Tag Archives: thief

Assassins & poison

Max Klinger, Rivals (source)

Max Klinger, Rivals (source)

Recently, when compiling a document of Finchbox classes, I noticed that, especially after basic house-rule adjustments, the assassin and thief classes seemed awfully similar. Both had d6 HD, light armor skill, backstab, low attack bonus, and a (slightly different) collection of skills. The only significant contrast was that assassins had disguise and poison-craft whereas thieves had the troubleshooting skills (search, find/remove traps, open locks, etc).

This is not enough to justify two separate classes for me, so the choice is to either reformulate the assassin or drop it. Another approach, I suppose, would be to replace both classes with something like the LotFP specialist, which can be customized, but I already know I don’t want to do that. For these rules, I prefer to have more focused, atmospheric classes. And I do want to keep the assassin as an option. So here is a modified S&W assassin, focused more on the ideal of single-shot kills (compared to the opportunism and utility that comprises the essence of the thief). Both classes still have backstab, but the increased martial focus of this assassin, along with the added poison-craft subsystem (described below), and lack of dungeon utility skills, distinguish the two classes. Max level in this game is 10.

The poison-craft description is still somewhat wordy, and I hope to tighten it up in the future, but for now this should be good enough to communicate the rules. I, of course, reserve the right to modify the poison rules if they don’t satisfy me in play. More poison recipes will be added later to bring the total above 10, so that high-level assassins don’t converge in poison knowledge.

Edit: added PDF version.


  • Hit die and weapon damage: d8
  • Starting saving throw: 15
  • Armor training: medium
  • Attack bonus: medium

Special abilities & restrictions:

  • Backstab: +4 to attack from surprise, +HD damage (5th: +2HD, 9th: +3HD)
  • Poison recipes, one per level (odd: random, even: pick)
  • Ambusher: a party with an assassin is more likely to surprise enemies (usually, 4 in 6)
  • Skills: disguise, poison-craft, stealth (as thief of same level)
  • Optional: vow of guild loyalty and guild connections


A flask of poison may be concocted as a downtime action for 100 SP. Applying poison to a weapon requires a poison kit (which is a significant item), an exploration turn, and a poison-craft check to see if the poison is used up. Each time the assassin hits with a poisoned weapon, another poison-craft check should be made to see if the poison application has worn off. In any case, a poison application will not last longer than a single excursion. Poison may also be extracted from a poisonous slain creature with a successful poison-craft check (this requires a downtime action, but doesn’t involve any expense). Any number of poisons may be carried in a poison kit without consuming further encumbrance slots.


  1. Affliction: +1d6 damage
  2. Anticoagulant: if further wounded, takes 1d6 bleed damage per round (save ends)
  3. Blindness: target is struck blind (new save allowed 1/day)
  4. Debilitation: -2 physical penalty, +1 damage from any attacks
  5. Delirium: unable to focus, hallucinations, actions have random targets
  6. Doom: death after one exploration turn
  7. Mage-bane: unable to cast spells (new save allowed 1/day)
  8. Paralysis: unable to move (new save allowed 1/exploration turn)
  9. Sleep: slumber for 8 hours (new save allowed if damaged)
  10. Suggestion: groggy, will obey general commands (charisma check needed)

All poisons allow a save to avoid the effect, and generally work only on living creatures approximately human-sized or less. Effects on other creatures are by referee ruling.

Rogue class

Here is a draft of a new rogue class I developed recently. It uses the Gravity Sinister skills, though skill improvement is simplified into categories of untrained, trained, and mastered (which translate into chances of success on a six-sided die).

One improvement option is chosen each time a character gains a level. A character must have training in something before mastery.

Regarding experience tables, my inclination recently has been to use the fighter progression for everyone. The rogue uses the medium rationalized hit dice progression, and attack bonus is also derived from hit dice.

The omission of a separate sneak attack or backstab ability is intentional. I am thinking that surprise attacks are probably better handled independent of class in terms of effect (an extra die of damage seems reasonable), and the stealth skill grants rogues a better chance of setting up a surprise attack in any case. A separate backstab-type skill also focuses too much on damage per round type calculations for my taste.

Rules for simple and light weapons are also included with streamlined weapon properties for ease of reference at the bottom of the post.


