[Analysis] Knave 2e

Mechanism Design and Theory

Foreward

Systems Impact Behavior

My main role in my day job is being a mechanism designer (along with cryptographer, and back end developer in a pinch). Wikipedia explains it well:

The design problem is the "inverse" of traditional economic theory, which is typically devoted to the analysis of the performance of a given mechanism.

An economist is skilled in analyzing the resultant behavior for a given (economic) system, where a mechanism designer is skilled in proposing a new system that drives the desired behavior.

When we set out to create new systems, whether they’re TTRPGs, video games, or new currencies, we’re typically doing so because we’re attempting to influence player behavior in a positive direction.

This is easiest to understand in the context of small changes, which you’ll often see with house-rules.

Say that you house-rule OSE’s two-handed melee weapons to not have the “slow” trait; it’s one of the most common house rules out there. What has changed, and what behavior does it influence?

The first-order effect is that wielders of battle axes, pole arms, staves, and two-handed swords no longer act last in a round. In doing so, a few things happen:

• Two-handed weapons are strictly improved. This leads to more characters (both PCs and NPCs) using two-handed weapons.

• The lethality of the game increases. The only weapon that can kill someone with 9 or 10 HP is a polearm or two-handed sword, and now more characters use those weapons.

• The output randomness of the game increases. The standard deviation of a d6 damage is ~1.9 and the standard deviation of d10 damage is ~3. On average, a d10 weapon is doing ~1.6x as much damage as a d6 weapon. A d6 weapon has a 1/36th chance to do 12 damage in 2 rounds, whereas a d10 weapon has a 1/100th chance to do 20 damage in 2 rounds.

• Strategical depth decreases:

  • Characters could rely on always winning initiative against two-handed wielders, so if they were engaged with them, they could always retreat and be sure that they would not be hit.

  • They could always get attacks on to the two-handed users before getting attacked back, and so could plan around that.

  • Spellcasters could always cast their spells in the two-handed users faces and know that the spell would not get disrupted.

• The rules are simplified. We no longer have 4 phases to combat (init winner, init loser, 2h init winner, 2h init loser) and just have 2 (winner/loser).

I’m sure there’s more, but I think that’s illustrative. If those above properties are ones that we, as the game designer, want then it’s a good change. Often times, it’s a mixed bag, and we think that the having simpler rules and more-fun two-handed weapons is worth the increased lethality, and decreased control over combat, then we do it anyway.

The point I’m attempting to make is that this is the context in which one analyzes systems. What behavior does it drive; what downstream effects does it create, and what behaviors do those downstream effects create?

I often see it said that “system doesn’t matter”. Maybe I have a bunch of bias here, but I couldn’t disagree more! The same player will behave differently in different systems, especially the ones that actually read the rules.

Rules as Written

In most modern legal systems, we have both legislation (rules written by law-makers) and common law (precedent created by judges). In a court case, typically what happens (at a high level), is that the judge will hear the case, then interpret and apply any relevant legislation, then look at the precedent created by past judges ruling on similar cases, and finally decide if they want to create new precedent or side with history.

This is exactly what a GM does in a TTRPG. The system designer creates a rulebook that lays out the legislation. When a situation arises, the GM interprets and applies rules from the rulebook. If the rulebook doesn’t cover the situation (or does so badly), the GM attempts to interpret the spirit of the rulebook to create a new ruling. This creates a precedent for future rulings (that may be overturned if it’s for the best).

It is the aim of some rulebooks to provide legislation for as many situations as possible. The idea is “it’s there if you need it, but you don’t have to use it”, which is how we get exhaustively detailed math for how long it takes to pickaxe through the various thicknesses of different materials. The downside is that much of the rulebook becomes “noise” (it’s harder to find one rule in a book of 10000 rules than a book of 10 rules), and that as a GM, it feels very compelling to use a rule if it’s there, so much of the “play” becomes remembering/interpreting legislation rather than imagining fiction.

The aim of other rulebooks is to provide way less specific legislation, but for each of the rules to be broadly applicable and for it to be easy to interpret the spirit of the rulebook to easily create rulings with its framework.

The best written example of this I know of is Advantage and Impact - Dreaming Dragonslayer, where the author analyzes a hypothetical game where situations are always resolved with coinflips. The “system” would be “when in doubt, flip a coin”, and rulings are just “do I need to flip?” and “what happens if heads, what happens if tails?”

Context

Hopefully, I’ve done a good job setting up where this going! I’m going to be analyzing Knave 2e with respect to how its changes (relative to BX, since I think that’s the zeitgeist) influence behavior (for both the players and GM), how clear and helpful it’s legislation is, and how well the legislation provides a spirit-of-the-game and framework to create rulings in.

Additionally, the at-the-table context is that we’re using the game to run OSR modules, like B2 - Keep on the Borderlands or The Black Wyrm of Brandonsford.

I think this is a fair context, given that this is the book’s opening paragraph:

Knave is an old-school fantasy roleplaying game in the tradition begun by David Wesely, Dave Arneson, and E. Gary Gygax in games such as Braunstein, Blackmoor, and Dungeons & Dragons. As part of that design tradition, it is broadly compatible with the monsters, items, and adventures created by thousands of hobbyists over the last 50 years.

I will specifically not be analyzing the d100 tables (many of which I find fantastic resources) or the art (which I think is stunning). These are not my domains of expertise!

