Tag Archives: The Final Castle

Hazard System v0.2

The Hazard System is the gameplay engine behind The Final Castle. Though it assumes games of fantasy adventure and exploration, it is modular and should be easily adaptable to many kinds of tabletop RPGs. I use some variation on this approach for pretty much every game I run now and I don’t think I could go back to any other method of pacing or timekeeping. It is the natural outgrowth and generalization of the overloaded encounter die.

Gameplay proceeds in turns at different fictional timescales with each turn accompanied by the chance of a hazard. What “hazard” means will vary based on the context. For example, a hazard during a haven turn, when characters are recovering between adventures, might be a natural disaster, while a hazard during a dungeon turn, a much shorter period of fictional time, might be an encounter with a wandering monster. The hazard system also serves as a general timekeeping and resource tracker.

The text below the divider is released under a Creative Commons Attribution 3.0 Unported license. Attribution:

Necropraxis Productions Hazard System v0.2 http://www.necropraxis.com/hazard-system/

A PDF is also available.

Currently, several references to undefined terminology from The Final Castle remain, such as Ability Tests. It should be relatively easy to interpret these in light of whatever system you are using, but the final Hazard System text will likely be entirely system agnostic. Additional turn types will also likely be included (Domain Turns and Generation Turns, particularly).

Hazard System v0.2

The game proceeds in turns of several different types. The turn types are haven, wilderness, dungeon, and combat. Each represent a progressively smaller amount of fictional time within the game world, though the exact durations are usually abstract. The passage of time within each turn is a resource to be spent wisely, as the hazard die is rolled for every turn that passes to represent potential danger.

The Hazard Die

The six-sided Hazard Die deploys threats, manages resources, keeps time, and tracks light. In short, it is the engine that drives gameplay forward and the heart of the Hazard System. Every significant action, whether in town recovering, traveling through the wilderness, or searching a dungeon corridor for traps, takes a turn. Every turn is accompanied by a roll of the Hazard Die. The exact interpretation of the die result varies by turn type, but the outcomes are conceptually similar. A haven hazard might be a shortage of supplies while a dungeon hazard might be a wandering monster.

Players other than the referee should roll the Hazard Die to make the time cost salient. After rolling the Hazard Die, hand it to another player so that everyone gets a chance. If playing in person, rotating clockwise around the table works well.

Beginning the game

Start a new campaign with the first Wilderness or Dungeon turn of an adventure. Players should choose a clue or quest from the Tavern to pursue. Roll HP following Recovery guidelines to determine initial HP.

Starting and ending sessions

Sessions should begin and end in a Haven if possible. This allows the the players and PCs to vary between sessions. After a session, players should tell the referee if they are going to follow a different clue or Quest during the next session so that the referee can do any preparation required beforehand.

Haven turns

To recover and replenish resources in a civilized refuge, take a Haven Turn. The exact fictional duration of a Haven Turn can be anything from a few days to several weeks. It is rarely necessary to interrogate the details.

  1. Roll the Hazard Die and resolve any hazard
  2. Pay upkeep: accommodation, retainer, property, and so forth
  3. Recover (roll HP, applying accommodation modifiers)
  4. Process retainer loyalty
  5. Reckon XP gained and level up if appropriate
  6. Buy or sell items, repair damaged gear, or recruit hirelings
  7. Optionally, take one Haven Action (scribe scroll and so forth)
  8. Prepare spells
  9. Review rumors and news

Haven Hazard Die results

  1. Complication (introduce at any point during the turn)
  2. Clue about next complication
  3. Abatement of one or more (by referee whim) haven conditions

Ignore results of 4 – 6. Starred complications persist as conditions.

Haven complications
d20 Complication d20 Complication
1. Assassination 11. Insurrection *
2. PC challenged 12. Invasion *
3. Curse * 13. Jailbreak
4. Earthquake 14. Mobilization *
5. Flood 15. Monster attack
6. Falling star 16. Murderer on the loose *
7. Famine * 17. Pestilence *
8. Fire 18. PC slandered
9. PC impersonated 19. PC item stolen
10. Inflation * 20. Winter *

Wilderness turns

Wilderness turns alternate between day and night. Characters taking two non-camp wilderness actions in a row suffer 1 damage and gain a point of Exhaustion. Choose a wilderness action: travel, search, explore, hunt, track, or camp.


Move the party into an adjacent area or access a known landmark such as a haven or dungeon.

Search, Explore, Hunt, or Track

The party leader makes a Search Test to locate (and enter, if desired) a hidden feature. To Explore, Search without a stated goal. Success reveals a random hidden feature. Track is a Search to follow a quarry. Hunting yields 1d6 rations (adjust for terrain) per hunter. Night applies a -1 penalty to Search, Hunt, or Track.


Camping requires a bedroll and consuming 1 ration per character. One person may stand watch for each four party members without impairment. Ignore Hazard Die results above 3.