Initial training:

  • Two weapons from the simple or light weapons lists.
  • Light armor.
  • Three skills.

Improvement options: skill training, skill mastery.


Skills are divided into basic and expert categories:

  • Basic skills: Climb, Listen, Search, Stealth
  • Expert skills: Devices, Locks, Steal

Skills that require general agility (climb, stealth, and steal) are penalized by one if wearing chain armor and two if wearing plate armor. Using a skill often takes some time and thus may require spending an exploration turn in focused application.

The climb skill allows allows the climbing of surfaces such as rough walls. Climbing a rope or ladder does not require a skill check. There is a penalty of one when attempting to climb smooth surfaces and a penalty of two when attempting to climb slippery surfaces. Climbing gear imparts a bonus of one on climb checks.

The devices skill can be used to disable or manipulate small mechanical traps and mechanisms. Failure does not trigger traps.

The steal skill allows something to be taken without being noticed. Steal can even be used in melee. On failure, the attempt is not noticed but the desired item is not acquired. Items held directly by others may be stolen, but this may not be done secretly.

Consider adding more expert skills if they fit your campaign. Some possibilities include tracking, poison-craft, herbalism, leadership, and chirurgy.

Untrained Skills

The chance of success when using an untrained basic skill is 1 in 6 for characters of any class. There is no chance of success when attempting an expert skill if untrained.

Trained Skills

The chance of success when using a trained skills is 3 in 6.

Mastered Skills

The chance of success when using a mastered skills is 5 in 6.


Simple Weapons
Weapon Properties Trained Mastered
Club bludgeon stun
Dagger throwable auto-hit after grapple
Spear reach throwable interposing
Staff two-handed, bludgeon +1 AC parry (melee)
Light Weapons
Weapon Properties Mastered
Short sword   +2 attack in formation
Short bow   +2 attack with aim
Sling unencumbering,
versatile ammo

Rogue, sorcerer, warrior

Why this split? This began as a comment on a Google Plus conversation, but I think it’s worth a blog post. For me, the split is based on two things: problem solving tools and archetypes. For archetypes, the inspiration is swords & sorcery. This, in my opinion, is uncontroversial and does not need further elaboration (other than to remark that the cleric, if taken too far away from the original Van Helsing and Solomon Kane inspirations, does not fit so well aesthetically or culturally).

Clerics are really a hybrid class in terms of problem solving, and could potentially be either fighter/mages (for the trad crusader vampire hunter that also has some magic) or thief/mages (a version less often seen, but just as thematic for zealous witch hunters or hashashin characters). However, the hybrid nature of the cleric means that it can be understood based on the other three main classes, so no more need be said about the cleric independently.

The primary problem solving qualities of the core classes are: combat/renewable resource (fighter), combat/consumable resource (magic-user), utility/renewable resource (thief), and utility/consumable resource (magic-user). Thus, the magic-user is more versatile, but resource-limited (and in most incarnations, more fragile). Obviously there is some bleed between the approaches when you consider the actual implementation (everyone can make melee attacks, fighters can still use some magic items, etc). So that’s where the split comes from in terms of OD&D game mechanics.

Edit: I should also link to Talysman’s post on classes and problem solving here.

JRPG Basic Skills

An X in 6 skill system fits very nicely with the 3 level cap of a “complete” basic game, because level can be directly incorporated into the probabilities without overwhelming the possibility set. No extra system or math is required. You will notice a simple nod to difficulty classes with the climb skill (two potential levels of penalties, for smooth and slippery surfaces) that I think works well, fits into the thief skill math, and is easy to remember. I may add similar tiers to the locks skill as well. In general though, I want to avoid the idea of difficulty classes, which is a concept that I think rarely works well outside of armor classes or other, similar combat targets.

I plan on including LotFP-style “fill in the dice” skill boxes on the character sheet, to remind players about the mechanical interfaces available to them (for example, X in 6 search in exchange for spending an exploration turn and risking a random encounter check). More details about how the skill system plugs into gameplay is contained in the section on movement and turns.


