Tag Archives: Hexagram

Fight off, dodge, or block

Some Dark Souls dude

Some Dark Souls dude

The combat system for my Stonehell Dark Souls game has drifted steadily away from traditional monster attack rolls toward monological combat (in short, players always roll, similar to Apocalypse World and Numenera).

Brief recap. In the initial December formulation, players chose between blocking or dodging (resolved using ability checks but also risking running out of stamina) or having the ref make a traditional monster attack roll versus character armor class. The trade-off was between relying entirely on ability scores or pitting character armor class against a monster’s potentially high attack bonus.

Making rules stick. In the past, I have sometimes had trouble getting combat house rules to stick because it is so easy to fall back on a familiar procedure, even when new rules result in more engaging outcomes and are advantageous to players. However, from the start of experimenting with the Dark Souls inspired active defense options, and across several groups of players with varying levels of tabletop roleplaying game experience, the active defenses seemed to remain top of mind. During the most recent session, players only ever defended actively, never letting the monsters make attack rolls, even with the risk of running out of stamina, which is punishing. I draw several lessons from this experience.

Choice prompts. First, the explicit choice prompt is an effective and low-maintenance way of communicating formal rules without needing non-referee players to read any rules (“zero homework” requirement). This is huge. Making such prompts habitual . This does place some constraints on potential rules, since the procedure must be fluent enough to survive being deployed all the time. That opposes complexity bloat which is positive more often than not.

Active options. Second, active options, assuming equal player effort requirements, have an advantage over passive options (such as submitting to a monster attack roll). Risking overgeneralization, I suspect this is universally true because players prefer a sense of control keeping all other factors constant.

Proposal. What follows is the procedure I am now considering, with parts that have not been play-tested in bold. Previously, armor class was a traditional passive defense score, but the approach below requires damage-reducing armor.

Resolve Monster Actions

  1. Determine actions for each monster.
  2. Match groups of monsters with defenders.
  3. Resolve defenses.


When monsters attack, to the defending player ask:

“Do you fight off the attack, dodge, block with a shield, or react in some other way?”

Resolve as specified below or by using the most relevant ability check.

Fight off. To defend using a melee weapon, roll the weapon’s damage and add the result to armor rating this turn, then suffer monster damage. In effect, this defends by comparing damage potential between player character and monster.

Shield block. To defend using a shield, make STR check (success → suffer no damage, failure → suffer ½ damage) and CON check (failure → out of stamina).

Dodge. To avoid monster attacks, make DEX check (success → suffer no damage, failure → suffer monster damage) and CON check (failure → out of stamina).

Suffer Damage

Add the damage from all monsters threatening a player character together, subtract armor rating from the total, and then suffer this amount of damage.


  • The fighting off procedure uses one roll per adventurer no matter the number of monsters. This means that the fight off option is more easily overwhelmed by multiple monsters, since the player character damage roll opposes multiple monster damage rolls added together.
  • Not needing to make tons of attack rolls for a host of enemies is a nice added bonus.
  • To simplify presentation for this post, I left out one small step, where player characters can intercede to block for an ally assuming the positioning makes sense fictionally. This can happen during the matching of monsters with defenders (step 2).
Playtest results

Playtest results

Let It Ride or Push Your Luck

Following is a designer note from the current working draft of Hexagram, the ruleset I have been working on.

The game Burning Wheel has a principle called Let It Ride:

A player shall test once against an obstacle and shall not roll again until conditions legitimately and drastically change. Neither GM nor player can call for a retest unless those conditions change. Successes from the initial roll count for all applicable situations in play (Burning Wheel Gold, page 32).

This means that once the players agree upon a particular test to resolve an uncertain outcome, the result of that one test fully determines the outcome. For example, a player may roll to determine if a character is able to open a lock. According to the Let It Ride principle, the player gets only one try to accomplish this goal using this means. Spending more fictional time for another attempt is not possible. Players must consider other means to get past the lock, such as smashing it with a hammer that may come with additional unintended consequences.

Hexagram play is based on a different game design principle: Push Your Luck. In Push Your Luck play, the number of attempts is not limited but risk attends each try. Additional tries tempt fate. In Hexagram, making Moves requires taking a Turn and taking a Turn requires rolling the Hazard Die and possibility of setbacks. In other words, potential mechanical reward entails potential risk. Part of the risk in taking another Turn comes from advancing fictional time. For example, taking a Haven Turn to recover could result in opponents gathering reinforcements, weather taking a turn for the worse, a political crisis, or a natural disaster. Though Adventurers may be making the same Moves, the setting does not remain static in response.

