OSRPGtalk

NPCs: Central Casting

March 30, 2026 / TAFL / 0 Comments

I’ve been struggling with crafting a cast of characters for a campaign start area. Then I realized I could approach it in a different fashion and suddenly the NPCs started springing up left and right. I’ve no idea why the change let the creative juices flow where the approach I’ve used for decades suddenly dried everything up. Ah, well, ideas are bubbling up and I figured it worth sharing.

I was thinking about a friend who has been an inspiration for an important NPC. I realized that I wasn’t really trying to put her into the setting, I was using her as an inspiration and it was much more like I was casting her to play that character on the world stage. That brought the epiphany — I’m a director casting people in parts for a play.

I found that this approach works well in two directions. That is, taking an NPC role and figuring out who I’d cast to play that part; and also considering a person I know in some fashion and figuring out what sort of role I’d cast them in. Note that I’m not considering basing the NPC on the person’s actual personality, I’m just considering how I’d cast them in a play based on how I think they’d play the NPC.

I’ve never overtly considered this as an approach to NPCs prior. I’ve rarely even thought of correlating a known person with an NPC in my games. I expect there are people who’ve taken this approach, though I’ve not encountered discussion of it anywhere.

Layout: Tables

March 26, 2026 / TAFL / 0 Comments

One layout feature found in abundance in RPGs, especially in old school systems, is the table. A table can organize information in a fashion that efficiently conveys the ideas involved, providing concise statements of information. The Turning Undead table, for example, lays out the chances for clerics to turn or destroy undead creatures while also providing a gauge for how tough the different undead types are to encounter.

Tables can also help decide among a variety of options via dice rolls, as with encounter tables for wilderness travel or random dungeon encounters. Those are the tables I want to look at here. I’m interested in taking the basic tables long used in rules systems and expanding on them a bit.

I’ve written before about wanting to use some large table types. Reality sat me down for an intervention and showed me that laying out such large tables wasn’t quite practical, even using a two-page spread. The cells on the table shrink to barely usable with such a size. So, I’m here to press more conventional tables into service.

The many systems that have been published over the decades all use tables of some sort, so the usage of tables to provide information in RPG systems isn’t new or unusual. D6 tables, D10, D20, D30, D100 tables abound in the published literature. While they all can prove useful, I’m mostly interested in those found in old school and OSR materials, specifically those using multiple dice in some fashion.

The venerable reaction table, for example, calls for 2D6 rolls. An OSR system favorite is the D66 table. Each of these types can be expanded to increase the range of results in ways that can serve a setting better and provide a bit more flavor to the rules.

The reaction table is a a single-axis table, meaning the roll indicates which row on the table to use to find a result. We can, however, expand the number of rows to add some interesting utility to the table. We can expand both the top and bottom of the table, so the lowest result is less than the lowest dice roll of 2, and the greatest result is greater than the highest dice roll of 12.

What this does is allow for a wider range of results and move the more extreme results — hostility resulting in immedate attack and all-in friendly helpfulness — outside the bounds of a default roll. To garner a result from the expanded rows, the roll has to be modified and that provides a chance for the players to wrangle modifiers in their favor. A look at a potential reaction table:

The five results from the standard table have now expanded to nine results, which allows a bit more nuance in play. It also allows for added interactions to shift results for better or worse, facilitating more interaction before either extreme can occur.

Now let’s look at an encounter table. By expanding the number of rows, we can reflect increased danger in deeper levels of a dungeon or risk increasing the more time is spent in a location. The more dangerous the area or longer the PCs have been in a risky place, the greater the modifier to the roll and the more dangerous the entries on the table become. A 2D6 table that has goblins as the most likely critter encountered now has gnolls as the most likely with a modifier shifting roll results, and the possibility of owlbears as the most powerful.

Now, we can also expand the number of columns available. Using a D66 roll, or adding a D6 alongside a 2D6 roll, allows us to shift things in a different direction. The D66 becomes a possible D68, so to speak, as a modifier of one or two column shifts moves us into options not part of default rolls.

Expanding a table in both directions allows us to color the results in more than one fashion. If we add rows, we can tie modifiers for that to one element, say level of dungeon in a dungeon crawl. Then, we can add columns to the table that reflect the time spent in the current level and how that increases the chances of stirring up more of the inhabitants. Or we can have the expansion in one direction reflect moving into a different area of the level that has more active or dangerous inhabitants. Or an increased chance of cave-ins or geothermal vents spraying steam or whatever.

