The Core Mechanic

I think I am close to settling on a core mechanic that I like. And by that I mean one that I often find myself using in my nugget to think about game problems or odds ("How hard should it be to scale that cliff, or to smack that orc and take his pie?"). With that, I present:

SEPTIMUS CORE MECHANIC

Roll 1d6. If you roll equal to or above the TN, you succeed. The TN, by default, is 5. Only roll when there are consequences for failure, when the action is contested, or when dramatic.

COMPLICATIONS

- DIFFICULTY OF TASKS:
-- HARD TASKS: If the task is harder than usual (or the character is lower level than the challenge, or circumstances are poor, or whatever), you increase the TN (6, then 7, then 8, etc).
-- EASY TASKS: If the task is easier than usual, you decrease the TN (5, 4, then 3, etc).

- PROFICIENCY OF ATTEMPT: Roll multiple dice if circumstances favor the player or if the character is particularly proficient. Retain only the highest die. If you get multiple 6s, each 6 beyond the first counts as +1 to the total. So, [5, 5, 6, 6] = 7.

- THE SPELL MATRIX: This mechanic is used only for spell casting.
-- The purpose is to give spells their own unique subsystem which makes them feel different. Also, it intentionally obfuscates the odds a bit, making magic more dangerous and mysterious.
-- Rolls may be carried over from round to round, thus a dice pool gradually is built. The player may stop rolling at any time and judge success or failure.
-- Determine three derived characteristics of the dice pool.
LENGTH: Equal to the highest die. Multiple 6s count as +1 to the total (same as above).
HEIGHT: Any number showing on any die in the pool.
WIDTH: The number of dice showing with the number selected as the height.
SAMPLE POOL: The pool 1, 1, 6, 6, 1, 3 has LENGTH 7 (6+1), HEIGHT 1 (or 6, or 3), and WIDTH 3 (there are three 1's showing)

BEHIND THE CURTAIN: Odds and Reasoning

Here's a table showing the odds of getting at least one success above a given TN as a function of the number of dice in your pool, with the average result thrown in just for good measure.

	AVG	TN 3	TN 4	TN 5	TN 6	TN 7	TN 8
1d	3.500	66.68%	50.01%	33.34%	16.67%	0.00%	0.00%
2d	4.500	88.89%	75.00%	55.56%	30.56%	2.78%	0.00%
3d	5.037	96.29%	87.49%	70.36%	42.12%	7.40%	0.46%
4d	5.394	98.76%	93.74%	80.24%	51.77%	13.19%	1.62%
5d	5.666	99.59%	96.88%	86.84%	59.82%	19.63%	3.55%
6d	5.895	99.86%	98.43%	91.21%	66.50%	26.31%	6.22%
7d	6.100	99.95%	99.21%	94.14%	72.08%	33.01%	9.57%
8d	6.290	99.98%	99.60%	96.09%	76.74%	39.53%	13.48%
9d	6.472	100.00%	99.81%	97.40%	80.62%	45.73%	17.82%
10d	6.648	100.00%	99.90%	98.26%	83.84%	51.54%	22.47%

The color coding is as follows:
- Dark Green = Sure thing (95% -- that is, you only "fail on a 1" in D20 terms)
- Bright Green = Excellent (>80%)
- Pale Green = Good (>66%)
- Yellow = Fair (between 34% and 65%)
- Orange = Poor (33% or less)
- Red = Hope for a miracle (<5% -- that is, you only succeed on a 20)

WHY TN 5?

Note that using a TN of 5 keeps the game in a statistically pleasing "sweet spot," assuming that an expert character regularly rolls no more than 3 dice or so. TN 5 allows the widest range of results, from poor all the way on up to a sure thing.

One design philosophy is to make challenges within 1 tier of the characters very feasible. This is doable by manipulating the TN. Say some Tier 2 characters come across a Tier 3 monster (such as a powerful giant or some sort of extraplanar being). The TN for this fight will be 6. Even the unskilled folks rolling 1d6 can hit it, but its a long shot. The experts rolling 3d6 even have fair odds, and when using some sort of limited resource to get up into the larger dice pools occasionally they'll even have good odds. So, it'll be a tough fight, but its possible.

The same Tier 2 folks come across some Tier 1 monsters, say, some goblins. The TN is 4 -- easy! The unskilled folks have fair odds, and the experts with 3d6 still do not have a "sure thing" (that is, they miss on more than a "20"). It will be an easier encounter to be sure, but still one worth rolling out.