From a general perspective, Let It Ride and Push Your Luck can be seen as two poles of a bipolar resolution finality spectrum. Let It Ride specifies that a resolution is final after one iteration while Push Your Luck specifies that resolution may be indeterminate. An Adventurer may fail to open a lock, take the outcome of the Hazard Die in stride, and then try again, repeating this procedure as many times as desired assuming the Adventurer remains capable. Various intermediate principles are also possible along this spectrum. For example, limiting the number of potential retests to some arbitrary number or requiring players to spend some consumable game resource to try again.

Neither principle is inherently superior, but they do have different properties and structure play differently. In Burning Wheel, the purpose of Let It Ride is to continuously push the fictional narrative forward. Additionally, Let It Ride may encourage more diverse problem solving over time as probability suggests that a given means will be insufficient at least some of the time, forcing players to use alternative strategies. Push Your Luck leverages the psychology of temptation, assuming the uncertainty in question stands between players and their desires. By allowing players to take on greater risk in pursuit of outcomes judged important, Push Your Luck also lends weight and consequence to player decisions.

Tactical Hazard Die


NES Dragon Warrior

The current unreleased working version of the Hazard System uses six potential outcomes which are then interpreted relative to the current turn type. The four turn types, from most abstract to least abstract, are Haven, Wilderness, Dungeon, and Combat. The six outcomes, mapped to the sides of the 1d6 Hazard Die, are 1) Setback, 2) Fatigue, 3) Expiration, 4) Locality, 5) Percept, and 6) Advantage. This unifies the set of potential outcomes so referees need learn fewer exceptions. Additionally, the order roughly ranks the outcomes from most negative (Setback) to most positive (Advantage) taking the perspective of player characters.

Previous versions of the Hazard System only used the Hazard Die for Haven, Wilderness, and Dungeon Turns, not Combat Turns. This makes sense genealogically given that the Hazard System was adapted from the Overloaded Encounter Die which was inspired by traditional random encounter checks. However, there are uncertain outcomes that require dice resolution during combat regularly, such as initiative, so perhaps the Hazard Die can subsume the resolution of uncertainty at all levels of abstraction.

It is not hard to find analogues in combat for most of the Hazard Die outcomes. For example, Setback could mean that reinforcements arrive or the opponents act first. Fatigue could be general attrition, such as all engaged combatants taking a point of damage. Locality could be any kind of change on the battlefield, such as a door being locked or a table overturned. Percept could be information telegraphing an opponent’s future strategy. Advantage could be an additional move per player character or a forced morale check for the enemy.

This set of outcomes does not replicate the probabilities of initiative in the same way that the wilderness travel or dungeon exploration applications of the Hazard Die replicates the chance of having an encounter. As described above, opponents have only a 1 in 6 chance of acting first, compared to the traditional 50/50 odds. Whether this is a problem will depend on how one sees the purpose of initiative. If the point of initiative is to inject some regular uncertainty and tension into combat, then it seems like the set of abstract Hazard Die results should serve the same purpose while also increasing combat dynamism through variety of events.

Another benefit I see of overloading the initiative die is that some other aspects of combat, such as morale, which are easily overlooked but quite beneficial to the dynamics of play, can be potentially built into regular game procedures. I am not sure if a 1 in 6 chance of opponent morale check (on the Advantage Hazard Die result) is the best way to do this but it seems promising. Are there any other combat events that deserve a place in the Combat Turn Hazard Die interpretation guidelines?

As with most systems that replace bookkeeping with probabilities, such as tracking ammo abstractly, there are absurd edge cases. What if your torch runs out on the first turn in the dungeon? What if you run out of arrows immediately? I see three solutions to this sort of problem: 1) use rulings based on fictional appropriateness, 2) use illogical results as a kind of oracle demanding explanation, or 3) make the system more complex to handle such edge cases reliably. I lean toward option 1 and away from option 3. In my opinion, it is no particular shortcoming in the system to rely on the referee to determine whether it makes sense fictionally for reinforcements to arrive in any given instance. Illogical results can also just be ignored occasionally given that doing so just falls back to the traditional mode which works reasonably reliably.