Skills represent the chances characters have at accomplishing certain common adventuring tasks within the game world, such as sneaking up on enemies or picking a lock. They are not meant to restrict potential actions, but rather to provide an impartial system for resolving actions with clear risks and rewards. Creative play may often allow characters to accomplish a task without recourse to a skill roll, and thus avoid the potential risks associated with a skill roll (such as the noise made by attempting to use the force skill on a door or chest). All characters begin with (at least) a score of 1 in 6 for all the basic skills. Expert skills are only available to characters of certain classes.

Thieves have access to the following expert skills: devices, locks, and steal. Thief skill chances start at 3 in 6 and increase by one point at each level gained. Thus, thieves have a chance of 4 in 6 at second level and 5 in 6 at third level.

Basic Skills


This skill is used to break open things like stuck doors and locked chests. Add +STR to this skill roll. Using a tool like a prybar or hammer may add another +1 depending on the object being forced. Because forcing something is noisy, it requires an dungeon turn. Note that thieves do not increase their force skill when gaining levels.


The climb skill allows you to climb surfaces such as rough walls. Climbing a rope or ladder does not require a skill check. There is a penalty of -1 when attempting to climb smooth surfaces and a penalty of -2 when attempting to climb slippery surfaces. Climbing gear imparts a +1 bonus on climb checks. Climbing requires a dungeon turn.


It is often useful to know what is behind a door before you break it down. Two characters may listen at a standard sized door simultaneously. A good effort requires quiet and time, so using the listen skill requires a dungeon turn.


If you don’t know exactly what to examine, you can just search an area from top to bottom. Using the search skill requires an dungeon turn, and allows a character to examine an area about the size of one skirmish melee (or a medium sized room). Note that examining specific features and interacting with them descriptively often does not require a full dungeon turn, and may not require a skill roll at all, so descriptive interaction is usually advantageous, as you may be able to avoid a random encounter check.


Sometimes you want to avoid danger rather than face it head on. Perhaps this is to set up a surprise attack, or to just avoid detection entirely. In any case, succeeding on a stealth roll allows a character to automatically gain surprise if it comes to combat. All characters attempting to sneak in this way must make separate stealth checks, but discovery of one character does not automatically reveal all sneaking characters. Stealth checks may not be attempted by characters in melee, but might be possible for those outside of melee during combat depending on the specific situation.

Expert Skills


The devices skill can be used to disable or manipulate small mechanical traps and mechanisms. Using the devices skill requires a dungeon turn. Failure does not trigger traps.


This skill is used to pick locks, and requires tools. Attempting to pick a lock requires a dungeon turn.


Take something without being noticed. Steal can even be used in melee. On failure, the attempt is not noticed but the desired item is not acquired. Items held directly by others may be stolen, but this can’t be done secretly.

Thief Roll

Jack from TOTGAD has an attractive house rule whereby he uses the “hear noise” d6 chance as the system for resolving all thief skills. This seems pretty reasonable to me, especially assuming that the alternative is the fixed percentile progression given in OD&D or B/X. All of the percentile skills start off rather low and slowly increase as levels are gained, with arbitrary differences between the various skills and no option to specialize in one skill over another (unlike the point buy systems of, for example, LotFP or Second Edition AD&D).

Looking at the OD&D thief (in Supplement I: Greyhawk), why does open locks start at 15% and move silently start at 20%? Do we really care about this distinction, given that all the skills start out at roughly the same level and increase at approximately the same rate? The one exception is climb walls, which starts out at 87%. But “always use hear noise” with perhaps one special case for climb walls is still far simpler than the official multiple stat percentile system, with functionally similar outcomes.

The schedule of thief hear noise improvement (from Greyhawk, page 11) is as follows. (Thanks to ODD74 for discussion about interpreting hear noise in the OD&D context.)

OD&D Thief Hear Noise
Level 1–2 3–6 7–10 11–12 13–14
Hear Noise 1-2 1-3 1-4 1-5 1-6
Effective d6 bonus +1 +2 +3 +4 +5

However, just when I was about to throw in my lot with this method, I realized that it does not allow for the variable degrees of success that I use with the old percentile system. That is, using d6, and without another roll, how do I distinguish between “no progress” and catastrophe? 16% (1 in 6) is too high of a chance for critical failure. In my current approach to thief skills, this is pretty important. Basically, if you fail your thief skill roll, but don’t roll 96 or higher, you don’t obtain your objective, but you can try again. The top 5% is the equivalent of rolling a natural 1 on an attack roll.

