Press The Beast

Thief Design & The Super Thief Class

Recently, I posted my thoughts on spell casting characters and shared what I intend as a solution for the specific bugbears I’ve had with these classes in my own games. However, among the various character classes the one I find the most fun to facilitate as a referee is the Thief. I also think I’m quite good at it—it’s a weird thing, but I pride myself on running Thieves particularly well. Unlike Wizards, which I have flipflopped and struggled with for years, the Thief has been a relatively straight course to my present understanding of the character. In this blogpost I’ll provide a blogerature review on the Thief Class, present some solutions which have been used in various games, describe my current Thief class, and then provide guidance on running the Thief from the referee seat.

The Thief class has been problematized or criticized in hundreds of blogposts, and implicitly in many games. I believe GROGNARDIA's fat thumbs helped tip the scale toward the Thief’s present notorious reputation. Both it and the Cleric are viewed as awkward appendages of the “original game” though, as Delta puts it in the linked blogpost, people take issue with them for different reasons: Clerics have a “flavor-setting” problem, while Thieves have a “mechanical” problem.

This mechanical problem is usually understood like this: with the Thief came a skill system that provided codified mechanics for tasks such as climbing or hiding that were previously resolved via rulings. The reason this is bad is because it can limit the field of play; if the Thief can “hide in shadows,” are other characters able to do so? This is typically the most common objection, which you will see noted again and again on various blog posts. Relatedly, there are questions of how these skills ought to operate in the fiction. A common solution is to present the Thief skills as supernatural as a way to reconcile the Thief's codified access to these abilities compared to the other classes. The Supernatural Thief is a rationalization of the inherent problems of the class. An interesting solution, though one I often find aesthetically twee in practice. The most interesting approach I've seen is Grognardia's comparison of the Thief to Jack of Shadows.

For my part, I think this inconsistency is basically true—the skill system really does clash with the assumptions of the “original game,” and doesn’t properly slot into the existing system in immediately intelligible ways, hence the profluence of blog posts—, but I also think that this inconsistency doesn’t matter much. It’s hardly irreconcilable, the Thief uses their fixed hiding mechanic and is presumed to be better at it whilst others use referee judgements (I provide notes on resolving this issue later). If you can't figure this out then you have poo brain.

Another major complaint, also valid, is that Thief skills begin at very low percentages of success. This is made even worse in Delta’s Target20 where the Thief is, basically, unusably bad. Delta's urge to use the d20 relates to the excessive granularity of the d100 used for Thief skills, another common complaint (Not an actual problem).

There are many solutions presented to the Thief class (Seriously, there’s a gobsmacking amount). The first line of attack is the conversion of the d% skills to the d6 skills, used in Dolmenwood and other sources. This is basically inoffensive, and tends to make the Thief slightly more likely to succeed at tasks at low levels while making the dice used more coherent with the rest of the game (Not an actual problem). A close cousin of this is the method used in Lamentations of the Flame Princess, where the Thief is provided various skills on a d6 which they choose to invest in. This method is very frequently borrowed, providing the benefits of the prior method alongside customization, a virtue of dubious quality that gamers love and crave. These methods have proven generally popular, though neither resolve the basic incoherence of Thief skills as part of the field of play.

In practice, the Customizable Thief tends to create lopsided characters, who immediately excel at one or two skills and then slowly fill out the rest. It is a method I’ve considered and tried out, and have found to be generally distasteful for more or less ideological reasons. While GROGNARDIA and Delta describe a mechanical issue with the typical D&D Thief, the Customizable Thief is imaginatively lazy and contradicts a fundamental joy in class based systems. The primary virtue of character class as a game element is to provide a simple package with progressive abilities (Meaning, “characterized by continuous improvement or advancement,” which I bother to specify because people have misunderstood this use of progressive in the past) that can then be interpreted by the player as they see fit. Indiana Jones could fairly be a Thief or a Fighter, and the joy of depicting this character is in using the basic canvas of the character class, perhaps eating the debuff (https://bluemountain.bearblog.dev/todays-advice-eat-the-debuff/) along the way, and playing the character. The mechanics are merely used for adjudication, they impute little beyond that and nothing about the character themself. To make a unique Archaeologist class is cool and maybe beautiful, but this risks making a vestigial appendage that does not properly slot into the assumptions of play (Take the risk if you are brave). To make an open ended “build-a-Thief” class is basically cowardly; it’s having the cake, eating it, and leaving nothing but the aftertaste of bland babyblue icing. Give a Thief a whip and a hat: you’ve made an Indiana Jones.