More concretely, my current play test interprets Combat Turn Hazard Die outcomes as:

  1. Setback: opponents act first or reinforcements arrive
  2. Fatigue: combatants engaged in melee suffer 1 point of damage
  3. Expiration: some or all ongoing effects end (such as burning oil)
  4. Locality: the battlefield changes in some way
  5. Percept: players gain some clue to opponent strategy
  6. Advantage: players choose extra action or forced morale check

(Post image is only mildly relevant, but hey it’s combat right?)


Traditions and corruption

2016-07-08 19.15.49 copy

Personal photo of Symbaroum core book

Several different traditions of magic coexist in the world of Symbaroum, including wizardry, witchcraft, theurgy, and sorcery. All forms of magic entail the risk of corruption, but the risk can be decreased somewhat by following the rituals and practices of a given tradition. Each tradition grants access to a set of spells and casting these spells in the traditional manner avoids some of the dangers of raw magic.

Wizardry is highly codified arcane knowledge as set down formally by the Ordo Magica. Learning wizardry requires long, systematic study and extensive formal training.

Witchcraft follows older ways from the great forest Davokar. The witches serve as spiritual advisors to the barbarians living on the outskirts of the great forest Davokar. Many Ambrians are highly distrustful of witchcraft and see witches as little different than demon-worshipers or sorcerers but witches have elevated social positions within their own tribes.

Theurgy channels the power of the gods, most commonly the sun god Prios. Covenants with other lesser-known powers are becoming more common following the exodus.

Sorcery is the least formal of all the traditions, though there are many secret lineages. Some sorcerers come to the art by pact with occult beings while others discover ancient proscribed treatises and are self-taught. Despite the dogma of the Ordo Magica, sorcerers do follow rules, just highly idiosyncratic rules. To someone trained in one of the other traditions of magic, sorcery can seem pure chaos. Sorcery is forbidden according to the law of old Alberetor, but that has less force in the frontier of Ambria.

Characters on the Path of Wonder choose a tradition before play. Other characters with Magic stat greater than zero may enter into a tradition during play. Characters may not belong to more than one tradition.

When a character casts a spell within the bounds of tradition, there is no immediate chance of catastrophe or abomination, though the character accumulates a point of corruption for each spell cast. Once the number of corruption points equals the character’s magic stat, however, the safeguards of tradition become less able to control the mystic power unleashed. Characters reset corruption points to zero during each Haven Turn. See the Hazard System for details about Haven Turns.

If a character casts a traditional spell when corruption is equal to the Magic stat, there is a chance that they are unable to control the arcane power. Reality objects to being ungently used and reacts proportionally. The spell caster must make a Magic Test or acquire a permanent stigma, a physical mark of arcane corruption. Determine stigma randomly.

Casting a nontraditional spell when corruption is equal to the Magic stat follows the same rules, but failing the Magic Test results in a catastrophe in addition to a permanent stigma. This is why untrained magicians are so feared and traditionally punished with exile or death. The Ordo Magica is often blamed for any magic disaster and so is particularly harsh in hunting down and punishing renegades.

Once a character has accumulated a number of stigmata equal to their Magic stat, their humanity hangs in the balance. The next time that character would acquire a stigma, instead they are fully transformed into an abomination. At this point, the player must make a new character and the abomination becomes a monster under the control of the referee.

Characters with Magic stat greater than zero and no tradition may still learn and cast spells or use enchanted objects following the magic rules, but have none of the safeguards against corruption that the traditions provide. Attempting to learn a spell outside of a tradition and failing also causes either a stigma or a catastrophe (the player may choose).

Spells marked as rituals take a full Dungeon Turn to cast. See the Hazard System for details about Dungeon Turns. Other spell can be cast as a combat move.

The four magic traditions and the progression from stigmata to abomination are based on the Symbaroum setting.


Spell attunement

2016-05-02 19.28.37 dark souls 3I want spell rules that:

  1. Do not require regular spell preparation (to decrease complexity)
  2. Avoid locking players into a very small spell list (for variety)
  3. Support acquisition of new spells through adventuring

Several other rule sets have systems that feel to me like they were built to satisfy similar goals. ACKS (2012) differentiates between repertoire, which are the spells available for casting, and formulae, which are all the other spells that a mage has access to (in something like a library). Mages in ACKS can swap spells in the repertoire with formulae, but only at extremely high cost (ACKS core, page 67):

An arcane spellcaster who already has a full repertoire of spells may sometimes wish to replace one spell in his spell repertoire with another of equal level. It costs 1 week of game time and 1,000gp for each spell level to replace a spell in the repertoire with another.