According to Greyhawk (page 11), pickpocket or move silently is the most favorable thief skill column, so one could just use that as a general thievery skill similarly to how Jack uses hear noise and preserve the 5% fumble chance.

Spell competency and other competencies

Spell competency is the highest level of spell that the magic-user can prepare. This is usually equivalent to experience level divided by 2 (round up). For example, a fifth level magic-user has a spell competency of 3. The original magic-user spell progression followed this pattern up until 6th level spells are considered, which are not gained until 12th level (rather than 11th, as the pattern would require).

I find treating this number as a separate stat is useful. For example, in the spell-casting roll. I imagine it could also be rather useful for some sort of magic duel as well. Relatedly, I read something by Jack on Google Plus about how he has spiced up the fighter class and simplified the thief for his Labyrinth Lord game. One of the abilities he added was a bonus to weapon damage every few levels (not exactly level divided by two in his system, but it easily could be with similar effect).

Does this risk the numerical inflation and illusionism of 4E? Regarding combat, as long as it’s not applied to the accuracy part of the equation, I don’t think so. I probably wouldn’t use this in OD&D, which has a very low numerical baseline (especially if you avoid +1 style magic items, like I do). But I could see it working well in a game that used B/X power assumptions perhaps.

As discussed by Talysman, this kind of “level divided by 2” number also fits well into the reaction roll when opposed by a similarly scoped difficulty number. He uses this as the basis for his X without spells class design paradigm:

The “without spells” approach is based on the reaction roll, and the interpretation of Turn Undead as a reaction roll: subtract the HD of the creature being commanded from the level of the character making the command, double the result, and use it to modify a 2d6 roll. On 9+, the command works.

The math is not exactly the same, but the similarity should be clear, I think. What if we generalized this kind of “reaction roll power” to every class? For example:
  • Cleric: turn undead (opposed by undead hit dice)
  • Fighter: intimidate or rally (opposed by enemy hit dice)
  • Magic-user: casting roll (opposed by spell level)
  • Thief: misdirect or dissemble (opposed by interlocutor hit dice)
And, if almost all capabilities are based on level divided by two, why not cut out the middle man and collapse level into that competency number? Attack progression does not happen every level in OD&D or Basic D&D. Saving throws do not increase every level. The only thing that increases every level is hit dice, and I think most of us can agree than D&D characters end up with too many hit dice anyways. An added benefit of this approach would be to do away with some of the ambiguity around the level numbers (as, for example, third level spells would be useable by third level magic-users).

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.)

Thief skills

Cropped image from Wikipedia

Eric recently did a clarification post on thief skills for use with B/X D&D, based on the idea of leaving the basic percentages as is (that is, working with the basics of the traditional system without completely rewriting it). The thief class in my Pahvelorn game is a version of the Greyhawk thief, modified slightly to fit the mechanics of the 3 LBBs. My general approach is similar to Eric’s, though I have not committed to writing how the skills are resolved (despite much rumination on the thief class in general). Here is an attempt at guidelines for thief skill use.

Thief skills include Climb Walls, Hear Noise, Hide in Shadows, Move Silently, Open Locks, Pickpocket, and Remove Traps. All of these are percentile skills, other than Hear Noise.

The following general principles apply to all percentile thief skills.

  1. Failing with a roll of 96, 97, 98, 99, or 00 means something goes wrong. The thief falls, the trap goes off, the lockpick breaks. Depending on the circumstances, the consequences of something going wrong may be dire (though a further saving throw might apply).
  2. Failing with a roll of 95 or less means the thief makes no progress, but does not suffer any other negative effects. Another attempt may be made. Most attempts take one turn, but see below.
  3. Succeeding by more than half means the skill use is quick. For example, if a thief has a 30% chance of success and rolls a 15 or less, the task was accomplished with alacrity. The exact amount of time required is up to the referee, but it should take much less than a full turn.

Move Silently is an attempt to approach or move past an enemy without being detected. Any action taken while moving silently automatically gains surprise.

Hide in Shadows allows a character to remain hidden even if someone is searching. Any action taken while hidden in shadows automatically gains surprise.

