Dice and Probability

[BCOVertigo]

I am considering the implications of two changes:

1. Replacing d20s rolled with d10s (retooling all DC formulas to reflect this change)
2. Making all aggressive actions (Attacks, spells, skill checks, etc) dynamic (rolls) and all defenses (saves, AC) static (formula derived with no rolls made).

I would like to get the playground's input on what this might do to a 3.5/pathfinder game at low, mid and high(least interesting tbh) level play.


Points I would like to direct discussion towards:
The effects on power attack.

The value of feats like Weapon Focus and the possible implementation of scaling (maybe +1/2hd, max +3?)

The effects on spell caster's (reduced, but still random) DCs vs saves scaling at normal speed.

The effects of low bab characters attempting (ranged) touch attacks.

The relative power of effects that are not affected by this change (no rolls).


What have I not thought of?

[DementedFellow]

More chances of critting means more crits which can prove disastrous at lower levels.

I hate to be that guy, but why do you feel a change is necessary? Chances are unless your group is willing to accept some unforeseen problem, this rule change could prove more problematic than it is worth.

[Anecronwashere]

The lesser dice has several effects:
For one making hitting DCs less swingy and making crits (both 1s and 10s) happen more often. (you have a 10% chance of fumbling each roll as opposed to a 5%).
It also makes smaller bonuses more meaningful. A +1 bonus raised your chances by 5% in the old system, in this one it raises by 10%, or a 2x increase

[Thomasinx]

2. Making all aggressive actions (Attacks, spells, skill checks, etc) dynamic (rolls) and all defenses (saves, AC) static (formula derived with no rolls made).

The reason the current rules have the players roll saves, is because it feels like crap for the DM to simply say "You failed your save, you are dead", without even giving the player the opportunity to roll. If the player can't roll, they'll blame the DM, which is bad. (Right now, the players blame the dice.)



1. Replacing d20s rolled with d10s (retooling all DC formulas to reflect this change)

Do you not own any D20's? The d20 system works pretty well as it is, with the way nat-1's and nat-20's are set up at 5% probability. Doubling the chance of a failure, and doubling the chance a player gets crit just seems to be asking for problems.

Even though it also doubles the crits the player rolls, considering the number of dice the players roll, this just increases the likelihood of failure.



The effects on power attack.
None really. The only visible change is that people might just say 'screw it' and full-power attack more often hoping to hinge on the 1/10 chance of an automatic success. Although, this would be frustrating in the long run for the player.



The value of feats like Weapon Focus and the possible implementation of scaling (maybe +1/2hd, max +3?)
None really... You are basically taking a D20 system, and rolling a d10 for it and doubling the value to get the D20 value. These feats aren't affected by this.



The effects on spell caster's (reduced, but still random) DCs vs saves scaling at normal speed.
As I said before, this takes a player's survivability out of his own hands when it comes to rolling his saves. If you want to make the players roll a d20 for target monster's saves, that's one thing. But you should never take rolling the save away from a player.



The effects of low bab characters attempting (ranged) touch attacks.
Nil..



The relative power of effects that are not affected by this change (no rolls).
Honestly, the dice change only increases the likelihood of a screwed up situation outside the DM's control. The save thing also takes the responsibility away from players...

Why do you want to implement this? I honestly cannot see any gain to doing this.

Edit: Fixed a dropped parenthesis... I hate those things...

[BCOVertigo]

This is more a thought exercise right now than anything. By default I will not be using this, unless I feel an acceptable level of forethought has been given, and then a discussion with players will follow. Lastly it would probably be voted on.

To be frank I like the idea of less randomness in the game. I don't want a commoner to have a snowball's chance in hell of harming a knight in full plate armor with a tower shield, and I don't want a poor roll to so drastically ruin a rogue's ability to hide that he essentially trips and tumbles into view of the drunk guards. I'm looking for a tad more consistency in performance. I always thought it was ridiculous that even at relatively low levels a character was supposed to be pushing the bounds of what is humanly possible, but some jerk could still outperform them because they got lucky. Why should a commoner every be able to beat an ogre on a strength check? That is patently idiotic. I want that level of luck to go away starting at about level 3, and by 8 to be nonexistent.