D&D 5E (2014) uses a similar approach where there is a difference between known and prepared spells but then adds an additional layer of complexity with spell slots, which are different in 5E than in previous editions, and function essentially as level-rated mana or spell points. To learn a new spell, 5E wizards must write the spell in a spellbook, which serves a similar function to formulae in ACKS, though there is only cost involved in the initial copying step, not when making the new spell available for preparation (5E Player’s Handbook, page 114):

For each level of the spell, the process takes 2 hours and costs 50 gp. The cost represents material components you expend as you experiment with the spell to master it, as well as the fine inks you need to record it. Once you have spent this time and money, you can prepare the spell just like your other spells.

Abstractly, both systems represent two pools of spells, essentially online and offline. I find both systems somewhat cumbersome to use and difficult to explain.

The traditional approach, probably represented most paradigmatically in AD&D (1978), also uses spellbooks as collections of offline spells, along with a complicated host of intelligence-based limits and checks to learn new spells (AD&D Players Handbook, page 10):

Acquisition of Heretofore Unknown Spells: Although the magic-user must immediately cease checking to determine if spells are known after the first complete check of each spell in the level group, or immediately thereafter during successive checks when the minimum number of spells which can be known is reached, it is possible to acquire knowledge of additional spells previously unknown as long as this does not violate the maximum number of spells which can be known. New spells can be gained from captured or otherwise acquired spell books or from scrolls of magic spells. In the latter event the scroll is destroyed in learning and knowing the new spell or spells.

Actually following all the AD&D procedures results in nicely differentiated magic-users that can acquire new spells from adventuring, but the overhead is rather high and the various rules are scattered all over at least the Players Handbook. The AD&D approach also warmly embraces and rewards high-maintenance spreadsheet-assisted play, which is not what I am looking for.

Below is an approach I have been working on, encoded in two rules: Attune and Scribe. The rules are written in cryptic Hexagram style, but for D&D application, replace Magic rating with class level or your favorite determinant of spell capacity. The transaction cost of swapping spells between equipment slots and spell slots is represented by magic ink, the cost of which needs to be squared with the other relevant economies of gameplay.

(For anyone that does not catch the allusion, this approach is inspired by Dark Souls.)

Attune. To attune a Spell, consume a Spell Scroll and add the Spell from the Scroll to the list of Attuned Spells. Attune no more Spells than the Magic rating. For example, an Adventurer with Magic rated 3 may attune no more than three spells.

Scribe. To scribe a Spell Scroll, consume magic ink, add a Scroll of an Attuned Spell to the Gear list, and optionally remove the Spell from the list of Attuned Spells. Like all items, each Spell Scroll occupies one Gear slot.


Weapons of unusual size

Young Guts from Berserk

Young Guts from Berserk

Hexagram characters begin with stats rated from 0 to 3, using the arrays I originally developed for Gravity Sinister. (There is a random determination table for players that do not like to bother with making choices.) Then, each level, including first, players choose one stat to improve. The same stat cannot be improved two levels in a row. The max character level is 10, which means that the highest a stat can be naturally is 8 (3 initial + the 5 for every other level increases).

Among other benefits, characters with higher strength scores can wield ever more obscenely scaled weapons. There are three size categories beyond standard: huge, giant, and colossal. They require, respectively, strength scores of 4, 6, and 8, to wield effectively. (Category names are subject to adjustment.)

For normal weapons, strength adds to melee damage, up to +3. Larger weapons can express strength beyond this limit. Huge weapons allow up to +5, giant up to +7, and colossal up to +8. (In general, the max bonus is one less than the ability threshold for the next largest weapon category.) For simplicity, there are no special encumbrance considerations for oversized weapons. Each counts as one significant item. They do, however, cost more to repair (an additional 1d6 * 10 SP per exceptional size category).