Regarding the two stealth skills, as specified above in principle 2, failing a Move Silently or Hide in Shadows roll does not mean that they are noticed (unless the roll is particularly bad, as laid out in principle 1). Neither Move Silently nor Hide in Shadows may be used in combat.

Characters other than thieves (or thieves that fail a stealth-related skill check) still have recourse to the standard chance of gaining surprise. This is 2 in 6 by default per encounter. It may be adjusted up or down based on the specific situation and character preparation.

Picking locks and removing small mechanical traps require tools and the special training of the thief. Larger traps must be disabled or avoided by player ingenuity. Traps may be discovered by using the same procedure for secret doors: 1 in 6 chance per turn (2 in 6 for demi-humans) given a 10′ x 10′ area, or by engagement with clues and explicit description.

Hear Noise functions exactly as the standard 1 in 6 listen at doors action, but with better chances.

Armor penalties apply to all percentile skills and are -20% if wearing chain and -30% if wearing plate. Hear Noise may not be attempted if wearing a helm.

I think this also reveals an interesting potential taxonomy. Listening and searching (the d6 checks) have to do with the state of the environment external to the thief. Whether or not a trap is present or a monster is beyond the door is not a function of the character. It’s either there or not. In contrast, all the other skills represent something about the character. Climbing, picking locks, removing traps, etc — these are all things that the character in some sense controls (particular task difficulty notwithstanding). Further, they are things that a character experientially can perceive the success of. They know whether they have failed to make progress climbing the wall or have fallen in a way that is different than a failure when attempting to use Hear Noise. Is there nothing there or did I just not hear it? This also explains why there is a Remove Traps percentage, but no Find Traps, as that is covered under the search action.

Thus, I propose the following final generalization: the d6 checks are rolled by the referee (to represent the objectivity and externality of the environment) and the percentile checks are rolled by the player. Even the Hide in Shadows and Move Silently skills, if looked at in the proper light, are not about being perceived by others, they are about the thief’s talent. Why shouldn’t the thief know whether or not they have successfully hidden in shadows? Thus, the thief can use the skill before they need to depend on it, unlike the surprise roll, which always happens when the thief is already potentially face to face with danger.

Note that this approach is more forgiving in several ways than the guidelines in Supplement I: Greyhawk, which specify (page 5):

The ability of a thief to climb is also a function of his level. There is a basic chance of 13% that a 1st level thief will slip and fall in climbing. With each higher level attained by the thief this chance is reduced by 1%, so that a 10th level thief has but a 4% chance of slipping.

And, regarding Open Locks, Remove Traps, Pickpocket, Move Silently, and Hide in Shadows (page 11):

A score above the indicated percentage means failure, and no further attempts may be made.

Hexagram Path of Guile Draft

The path of guile is about using misdirection and cleverness. The primary mechanic behind most path of guile traits is the saving throw, as many of them focus on avoidance of bad outcomes (climbing is the avoidance of falling, stealth is the avoidance of being noticed, tumbling is the avoidance of gravity, etc). Most tasks associated with the path of guile may be attempted by all characters (though a few cannot be attempted without training, like picking complex locks or interacting with some ancient technology), but characters with path of guile traits have several important advantages.

Tasks that benefit from guile traits make use of the general saving throw, which in most cases is a character’s worst saving throw value (considering all five saving throw categories). Characters with guile traits, however, use their best saving throw value for these tasks.

Additionally, characters with points in guile traits may size up and prepare for a situation before actually attempting to use a guile trait. The roll happens after the preparation is complete, but before the character must decide whether or not to actually attempt the task. In other words, the player will know that something will be successful or not before trying (and presumably will not try if it has been decided that the task will fail, though the time spent preparing is still expended).

For example, a character may be confronted with a sheer mountain face. They could spend some time carefully examining the obstacle, and then make their roll. The sizing up and preparation only remain valid as long as conditions do not change. Thus, if a character is halfway up a cliff when enemies begin firing arrows, another climb saving throw will likely be required to avoid falling.

In the climbing example, all characters would be able to attempt the action, but they would not be allowed the benefit of the sizing up procedure, and they must use the standard general saving throw rather than the best general saving throw. They must roll their saving throw after taking action, and let the dice fall where they may.