In this fashion, the regular 36 possible results of a D66 table can be expanded to, say, 64 possible results by allowing for modifiers of +1 or +2 in each direction. I’m looking at using a 2D6x6 table format to take advantage of the bell curve, so that involves 66 possible results of varying likelihood. (The 2D6^2 table format I wanted to use — 121 possible results by default and bunches more with expansion — just doesn’t fit well in a rulebook, alas.)

This is, of course, just a cursory look at expanding tables and how they could be used. The true utility has to be worked out in game materials and used at tables to see how they’re best used. I’ll be hashing out expanded tables in the materials I’m preparing, so concrete examples will be found therein.

The Old School: Hallmarks II

March 19, 2026 / TAFL / 0 Comments

More thoughts on what makes for an old school experience. Some of this is simply expansion on a topic mentioned in Hallmarks I because I think the added commentary adds something to the discussion. As I return to thoughts of playstyles, I expect more details and notions to arise and I’ll toss them up here. Perhaps I’ll get around to pulling all of it together, properly organized, and then post it.

I’ve offered that settings are what define the world in which PCs operate. Any exceptions to normal physics are delineated so that players can interact with a consistent basis of understanding; gravity is going to make a long fall quite detrimental to a PC’s health…although here are the specific ways in which that can be ameliorated. Players are able to predict, in general terms, how things work and thus gauge how risky a course of action will be for the PC.

Settings also tie in with other elements. PCs are expected to be grounded in the setting, with some nominal ties to specific elements. The most basic is likely that of origin; where the PC comes from establishes them as *part of the setting* from the outset. The PCs aren’t foreign to the setting, they’re just another one of many elements that are part of the setting.

That’s important because old school campaigns are about the setting. The system rules didn’t include anything about pre-planned “character arcs” or long, involved background stories because the campaigns were expected to be about what’s going on in the world and how a group of random characters engage with it. With everything involved being part of the setting, it’s the setting that is featured in play. It’s this that makes every PC replaceable in a campaign, with it not unusual that a player could begin a campaign and then change PCs over time without that campaign ending.

A system begins to lose old school flavor when the approach to PCs loses tight definition in an effort to provide a lot of customization. As the ancestral wargames offered up a “fighter is a fighter is a fighter” approach, the old school RPGs proceeded with that idea, I reckon, because it kept the PCs grounded in the setting and helped to make them replaceable. The expectation was that a PC becomes memorable based on how it’s played with the choices the player makes defining the unique nature of the character. Thus, a character is only fully defined through play and that’s where the important differences should arise.

I’ll even offer that this is part of developing player skill, in part — take this standard-issue archetype and make something more of it *in play.* So much of the play experience and character definition happens “off-sheet,” as is often said. Players can’t rely on special combinations of customization options to provide “easy button” solutions to play challenges, a build isn’t a substitute for creative solutions the player cooks up. The character sheet acts as a toolbox for the player, not a source of solutions; interrogating the fiction by asking questions to find elements that can possibly be exploited to help overcome the challenge at hand is far more important than looking to the sheet for an answer.

Old school play is very much in the “play to find out” camp. Even in tournament play, while each round of the tournament provided a continuation of a general plot, the individual installments often didn’t dictate an explicit sequence of encounters and events. G1-3 Against the Giants, for example, moved the overall plot from round to round from G1 with hill giants to G2 with frost giants and then to G3 and fire giants. When playing through G1, however, there was no expected order of encounters as to how it would play out, nor do I recall any instances of “the PCs have to do X then Y” or similar to finish the module.

Forgotten Temple of Tharizdun, with its order of battle roster, makes this even more apparent. PCs can have denizens appear during a fight with others and then, when arriving at those critters’ regular quarters, there won’t be any critters to be found. The module doesn’t decide if the critters will be encountered during the large melee, in their own quarters, or somewhere in between. That lack of expectation and reliance on the GM to sort the happenstance is representative of an old school approach.

Old school systems also expect exploration as an important component of play. this hearkens back to the importance of setting, because finding out what is in the setting and what’s happening in the setting is core to the experience. Exploration becomes important as a way to find out about the setting, as the setting provides for play. A solid procedure for exploring sites thus seems to be a requirement in old school playstyles.

