolling them.

• And in open ended systems, you don't know if you need to roll another die (or how many other dice you will need to roll) until you know what the outcome is on a previous die or dice roll, e.g. rerolling a D6 on sixes only occurs if there is a six. And this could occur over and over. It's not very amenable to rolling multiple contingent dice at once.

Quote:-) There are other things in the system that make it feel very much like something only an engineer or mathematician (or computer scientist - I got my family friend who became a professor at my college hooked) could appreciate.

Care to share?

Quote from: Bren;1113778Yes, bell curves do have different properties compared to the linear distribution of a d20 or d100 which make them nice for modeling natural processes. I know that's one reason some people prefer GURPS since you get results closer to a normal distribution.

Doh! 20 is what I meant, but not what I said.

Rolling dice gives integer results so any choice will give you a step function. What changes are how far apart the steps are and whether the steps are of uniform size or not. D20 gives you 20 steps, d100 gives you 100 steps and each step is the same size (i.e. equally likely). 3D6 only gives you 16 steps and each step is not the same size (i.e. not equally likely). The graph of the results is symmetric about the mean and probabilities increase from 3 to 10 and decrease from 11 to 18 with P(10) = P(11) and more generally P(N) = P(21-N) where N is a natural number between 3 and 10 inclusive.

It seems that your friend wanted to keep the same sort of symmetric step structure with probabilities increasing towards and decreasing away from the mean as 3D6, but with open tails on either end. And he was willing to put up with up with both a table lookup table and the possibility of sequential dice rolls to get that result. I have two issues with the method.

I'm not especially fond of table look up tables. There's three reasons for that. The first is clutter. As the GM a lookup table is one more bit of paper for me to keep track of or place on a GM screen. Ditto if the players are going to have to access the table. The second is related to clutter. Needing to use a lookup table increases the handling time for resolving an action. The third is aesthetic. I prefer simple formulas and symmetry. (I note that your friend's system does yield symmetry and it is formulaic, but the formula isn't what I would describe as simple.)

• Simple attack tables are OK. Attack tables used to be common in board game versions of wargames. They needed to be small tables printed on a single card and the game usually didn't require the players to have many (or any) other bits of paper to keep track of or shuffle through. Everything else was out on the map board.
• Tables in OD&D are OK. There were only two look up tables (To-Hit and Saving Throw). The former was a simple linear formula once you knew what it took for a character to hit AC 2 (or AC 9) so you needed a single lookup at the start of a combat and the latter only applied to spells, poison, and special monster attacks so it wasn't needed for most of the rolls required in a combat round and sometimes it wasn't needed at all.
• Lookup tables are no problem in Runequest 1-3. Since the Resistance table was based on a simple linear formula that compared Attack Value vs Defense Value I never needed a look up table. Probability of success was 50% if AV = DV, with a 5% increase/decrease for every point of difference. So 17 vs 15 was a 60% chance of success for the 17 and a 40% chance for the 15. Similarly P(critical hit) and P(special hit) were based on simple linear formulas so again no need to look things up if you can do the arithmetic in your head.

Sequential dice rolling increases handling time. For me this is a pretty big issue.

• For known die roll sequences, this can be mitigated if one can get players to roll multiple differently marked dice at the same time e.g. in D&D one could roll the attack and damage rolls at the same time similarly in Runequest one could roll the attack roll, the hit location roll, the damage roll, and even the parry roll all at the same time. In practice I've had minimal success getting players to do this.

• Part of that is psychological, in part based on our common understanding of cause and effect. Unless I hit, I won't do any damage, so why would I roll damage before knowing whether or not I hit.
• Part of it is players don't always have enough differently colored dice to do this.
• Part of it is for that to work the player has to leave the dice on the table while the various outcomes are figured. Players often seem to want to pick up the dice shortly after rolling them.

• And in open ended systems, you don't know if you need to roll another die (or how many other dice you will need to roll) until you know what the outcome is on a previous die or dice roll, e.g. rerolling a D6 on sixes only occurs if there is a six. And this could occur over and over. It's not very amenable to rolling multiple contingent dice at once.

Care to share?

Well thought out and worded response.