If you think that such knowledge would not require a full turn, consider if the task itself is really deserving of a check at all. Is this something that anybody would reasonably have trouble with? Is the chance of failure an interesting hazard within the context of the game? If the answer to either of those questions is no, the action should just succeed.

At first glance, it might seem more logical to use ability checks as the basis for guile traits rather than saving throws. However, that would make ability scores too important, and greatly increase the variance of starting competency. By using saving throws, improvement is gradual and level-based.

I’m on the fence about the name of the perception trait. That really is the correct name, but I’m tempted to call it listen just to distance it from the 3E perception skill. I’m also still on the fence about including the last chance trait at all. So let me know if you love it or hate it.


  1. Perception. Save +T to notice details such as noise behind a door.
  2. Stealth. Save +T to move without being noticed.
  3. Devices. Save +T to manipulate small mechanical devices.
  4. Climb. Save +T to climb a sheer surface.
  5. Assassination. +T surprise attack damage dice. Poison use.
  6. Tumbling. Save +T to avoid falling damage if < T x 10'. Free unarmed parry.
  7. Antediluvia. Save +T to utilize artifacts from before the deluge.
  8. Tracking. Follow trails left by creatures. Poison extraction.
  9. Last Chance. T + level % chance to pass a failed catastrophic saving throw.

Perception. This allows characters a better chance to notice details, such as listening for movement behind a door or searching a room for secret doors. Using the perception trait in this way always requires 1 exploration turn (10 minutes). Note that this trait should not be used to decide which clues or details to reveal. Instead, it is a measure of a character’s thoroughness. Perception also grants a saving throw to avoid surprise (though this only works for the character, not companions as well).

Stealth. Save +T to avoid detection when hiding or moving silently. A situation may be sized up for stealthy action beforehand. Hidden characters may attack with surprise.

Image from Wikipedia

Devices. Save +T to pick a lock (requires tools), manipulate a mundane mechanical device, or disable a small mechanical trap (such as a spring-loaded poison needle).

Climb. Save +T to climb sheer surfaces. Note that no saving throw is ever required to climb something like a ladder, unless under great stress. A climb may be sized up.

Assassination. +T to hit for attacks from surprise. If such an attack is successful, T extra damage dice are rolled. In addition, characters trained in assassination may apply poison to weapons without danger to themselves.

Tumbling. Save +T when falling up to T x 10 feet for no damage. On a failed save, falling damage is halved if the distance fallen is less than T X 10 feet. Additionally, tumbling grants one free unarmed parry per turn (which does not stack). Tumbling also includes general training in acrobatics (see skills and ability checks).

Antediluvia. Save +T to activate or use artifacts from before the deluge. Some such artifacts require a successful check for every use, some require only one check to decode, and some are totally incomprehensible without a minimum degree of antediluvia.

Tracking. Save +T to follow the path left by others previously. Tracking also allows up to T doses of poison the be extracted from poisonous monsters up to T hit dice in strength.

Last Chance. T + level % chance to succeed on a failed catastrophic saving throw, which includes things like dragon breath, poison, or the gaze of a medusa. Last chance may not be used for trait saving throws. For example, a 15th level character with last chance 2 will have a 17% chance to succeed on a failed saving throw. (Thanks to Ed Dove for the name suggestion here.)

Miséricorde from Wikipedia

3 LBB Thief

Hokusai Ninja

The thief class was not included in the original Dungeons & Dragons boxed set. The only classes at the beginning were the cleric, fighter, and magic-user. The thief was introduced in the first add-on product, Supplement I: Greyhawk (which, despite its name, is a collection of new game options rather than a setting). (Note regarding the image to the right: the oriental style is not really appropriate for Pahvelorn, but it’s really hard to find a good public domain image that evokes the thief archetype. Submissions welcome!)

Greyhawk also introduced a whole host of rules which will be familiar to players of later D&D (different hit dice for different classes, difference dice for different weapons, more influential ability scores) but which differ rather drastically from the game as presented in the three little brown books. If you play with all the rules changes in Greyhawk, the game begins to resemble proto-AD&D.