Page By Page Stream of Consciousness

Nothing in these rules is sacrosanct. The rules at your table will evolve over the course of a campaign, as you and your players tailor them to fit the way you like to play. This is a good thing. Altering the rules and writing your own is a time-honored part of the hobby and a critical part of becoming a great Game Master (and, in time, a game designer).

I’m quoting this up-front because I want to get out ahead of it. Just because the rules aren’t sacrosanct, and just because tables can (and are encouraged to) change and tailor them, does not mean they’re immune to criticism or praise. If I release a book with nothing but hot garbage between the bindings, it’s not okay just because I say “they were just suggestions man, change them if they’re not working for ya”. Likewise, if I release the most platonically ideal set of TTRPG rules, they deserve praise, even if I make this disclaimer.

I’ll be analyzing these non-sacrosanct rules!

Game Master Duties

Don’t plan out a plot for the players to experience. Each session’s outcome should be a surprise to everyone.

I find this to be unintuitive advice. I totally agree that there shouldn’t be a plot, but as GMs we definitely want to be prepping situations. Normally that means we know ahead-of-time what the relevant actors in those situations motivations are, and what tools they have to accomplish their motivations. If we start materializing tools after-the-fact to deal with the player’s plans, then it starts to feel grimy.

Similarly, if you’re perpetually surprised by your players, this means that there’s some sort of calibration problem happening, or some sort of social contract is being violated. For instance, when my players were deep into hole-in-the-oak, and we discussed at the end of the session their plans to continue exploring to the northeast, I was not especially surprised by how the session played out (they explored the northeast). Sure, different rooms didn’t play out exactly as anticipated, but broadly, things seemed pretty normal. I think that’s totally fine.

If your players are constantly setting things on fire or mining through walls or trying to befriend unlikely allies, expect that and stop being surprised!

Player Duties

Treat the game world as if it is real and work to turn every aspect of it to your advantage. When simulating a living world, no detail is simply “flavor”.

The other side of this is How to Provoke Any Dungeon Master. I get what Ben’s saying here though. If you walk into Brennan’s room with “a musty old bookshelf”, and “a table in the corner that has some goblets on it”, that bookshelf can be used as kindling, or to barricade a door. Those goblets can be filled with poison or whatever.

From the GM’s side, however, so much of a dungeon is set dressing, and the quicker that a player learns to distinguish the set dressing from the key features, the more of the session will be spent doing interesting things.

I’ve been known to literally say “this is set dressing” to re-focus the game quickly!

Ability Scores

Getting into the meat now. We keep the 6 original stats, but shuffle around their duties. Saving throws have been removed entirely, and their job has been folded into the 6 stats.

This creates some compatibility issues. There are plenty of creatures in OSR adventures that harm stats directly. For instance, a lot of books stat shadows or vampires and whatnot as draining strength. An OSE character might have 11 STR, and then get drained for 4 down to 7. This shifts their modifier from +0 to -1. What happens if a Knave character gets drained for 4 strength?

Changes wise, we have:

• STR does not improve melee damage (only accuracy)

• DEX loses missile attack bonus (moved to WIS) and bonus to AC

• CON increases item slots but does not increase HP (though, slots are effectively HP via wounds), saves against forced march

• INT loses influence on language, but increases the number of spells a character can cast and increases the effectiveness of those spells, helps with brewing potions, helps with picking locks.

• WIS gains missile attack bonus (moved from DEX) and now governs survival checks (previously flat, class-based rolls), surprise, and being sleep deprived.

• CHA gains a bonus to initiative, limits the number of hirelings a character can have in their lifetime.

Characters start out with 3 points allocated to stats. Someone trying to play a heavily armored front-liner might choose to allocate 2 points to STR (to improve chance to hit) and 1 point to CON (extra slot for armor and wounds).

Relative to BX Fighters, such a character:

• Has much more effective health before the die. Fighters have an average of 4.5 HP, whereas this Knave character can suffer up to 14.5 (3.5 hp + 11 wounds) damage before dying. This makes the Knave character ~3.2x as resilient.

• Does much less damage. A BX Fighter can wield a d8 longsword and gets their STR as a bonus to damage. A Knave fighter with a longsword does an average of 3.5 damage. A 16 STR fighter with a longsword does an average of 6.5 damage. The BX fighter is hitting ~1.9x harder.

Saving throws are now split between STR (paralysis), DEX (wands and breath), CON (poison/death), and WIS (spells), so characters feel dependent on multiple attributes. A 7th level fighter in BX will have +5 to hit, 65% to save vs death, 60% to save vs wands, 55% to save vs paralysis and breath, a 45% to save vs spells, and may have 4 hirelings.

For a Knave character to achieve the same thing as a BX fighter with all +0 stats (via 1d20 + stat >= 16), they would need 5 STR, 8 CON, 6 DEX, 4 WIS, and 4 CHA, which is already more points than a level 7 character can have.

Thieves have it even worse; they need STR to stab people, DEX to sneak, CON like everyone else, INT to pick locks, and WIS to perceive (?), use ranged weapons, and resist spells.

Especially as we get into the upper tiers of play (try playing the G-Series), this starts to hurt pretty badly.

We also still run into linear fighter vs quadratic wizard. A Knave fighter with 5 STR hits a bugbear 60% of the time for 1d6 damage. In Knave, we have:

When generating a damage dealing spell, a good rule of thumb is that it deals INT × d6 damage, although this can be reduced if it has other beneficial effects.

None of the pre-listed 100 spells are damaging, but we do get (vague) guidelines for making damaging spells. This is a significant hit to a party’s ability to handle mobs (fireball is really good in BX), but INT•d6 damage still means that the Wizard can pump out 5x as much damage as the fighter 5 times a day (on top of all of it’s other spell-based utility).