I've considered the 3d6 bell curve, but it still just seems like a bandaid, and doesn't address the issue of skill being downplayed to rolls at a level of power that in real life, is never decided by luck. I can not out swim Micheal Phelps, regardless. I think having a system that mirrors that is a reasonable goal.

I may be barking up the wrong tree, it may be more sensible to search out another system, but if I could mod the one my friends and I all love to do what I want I'm more than willing to work out the kinks.

Thank you for the valid point though. Critical hits would definitely need to be worked on, as having a crit range 3-5 wide on a d10 is... well disastrous seems like a good word for it.

[DementedFellow]

Not for nothing but the Basic Roleplay (BRP) system is one of the easiest and more functional systems to pick up and play. It allows for a little more variation in battle, and I would argue that it cuts down on the randomness. You essentially would be rolling percentile die and if your rank in some skill is 37, if you roll below it, YAY you did it. If you roll above it, YUCK, you didn't do it.


I know it sounds random, and to a certain degree it is. But if you are going to change a d20 for a d10 you are actually increasing the randomness of the game, not decreasing it.

[Fish]

I don't know how you plan to replace d20 with d10, but simply doubling the result would give you some oddities. It would no longer be possible to roll any odd-numbered result. This would have the paradoxical effect of making a Ring of Protection +1 useless to any character with an odd armor class value (since any roll that would hit his current AC without the ring would also hit him with the ring).

[Thomasinx]

To be frank I like the idea of less randomness in the game.

Then you really dont want to do this with a D10. A d10 causes a bigger chance of randomness occurring than a D20.



I don't want a commoner to have a snowball's chance in hell of harming a knight in full plate armor with a tower shield, and I don't want a poor roll to so drastically ruin a rogue's ability to hide that he essentially trips and tumbles into view of the drunk guards.

The thing is, people like rooting for the underdogs. Players tend to enjoy it more when surprising things happen (such as the commoner getting lucky when the knight trips over a rock in the road, and stabbing the knight when he's distracted). It isn't really that fun when the big guy wins. Even more so when the players are the little guy.


Why should a commoner every be able to beat an ogre on a strength check? That is patently idiotic. I want that level of luck to go away starting at about level 3, and by 8 to be nonexistent.

Ability score vs Ability score checks are one of the few places where such a situation can occur. If you dont want that to happen, use a different roll. Arm-wrestling? Tell each side to make a grapple attempt instead of a straight strength-check. Tug of war? take turns with bull-rush attempts. There are options here...


I can not out swim Micheal Phelps, regardless. I think having a system that mirrors that is a reasonable goal.
Such a race shouldn't be based on a single dice roll. You'd make multiple swim checks, the result of each would determine how far you go. Unless he basically rolled straight ones, you would lose horribly. (And if he did roll straight ones, with such terrible luck, he probably had food poisoning).



I may be barking up the wrong tree, it may be more sensible to search out another system, but if I could mod the one my friends and I all love to do what I want I'm more than willing to work out the kinks.

You don't have to mod anything. Figure out how you want the rules to work within the system, and work with it. It isn't as random as you think.

[SiuiS]

I'm seeing a lot of confusing responses so I will just ask this now;

Will you replace 1d20 with a single d10, or with 2d10, keeping the scale from 1-20? I assumed you would use 2d10 and were asking how the bell curve would affect things.

If using two dice to create a bell curve, then you'll notice that until they are large enough, bonuses are better for weak characters than strong ones. The difference between a +1 an +2 is going to be bigger for a character than the difference between a +5 and +6. I think I'm mangling that, but it's one I the key points of GURPS.

How would you handle critical chances? Do both dice need to come up 1s to critfail? Do both dice need to come up 10 to crit?

If you use this, and have enemies roll against a static "save Defense" that the players have, you could just add 10 to their save bonus and that becomes their "will defense" or "reflex defense". Enemies would try to hit that like they do an AC, but it's a definite 4e vibe and I could see that being bland. Saving throws have a definite feeling of the player doing something. Takin that away will change how they are approached on a visceral level.

Using this system, you are way more likely to roll 11 than any other one number, so the odds of a critical failure an critical success become 1/100 instead of 1/20. This makes axes terribly bad and falchions an such terribly good - getting a critical on a natural 18, 19 or 20 isn't 3 numbers, 3/20, it's (and bear with me, it's 4:00 am) 7/100, which is a biggish boost. I mean it's only 6% more likely, instead of the 10% more likely on a d20, but you'll see a big difference in critical trends with the two dice. The disparity between a player with a rapier and a player with a long sword would be noticeable without keen twinkery.