You're right, any gaming probability curve is stepped, that's why I pointed out the d20 roll of 20 roll again and add has wide steps whereas even a 3d6 probability shown with a bar chart looks like a curve.

On the tables - I have a single page cheat sheet for Cold Iron, so the clutter issue isn't too bad. Yes, it does mean players need two sheets, but that single sheet has a lot of stuff on it. I even have a wallet card cheat sheet with the absolute minimum required tables on it (in 6 point type - doesn't work for my AARP qualifying self as well as it did in college). The table lookup handling time isn't too bad for a couple reasons:

1. Most of the time, you need more than a +0 (0.50 or better die roll) to succeed.
2. Given that, a large enough percentage of the results fall in a range of 7 or so steps on the table, which is possible to memorize (I have in the past spit them out from memory having not looked at a table in months to years, and not really played for many years so over a long period of play they can be lodged in memory pretty good).
3. The extra dice isn't too bad, it only comes up on 20% of rolls (if you roll a leading 0 or leading 9), and the extra time there isn't too detrimental because it's likely to lead to excitement (so no worse than any other open ended die roll system).

Also on the handling time, since an attack is OCV + Chance Adjustment vs DCV, there is no separate parry roll. Damage does have to be rolled after determining margin of success, though rolling base damage with the attack roll would not be an issue, you would just add additional damage rolls for multiple damage if the margin of success was high enough. A big crit can have some handling time from rolling and counting a big handful of dice, but I have never seen anyone frustrated by that (and I can roll and add pretty quick so NPCs scoring crits don't slow things down too bad).

Back to the clutter: There's another page of charts needed for character creation and advancement and a page of weapon references. As a GM, I found the clutter not really a problem. My favorite gaming setup was the living room in one house I lived in during grad school. There was an arm chair with very wide (9" or so) nice flat arms. My GM scratch sheet and reference charts would easily rest on the arms. Other materials I had at hand was a three ring binder with the rules and my monster writeups. Then there would be some other reference stuff for adventures and the Blackmoor map (originally from Judges Guild, later one of the color ones from the TSR DA series modules). On the table (a board I had laid on milk crates) was a battle mat, counters, and other stuff to run battles. At the time, I had a nice little wood tray that had originally held cigars that kept my dice handy and made a nice surface for rolling. The players would have their character sheet and the reference charts with booklets for rules, spells, and price charts (with several pages of magic item prices - most Cold Iron GMs had magic shops, most magic items are single use or a few charges).

As to more "engineeringness" of the system: There were a lot of spells that came at different levels (Fireball III does d4s, Fireball IV does d6s, Fireball V does d8s for example). The way magic items are handled doesn't feel very mystical compared to say D&D (but I have in the past thrown D&D items into the game for more wondrous items). There's a chart for using the square cube law to derive a creature's strength from it's size (reflected simply as mass).

A lot if it probably stems from a bunch of engineers running and playing the game. The same mechanics in the hands of someone else might feel different.

Quote from: Omega;1113685This has pervaded board gaming as well slowly to the point now several designers and players are very against math in games at all. Apparetly 1+2+3 is just too complex for "modern" gamers. You have to dumb it down to the point a preschooler is mocking you.

You're clearly an ableist and a bigot!  

Quote from: ffilz;1113785A lot if it probably stems from a bunch of engineers running and playing the game. The same mechanics in the hands of someone else might feel different.

I'm an engineer and played D&D in grad school with a bunch of dudes, three of which were engineers. There is no way we would have used such a system...pretty sure you mean "math nerds".

Quote from: Omega;1113687That was because intitially a d6 was all they had to work with, jackass.

I guess D6 + 1 was too complex, D6 + 2 inelegant? 2D6 just too much damage?
I think the statement that all weapons doing D6 was the worst thing ever was over the top but an  unmodified D6 was not their only option.

On a 20, rolling again and adding does leave out the possibility of an exact 20 but "roll again and add the amount of the new roll that is over ten" leaves the possibility of exactly 20.

Quote from: David Johansen;1113790You're clearly an ableist and a bigot!  

Math, the most insidious gatekeeping tool of all time.

Quote from: WillInNewHaven;1113794I guess D6 + 1 was too complex, D6 + 2 inelegant? 2D6 just too much damage?
I think the statement that all weapons doing D6 was the worst thing ever was over the top but an  unmodified D6 was not their only option.