In terms of player behavior, people need a firm understanding of the rulebook in order to make informed decisions each level about where to allocate their 3 points every time they level up. The things explicitly listed as checks are very spread out, which makes the task more difficult.

Taking melee damage bonuses away and damaging spells away also makes combat drag. Enemies, which use BX stat blocks, do not have their HP values decreased, and so combat length increases. This makes combat likely to be more taxing to players, and combat overall to be a worse option.

Item Slots and Wounds

PCs have 10 + CON item slots to record their gear. Most items, including groups of small items that could fit in one hand, take up one slot. Two-handed items take two slots. 500 coins use a full slot.

Not noted here: each point of AC also takes a slot. Level 1 characters have, at max, 13 slots. PCs can wear 7 pieces of armor, so getting to 18 AC (higher than plate+shield in BX) takes 7 slots. That gives us 6 remaining slots for gear. If we have two weapons (one melee, one ranged), we have 4 slots left.

Slots are the game’s main competing resource. Armor, weapons (which break), light, rations, and treasure all take up the same slots.

I think the design intent here is that players “have fun” managing their slots, especially during the pre-dungeon expedition planning. When the players find treasure, they “have fun” figuring out how to distribute their coins and whatnot so that they can walk out with it.

In practice (I briefly implemented this exact system in my OSE game), it feels like a hassle, especially with the introduction of porters. Compare to the default (optional) encumbrance system in BX: unarmored folks move at 40’, leather moves at 30’, chain/plate moves at 20’, and if you’re encumbered (GM’s discretion) you move 10’ slower. In my game, I say that you can fit ~40lbs in your backpack, and anything past that you’re encumbered.

I think the problem here is that the lion’s share of folks don’t enjoy logistics planning. I don’t think it’s fun to figure out if you want to carry 2 rations and 6 torches or 6 rations and 2 torches. I don’t think it’s fun to figure out if you should use a lantern or a torch. Whether you should bring along shoe polish or beeswax.

I think what ends up happening is that the players take some of their starting ~110c a piece and buy hirelings to carry stuff around for them, more-or-less alleviating slot scarcity for adventuring supplies, and leaving the actual adventurers to carry combat equipment, mostly.

It seems very effective to stack everyone up with 18 AC and then have them do their class stuff with their remaining slots.

Damage a PC receives is subtracted from their HP. Once their HP reaches 0, each point of damage fills an item slot with an appropriate wound (stabbed, frozen, burned, etc.), from the highest slot to the lowest. Items in a wounded slot must be dropped.

Ah, not only must players manage what they fill their slots with, they need to properly order them. You want to place your least important/least expensive items at the top of the list so that they’re the ones you drop first.

A conversation I had on the subject made me realize that the conditions might (the rules are unclear) not just be flavor, but might have in-fiction consequences, sort of like aspects in FATE. For example, say that you take 1 direct damage from an arrow show, and so you get a wound that says “arrow to the knee”. Do you have reduced speed? Should you get lowered DEX? Totally unclear.

PCs’ HP returns to maximum each morning, as long as they slept for two watches and ate a meal the night before. If they are in a safe haven they also heal one wound.

This is massive compared to the BX healing rules. In BX, natural healing is 1d3 hp per day of complete rest (no traveling, no exploring, no fighting, etc). Here, our party of level 5s (17hp on average) all lose ~11 hp. In BX, this would take ~5 days of rest to recover (or a cleric casting cure wounds a bunch). In Knave, they’re back to full in the morning.

This doesn’t feel that different in a party with ample clerics, or a party near a town (that has ample clerics), but will make it much harder to get proper attrition going on a long delve or expedition.

Leveling Up

PCs are awarded 1 experience point (XP) for each coin (c) worth of treasure recovered from remote, dangerous locations like dungeons and returned to civilization, split evenly between all PCs who assisted. If you are using a pre-made dungeon from another RPG that uses copper, silver, electrum, gold, and platinum coins, then convert the total to gold coins and gain that much XP.

Okay, so 1g = 1c in terms of converting economies. If the party finds an urn filled with 3000 copper, that’s worth 30c and 30xp in Knave-land. Does it take 6 slots? Unclear.

Additionally, this removes the ability for PCs to gain XP from defeating monsters. According to my sims, XP from treasure make up somewhere between 80% to 90% of a character’s total XP, so this is anywhere from a 10% to 20% nerf to XP rates. It strongly discourages fighting things and makes random encounters more punishing.

At certain XP thresholds, PCs gain a level, which adds 1 to three different ability scores. Do not reset XP to zero. The three scores can be picked by the player or chosen randomly. Each level also allows the player to reroll their PC’s HP maximum using one additional d6. If the rolled total is not greater than their last maximum, add 1 to the last maximum.

I love the wording here; makes it abundantly clear that XP accumulates rather than each step being the amount of additional XP you need to reach the next level.