Larger weapons retain any type benefits. Thus, a giant axe can express up to +7 melee damage from strength and also provides a sunder bonus to damaging enemy equipment. Oversized missile weapons apply strength to damage rather than perception, but are fixed at +4, +6, or +8, depending on the size category. For example, a huge elephant gun deals +4 damage even if the wielder has 5 strength. Such weapons still use perception for attack tests.

Though this system is designed with big weapons in mind, it would be easy to adapt to enchanted weapons that would only serve worthy warriors (that is, those strong enough or with large enough attack bonus for D&D), and so could be another way to explain and manage the traditional restriction that only fighters can use magic swords.

For AD&D (1E and 2E) ability scores, use the strength damage bonus rather than the Hexagram strength ability. For something like D&D 3E or 5E, use the ability modifier. The mappings are not perfect, but they should be good enough. Some other rulings may be required, given that HP quantities in 3E or 5E are higher that the OD&D standards I tend to assume, so adjust accordingly.

Edit: though above I noted that there are no special considerations regarding encumbrance, I am not fully convinced that is the right way to go. I think as written there may be insufficient incentive for diversity of weapon choices (that is, anyone with high strength would prefer an oversized weapon), which is perhaps uninteresting. I will need to see how this plays at the table, but one potential modification would be for each extra size category to count as a significant item, though I am wary of slipping graduated encumbrance in via the backdoor.


Pursuer's Ultra Greatsword from Dark Souls 2

Pursuer’s Ultra Greatsword from Dark Souls 2

Guts from Berserk

Guts from Berserk

Monster Hunter concept art

Monster Hunter concept art

Cloud from Final Fantasy 7

Cloud from Final Fantasy 7

Saw spear from Bloodborne

Saw spear from Bloodborne

Monster Hunter concept art

Monster Hunter concept art

Bow from Monster Hunter

Monster Hunter concept art

Hexagram crafting

Several times over the past few years I have tried to create a Diablo style loot and gear generator that would work elegantly with a traditional tabletop fantasy game. This is hard, and I have not yet come up with something that I find satisfying. It is hard largely because I feel torn between writing something overly general and something overly specific. Ideally, such a generator should interface with the setting monsters, but then those need to be nailed down and the generator becomes less useful to any other setting that deviates from the setting implied by those monsters. That realization leads to a turn back toward generality, which tends to be too schematic and not particularly atmospheric.

VRA 1 mentioned the old 3E era tabletop supplement Diablo II: Diablerie which has such a generator. It is actually not bad. The prefix-item-suffix approach has potential and the volume of content alone is enough to create interesting sounding item names. Many (though not all) of the attached mechanics, however, are mostly just numerical inflation. Bonuses to attack, bonus HP, etc. It could be used, but would require so much at the table ruling and interpretation that it almost does not feel worth it.

So I have decided to try approaching the problem from a different angle. Rather than building the generator beforehand, instead I will build a game system that will build such a mechanism naturally as the campaign progresses. This sidesteps the problem of bland generalness versus specific atmospherics completely. The first time a component such as a mineral or monster part is used for crafting, a particular augmentation becomes associated with that component. Augmentations can either be taken from a list of examples or improvised using that list as guidelines to appropriate power levels. Harvesting monster parts takes one Dungeon Turn. The material and augmentation will also be added to the general gear generator table. I still sense a small amount of hand-waving here, but I think this is close enough to a workable mechanism that the remaining details will naturally reveal themselves through play tests. I can see this easily applying without fuss to just about anything, even things like gems, leading to an interesting gamble: sell that ruby for coin, or have it forged into some piece of gear?

Power inflation should be possible to keep in check through a combination of the always operative gear degradation rules (all items not made of adamant still have at least a 5% chance of being damaged during use depending on quality rating) and abilities that refresh only during haven turns. For example, I could see adding a lightning burst augmentation which would allow an adventurer to add a die of lightning damage to one successful attack per excursion.

Crafting is something you pay a Haven NPC to do for you during a Haven Turn. Though I would not say no to a player that wanted to take a crafting Expert Skill such as Smith or Brew, I am not going to include it as an option in the text because 1) I suspect (though am not sure) that it would feel suboptimal compared to skills usable within a dungeon and 2) really I am not trying to write Shopkeepers & Spreadsheets.

Specifically (now for the player-facing Haven rules):


Repair: To repair a Damaged item, pay a smith 1d6 × 10 coins. Once a particular repair cost has been determined, it will not change on subsequent Haven Turns.