Wilderness Hazard Die results

  1. Encounter (may differ between day and night)
  2. Percept (regarding next encounter)
  3. Locality (mechanical change in environment)
  4. Percept (regarding hidden feature)
  5. Resource exhaustion
  6. Lost


Travel is no longer an option if a party is lost. Search must be used to locate a landmark before travel can be resumed.


Each point of Exhaustion imposes a cumulative -1 penalty on all physical Ability Tests. This adds to Encumbrance penalties.

Dungeon turns

Some actions that require a Dungeon Turn include climbing, forcing a door, guarding the party, listening at a door, moving to a new area, searching the current area, and other tasks of similar scope. Each player may take a different action during a Dungeon Turn. Dungeon Turns can represent a fictional amount of time anywhere between a few minutes to an hour, though most commonly are about 10 minutes long.

In practice, the passage of Dungeon Turns can be more fluid than selecting actions by name, resolving any hazard, and iterating. However, always remain cognizant of lurking dangers and call for the Hazard Die whenever significant actions are taken.

Free dungeon actions

Minor actions often do not consume a full Dungeon Turn. Interacting with particular features such as looking under a rug, opening doors that are not stuck, and pulling levers are all free actions. Clever use of free dungeon actions can forestall the Hazard Die and thus decrease risk.

Dungeon Hazard Die results

  1. Encounter
  2. Percept (regarding next encounter)
  3. Locality (mechanical change in environment)
  4. Fatigue (take 1 damage unless the next turn is spent resting)
  5. Resource exhaustion (spell duration, etc)
  6. Light source exhaustion (any or all light sources go out)

Reasonable resolution

Fatigue and light source results may be ignored when they do not make fictional sense, such as during the first few turns of an exploration. Exhausted light sources rarely all go out at exactly the same time, but instead dwindle over the course of the turn, and may be relit given sufficient PC resources.

Released under the Creative Commons Attribution 3.0 Unported license


Necropraxis Productions Hazard System v0.2 http://www.necropraxis.com/hazard-system/

Stats and magics

Playing Dark Souls has helped crystalize in my mind a lot of how I want the character advancement options to work in The Final Castle. Back when I was working on the Hexagram rules, one of my main goals was to support flexible cross-class abilities without complexity or undermining traditional class archetypes, though now I find the particular approach I was working on somewhat unsatisfying. There was too much discretion, not enough structure, and the lists of system options were too long.

The Final Castle has a far simpler, and more elegant, method of advancement which I believe satisfies my original requirements. Characters may advance potentially to level ten, and each level gained allows the increase of one ability score* (though the same score may not be increased over subsequent levels). The ability scores are combined with a class bonus (which is half level, round up) to determine most action resolution. So, for example, a character is going to roll something like 1d20 +dexterity +fighter (shorthand here for fighter class bonus) when making a combat roll. Starting stats range from 0 to 3, a given stat can be increased up to +5, and the class bonus rises to +5 at most, yielding a nice range of bonus for even the luckiest and most focused character (up to +13 on the d20 scale at level 10). Such specialization comes at the cost of flexibility, as will become clear momentarily.

It may seem at first glance like this does not have much to do with the previous discussion of Dark Souls. However, like Dark Souls, the magic rules apply to characters of all classes. That is, a fighter, for example, rolls 1d20 +magic +magician when casting spells, and the number of spells that can be prepared is also governed by those numbers. (Recall that intelligence has been replaced by magic.) Now, in the case of a fighter, +magician (the class bonus) is always going to be zero, but +magic may be increased (if the fighter wants to dabble in magic) during level up rather than one of the physical stats. Magicians have access to more methods for learning spells, but any character with sufficient stats can at least learn spells from a teacher, and any character has the potential of sufficient stats through level up choices.

Cleric magic (called boons), is handled similarly, with +charisma and +cleric taking the place of +magic and +magician. Rather than learning spells one by one as does a magician, clerics are granted access to a full suite of powers upon making a covenant with a given immortal. The default covenant available to clerics at first level is with The King of Life**, but other covenants may be discovered during play and accessed by any character that has sufficient charisma score. Most immortals will not covenant with characters that use magic though, as such is considered presumptuous and hubristic. More than one covenant at a time is impossible, and breaking a covenant may come with serious consequences.

* Oversimplifying slightly for clarity.

** Inspired by Dogs in the Vineyard and used with permission.

Combat and maneuvers

Combat actions other than the standard, damage dealing attack can be resolved in many different ways. In the past, I have used several approaches, such as requiring ability checks instead of or along with attack rolls, or using ability score contests similar to the opposed skill rolls suggested by D&D 4E. However, recently I have come to think that using ability scores in this area is not the best approach. It requires generating ability scores for monsters regularly, which granted is not that cumbersome, but is nonetheless suboptimal. Further, it plays oddly with the primary measure of combat skill, which is the attack bonus (or combat tables), which seem better suited to resolving most kinds of nonstandard attacks. The system below is from The Final Castle rules, but works just as well with most traditional fantasy games, whether or not monological combat is being used.