I think it’s also a good time to run a monte-carlo on this method of calcuating HP, since it’s quite popular.

function rollHp(previous, n) {
  let newHp = 0;
  for (let i = 0; i < n; i++) {
    newHp += Math.floor(Math.random() * 6) + 1;
  }
  return Math.max(newHp, previous + 1);
}

function monteCarlo(simulations) {
  let hpPerLevel = {};
  for (let i = 0; i < simulations; i++) {
    let hp = 0;
    for (let level = 1; level < 11; level++) {
      hp = rollHp(hp, level);
      hpPerLevel[level] = (hpPerLevel[level] || 0) + hp;
    }
  }

  for (let level = 1; level < 11; level++) {
    hpPerLevel[level] = hpPerLevel[level] / simulations;
  }

  return hpPerLevel;
}

console.log(monteCarlo(10000));

You can drop that code into any javascript repl. My output was:

1: 3.5, 2: 7.3, 3: 11.1, 4: 14.9, 5: 18.7, 6: 22.5, 7: 26.2, 8: 30, 9: 33.8, 10: 37.5

Compare this to the average for d6 progression:

1: 3.5, 2: 7, 3: 10.5, 4: 14, 5: 17.5, 6: 21, 7: 24.5, 8: 28, 9: 31.5, 10: 35

At each level, we have extra health. Here’s the breakdown:

So it’s around a ~5 - 7% hp buff. I’m personally a fan of this change in general; really helps with removing some (typically unfun) hp variance while not being a to crazy of a buff to hp.

XP Chart

This follows the BX Fighter progression exactly to level 7. Then, we use 125k instead of 120k, 250k instead of 240k, and 500k instead of 360k. At each level, it takes roughly twice as much XP to level up as the previous level. This has a really cool property where in the course of gaining enough XP for a full level, any fresh characters will only be 1 level behind.

For example, a level 7 character gains 64k XP (enough to put them to level 8). That’s enough XP for a level 1 character to become level 7, so freshly created characters get caught up quickly.

The table is also pretty harsh relative to thieves and clerics (but a buff to magic users). Assuming that 20% of XP in BX comes from treasure:

So a Knave thief needs to find a little more than twice as much treasure as a BX thief to level up. Ouch! This hurts compatibility with larger adventures, where the adventure places enough treasure in it’s early and middle sections to get characters to the appropriate levels for its later sections.

Further, since there is no real bestiary, and smattering of entries in the bestiary lack a treasure type, there’s no implied treasure placement.

For instance! In BX, you might stock a dungeon by rolling on the Random Room Stocking table. In a 36-room dungeon, ~6 rooms are going to be monsters-with-treasure which we can imagine as being a Dungeon Lair. We can also roll on the Dungeon Encounter table for level 1. Say we roll Orc, which says that a Orc Lair has 1d6•10 Orcs (average of ~35) and Treasure Type D, which in turn has an average of 3900g and a 15% chance of 2 magic items and a potion.

A GM following this procedure creates a world where the amount of treasure roughly corresponds to the strength of it’s guardians. Players then choose their risk level. It also implies “if you want to take a Party of 5 PCs from level 1 to level 2, they’re going to need to loot ~2 Orc lairs”.

Each level 1 character needs about 2000 XP on average, so a party of 5 needs to earn 10000 XP. Around 80% (in BX) of XP comes from treasure, which is almost exactly what would be contained in 2 Orc lairs.

Knave totally lacks this guidance!

If Knave wanted to avoid creating treasure types, then I think a great alternative is something like what BXBlackrazor recommends. Put enough treasure for the party to level up spread across ~20 rooms. Put ~half of it in one well-guarded hoard, 1/4th in a second location, 1/8th in a third location, 1/16th in a fourth and 5th.

For example, a party of five 4th level Knaves needs collectively 40000 XP (coin) to level up. So, in a ~20 room dungeon, put ~20000 coin in the vault, 10000 coin in the lieutenant’s stash, 5000 coin as the secret treasure behind the false wall, 2500 coin in the rhadogessa’s nest, and 2500 coin in ancient urns at the shrine.

Checks

When a creature attempts something risky, they make a check by rolling a d20 and adding one of their ability scores. If their total meets or exceeds a target number set by the GM, they succeed. If a creature doesn’t have ability scores, the GM can use its level, half its level or zero based on how good it is at the task. GMs should not call for checks for situations that could be solved with critical thinking. Some actions may be impossible unless the PC has the proper tools or careers.

Setting the target number: Start with 11 and then add a difficulty rating from 0-10 (5 by default). If the check is against another creature, the difficulty rating is equal to their relevant ability score or level. In an attack, the difficulty is the defender’s armor points (the target number of 11 + AP is called Armor Class).

Modifiers: The GM can apply -5 penalty for each disadvantage and a +5 bonus for each advantage that the rolling player has on a check. (e.g., a related career, a clever approach, extra time, the right tools, etc.).

Okay, this is our core resolution mechanic and framework for making rulings. Similar to D&D after 2e, the idea is to pick a DC for the task (between 11 and 20) that’s is independent of the approach and relevant character and specify a relevant ability.

For example, if someone is someone is trying to jump a 15 foot gap, we might call that a DC 17 strength check. If they brought their springboard, we’d give them a +5.

This is setting some common law that jumping is going to be STR related, and that 15-foot-gaps are ~DC 17. Players can expect in the future that smaller gaps shouldn’t be higher than DC 17, and larger gaps shouldn’t be lower than DC 17.

(As an aside, I find the task of generating internally consistent DCs to be difficult).

There’s also the implied probabilities. By making it a DC 17 jump, we’re saying that there’s a 20% chance a 0 STR character can make it, a 45% chance that a 5 STR character can make it, and a 70% chance that a 10 STR character can make it. We’re saying that a 0 STR character goes from 20% to 65% when they use a springboard, and a 5 STR character goes from 45% to 70% when they use a springboard.

In terms of player behavior, the default DC of 16 combined with the low early game stats (0-3) means that most rolls have around a ~30% chance of success, or ~55% if you can negotiate a positive modifier.