The Customizable Thief is the first step toward recreating RuneQuest, which has already been done for you in OpenQuest and Mythras with varying degrees of granularity. What I mean is that if you follow the logic of build-a-classes that have selectable advancement to the furthest extent you get a skill based system, and you ought to follow where your instinct leads and play one of those instead of a half-in half-out abomination. I find this problem more galling than the mechanical incongruence people typically struggle with, and this is related to my distaste with the skill based systems of 3e and beyond, or the bizarre abortive skill system present in AD&D.

Ahem. Another solution, one which I used in my Devil’s World Heroes campaign (after a rough start with the Target20 Thief) is the one presented in White Box: Fantastic Medieval Adventure Game. Here the Thief has a single Thievery skill that begins at 2-in-6 and goes up to 5-in-6 at the highest level. This one skill is used for all typical applications of the Thief. It’s a great solution, as it provides more ease of use than discrete skill systems at the table and makes the Thief much stronger at low levels. Delving Deeper uses a similar mechanic, though it stays fixed at 3-in-6, which works more or less equally as well.

Justin H. recommends this iteration of the Thief, and in practice it has been my favourite of the “appendage” Thieves which include skill systems (Idea for GLOG class: a Thief that uses prosthetics for their various skills… Sekiro appendage Thief; a task only Garamondia can do). It works very smoothly, and the general “Thievery” skill implies more latitude to interpret flexible applications compared to more discrete skill systems. I also recommend this approach if you want something approximating the classic Thief, but are scared of the d% and also want Thieves to be able to succeed at stuff for some reason.

Finally there are numerous other approaches which, broadly, fit into three categories:

A note on GLOG Thieves, of which there are a stunning variety: the Luck Thief was once very common in GLOGs due to the influence of an old GLOG Thief by the GLOGriginator. The Luck Thief generally relies on a universal skill system of some kind, so I must dismiss it out of hand (though you could get it to work with x-in-6 checks, a thought I am having at this very moment… In fact ,you could just have 1 luck point convert to the d6, d20, d100, etc, in different amounts, thus letting it influence whatever procedure the Thief is interacting with…)

Otherwise it is common for GLOGs to use selectable or randomized talents as Thief skills, which more or less resembles the pick list problems of the Lamentations Thieves. I'm not sure who started the Talent Thief, though I'm sure a GLOGster knows.

Of these options I find the Item Thief most novel and appealing, and I’ve tried to make it work numerous times, though I’ve struggled with properly establishing “the fiction” for these items. Jenx over at Gorgon Bones has the best attempt I’ve yet seen, though he provides only a starting point: something like 5 or 6 items with the same kind of iconic clarity would be required.

(Another approach worth noting is one from Phlox which approximates the Skills-as-Spells seen in an unpublished iteration of the original Thief class. It’s compelling as hell! though a bit heavy weight for my liking.)

In summary! the Thief has had so much written about it because the central contradiction between a set of codified skills limited to a single class and the tactical infinity present in RPG wargaming is intractable and therefore generative. I also think it is worth solving, which brings me to one final approach to the Thief. Philotomy’s Thief (Scroll down very far in this link to find it) integrates the various skills into existing mechanics such as searching for hidden doors, ambushing enemies, and avoiding surprise. x-in-6 skills for disarming traps or picking locks can be bolted on without too much issue—even if only the Thief can do those things, it doesn’t step on other classes’ toes in the same way as “Hide” or “Climb,” which are much more universal actions performed even by monkeys with alarming frequency. These abilities then grow when the character achieves 4th and 8th level.

What this means is that the Thief is just run as normal with no need for the player to press the “I use a skill” button, and the referee processes the odds of springing an ambush, avoiding surprise, etc, differently for the Thief than other characters. At first I worried this might be difficult to remember, but the mechanics are simple enough in Ye Olde D&D that it proved trivial to apply these exceptions even with 10+ players in an online game.