Power attack woul probably become all-or-nothing, because unlike a d20 you can pretty much always bank on getting a 10. If you'll hit on a 10, drop a bunch of points into power attack. If you'll hit on a 13, you'd be much more careful.


I don't think this would above anything really. Base attack bonus, weapons, stats, circumstances, buffs, magic, all add together and become a pretty damn big modifier on their own. Your average rogue is swinging with a +8 at low levels, which makes the dice roll obsolete to a degree. Taking a lot of the chance out and nrmalizinf the roll around 10.5 wouldn't affect things in the long run. When modifiers to the die become bigger than the dice roll, the dice stop really mattering. That happens as early as level six.

[limejuicepowder]

This is just a minor response to your much larger question, but I completely agree with you on the strength check. More then any other ability check, making opposed strength checks always rankled with me.

A power lifter/defensive lineman type of individual with 18 str gets challenged by a totally average dude with 10 str in a strength contest. This shouldn't even be a contest, thus it shouldn't even be a roll.* A +4 bonus vs a +0 bonus on a d20 does not begin to express the difference between 18 str and 10 str. For those reasons, I usually forgo the roll and just say who wins.



*This would apply to a "strength only" contest where skill doesn't play much of a part, like arm wrestling, or lifting, etc.

[SiuiS]

This is just a minor response to your much larger question, but I completely agree with you on the strength check. More then any other ability check, making opposed strength checks always rankled with me.

A power lifter/defensive lineman type of individual with 18 str gets challenged by a totally average dude with 10 str in a strength contest. This shouldn't even be a contest, thus it shouldn't even be a roll.* A +4 bonus vs a +0 bonus on a d20 does not begin to express the difference between 18 str and 10 str. For those reasons, I usually forgo the roll and just say who wins.



*This would apply to a "strength only" contest where skill doesn't play much of a part, like arm wrestling, or lifting, etc.

But there are other factors. How defensive you are at lining mans doesn't affect your arm wrestling skill at all. As far as pure strength, an 18 defensive lineman, an 18 fat lazy dude and an 18 professional boxer are equal. The differences are elsewhere; the lineman has differing reflexes, feat choices, skill allocation, equipment bonuses and tricks. The lineman can auto-win by lifting the STR 10 guy off the ground, where the STR 10 guy cannot do the same to the lineman. Your example of why strength checks don't work isn't a strength check.

Skill ALWAYS matters. There are several tricks to winning arm wrestling, for example. Wrist angle, elbow, body use, distance and placement. My old mate IS a 200 pound heavyweight straight out of the US military, and straight out of a powerlifting magazine. I'm a magical unicorn. But I can still hold my own, and that's about a 10 to 18 disparity.

[Zaydos]

The way I'm reading this, you'd drop DCs and AC by 5 and use a d10 instead of a d20 roll for things. Okay, so crits need to be reworked (reduce x3 and x4 crits by 1 and 19-20 and 18-20 lose a number... still makes 9-10 a really high chance of critical). On to other things. You have a 20% chance of your number not even mattering, but otherwise you're going to get closer to the center of things just because there is a smaller range. Ranged touch attacks would be even easier or twice as hard, as in my experience you generally either hit on a 6+ (now a 2+) or only miss on a 1. Instead of needing a bonus +8 higher than the difficulty of the task to virtually ensure success it's now +4. This means a mage who focused on save DCs will be harder to resist, but easy saves will be easier to make. Melee that keeps their to-hit up will hit on their first two attacks with something like 90% accuracy (instead of 95, 80 which are the numbers that stick in my head from building a few high level melee characters) with a roughly 60% accuracy on their third (instead of 55) and almost never hit on their last (10% instead of 30%), which means fighters get hurt fairly bad. Unforeseen consequences are in the skill checks. You'd have to rework them more heavily than just -5 to DCs because they were built to have a full range of DCs.

If you're just doing this to decrease randomness, check out Unearthed Arcana's bell curve rules for using 3d6 instead of 1d20, they modify crits and the bell curve will actually increase the chance of getting an average roll instead of just changing the range.