I hope my thoughts here can spur thought and spark ideas from other people. I can’t say that my idea of the hallmarks of old school styles are necessarily the most accurate or thorough, certainly. I can say that thinking on the subject has helped me deepen my understanding, certainly.

First Look at a Witch Class

March 5, 2026 / TAFL / 0 Comments

I’ve roughed out a Witch class for the DA project. The class will be quite similar when it appears as an LJ class.

I’d love to hear ideas on more powers for Witches to be able to develop, as the list I have seems a bit short. I’ll be basing the spell descriptions and lists on those powers.

Feedback on the abilities described also welcome…shoot, feedback on all of it is welcome, at this point. I think I’ve conveyed the gist of the class reasonably well and I’d like to know how it comes off to you.

https://osrpgtalk.net/wp-content/uploads/2026/03/DA-Class-Witch.pdf

Oh, yeah, I can put documents up for download….

March 3, 2026 / TAFL / 0 Comments

I realized that I might be able to post other media in addition to images. Sure enough, I can put PDFs and other types of file up. Yeah…I’m not quick to check on this sort of thing.

Anyway, I still have no idea how to go about accessing those through the guest interface, so I’ll check to see how difficult that is.

Design: Travel Navigation

January 29, 2026 / TAFL / 0 Comments

I’ve been wrestling with how to make travel the equivalent of dungeon exploration in activity (my bespoke system is entitled “Legendary Journeys,” after all). I think I cracked that nut.

It’s typical to have a navigation check when parties are wandering in the countryside. Their fearless leader checks to see if they’ve wandered off course during the day and are lost. Well, what if we change that up a bit?

Suppose the navigation roll does two different things. First, it slows the party by some amount, measured in time or movement points or whatever your system uses to measure progress when traveling. Assume the party gets it sorted out by discovering the error and getting headed in the correct direction from where they are. This, I think, can offer that “wandering in the wilderness” feeling without the wholesale “getting lost” condition.

What it also does, secondly, is provide an opportunity for the party to find/stumble upon Points of Interest they would otherwise have missed. So, if we make the navigation check a bit more difficult so the party wanders a bit more often, they can find a lot more stuff to eat their faces…er, explore.

Each time the party travels through the hex, the navigation check can result in the party traipsing along a different path and finding something new. If we want to, we can restructure the navigation roll to reflect that each trip through is likely to be a bit different which allows for new discoveries. Indeed, we can structure things so each trip through a hex is just different enough that new discoveries are possible.

That would provide more possible interaction and greater interest in travel situations. The wilderness and travel should just be something to endure before getting to the good stuff at the destination. Traveling should be an adventure of itself. It should be possible to have the journey to a location to be more involved than what happens at the location, which upends the usual circumstances of play.

Update:

I’ve been stewing on this topic for a long time, wanting to end with a sub-system that offers up travel-as-adventure instead of travel-to-reach-adventure. What I wrote above is just a part of that. Yes, offering variation in trips through a given wilderness space increase the chances of discovering new points of interest. That, of itself, doesn’t make a wilderness trek feel like a dungeon foray, though.

What that doesn’t offer is all of the interaction that can be expected in a dungeon setting. The tricks and traps and puzzles we find in dungeons are generally missing in the traditional wilderness travel rules. I think that’s a element we have to introduce to bring wilderness adventure up to the same level as dungeon adventure.

Providing more chances to find POI is just the beginning. Having multiple small dungeons to be found during travel doesn’t provide the same flavor as full-size dungeons. I think we need to have our encounter tables expanded to include more, plus add in pre-planned material for any given trip.

I think of the wilderness in a fantasy game setting as wild, in the sense that the kinds of things and situations found in folk tales/fairy tales are not only possible, they’re not uncommon. Spend much time traipsing about in the wild and you will encounter weirdness of some sort — fairies come through the Hedge, elder beasts from the Greenwood, people from other worlds, weather anomalies, malevolent spirits, enchanted pools, ancient shrines radiating power, and on and on. These are the things that can provide the same flavor in the wilds that a variety of rooms and chambers provide in a dungeon, especially when many of them invite direct interaction with the PCs.

To that end, I think the encounter rules have to provide as many opportunities for something to interact with as a party could expect to have in a dungeon. That’s not to say that every ten-minute turn of travel should include a check or reaching a new “room.” It does mean that there should be fairly regular instances of interesting things to do, whether measured by hour-long turns or opportunities per hex or in some other fashion. A trip should be filled with chances to engage with the unusual and interesting bits of the setting.