Such low odds encourage players to attempt riskless strategies (ones that don’t require dice rolling), or to seek advantage. This is good for structuring and reinforcing the author’s desired model of play (apply tactical infinity, scheme, fight dirty). I like it a lot for the low levels.

A concern is that in a full party, the same design in combination with the level up rules is going to mechanically encourage specialists. Each party member and companion can focus on 1 attribute and pump it up to 5 for 4000 xp. In a party of 6 characters, that means someone in the party has a 50% chance to succeed at any particular thing by level 3.

Traveling

While traveling, days are divided into six four-hour watches: three for day, three for night. Most major actions take one watch to complete. PCs can move one six-mile hex per watch, up to three times per day.

That’s elegant! In BX, characters can travel a number of miles per day equal to 3/5ths of their combat speed. So, Unarmored characters (MV 40) can travel 24mi, (4 hexes), leather (MV 30) can travel 18 miles (3 hexes), and chain/plate (MV 20) can travel 12 miles (2 hexes). Encumbered characters get reduced by 10ft, so those distances would be 3 hexes (unarmored), 2 hexes (leather), and 1 hex (plate).

Here, encumbrance is binary; either you move at full speed or you can’t move at all (because you’re carrying too many things), so there’s no napkin math to be done or logistics for moving at different weights or shuffling things around so that the armor-wearers can be unencumbered and so on. This is good!

If the terrain or weather while traveling is disorienting, the GM may require a WIS check of the party’s leader (which the GM rolls in secret) to see if they move to a random adjacent hex.

It’s unclear what “disorienting terrain” is, but “disorienting weather” is defined. The next page has a weather table with weather descriptions! Rain, hail, fog, and heavy snow are all explicitly “disorienting”. Those happen on a 2, 3, 5, and 12, which occur ~22% of the time, and last until the next travel shift (~1 day).

With 3 WIS, the chance to succeed the WIS check is 40%, so they’re getting lost ~13% of the time they move hexes (terrain excluded).

I’ve talked about it before, but it’s very unclear to me what getting lost actually plays like at the table. I think the assumption is that the players aren’t filling in their own hex map (making them erase and rewrite a bunch of hexes would be beyond cruel), so the GM is looking at their own hex map, and describing the nearby terrain.

Here’s what I mean

Say that the party is current in <3,5> and they’re trying to go north to <3,4>. We roll (secretly) that they get lost, so we roll a d6 and determine that they’re in <2,6>. When do they find this out? Presumably, they think they’re in the right place, so they take the ‘exploring’ action to find the thing they were looking for. They don’t find it. Is fun happening? Since their previous hex is now to the north east, and we’re supposed to be describing adjacent hexes, can they see that’s the case?

I just really don’t know how this is supposed to work and I’ve never seen a compelling answer or example of play.

Foraging (major action): Finding food takes a watch and requires passing a WIS check, with modifiers for weather, terrain, etc. On a success, a PC collects d6 rations.

The modifiers for weather, terrain, etc are undefined. Maybe this means that if the terrain is “bad" (which terrain is bad?) you get a -5 per the modifier rules, and that the default DC is 16?

It’s also unclear whether or not everyone can forage independently or if this is a party activity (like it is in BX). Let’s be nice and say that everyone can do it.

Travel Hazard Die

At the end of each watch, roll a d6. 1: Encounter, 2: Fatigue (take 1 (direct?) damage unless you spend the next watch resting), 3: Depletion (roll 1d6 for each perishable, on a 1 it goes bad), 4: Travel shift (weather, etc), 5: Sign (to the next random encounter), 6: Nothing

Holy moly.

This feels so heavy. Let’s do a worked example from Silveraxe.

For the sake of simplicity, we’re going to say that this is all difficult terrain (mountainous), canceled out by there being a road, so we’re moving at 3 hexes a day in normal conditions. We’re going to say that foraging is normal unless it’s “disorienting” weather, and then it’s a -5. We’ll set the foraging DC to 16 (the default).

We’re going to say that there’s 5 characters who are each carrying 7 rations each (whether or not a “ration” takes a slot, and how many days a ration is worth is undefined; i’m going to say 1 day’s ration is 1 slot and that’s what the mechanic refers to). We’re going to say that each party member has a WIS equal to their number (so #1 has 1 WIS).

• Day 1 Watch 1: Travel to B5. Roll 1d6 (3). Depletion. Each of the 5 PCs now rolls a 7d6, on the lookout for 1s. Here’s the results: 1: [3 1 4 1 3 3 1], 2: [1 5 3 5 1 1 6], 3: [2 5 4 5 1 4 2], 4: [1 6 2 4 6 1 5], 5: [3 2 2 1 5 6 4]. So #1 has 4 rations, #2 has 4 rations, #3 has 6 rations, #4 has 5 rations, and #5 has 6 rations.

• Day 1 Watch 2: Travel to C6. Roll 1d6 (3). Depletion. Each of the 5 PCs now rolls a d6 for each of their rations. Here’s the results: 1: [5 4 1 3], 2: [2 4 3 2], 3: [6 1 1 6 6 2], 4: [4 5 3 5 6], 5: [6 4 1 3 6 1]. So, #1 has 3 rations, #2 has 4 rations, #3 has 4 rations, #4 has 5 rations, #5 has 4 rations.

• Day 1 Watch 3: Travel to D7. Roll 1d6 (1). Encounter. Roll 4d6•30 (450ft) for distance, 2d6 (5; Insult, threaten, or command the PCs), 1d12 for type (3; 1d8 brigands), and amount 1d8 (7). Have a grand ol’ fight against 7 Brigands that are insulting, threatening, or commanding the PCs from 450ft away.