As for the Str check, thing, the PHB says not to roll for arm-wrestling and lifting unless you have tied Str (meaning 17 always beats 16 :smallconfused:). That said both of those are in fact competitions where skill plays a part (when my form resembles proper I can lift almost twice as much as otherwise), but yeah if you're 3 times as strong (difference in carrying capacities) then +4 doesn't do it justice.

[KillianHawkeye]

The reason the current rules have the players roll saves, is because it feels like crap for the DM to simply say "You failed your save, you are dead", without even giving the player the opportunity to roll. If the player can't roll, they'll blame the DM, which is bad. (Right now, the players blame the dice.)

... ...

If you want to make the players roll a d20 for target monster's saves, that's one thing. But you should never take rolling the save away from a player.

Wow, you would REALLY hate to play 4th Edition then. Or Star Wars Saga Edition.

Both of which are really fun games BTW. :smallwink::smallamused:

[Sewercop]

The lesser dice has several effects:
For one making hitting DCs less swingy and making crits (both 1s and 10s) happen more often. (you have a 10% chance of fumbling each roll as opposed to a 5%).
It also makes smaller bonuses more meaningful. A +1 bonus raised your chances by 5% in the old system, in this one it raises by 10%, or a 2x increase


Id love for you to show me where the fumble rules are written down...

[Lonely Tylenol]

Id love for you to show me where the fumble rules are written down...

From the SRD (http://www.d20srd.org/srd/combat/combatStatistics.htm):


Automatic Misses and Hits

A natural 1 (the d20 comes up 1) on an attack roll is always a miss. A natural 20 (the d20 comes up 20) is always a hit. A natural 20 is also a threat—a possible critical hit.

I'm assuming "fumble" means "automatic miss". If it means anything else, I refuse to defend it, and disregard this post.

OP: If you want less randomness (in extremes), then I should have you know that going from a d20 to a d10 does the opposite; in fact, as has been demonstrated before, it increases the chance of automatic hits and automatic misses, as per the "critical" rules.

What you want is not a single smaller dice, but instead more dice. One variant that Unearthed Arcana suggests for skill checks (that can be more broadly applied elsewhere in your games, if you feel it's appropriate) is the 3d6 method, which replaces the 1d20 roll with 3d6. Right off the bat, this eliminates the possibility of a 1, 2, 19, or 20, as 3d6 can only result in a number between 3 or 18. Further, the numbers 3, 4, 5, 16, 17, and 18 are all relatively unlikely (a 1 in 216 chance of rolling a 3 or 8, a 3 in 216 chance of rolling a 4 or 17, and a 6 in 216 chance of rolling a 5 or 16--all of which are less than 5% by a statistically significant amount), which means that the top 5 and lowest 5 numbers traditionally found in a d20 roll are either less likely to happen as they would with a d20, or don't exist. Some degree of randomness does exist, but because of the way that a bell curve shapes, most rolls will tend toward 8, 9, 10, 11, 12 and 13, which should make up roughly 2/3 of all of your rolls.

Do note that broadening the bell curve model makes inherently tougher monsters significantly tougher, and runs contrary to at least one of your examples: using the bell curve model, a peasant will likely never strike a telling blow against a plate-armored knight before being struck down himself (because the system favors averages, and the peasant will never hit a plate-armored knight on an average attack). It will also reduce the odds that a rogue hiding from the guard will simply trip and stumble, or make some loud noise, however.

It's possible that this extreme doesn't even need to happen, however, at least for out-of-combat, as the Take 10 and Take 20 rules eliminate a great deal of randomness themselves.

From the SRD (http://www.d20srd.org/srd/skills/usingSkills.htm):


Checks Without Rolls

A skill check represents an attempt to accomplish some goal, usually while under some sort of time pressure or distraction. Sometimes, though, a character can use a skill under more favorable conditions and eliminate the luck factor.

Taking 10

When your character is not being threatened or distracted, you may choose to take 10. Instead of rolling 1d20 for the skill check, calculate your result as if you had rolled a 10. For many routine tasks, taking 10 makes them automatically successful. Distractions or threats (such as combat) make it impossible for a character to take 10. In most cases, taking 10 is purely a safety measure —you know (or expect) that an average roll will succeed but fear that a poor roll might fail, so you elect to settle for the average roll (a 10). Taking 10 is especially useful in situations where a particularly high roll wouldn’t help.