As there are many folks who approach dungeon settings as the Mythic Underworld, I think it just as useful to approach at least parts of the wilderness as the Mythic Wilds. Stretches of forest or swamp or plains that feel as if they have an interest in the PCs traveling through and aren’t taking kindly to the intrusion. I’ve already been thinking of the Little Gods, spirits of the place, that can take notice of travelers; this extends that to the entirety of a locale taking notice. An increase in encounters with hostile critters, say, with those critters being the most dangerous in the locale. Those critters trailing the party and causing issues. The party getting slowed repeatedly by fresh obstacles the terrain throws up in the way. Stopping for rest breaks or camping resulting in getting constantly harried by denizens of the forest that make rest impossible.

Even locales that aren’t Mythic Wilds can offer similar experiences. The lands immediately surrounding an ancient barrow that holds a malevolent golem possessed and animated by the spirit of a long-ago tyrant and warlord. Hauntings a-plenty, undead clawing out of the ground and shambling after the PCs, shadows flitting about out of direct sunlight and chilling the PCs to the bone. The PCs can figure out what’s at the area’s heart and try to put an end to it, though that just may result in the golem escaping it’s chambers and roaming the wilds to create yet more havoc.

Encounter tables should provide for much more in the way of flavor and challenge. I can see each area having designation of danger that affects the results on tables, or specific tables for each such area. As I mentioned previously, I also think the regular addition of pre-planned POI can make a major difference. Even if it’s a drop-in selected from a list of possible drop-ins, knowing that on a given stretch of travel something a bit more involved is going to crop up that’s already prepped makes GMing a trip a bit easier, and the players will get regular doses of more substantial and demanding play. Prepping a handful of events/encounters/sites for a given stretch of wilds in advance can go a long way to filling out travel in an interesting fashion.

Planning a trip thus gains quite a bit of flavor. Planning a route now involves deciding what known stretches of dangerous ground it might be better to avoid, though at the expnse of taking more time to reach the destination. PCs can try to find good sources about what to expect along the way and prepare for it. Planning a trip becomes the equivalent of planning a dungeon foray, as does the actual experience of it. An exercise in survival becomes an exercise in adventure.

On Evil Monsters

January 19, 2026 / TAFL / 0 Comments

There are topics that seem to rise again and again in RPG fora like undead horses to be flogged. This comes about because of the regular influx of gamers to discussions, so evergreen discussions are to be expected despite the eyerolls they garner from veterans. One such horse carcass is that of evil monsters, particularly humanoid monsters.

The argument against having such moves along the path of “innately evil humanoids don’t stand to reason because there’s going to be good individuals even in an evil culture and it’s horrible to say they’re all evil” and then to move on to how it’s inherently racist to have evil non-humans.

Stop. Stop right there.

The problem I see is simple: these claims rely on the notion that non-human critters are simply humans in funny suits. That what one can expect from human cultures is what one can expect from non-humans. That non-human thought and feeling and understanding is exactly the same as with humans.

I’ll offer up that non-humans aren’t humans in funny suits, that they are, indeed, non-human and that expecting the human experience to map onto the non-human experience is foolish and leads to bad fantasy. The non-human experience should NOT map directly onto human experience.

Now, for us, as humans, to be able to use non-humans in play, yes, we have to be have enough similarity between the two for us to use — for us to understand so we can portray elves and dwarves and so on — and still be different enough to set them apart as not human. (Yes, RPG play has long suffered from elves and dwarves as simply humans in funny suits.) The Venn diagram humans and non-humans should never be a perfect circle, even if there’s relatively little space separate.

So I’ll observe that we humans can’t comprehend all of the non-human experience, so we’ll never understand why some non-humans are what we find to be irredeemably evil. They’re not poor, downtrodden souls being wrongly persecuted for being in the wrong place, they’re inherently evil creatures creeping into human domains and a threat to all humans.

Even if they’re bipedal and human-like in form, they aren’t human. They’re not stand-ins for oppressed peoples. They’re non-human monsters and a threat to us all.

Non-human creatures that aren’t evil should also be viewed as not being fully understandable by humans, and thus by the players at the table, for much the same reason. They may not be the inherent threat the evil monsters are, yet some of their behavior should certainly be inscrutable to humans to preserve the weirdness of the fantasy.