This Procedural Thief is simple, slots into the basic mechanics of D&D, and doesn’t step on the toes of any other characters; the difference between Player Button and Procedural Adjustment is enormous in practice, improving both the smoothness and effectiveness of Thieves, particularly in a scouting role. After using percentile Thieves, x-in-6 Thieves, Whitebox Thievery Thieves, Target20 Thieves, various GLOG-esque Thieves, 5e skill monkey Thieves, and more in the games I have played, this has thus far resonated most. While they all more or less work (except the Target20 Thief), the Procedural Thief is the only one that coheres neatly with the rest of the game. Like the Fighter, it interfaces with extant mechanics and feels "of a kind" with the rest. This integration and elegance is very important for me in design, especially in RPGs where, at the table, an element of a game is only as good as its application; and application is a function of rememberability.

—

Super Thief

Assume your OD&D/BX progression of choice for HD, attack bonus, and Saves, replacing the skill system with the following:

Thieves are particularly adept at dungeon crawling, with various talents that grow as they gain in level.

Their furtive feet allow them to ambush foes with the following odds:

Level Ambush
1-3 3-in-6
4-6 4-in-6
7+ 5-in-6

When striking from an ambush, they add 4 to hit and deal double damage. At 4th level, this bonus increases triple damage, and again to quadruple damage at 8th level. Defensively, they are able to perform a backflip or another evasive maneuver to disengage from combat at full speed with the following odds:

Level Backflip
1-3 3-in-6
4-6 4-in-6
7+ 5-in-6

Animal-esque senses heightened by years of urchin living mean they are surprised on a 1-in-6, rather than a 2-in-6. Furthermore, they may hear noises or eavesdrop, as well as search for secrets, with the following odds:

Level Hear Noises Find Secrets
1-3 2-in-6 2-in-6
4-6 3-in-6 3-in-6
7+ 4-in-6 4-in-6

Fingers quick and slick as darting eels excel at picking locks, disarming small mechanical traps, and performing sleight of hand—this encompassing the sundry skills of the Thief: palming objects, card tricks, hiding a rose in a sleeve, applying poison to a rose thorn, etc—with the following odds:

Level Locks & Traps Sleight of Hand
1-3 2-in-6 3-in-6
4-6 3-in-6 4-in-6
7+ 4-in-6 5-in-6

Both picking locks and disarming traps take 10 minutes to accomplish.

They are also unparalleled climbers, with a 5-in-6 chance to scale any surface. The odds may be adjusted down based on various factors, each decreasing the odds of success by 1. Consider the following adjustments:

A similar check may be made for acrobatics, brachiation, tight rope ambulations, rooftop hops and vertical absquatulations or, perchance, auto-defenestrations, etc. Should a Thief pass the check they are able to fight with a one handed weapon or throw daggers while climbing.

Thieves excel at retreating in dungeons, though only while alone or in a group composed entirely of Thieves. A pursuing monster will only continue to chase on a 1-in-6 after losing sight of the Thief, rather than the standard 2-in-6 odds.

While in the wilderness, a Thief serving as a scout may use their expertise to increase the party’s odds of evasion by a percentage equal to their level. If there are multiple Thieves scouting, only the highest bonus applies. Likewise, they may use their “find secrets” odds to look for hidden features in a given outdoor tract.

A habit of caution means that Thieves always carry what they need. They have 3 Gear Bubbles which may be checked off to produce a mundane adventuring item worth no more than 25gp. All Gear Bubbles may be cleared by spending 25gp at a village or settlement.

Finally, practice deciphering treasure maps allows Thieves to use spell scrolls.

Design Notes

The main addition to Philotomy’s Thief are the rules pertaining to retreating whilst in a dungeon and pursuit and evasion rules. The rules assumed are sourced from FMC and BECMI respectively. The only thing difficult to remember here is the percentage bonus to wilderness evasion, which a good player can help remind their ref of; while the bonuses are relatively minor there aren't typically any other ways to improve this score, so it is significant. You may have noticed that the Thief gains some of the wilderness skill associated with a Ranger: while I don’t hate the idea of an outdoorsman-type, the Thief is in an awkward position where its skills are best used in the dungeon, then diminish in value in the wilderness. Giving some wilderness expertise to the Thief is, I think, totally fine, and helps to preserve the Thief’s utility outside of the dungeon.