Forge: To forge a weapon or piece of armor from special materials, pay 1d6 × 100 coins and consume the materials.


Brew: To brew a concoction from special materials, pay 1d6 × 100 coins and consume the materials.

Smiths and Alchemists can each carry out up to 1d6 tasks per Haven Turn.

Brewing Potions

The Love Potion (from Wikipedia)

Magic-users (and, to a limited extent, clerics) can brew potions. The same game systems also apply to creating poison (for thieves) and incendiaries (for fighters). These items have no special use requirements, though some classes are better at creating them than others. For example, anyone can imbibe a potion of gaseous form brewed by a magic-user or coat a weapon with poison created by a thief.

To brew a potion, a recipe is required. Every recipe specifies one or more special components that are required, in addition to mundane ingredients and procedures. There may be more than one recipe for the same potion (each making use of different special components). Recipes can be discovered in play much like scrolls or purchased from specialists such as apothecaries (who tend not to share the secrets of their livelihood) or sages (who often charge ungodly prices).

Potion recipes have a level, just like spells. In order to brew a potion from the recipe, the character in question must be able to cast spells of the equivalent level. Potion components cost 500 GP and one week per level (so a second level potion would cost 1000 GP of ingredients and require two weeks of work). Like magic research, brewing potions may be done during downtime punctuated by adventuring, as long as too much time (by referee ruling) does not pass. Characters do not need to spend money separately to establish a laboratory. It is assumed that as items are created, the character naturally accumulates the paraphernalia required, and this is abstracted into the cost of ingredients.

Fighters and thieves should use the magic-user spell progression to determine if a given character is skilled and knowledgeable enough to create a particular item. Costs are identical (500 GP for level 1 poison, etc). The only significant difference is that each “brew” of poison results in 1d6 doses (unless otherwise specified in the recipe). Different poisons may also have different application methods (also by recipe), so one poison may be contact, one poison may be injected (i.e., for coating a weapon), another poison may end up being a beaker full of gas that may be hurled like a grenade. Mutatis mutandis for fighters. In addition to incendiaries, fighters can create (or oversee the creation of) siege engines and siege works. Schematics for these work exactly as other recipes, and are rated similarly by level.

Note that though the ability to brew potions is available to characters of any level (given appropriate class), the costs involved (along with the fact that spending GP results in XP) means that characters that craft several items (be they scrolls, potions, or something else) will naturally end up becoming higher level, with no other constraints required.

Optional rule: cross-class brewing. One kind of class may create the type of recipe items appropriate to another class (assuming a recipe and special components are available), but the the costs are doubled due to unfamiliarity and the crafting is only successful on a d20 roll less than or equal to the intelligence score. Upon failure, the components are not wasted, but another week must be spent (and another check made) until either the brewing is successful or the task is abandoned. In any case, the spent GP results in XP (learning from failure!).

I’m thinking that maybe each class should begin with one basic first level recipe (love potion, healing potion, minor firebomb, and minor poison, perhaps).

(In Hexagram, provisionally, the ability to brew potions comes with the alchemy trait, the ability to brew poison comes from the assassination trait, and the ability to craft incendiaries or do siege-work comes with the ranged combat and melee combat traits, respectively.)

Ghost Traps

Sapphire image from Wikipedia

Ghost traps are items (most commonly gems) that have been enchanted to serve as prisons for spirits and incorporeal undead. Clerics of any level may create ghost traps with hit dice capacity no greater than their level. Ghost traps may contain no more than one spirit, even if the trapped spirit has hit dice less than the capacity of the trap. When occupied, a miniature version of the spirit can be seen leering within the gem, seemingly contorted and tormented.

To create a ghost trap, a cleric must have a gem of value equal to the desired hit dice capacity multiplied by 100 GP. For example, a 3 HD ghost trap requires a 300 GP gem. In addition, hit dice capacity * 100 GP must be spent on ritual components, and the whole process takes 1 week. There are some situations described below where a ghost trap is destroyed. In these cases, the gem crumbles to ash.