• Day 1 Watch 4: Forage. 5 forage attempts: 1d20+1 (14), 1d20+2 (4), 1d20+3 (10), 1d20+4 (12), 1d20+5 (15). No successes. Roll 1d6 (2) for the hazard. Fatigue.

• Day 1 Watch 5: Rest. Roll 1d6 (2). Fatigue.

• Day 1 Watch 6: Rest. Roll 1d6 (4). Travel shift. Roll 2d6 (5). Light rain (hurts foraging). Each PC consumes a ration. #1 has 2 rations, #2 has 3 rations, #3 has 3 rations, #4 has 4 rations, #5 has 3 rations.

• Day 2 Watch 1: Travel to E7. Roll 1d6 (6). Free.

• Day 2 Watch 2: Travel to F8. Roll 1d6 (3). Depletion. Each of the 5 PCs now rolls a d6 for each of their rations. 1: [3 1], 2: [3 3 2], 3: [4 3 3], 4: [1 2 6 4], 5: [4 4 3]. So, #1 has 1 ration, #2 has 3 rations, #3 has 3 rations, #4 has 3 rations, #5 has 3 rations.

• Day 2 Watch 3: Travel to G8. Roll 1d6 (2). Fatigue.

• Day 2 Watch 4: Rest. Roll 1d6 (2). Fatigue (ignore).

• Day 2 Watch 5: Rest. Roll 1d6 (5). Sign. Roll an encounter and improvise a sign for that encounter. Use that encounter next if it’s still appropriate. The encounter is 1d12 (7; 1d4 (3) wolves) that have a reaction of 2d6 (7; ignore the PCs).

• Day 2 Watch 6: Forage. Roll 1d20-4 (14), 1d20-2 (1), 1d20-2 (2) 1d20-1 (17), 1d20 (2). One success, so roll 1d6 (6) for rations. Distribute them after everyone consumes a ration, so now #1 has 2 rations, #2 has 3 rations, #3 has 3 rations, #4 has 3 rations, #5 has 2 rations. Roll 1d6 (3) for hazard. Depletion. Each of the 5 PCs now rolls a d6 for each of their rations. 1: [6 2], 2: [4 3 6], 3: [1 1 5], 4: [2 4 1], 5: [4 2]. Now #1 has 2 rations, #2 has 3 rations, #3 has 1 ration, #4 has 2 rations, #5 has 2 rations.

• Day 3 Watch 1: Travel to H9. Roll 1d6 (5). Sign. Unclear what to do here; improvise another sign for the same random encounter?

• Day 3 Watch 2: Search H9; find the cave. Roll 1d6 (1). The PCs find the wolves that ignore them.

We’ve made it to the destination! It merely took 127 rolls to get there! Rolling to randomly make your rations perish multiple times a day is absolutely bananas (why are we focused on this so much?).

We had 14 watches and 4 depletion events (only expected 2.3), so that was slightly higher than normal, but this is a high variance mechanic, and 4 depletions isn’t crazy (~20% to get at least 4 depletions in 14 watches).

In terms of player behavior, rations now last a variable amount of time (as opposed to something they can plan for, like 1 week). On average, a ration will last 36 watches (1/6 chance to get a depletion, 1/6th chance to deplete on a depletion), which is ~6 days. If we compare that to a ration that always last 6 days, this is worse for the player. If it lasts longer, it doesn’t especially help them. Rations are cheap (probably? no price is provided), so a ration lasting for a whole adventure means they spend marginally less money to restock for the next one. A ration expiring early is potentially disastrous, so they need to hedge against the downside risk by carrying extra rations.

Previously, to have enough food for 7 days you carry 7 rations. How many rations do you need to carry to be 99% confident you have enough food for 7 days in this system? I don’t even want to calculate it =/

The chance for an encounter is massively higher than in BX. The default is one check per day, with a ~2-in-6 chance per check, so ~1 encounter every 3 days. In Knave 2e, we’re rolling 1-in-6 6 times a day, so ~1 encounter every day. We expect 3x as many wilderness encounters in Knave as in BX!

This means that relative to BX, wilderness exploration is more risky (and thus relatively discouraged).

Hazard Die Woes

I have beef with hazard dice. There’s a 1/6th chance that you’ll get the same thing as last time, every time you roll. 1/6th is high! Frequently, getting the same roll twice in a row doesn’t make a lot of sense. It’s especially bad with getting two depletions in a row, but also gets weird with signs.

Second, one of my favorite parts of the normal 1-in-6 encounter die is that I can roll a whole bunch to represent a trip, look for ones, and gloss over the rest. For instance, in BX, here’s how the same above journey goes.

1. Assume that one of the party members is wearing at least chain. They can travel 2 hexes per day.

2. Count the number of hexes to H9. It’s 7, so it’ll take 4 days of travel. Roll 4d6: [5 2 4 2] for encounters (no encounters), and have the PCs roll 4d6 for foraging [5 6 2 1]. Tell the PCs to mark each off 4 rations and narrate their smooth trip to the cave. The last day of foraging was successful, so they get back 1d6 (5) rations total.

We did it! This is very hard to do with hazard dice because each result does something, potentially affecting future results or travel duration.

Delving

Similar Hazard Die here. 1: Encounter, 2: Fatigue, 3: Burn (torches expire), 4: Delve Shift, 5: Sign, 6: Free.