To pay the Joesky tax, I offer up the Shadow Curs, as yet untested kobold substitutes:

NO. APPEARING: 10-100
ARMOR CLASS: 7
MOVE: 12″ / 9″
HIT DICE: 2-5 hp
NO. OF ATTACKS: 1
DAMAGE/ATTACK: 1-4 or by weapon
SPCL ATTACKS: none
SPCL DEFENSE: none
MAGIC RESIST: standard
INTELLIGENCE: low
ALIGNMENT: Lawful Evil
SIZE: Small (2 1/2′ – 3′)

The curs have two forms, that of a dog and that of a dog-faced humanoid, similar to a werewolf. When in the dog form, a cur has capabilities as a dog. In the humanoid form, a cur has opposable thumbs and can wield weapons accordingly. Transforming from one form to the other takes but a couple of seconds and doesn’t affect the cur’s ability to move and fight.

In either form, a cur looks like a mongrel dog, with fur and fangs accordingly. Small packs are families led by the sire and dame. Large groups are composed of multiple family packs and typically led by the largest of the pack leaders.

Design: Surprise

November 27, 2025 / TAFL / 0 Comments

Surprise is one of the bits of system that didn’t hold up under scrutiny and is worthy of a post on its own, I think. There’s a whole lot of tinkering going on with how surprise operates, beginning with how it feels and including what purpose it serves in the system.

Let’s begin with a look at how it plays out in RAW. I’ve long wondered why groups get surprised one-third of the time upon encountering another group or monster. Reading through all of the OG materials has offered no reasoning to support that rate, so this appears to be one of the things that has been perpetuated simply because of tradition. I decided that I’m not down with that; one-third of the time is simply more than I can sustain disbelief for–it just doesn’t serve my fantasy to have competent adventurers getting flummoxed by running into beasties that often.

To that end, I’ve moved to lesser chances, closer to 25% of the time. If I recall correctly, the odds are about 27%, using the dice roll I’ve landed on. I’m using the 2D6 roll that I’ve pressed into service for some other purposes.

I’m also all for PCs being able to bump the odds in their favor, whether decreasing the chance of being surprised or increasing the chance of surprising others. How the players play should matter and this is but another way their choices matter mechanically and fictionally.

How long surprise lasts has also not survived its viewing under the microscope. The initial measure being provided by the die roll, with a surprise result of 1 or 2 resulting in a matching number of suprise segments was an elegant way to find out long surprise would last. Two segments of surprise are also referred to as “full surprise,” which illustrates that two segments was the longest surprise would last, originally. Then, with the accretion of further rules, notions such as creatures that could surprise more frequently changed how many segments could be notched in that fashion. Tossing in the use of dice of a larger size — D8 or D10 — for surprise checks in some situations, and figuring out how long surprise lasts becomes muddled.

As I don’t think more than three segments of surprise is reasonable — especially with a one minute round — I’ve capped surprise at that. The number of phases of surprise is also established by the dice roll for surprise.

What surprising combatants are able to do during each segment of surprise also got limited. The notion that each segment of surprise garnering a full series of attacks, without regard to the number of attacks entailed, got dropped, too. Two attacks are certainly possible, depending on circumstance; anything beyond that stretched my sensibilities past breaking. And certainly no launching multiple arrows each segment, so only one readied arrow per phase, with a phase required to ready one. I’m also considering an added bonus to having surprised the opponents when the first full round of action fires up, which returns a bit of the overwhelming advantage provided by surprise in the RAW.

This has also resulted in dropping references to segments in the surprise rules, in favor of referring to phases. This is with the understanding that I’m also dropping the use of the term “segment” entirely, at this point, because of its long association with being one-tenth of a round; rounds will be apportioned in fewer parts when the whole is reassembled (five phases per 20-second round).

The urge to allow for PC actions (player choices) to affect the odds of surprise also extends to how long surprise lasts, measured by how much activity the surprising party can do before the surprised can engage fully and player choices can quicken response when PCs are surprised. As above, the number of melee and missile attacks possible will be limited, what movement is allowed carefully meted out, and what casting or other magical work can get started and/or finished. A surprise situation, under this approach, may not happen as often, yet it provides a major advantage to one side when it does happen.

Ambush

An ambush provides a different experience of surprise. Only the party being ambushed can be surprised due to the ambushing party knowing full well when it will act. How the ambush is set up and executed will affect how much activity the ambushing side will get to do before the victims can respond. A well-planned and -executed ambush can provide a bit more time for salvos, too.