All physical actions that might be taken in combat are handled with a combat roll. By default, this is the standard “roll high, hit a target number, do damage” that should be commonly recognizable. However, rather than inflicting damage, a character may attempt to cause any number of other reasonable effects, taking the effect rather than damage, as long as the intent is declared prior to the roll and is fictionally reasonable. I have predefined a number of common maneuvers which can be substituted for a standard attack, such as disarm, grapple, and disengage, but these are intended to be samples rather than a comprehensive list of moves. Undefined maneuvers should be negotiated between players and referees prior to any action declaration.

Additionally, I have a rule called overkill which says that attacks do +1d6 damage if the combat roll exceeds the target number by four or more. This is another way that the fighter’s increasing attack competency with level scales damage up, but it also applies to maneuvers. That is, if a combatant is attempting something like a push-back bull rush maneuver, if they succeed with overkill, the result is both the desired effect and 1d6 damage. Thus, doing an attack/maneuver at once is possible, but more difficult, and you might get just the effect without direct damage.

As a more extended example, consider the standard grapple attempt. If it is fictionally reasonable for a combatant to attempt a grapple (and note this is no more unambiguous than whether or not a standard damage-causing attack is fictionally reasonable), the grappling agressor makes a combat roll. On success, the target is successfully grappled, and can no longer move, though may be able to perform close attacks. That now-grappled target will need to attempt an escape maneuver to be free of the grapple, if that is desired, which will require a dedicated future action. Further, in this case, any appropriate side effects of a grapple automatically trigger, such as armor spikes or flaming body dealing damage. Were the initial grapple combat roll to achieve overkill, damage would also be dealt on the first round.

Since the resolution system uses the combat roll, fighters are better at maneuvers than other characters, but maneuvers are not limited to fighters. Like thief skills, I prefer for creative actions to be available to all characters, rather than being limited to fighters for niche protection. The trade-off is clear: give up damage in return for an effect. And the system is trivial to remember, in all cases: make a combat roll. In general, I think this maneuver system can be used for many actions that might be considered stunts in other systems. Particularly tricky maneuvers could either be done with a penalty such as 5E-style disadvantage, or require an overkill result to get the basic effect (note that this naturally reduces then to the 2E standard of called shots being at -4, which seems like a nice result).

In The Final Castle, armor reduces damage and the target number represents a more abstract enemy threat level (as was planned for Gravity Sinister). This means that there may be advantages to grappling, or engaging with some other nonstandard maneuver, a heavily armored foe, as the damage reduction will be less likely to come into play. This is a difference from D&D, which would make the trade-off dynamic slightly different if AC is used directly as the target number in all cases, but is in any case easy to adjust for using a basic penalty or advantage/disadvantage scheme.

The rise and fall of the Harbinger

As previously mentioned, I have been working on a system which is the evolution of my earlier Gravity Sinister work, also informed by my experience running OD&D over the past few years. This is coming to fruition as an integrated coupling of setting and rules. I am calling the project as a whole The Final Castle, after the fortress/dungeon which is the center of the games I am running with it.

Here is the background summary, which is also the intro to the current Player’s Guide draft. This draft is actually a text-complete finished document that has already seen some play-testing but will need to see more before I decide on the best way to release it.

Scion of the immortal Basmophael, the figure known only as the Harbinger claimed to herald the next phase of reality. The Harbinger moved inexorably across the known lands, unleashing monsters previously imprisoned in the dark places of the earth, and promising great power to those that would recognize the new dominion.

The arrival of the Harbinger was always preceded by a fortress that would tear itself from the ground. Then, the legions would pour forth, destroying all defenders. If the settlement was not annihilated entirely, the leadership would be replaced with a loyal servant, and the fortress would wither, soon becoming little more than a looming husk. Then the cycle would be repeated elsewhere.

The sempiternal courts took no action, assuming that they would be able to endure any human turmoil. The agents of Basmophael, however, infiltrated their hidden strongholds. Some surrendered and were corrupted. Others fought and were reduced to shadowy echoes of their former glory. Few remain.

Using subterfuge, an alliance of champions entered the Harbinger’s castle by means now unknown. With great hubris, the Harbinger had taken insufficient precautions against such a daring attack, and was subdued, though not destroyed, as destruction was beyond mortal power. The Harbinger’s armor was found to be a locus of Basmophael’s power, though it too was impervious. Unable to destroy the panoply, pieces of the armor were divided among the champions for safekeeping, but the Harbinger’s will was underestimated. Many of the panoply bearers that escaped have since been driven mad, compelled to return with the armor to the Harbinger’s fortress. Thus, the divided panoply remains held by occupants within the castle, each scheming to acquire the entire suit, as it is said to allow full control of the castle’s terrible armaments.

Unlike previous castles, the final castle has not withered, though neither has it belched forth new armies since the armor was disassembled. Rumors say that the Harbinger’s generals and creatures have fallen to internecine struggle. The final castle broods darkly over surrounding battlefields, which seem to be slowly expanding, inch by inch, day by day, despite the absence of any true fighting since the daring infiltration that seemingly ended the Harbinger’s war.