I used this in play in my OSE hole-in-the oak game. We found the fatigue result to be lame and jarring. Wake up refreshed and go on a delve. Search the first new room; it’s a heretic gnome kitchen. Rummage through the cupboards and find some nice loot. Turn ends, and get a fatigue result. Tell the players that their characters are le tired, and will take damage next turn if they don’t take a quick 10-minute breather after that exhausting rummaging. They grumble and do so. Why are we doing this?

Same thing with torches burning out early - you light a torch and at the end of that turn it’s out. Players are not jazzed. Weird bookkeeping. Torches last ~6 turns on average, but players need to carry extra to hedge against bad luck, so this feels like another element of randomness they can’t control. How many torches do you need in BX for a 6 hour delve? 6. How many torches do you need to have in order to be 99% confident you’ll have enough torches for a 6-hour delve in Knave 2e? This one is easier because we’re not also eating torches, so it’s just a binomial distribution calculation.

So, we’re 99% confident that if we carry 13 torches we’ll have at least 1 left at the end of a 6 hour delve. Wheeeee!

I also had a lot of trouble running Dungeon Shifts. I had no idea what was supposed to be shifting in Hole-in-the-oak, which took my party ~90 turns to complete. I had to improvise ~15 dungeon shifts. I wanted to cry by the third one.

Crawling speed: 120’ per turn. PCs spend most of this time standing still and listening, testing surfaces, and drawing maps.

This is a large buff to armor, which in turn is a buff to character survivability. One of the downsides to plate armor was having to move at 60ft per dungeon turn (compared to 120ft for unarmored), so by choosing to be fully armored, you’re also choosing for your expedition to take more time and trigger more random encounters.

Here, all that heavy armor costs is slots!

At this speed, the PCs automatically detect traps and map their environment.

This is unequivocally great. The BX rules for traps make no sense to me, and I massively prefer Ben's approach to traps. In my auditing process, I make sure that each trap has a tell, per GFC. This tell is what players in my games automatically detect when they’re moving at exploration speed.

Walking speed: 2,400ft per turn. PCs are surprised by all encounters and spring all traps, but may still map the dungeon. Walking speed is usually used when backtracking through areas the party has already explored.

2400ft per turn is 14400ft per hour, which is ~2.7mph. That’s roughly the speed that actually walk at in real life, which feels quick for being able to map. Not a big deal though. My bigger complaint is that it makes counting turns to walk back to another room kind of a pain. Dungeon maps tend to be in 10ft squares, so counting in increments of 240 is… a lot.

Light

Candles: Candles reveal general shapes within 20’, and details within 5’. They last for 8 hours, or a whole dungeon crawling expedition. 10 candles fill a slot.

Lanterns: Lanterns are candles inside a glass and metal case. They can be constructed with shutters for more directed and controlled light. Like candles, they last a whole dungeon crawling expedition, but their case prevents them from being blown out. Candles and lanterns are great as emergency back-up lights when your torches run out.

Torches: Torches reveal general shapes within 40’, and details within 10’. The brightness of torches lets the party search areas faster (see below), but they take up a whole slot and burn out when a 3 is rolled on the Dungeon Hazard Die.

Searching: One turn spent searching a room reveals any non-obvious features (a statuette inside a drawer, a cracked tile, etc.) as well as any clues to secrets. In dim light like candlelight or lantern light, this takes two turns instead.

As far as I know, this is novel. Candles elegantly solve the issue of “what happens if we run out of light”, which, as far as I can tell, BX implies but never explores. I still have no clue how no-light gameplay is supposed to be performed at the table.

The intent here (confirmed in the commentary) is that this creates actual mechanical choice between a lantern and a torch, which isn’t something that exists in BX.

In BX, the difference between torches and lanterns is purely diegetic. Torches are strictly cheaper (1g for 6 hours of light) than a lantern (2g for 4 hours of light). Neither of expensive enough to be worth writing down. Torches can be swung as an improvised fire-club, and lantern oil has all sort of fire-and-not-fire based applications.

In Knave 2e, though, now we’re having to decide between slots and visibility/search time. Torches make searching twice as fast and let you see further into the room (being further away from things is less dangerous than being close).

I still maintain that none of this matters because the PCs will just hire a bunch of people to carry their stuff around for them after they get sufficiently annoyed by slots.

Encounters

Distance from party: In confined environments with limited visibility like dungeons, encounters appear 2d6 × 10’ from the PCs when their initial distance is unclear. In the wilderness, large caverns, or other wide-open spaces, they appear 4d6 × 30’ away.

This mostly follows the BX rules, except for one detail. In BX, everything in the wilderness is measured in yards not feet. So while wilderness encounters start 4d6•30ft away in both Knave and BX, movement speeds are 1/3rd of what they would be, and ranged weapons are massively lower range in Knave (120ft in the wilderness vs 630ft in BX for bows).

This means that ranged characters are significantly weaker (and therefore melee characters are relatively stronger) in Knave compared to BX.

Surprise: If an encounter occurs within 80’ of the party, the closest PC makes a WIS check against the WIS of the closest encountered creature. If one side wins by 5 or more, they have surprised the other side. The surprising side will act first in combat and gains a +5 bonus on all combat checks during the first round.

Couple of notes here.

The minimum distance of a wilderness encounter is 120ft, which means that surprise in the wilderness is impossible. Weird.

The closest PC is usually the frontliner, so now that frontliner is going to be making a lot of WIS checks.

Here’s an anydice view of this mechanic.