I really like the thief though, and want to include it in my otherwise “3 LBB only” setting. It only requires a few minor tweaks to fit in. Most of the following details come from Greyhawk unchanged. The divergences are noted.

  • Combat ranks: as cleric (steps based on 4 levels; 1-4, 5-8, etc)
  • Saving throws: as magic-user
  • Prime requisite: dexterity (bonus or penalty to XP like other classes)
  • Hit dice: as magic-user
  • Strike silently from behind: +4 attack, +1d6 damage per combat rank
  • 3rd level: 80% chance to decipher obscured treasure maps
  • May cast spells from scrolls with a successful save versus magic
  • 10th level: may use scrolls of all but the most powerful spells reliably
  • Name level is “Master Thief” at 11th level
Skills by level: climb sheer surfaces, open locks, remove traps, pickpocket or move silently, hide in shadows, hear noise. As per Greyhawk; just look the values up in the booklet or ask me. The progressions could probably be rationalized (I’ve seen several such approaches on blogs and forums), but my goal here is not streamlining so much as interpretation in the light of the original booklets (though I did make a few minor changes).
With the exception of picking locks and removing traps (see below), thief skills are not unique to thieves. Anyone may attempt to move stealthily or listen at a dungeon door. Thieves, however, are the only class that gets better at these things. Also, in most cases, the abilities function as a saving throw. That is, where a character of another class would fall, a thief gets a climb sheer surfaces chance. Where a character of another class would be noticed, a thief gets a hide in shadows chance.
WEAPONS (Greyhawk page 4):

Thieves can employ magic daggers and magic swords but none of the other magical weaponry.

Thieves may use any mundane weapon in my game. They may use magic daggers and swords to their full potential. Magic weapons other than daggers and swords count as magical for determining if certain creatures (like golems) can be hit at all, but do not grant any mechanical bonus to the thief. For example, an axe +1 would not get a +1 to attack or damage when wielded by a thief, but it would be able to hit monsters that can only be damaged by magic weapons.

ARMOR (Greyhawk page 4):

They can wear only leather armor and cannot employ shields. 

Wearing armor heavier than leather will result in penalties to thief skill rolls. Some skills may not be attempted or are penalized while employing shields (preternatural climbing and striking silently from behind for sure, and others by context).

TRAPS (Greyhawk page 4):

remove small trap devices (such as poisoned needles)

The thief ability to “remove traps” is not an arbitrary trap deactivation skill, but rather a limited skill to disarm small mechanisms.
SCROLLS (Greyhawk page 4):

Thieves of the 10th level and above are able to understand magical writings, so any scroll that falls into their hands can be used by them — excluding spells which are clerical in nature. However, with spells of the 7th level and above there is a 10% chance that the effect will be the reverse of that intended (due to the fact that even Master Thieves do not fully comprehend such great magic). This reverse effect can be known only after the spell is read.

Well, first thing, in the 3 LBBs there are no spells of the 7th level and above (there may be magic more powerful than sixth level spells, but it is not the kind of magic that can be prepared in a spell slot). So, by those rules, the 10% chance of failure would never come into effect. So I have decided to extend the use of spells from scrolls backwards to lower levels, given a successful save versus spells (failure miscasts the spell and consumes the scroll).

The ability to use scrolls (unreliably) at lower levels is the only substantial change I have made to the class. I think it is reasonable because it encourages fun play (“roll to see what fun way the thief is going to screw this spell up!”) and means that players of thieves will be more likely to get some use out of scrolls (since few characters reach name level). I don’t think this “save to cast from scrolls” steps on the magic-user’s toes because it will always be more reliable to give scrolls to magic-users (since they never fail when casting a spell from a scroll). At tenth level, thief scroll use also becomes reliable, though the thief never learns how to scribe scrolls and thus still must still find them or procure them from magic-users. Also, the same societal pressures regarding diabolism and black magic apply to thieves, especially since thieves don’t usually advertise any sorcerous power they may possess. Also, many magic-users will not look kindly on their secrets being stolen.

Though I have tried to stay within the parameters of the class as written in Greyhawk, my interpretations are heavily influenced by the following sources.

You can also check out my previous attempt at a thief class rewrite.

2012 10 30 edit: see also my clarification on thief skill use.