To use a ghost-trap, it must be strongly presented within 10′ of the spirit (this is a standard combat action). If the spirit fails its saving throw versus magic, it is imprisoned in the gem. A trap may also be thrown, but if the spirit makes its saving throw, there is a 1 in 6 chance that the gem will be destroyed when it hits the ground. Turned spirits do not get a saving throw and may be automatically imprisoned. All classes may use ghost traps, though clerics may set them as traps that will trigger when spirits draw near. An attempt to trap a spirit more powerful than the capacity of a given trap will automatically fail, and there is a 1 in 6 chance that the trap will be destroyed.

Trapped spirits may be released at any time, though they will generally be quite unhappy about having been imprisoned (-2 reaction). Releasing a spirit has a 1 in 6 chance of destroying the trap (otherwise it may be used again). Trapped spirits may speak telepathically with anyone touching the gem directly, will generally beseech any holder for freedom, and may promise all sorts of rewards (probably all lies). This contact is one-way, however, and it is impossible to communicate with a trapped spirit without using magic (see the magic-user ritual below).

Though only clerics can create ghost traps, many magic-users prize occupied ghost traps, as they believe that they are valuable currency in the land of the dead, and can also be used in arcane researches. Magic-users can also conduct rituals to communicate with trapped spirits. This requires 1d6 * 100 GP worth of components, and requires a saving throw versus spells. On a saving throw roll of 1, the spirit is released (or some other suitable magical mishap occurs). If the magic-user succeeds on the saving throw, the spirit will answer 1d6 yes/no questions to the best of its ability, and will be unable to lie.

(Though I know about Skyrim soul gems, the inspiration here is more Ghostbusters.)

(In Hexagram, the ability to create ghost traps comes with the banishment trait.)

Hexagram Treasure Overview

Excalibur the Sword (source)

There are two major categories of treasure: mundane and wondrous. Mundane treasure includes money, valuables, simple machines, NPC favors, and other such things. The most important quality of mundane treasure is that it is not connected to character traits in any meaningful way. There are no prerequisites for use. A sack of money can be used by a squire just as easily as by a powerful sorcerer.

Wondrous treasure is not necessarily more potent than mundane treasure (though it often is). The defining quality of wondrous treasure is that use is dependent upon character traits. Some traits may also be primarily concerned with using (or even creating) wondrous items. All wondrous items have a rating, from 1 to 6, which relates to a specific character trait. There are traits on every path which relate to different kinds of wondrous items.

Exactly how wondrous items interact with character traits varies by item. Some require a minimum level of a certain trait in order to function at all. For example, a certain magic sword may be inert (that is, function as a mundane sword) for warriors with less skill than 3 in melee combat. Other items may require a successful trait check (less than or equal to the trait value when rolling a d6) before the item can be used, or even per use. For example, an antediluvian gauntlet may allow reaching into stone as if it were water, but only with a successful antediluvia trait check.
The wondrous items related to the path of battle are magic weapons, which are governed by either the melee combat trait or the ranged combat trait (depending on whether the item in question is a melee weapon or a ranged weapon). The wondrous items related to the path of guile are antediluvia (artifacts left over from before the deluge). The wondrous items related to the path of wonder are spell formulae, potions or potion recipes, magic devices, and scrolls. Though anyone can use a potion, it can also be reverse engineered. Thus, potions straddle the line between mundane and wondrous treasure.
As wondrous items are keyed to specific traits, not paths, characters with points in the necessary off-path traits can potentially make use of any kind of wondrous item. For example, a sorcerer with melee combat 3 can make use of a potency 3 sword just as well as a character on the path of battle (though it would have been more costly in terms of experience for the character on the path of wonder to develop that martial skill). The most complex antediluvia might only be usable by a character on the path of guile (based on the cap for off-path traits), but other less complex items are usable by any character with the appropriate trait.
One method for placing wondrous items: roll d6 to determine if there is a wondrous item; on a 6 there is. Then, roll another 2d6 and take the lowest number to determine the potency or complexity. Finally, roll on the following table to determine the wondrous item type.
  1. Potion
  2. Antediluvian item
  3. Weapon of power
  4. Scroll
  5. Magic item
  6. Spell formula
All wondrous items should be unique to a given campaign. Though not all wondrous items need have drawbacks, they should all have quirks and idiosyncrasies. Some wondrous items may have no direct adventuring use, only being able to generate strange seemingly useless effects. Few wondrous items are both of unlimited use and without potential drawbacks.

I am building a system for generating campaign-specific wondrous items as part of the referee scenario design process.