Sneak Attacks

A sneak attack is a form of ambush, just on a smaller scale, so to speak. Sneaking up behind a guard to take them out or firing a crossbow from the darkness into a sentry’s back rely on the at-least-momentary lack of awareness and engagement, as much as springing forth from hidden positions along a road when attacking a caravan does.

A Rough Draft of Surprise Rules

Surprise: Roll For doom

https://osrpgtalk.net/wp-content/uploads/2025/11/DA-Surprise-v-0.8.pdf

Layout: Red Letter Edition

October 20, 2025 / TAFL / 0 Comments

An approach now appearing in game materials involves using bold print to highlight important info. Adventure products have it to highlight initial aspects PCs notice. Rulebooks emphasize basic rule statements. Colored type has been used similarly here and there over the years. Recently, I’ve seen game text using red letter print ala Biblical publishing practice–I’m intrigued.

So, I’m going to experiment with red letter editions of rules, at least into playtesting. I want to get feedback from users as to whether the use of color, in addition to bold type, helps make it easier to parse the rules in any fashion. I expect to experiment with the practice a bit, too, with using bold, colored type for one purpose and simple bold type for another.

Accessibility may be an issue for color blind readers. Using bold, colored type may overcome that with the bold type still standing out despite lack of color. And the colored type may appear a different shade than standard black type. As long as the colored type appears a different shade than the standard bold type, I expect using both would be workable.

The actual red tint to use is also an open question. I don’t think the standard red letter print used in Bibles provides the best flavor for a game, so I’ll be experimenting. Perhaps a darker red. I’ve seen blue in some game texts and found the shade used too light for my tastes, so all of the shades I experiment with will be dark, to start.

My hope is for the use of red letter print to make important information even easier to find and scan. As the texts will be used for reference, anything that helps with scanning words during a lookup will be useful.

Design Theory: What About Skills?

October 7, 2025 / TAFL / 0 Comments

One vexing problem that presents itself when designing an old school system now is the question of how to use skills. This is a design problem because the earliest D&D systems used an implicit skill system instead of an explicit system. That has lead to a lot of people thinking that the old school rules don’t have skills built into them, a mistaken position.

I’ll leave it to those who’ve spoken of it before me to lay out how skills in old D&D appear (see Lucy Blumire’s blogpost about old school skills) and simply assume that the skill system is present as fact. That skill system is implicit in the text, with instances of it only showing when specifics are needed; at no point are the skills presented as being a subsystem in the rules. One way in which this characteristic affects the mechanics directly is in the realm of PC abilities.

At no point in describing PC abilities is a general skill system laid out or referred to. Even when the thief class showed up in a supplement with its personal skill set described, there was no discussion of a general skill subsystem in the rules. The thief skills also differed from the general skill usage by requiring percentile dice rolls instead of an X-in-6 approach. It wasn’t until the appearance of expansion books for AD&D that an explicit presentation of a skill system debuted.

This all means that designers of systems taking an old school approach these days have to decide how to involve skills in the systems. Do we design implicit skills that only get described as a discrete dice roll used in a specific situation here and there, or do we lay out skills as an explicit part of the system? Are skills to be a one-size-fits-all measure of X-in-6 chances called out discretely for everybody or will PC competence vary by class and/or experience?

I’ve decided I want to take a hybrid approach to skills. PC skills will be described as best as possible as discrete instances of what a specific class can do, instead of a list of skills that can be learned and developed by any PC. I also. however, will describe how PCs of other classes use those skills, without them being able to develop any greater competence (much the same as with the X-in-6 approach). This is the same basic tack as thieves being able to develop their ability to climb sheer surfaces while non-thief PCs have just a general chance to climb and never develop that skill further, so it’s not without precedent. An example from my projects would be that of fighters being able to employ stunts in melee (and the development of those locked in through level advancement) and non-fighters being able to attempt the same with much worse chances of success (and no improvement possible).

This approach involves no setting up an explicit skills system that covers learning new skills and generating ratings, then rules on how to develop those skills over the course of play, and how players can go about choosing how many skills and what skills their PCs possess. It stays with “at this level, PCs of this class can do X and all other classes can only wish they could do that as well” approach of AD&D and avoids the explicit approach of Traveller or Runequest or GURPS. I just won’t be shy about calling out the skills.