It's over the top because it's retarded; there's nothing wrong with basically saying any weapon can kill a normal person in battle. Agree with other options, though. You can EASILY make a results table that only uses d6, a la Star Fleet Battles.

Quote from: WillInNewHaven;1113795On a 20, rolling again and adding does leave out the possibility of an exact 20 but "roll again and add the amount of the new roll that is over ten" leaves the possibility of exactly 20.

But that requires subtraction which is a more difficult operation than addition. This would bar many kindergartners and some first graders from calculating their own to-hit rolls. No game designer wants to lose that market segment.

Quote from: Antiquation!;1113796Math, the most insidious gatekeeping tool of all time.

A couple of kids who didn't finish high school learn how to figure (D4 + N) squared really quickly when it is how you calculate Lightning Bolt damage and their characters are casting the spell. Not only that, but "N" is a power of three mana and they figure that out quickly too. One of the other players, he finished high-school but is not a math student, figured out that N to the 0 power is one so he cast (D4) squared lightning bolts for one point each when short of power.

Quote from: ffilz;1113785Well thought out and worded response.

Thanks. It's occasionally fun to discuss die rolling at a somewhat deeper level than, "Use the blue D20, it usually rolls high."

Quote1. Most of the time, you need more than a +0 (0.50 or better die roll) to succeed.

I'd say that's both good and bad. It makes lookup easier, which is good.

However, the Cold Iron (CI) mechanics are designed to have a two-sided opened ended, symmetric range centered on the mean, which is something that roll again systems don't do. (They are only open ended on the high side and hence they aren't completely symmetric either.) You aren't using half of the CI range which means open ended at the bottom is unnecessary. And unnecessary mechanics are inelegant, and if the mechanics also add complexity they are bad.

Quote2. Given that, a large enough percentage of the results fall in a range of 7 or so steps on the table, which is possible to memorize

I wanted to ask a clarifying question or two about the table and the die rolling method.

What is the number in the table (which ranges from -25 to +25) used for?

As I understand it, in CI a player starts out rolling to 2D10 (which is read as a decimal) and this 2-digit decimal place number is then compared to the table to figure out the bonus or minus.

As I read the table a roll of 45-55 = +0, 56-61 = +1, 62-66 = +2, 68-72 = +3, 73-76 = +4, 77-81 = +5, 82-84 = +6, 85-87 = +7, 88-89 = +8.

For rolls of 90+ additional dice must be rolled to determine one or more additional decimal places.
So if a 9 is rolled for the 1st D10, an additional D10 has to be rolled to figure out what the digit is in the third decimal. If a 9 is also rolled for the 2nd D10, a 4th D10 must be rolled to find the 4th decimal place, and so on.

Did I get that correct?

If so, there's another issue. The probabilities that I calculate appear to be symmetric, but they don't follow the strictly increasing towards the mean from the left and strictly decreasing from the mean towards the right that we see in a normal distribution. Here's the probabilities I calculated for the two digit outcomes that don't require a third (or more) digit.

-7   3%
-6   3%
-5   5%
-4   4%
-3   6%
-2   5%
-1   6%
0   12%
1   6%
2   5%
3   6%
4   4%
5   5%
6   3%
7   3%

Is this correct or am I missing something?

Quote from: WillInNewHaven;1113799A couple of kids who didn't finish high school learn how to figure (D4 + N) squared really quickly when it is how you calculate Lightning Bolt damage and their characters are casting the spell. Not only that, but "N" is a power of three mana and they figure that out quickly too. One of the other players, he finished high-school but is not a math student, figured out that N to the 0 power is one so he cast (D4) squared lightning bolts for one point each when short of power.

So if I cast 1 mana, N=0. If I cast 3 mana N=1. If I cast 9 mana, N =2. Is that correct?

Then you roll 1D4 add N and square the result. So if the die roll was 2, damage would be 4 pts, 9 pts, 16 pts when casting 1, 3, and 9 mana points, respectively, right? While less nerdy than the table in construction, that's more nerdy in practice.

What happens if you cast mana that is NOT exactly equal to a power of 3?