Taking 20

When you have plenty of time (generally 2 minutes for a skill that can normally be checked in 1 round, one full-round action, or one standard action), you are faced with no threats or distractions, and the skill being attempted carries no penalties for failure, you can take 20. In other words, eventually you will get a 20 on 1d20 if you roll enough times. Instead of rolling 1d20 for the skill check, just calculate your result as if you had rolled a 20.

Taking 20 means you are trying until you get it right, and it assumes that you fail many times before succeeding. Taking 20 takes twenty times as long as making a single check would take.

Since taking 20 assumes that the character will fail many times before succeeding, if you did attempt to take 20 on a skill that carries penalties for failure, your character would automatically incur those penalties before he or she could complete the task. Common “take 20” skills include Escape Artist, Open Lock, and Search.

Ability Checks and Caster Level Checks

The normal take 10 and take 20 rules apply for ability checks. Neither rule applies to caster level checks.

[Fish]

If you are looking for a system where probability is done on a bell curve, look into GURPS. Its probabilities are built from the ground up where all dice rolls use combinations of d6.

Or just borrow some key concepts. Opposed strength check (Quick Contest) would look like this:

Each player rolls 3d6 and tries to beat (ie, roll under) his score (ST, DX, a skill). If both succeed, the one who succeeds by the biggest margin is the winner. If both fail, the one who fails by the smallest margin wins. Example: a commoner with ST 10 arm-wrestles an adventurer with ST 15. The commoner rolls 3d* and gets an 8, so he beat his score by 2. The adventurer has to roll anything less than a 12 to win (very likely: 74.07%).

Or the adventurer goes first; he rolls a 9, beating his ST by 6. In order to TIE, the commoner must roll a 4 (1.9% chance).

That system uses a different critical system. A roll of 17 or 18 is always a failure, and 18 a critical failure. Failing by 10 or more is also always a failure. A 3 is always a critical hit.

If you didn't follow any of that math ... then I recommend against building your own dice mechanics.

*In GURPS, all dice rolls are on d6 so there is no need to identify the die. 3d6 in D&D = 3d in GURPS.

[Hyde]

Because some people seem confused on what "randomness" means, since the proposed changes do, in fact, decrease randomness. The "increase in randomness" that they are referencing comes from the concept "because the DM rolls more dice than the players, he is more likely to roll any particular number (or combination of numbers) than the PCs. Therefore, any increase in randomness generally benefits the DM and screws the players". The reason this is true is that it is generally expected for the monsters to die, and so DMs don't actually have any consequences for poor rolls.

Anyway, this system still sucks for players for the same reason an increase in randomness does- the DM rolls more dice. In fact, it could be said that "any change in randomness basically screws the players".


I am considering the implications of two changes:

1. Replacing d20s rolled with d10s (retooling all DC formulas to reflect this change)

A static value of ten added to all DCs is the average roll of a d20, rounded down. Since DnD employs a "meet or beat" system, there are eleven numbers that check this value, and nine that miss it- a 55% success chance before modifiers.

A static value of five added to all DCs is the average roll of a d10, rounded down. In the same system, this would result in six numbers checking the value, and four missing, for a 60% success chance.

So your system would see slightly increased successes on both sides.


2. Making all aggressive actions (Attacks, spells, skill checks, etc) dynamic (rolls) and all defenses (saves, AC) static (formula derived with no rolls made).

This is essentially the way 4th edition is played. Typically, it means rolling slightly more dice for the DM than the players. If we assume the only change becomes taking the static value of the save DC and adding it to the players' defenses, the recipient of the spell is 10% more screwed in a d20 system and 20% more screwed in a d10 replacement because of the meet-or-beat rule. If this is confusing as to why this is, I can explain it in greater detail but I would be happy if you could just take my word for it, as that explanation is entirely too long. The short version is that meeting a number adds that number to the list of possible values increasing the odds of success by 5% (10%) for a d20 (d10). So in taking the meet out of the hands of the spell target (5%/10% success) and then adding it to the spell slinger (5%/10%) you essentially double down. 4th ed accounts for this, 3.5/PF does not.