The climbing rules also differ, being broadened out into a more general procedure, and beginning at 5-in-6 without any growth for the Thief as they level. This is basically a simple Thief only climbing procedure. You could also use Ms. Screwhead’s very famous parkour rules if you like Being Funky.

Other additions include the backflip, which I’ve included as a feature for ninja-esque classes before and really enjoyed. I like to imagine the Thief as a kind of switch-hitter, who can alternate between an offensive and defensive posture and apply pressure to enemy lines where needed. The backflip ability helps facilitate skirmishing and hopefully encourages players to take riskier engagements than they otherwise would; an important corollary to this is that I process movement before any attacks, so an effective escape can allow the Thief to avoid the consequences of a poor engagement.

While the above doesn’t provide any commentary on Thief HD, I’ve come to prefer Thieves with HD as Wizards (Either d4 per HD or the slow HD progression in OD&D). Using HD as Clerics is a very common house rule, and it isn’t wrong, I just think that the class is balanced enough with its various skills and very low XP requirements without the boosted HP, and actually becomes a little too meaty.

The main intention here is to give both the player and ref a bucket of procedural modifications to help adjudicate the Thiefly business players like to get up to.

In an ideal world, most of these rules would not be in the Thief class section itself but instead in the actual rules for exploring, ambushing, evasion and pursuit, etc.

Running Thieves

The Thief requires a generous ref to excel. In games with Thief skills, it can be extremely tempting for referees to offload all Thiefish derring-do to a throw of the bones: resist this urge! Be generous with rulings for stealth, climbing, sleight of hand, etc, only using dice where there is a real risk of failure or real stakes. This is, of course, pretty typical referee advice; it is also how the intractable problem of Thief skills can be… tracted, at least in play. While the theoretical issues of Thief skills remain, those don’t actually have to be a problem at the table, and haven’t been for many tables for many years. This is the main piece of advice: the game is a kind of shared dream, and this must be prioritized above rules or procedures.

The introduction of verticality in combat and dungeon environments is also a huge boon to the Thief. By adding ledges, balconies, chandeliers, overlooks, multi-story buildings, cliff sides, towers, etc, to the dungeon or wilderness the Thief’s climbing ability suddenly becomes extremely useful for both fighting and navigation. In my DWH campaign, the Thieves of the parties constantly climbed to get bonuses from plunging attacks, shut down enemy archers, intercept reinforcements get a line of fire, etc. Without exaggeration, this occurred in some form or another almost every session that had combat.

Becoming comfortable with using running skirmishes provides a chance for Thiefish stealth and mobility to shine, as they are better able to choose engagements than clumsier classes. By “running skirmishes” I mean, well, a skirmish in the proper sense, rather than in the D&D sense (which often just means “a fight where everyone kills each other to death”). For instance, if the party engages with a foe in a corridor and performs a full charge (parties often do this), the OPFOR may retreat and use projectile weapons, then follow as the players regroup. After an initial engagement, one or the other force may fall back, and here begins the running skirmish. I’ve had fights occur over multiple dungeon rooms, with encounter rolls and keyed encounters altering the engagement on the go, and in this situation Thieves are extremely useful. Simply taking a corner on a retreat can allow for a new back attack, and that could change the tide of a battle.

Likewise, consider movement seriously. As a light armoured class the Thief will have a higher movement rate than many foes. If you run this right, players can use the famous techniques of Zatoichi: retreat, draw out a smaller portion of the enemy force, then fight against this isolated element until the rest begins to close in; retreat and do it again, using terrain and movement to cut off lines of attack and take only favourable engagements. Canny foes may simple regroup and advance together, avoiding the poor engagement—more power to them, the Thief can just retreat! Havoc once used tactics like this in a Wolves Upon the Coast game to defeat a better armed and more numerous force after the rest of the party had already been slaughtered.

Consider the deeds of Gray Mouser from Leiber's Fafhrd and the Gray Mouser stories, Cugel's deeds from Vance's Dying Earth novels, and Borlbo Brorggin's deeds from Tolkein's The Horbott coughs.

Prison escape scenarios are awesome for Thieves. Same with heists.

What about out of combat? Combat is The Thing I Like, so I tend to think about it the most. Ask someone else about other stuff!!!

Thanks for reading <3 <3 <3