If our band of adventurers then comes across a Tier 4 challenge like a Balrog, the TN is 7. Things get ugly fast. The unskilled are basically helpless. They cannot hope to best it. The experts may land the occasional hit, but it will be quite rare. The group MAY be able to deal with one of these creatures (at least well enough to effect a stalemate or fighting withdrawal) but it will take lots of consumable resources.

A Tier 5 challenge would be overwhelming (TN 8). Likewise, a Tier 0 challenge (say, normal rats) is very easy. A tier -1 challenge (say, normal bugs) is absolutely trivial and should not be used either.

DIFFICULTY OF TASKS

Note that it is fairly trivial to judge the impacts of modifiers. Each modifier is about a 15% swing in the chances, or equal to a +/- 3 on a D20. Once you get above 6, you get a rapidly decreasing chance of success, and indeed, it is only possible for highly proficient users to succeed at this point. Hence, we get this table:

TN 9 = THEORETICAL USE ONLY: THE UPPER LIMITS OF POSSIBILITY
TN 8 = NOT ADVISED: NEAR IMPOSSIBLE
TN 7 = USE WITH CAUTION: LEGENDARY - APPROPRIATE CHALLENGE FOR TIER +2
TN 6 = CHALLENGING - APPROPRIATE CHALLENGE FOR TIER +1
TN 5 = BASELINE - APPROPRIATE CHALLENGE FOR TIER
TN 4 = STRAIGHTFORWARD - APPROPRIATE CHALLENGE FOR TIER -1
TN 3 = USE WITH CAUTION: EASY - APPROPRIATE CHALLENGE FOR TIER -2
TN 2 = NOT ADVISED: TRIVIAL
TN 1 = THEORETICAL USE ONLY: AUTO-SUCCESS

PROFICIENCY & CIRCUMSTANCES

The dice pool system I propose also allows for relatively easy gauging of the odds. +1d6 = about +1 (+16%). +2d6 = another +0.5. +3d6 = another +0.3. Diminishing returns continue as you might guess by the pattern. You effectively top out at a 3 die pool (which has an average of about 5.5), or maybe a 4 die pool, with more than that having small impacts, except in the case of (A) TNs above 6 and (B) the Spell Matrix.

GRANULARITY

A "traditional" D20 system has about 10 levels of granularity. That is, if players hit a typical monster on a 10 or 11 or so, then at most you can give a +10 modifier before die rolls become irrelevant.

This system has similar mathematical granularity but doesn't feel like it. For example, you have 5 usable levels of difficulty (TNs 3 through 7), and dice pools of 1, 2, or 3 dice. That's about 8 levels of granularity for a task (additive), many more when you start combining them. But, because the D6 has chunky results, it feels streamlined.

This is intentional, and allows psychological focus to be put on other elements of the game.

QUESTIONS/TBD

One issue "TBD" is exactly the breakdown between proficiency and challenge level. Specifically, I am thinking that each broad type of character get a significant bonus to their "prime" ability. So, strong types get a bonus to hit, smart characters a bonus when casting spells, etc. Should this be reflected by (A) giving a bonus die or (B) treating the character as one tier higher for setting TNs?

So, should a strong fighter roll 2d6 vs. TN 5 for a tier-appropriate threat, or should they roll 1d6 vs. TN 4? The odds of success are similar but the implications quite different.

Second, circumstance modifiers are easy (add dice to the pool). But how to deal with penalties?
- Once you are down to a lone die, start increasing the TN. So, say a character facing a level appropriate threat is at a significant disadvantage. They have no dice to give up. The TN moves from 5 to 6. The bad thing about this is that you can only go one step before failure is certain.
-0: 1d6 vs TN 5= 33%
-1: 1d6 vs. TN 6 = 16%
-2: 1d6 vs. TN 7 = 0%

- Use a negative dice pool. That is, roll a dice pool and retain the LOWEST result. My quick math tells me the following odds of success:
-0: 1d6 vs TN 5 = 33%
-1: 2d6 vs TN 5 = 11%
-2: 3d6 vs TN 5 = 3.5%
-3: 4d6 vs TN 5 = 1.1%

I like this because you can effectively go to -2, whereas the previous system only lets you go to -1. With a -2 situation, you basically succeed only on a 20 (a little less often actually). A -1 situation is bad but not impossible. I still don't love it, but it works.

All this points to another axiom: Septimus should be quick to reward players with bonuses but penalties should be rare. Because of the nature of the dice pool system, large numbers of bonus dice can be absorbed without breaking anything. But only one or two levels of penalties can be given before success is basically impossible.