I would like to get the playground's input on what this might do to a 3.5/pathfinder game at low, mid and high(least interesting tbh) level play.

The effects would likely remain static throughout levels of play, other than that PCs would likely not see high levels under normal encounter tables (hard encounters would be harder, less difficult encounters would likely become trivialized).



Points I would like to direct discussion towards:
The effects on power attack.

The value of feats like Weapon Focus and the possible implementation of scaling (maybe +1/2hd, max +3?)

At all levels, your static values are going to become more significant- and likely a more significant portion of your attack roll. Again, the short of it is that any +1 adds a number to your list of possible successes. in a d20 system, this means 5%, in a d10 system, this means 10%. Armor values do the same thing. What will likely happen is that against targets with a high AC, your fighter will be largely unaffected by the change- requiring say, a six or better (75%) now requires a 1 (100%), whereas your rogue that previously hit only on a 16 or better on a d20 (25%) now only hits on an 11 (0%). Basically, it makes things already likely to happen more likely to happen, thereby making challenging encounters impossible, easier encounters trivial, and perfectly level encounters unchanged.



The effects on spell caster's (reduced, but still random) DCs vs saves scaling at normal speed.

The effects of low bab characters attempting (ranged) touch attacks.

As above.



The relative power of effects that are not affected by this change (no rolls).

If they're not affected, then they're not affected- if the power produces an effect similar to anything affected by the change, then those effects are affected in a way appropriate to that effect.




What have I not thought of?

Crits. Keeping the same ranges makes the chance of a crit go from 5-30% (20 vs keen 18-20) to 10%-60%. The probabilities can be altered as needed- at the very least, the keen property should be done away with.

While the system makes critical threats more likely, actual critical hits become more rare or less rare in accordance with the changes to attack rolls, above.


the entirety of the system revolves around the concept of rolling a d20- changing that fundamentally alters the system in every conceivable way, in this case making the game virtually unplayable after a few levels.

You could argue that all of these changes could be accounted for, but the scope of the necessary adjustments would be such that it would be faster to just build an entirely new system.


[Edit: I would like to note that if the randomness of the game is a concern, the bell curve system as described is probably the best fix, though 2 d10 would result in a less curvy curve, making 20s still possible while eliminating fumbles.]

[Thomasinx]

Because some people seem confused on what "randomness" means, since the proposed changes do, in fact, decrease randomness. The "increase in randomness" that they are referencing comes from the concept "because the DM rolls more dice than the players, he is more likely to roll any particular number (or combination of numbers) than the PCs. Therefore, any increase in randomness generally benefits the DM and screws the players". The reason this is true is that it is generally expected for the monsters to die, and so DMs don't actually have any consequences for poor rolls.


Actually no. The D10 really does increase randomness, by definition.

Random, from the free dictionary is:

Of or relating to an event in which all outcomes are equally likely, as in the testing of a blood sample for the presence of a substance.

While there are more possible dice outcomes, what we care about is the relationship between success vs failure. With a D10 instead of a D20, the probability of success and the probability of failure are more equally likely, due to the increase in number of automatic successes and failures. This therefore makes the results more random, and less dependent upon inherent skills/stats.

Take a D2. This would be even more random, since everything would boil down to a 50/50 chance. There would be less possible numbers to roll, but the outcomes closest to equal.

Sorry. Had to argue semantics :)

[Hyde]

Actually no. The D10 really does increase randomness, by definition.

Random, from the free dictionary is:


While there are more possible dice outcomes, what we care about is the relationship between success vs failure. With a D10 instead of a D20, the probability of success and the probability of failure are more equally likely, due to the increase in number of automatic successes and failures. This therefore makes the results more random, and less dependent upon inherent skills/stats.

Take a D2. This would be even more random, since everything would boil down to a 50/50 chance. There would be less possible numbers to roll, but the outcomes closest to equal.

Sorry. Had to argue semantics :)
You seem to be confused between the idea of randomness vs probability.

True randomness means that any given outcome is just as likely as any other. in this case, rolling a 1 on a d10 is as likely as rolling a 2, or a 4, etc.

In point of fact, rolling any single die, any single number is just as likely as any other, ideally, so dice are "truly random".

Randomness is a property of probability (when all probabilities are equal). It doesn't matter if there are fifty options or fifty thousand.




Since either way we're looking at dice, the definition of "randomness" as it applies to any specific roll is irrelevant, so we have to look at the outcome of the die roll. In using a d10 vs a d20, the outcomes that were more likely before are even more likely now (representing a much larger probability than before) as demonstrated above, and thusly are LESS RANDOM than before.

In other words, don't tell a statistician he doesn't know what randomness is, noob.

[Thomasinx]

Since either way we're looking at dice, the definition of "randomness" as it applies to any specific roll is irrelevant, so we have to look at the outcome of the die roll. In using a d10 vs a d20, the outcomes that were more likely before are even more likely now (representing a much larger probability than before) as demonstrated above, and thusly are LESS RANDOM than before.

In other words, don't tell a statistician he doesn't know what randomness is, noob.

Statistician or not, you are missing the point. There are only two outcomes: Success or failure. The die is only the vehicle used to reach this outcome. It doesn't matter if you roll a 15, a 16, or a 19, it only matters whether it is a 'success' or 'failure'.



True randomness means that any given outcome is just as likely as any other. in this case, rolling a 1 on a d10 is as likely as rolling a 2, or a 4, etc.

Going from a D20 to a D10 reduces the difference in probability between successes and failures (ie, makes any given outcome closer to a truly random choice). In other words: The D10 is more random.

Edit: I just reread what you said, and found where your misconception comes from...

the outcomes that were more likely before are even more likely now
This is wrong. With a smaller die-size, automatic failures with '1's and automatic successes with '10/20's happen significantly more often. These were originally occurring with only a 5% chance, and this change doubles this. This is what makes the switch 'more random'.

[BCOVertigo]

Thank you everyone for the responses! I was expecting maybe 3.

I am particularly interested in the 2d10 idea, as it seems like a reasonable compromise and would eliminate the possibility of critical failures, which I will not mourn the loss of.

As an aside, I apologize for the poor Phelps analogy, but I also dislike the player having to make excuses for poor performance so often. Under this system it is possible for me to beat him in a fair contest where he did not in fact have food poisoning. That would impose a penalty, whereas his dice just hate him.

And on the subject of a d10 being more or less random, it has an equal chance of randomness as there are no weighted sides (I hope). Each individual side may be more extreme in terms of jumps between percentile, but the spread among possible results will still come out looking the same on a graph. The difference is the randomness is easier to overcome because a +1 bonus is now worth 10% of 10 instead of 5% of 20.

As for AC, DCs, etc, I must have been unclear. The formulas would replace that static 10+OtherStuff for 5+Otherstuff. On this subject I especially appreciate Hyde's response on the effect on Medium BaB characters. Although I have to say I am not so sympathetic to anyone but the Bard/Rogue in core, it's not like the Cleric can't ignore this and the Druid isn't a bear anyway. And to be honest Low BaB = Strongest, most Versatile characters, so punching them in the gut in combat isn't really making me feel so bad. :smallannoyed:

I would like to say that I will probably not be using this immediately because my players wanted to run a gestalt campaign, but I would still like to continue discussing what potential tweaks to WBL, or other factors may make this more workable. I understand that WBL increase is a bag of cats (sometimes a bag of awesome cats) but with accuracy in combat now at a premium, I think it is reasonable to say that players might prioritize it more.

Shelving the arguments on Crits suck with this (agreed, tbd later) and varying dice options (although I like the 2d10 idea a lot), lets focus the hivemind on how to make this work, and identifying intrinsic problems.

As an aside, let's try to involve math in the statistics arguments, as it's not a matter of opinion and will show plainly who is correct.

[SiuiS]

Wow, you would REALLY hate to play 4th Edition then. Or Star Wars Saga Edition.

Both of which are really fun games BTW. :smallwink::smallamused:

Matter o opinion unfortunately. While there have been times it streamlined things and made more sense, there have also been times where poor dungeons mastering made me wish I had a saving throw option instead.

-

On using a single d10, I have more experience with what Thomasinx says, where things divide quite quickly along "pass" and "fail". This reduces the fun for me, personally, and would require being careful with some feats and such. I don't have anything to say other than "I don't think it's fun, I suggest you try a bell curve to eliminate randomness". But try it out. Who knows?