Here’s how to read this: If the two sides have the same WIS, there’s a 30% chance that your d20 will be at least 5 more than their d20. If your WIS is 1 higher than there’s all you need is for the your die to be 4 better, so your chance jumps up to 34%.

If your WIS is 10 better, you can be at -5 on the dice and still get there, so you have a 74% chance of surprise. If your 10 WIS worse, you need to beat their roll by 15, which only happens 4% of the time. Monsters don’t have stats, so we’re told to use their level (HD).

Reaction Table

Here’s the BX table:

2: Attacks, 3–5: Hostile, may attack, 6–8: Uncertain, confused, 9–11: Indifferent, may negotiate, 12: Eager, friendly

Here’s the Knave table:

2: Kill the PCs, 3: Injure or capture the PCs, 4: Harass or rob the PCs, 5: Insult, threaten, or command the PCs, 6: Avoid the PCs, 7: Ignore the PCs, 8: Follow or observe the PCs, 9: Greet or question the PCs, 10: Share information with the PCs, 11: Perform minor favors for the PCs, 12: Ask to join the PCs’ party

I have the structure of the BX table memorized, and definitely do not have the Knave memorized, so it needed to be looked up each combat in hole-in-the-oak. I also found that it lead to incongruent results, like crab spiders performing minors favors.

I think the design intent is to be more evocative, but I found it ended up being overly specific and actually harder to use. That said, the original wording of the BX table is very ambiguous, and it didn’t really click until I read the ACKS interpretation.

Combat

Initiative: Combat is measured in 10-second rounds, during which each side has an opportunity to act. Decide which side acts first by making a CHA vs. CHA check between the sides’ leaders.

In BX, it’s a 50:50 to see who gets to act first in a round. In Knave, use the above surprise table (which models 1d20-1d20).

The rules don’t mention what happens in the case of tie. As far as I can tell, that means if your CHA is 1 higher than their CHA, you have a 52.5% chance to win initiative, 4.75% chance to tie initiative, and 42.75% chance to lose initiative. Winning initiative is a big deal, so you probably want your front-man to have high WIS and high CHA so they can surprise, lead the round, and then escape to the back.

When a side acts, all its creatures, in any order, may move (40' for PCs) and take one other action such as attacking, casting a spell, moving, maneuvering, etc.

So we get 1 move, 1 action as our action economy. The rules don’t specify anything about order, or whether or not you can spend some of your movement, use an action, and spend the rest of your movement. All sorts of armor move at 40ft in combat.

Can:

• An archer/wizard move out from behind cover, fire/cast, and move back into cover?

• A melee combatant attack a foe and then move out of their range?

• A combatant hold their attack until a foe comes into range/line of sight?

• A combatant move through a friend’s position?

• A combatant move through a foe to attack the back line?

Attacks: An attack is a check using the attacker’s STR (for melee attacks) or WIS (for ranged) trying to hit the defender’s armor class (armor points + 11). On a hit, the attacker deals damage to the target. If an attack roll’s total is 21 or higher, the attacker may choose to also succeed at a free maneuver of their choice. If an unmodified 1 is rolled, the weapon breaks.

Whew. So the math is the same as above with surprise. If your STR bonus is 1 higher than their AC, you have a 34% chance each attack to get a free maneuver (which likely grants some sort of status that counts as an advantage for a +5 modifier for your allies, which in turn leads to more free maneuvers and so on).

Additionally, weapons break in 1 out of every 20 rolls. I don’t even. Your fresh new shiny sword from the blacksmith is good for ~20 swings. The magic sword that you just quenched in the blood of a dragon is good for ~20 swings. The rulebook contains no mechanics for repairing weapons. I hate it.

Finally, “melee attacks” never seem to be defined; it seems like the author assumes that folks already know how to play TTRPGs on a grid (or similar), and to use the definition of “in melee” from those.

Sneak attacks: Melee attacks against unsuspecting foes always hit and deal direct damage. Attacks against truly defenseless foes automatically kill them.

When is a foe unsuspecting? It seems like this is left to common law (oh, how about make a dex check vs their wisdom. Does it take a round?), but it’s extremely important to get this right, otherwise you have characters spamming sneak attacks for auto-hit and 3x damage (since monsters that don’t have slots take triple damage from direct damage).

Ranged attacks: Ranged attacks cannot be made while in melee. If the target is in melee, the attack takes a -5 penalty.

I think this is a very sensible rule. Unfortunately, “in melee” is never defined, and as far as I can tell, nothing keeps anyone in melee. The archer can’t shoot in melee. So they step 5 feet away and shoot, right?

Power attacks: After a successful attack roll but before rolling for damage, PCs can decide to make a melee attack a power attack, which doubles the number of damage dice rolled but breaks the weapon.

Okay, so now we’re doing a sneak attack power attack for 6x damage. Additionally, we’re having our porters cart around piles of weapons so that we can power attack every combat right?

Modifiers: Attack and maneuver checks may gain +5 or -5 modifiers based on positioning, ganging up, weapon types, aiming, visibility, cover, foe size, range, surprise, elevation, etc.

What qualifies as “positioning” or “ganging up”? When is a “weapon type” enough to get +5 on a swing? Is there some sort of implied fire emblem weapon triangle (sword > axe > lance > sword)? How does “aiming” work? These are huge bonuses (bigger than wielding the most powerful magic sword in BX), so the play experience will massively vary at GM discretion.

Hazards

This is the one that stuck out:

Fire: 1d6 direct damage per round. On fire: 2d6 direct damage per round.

There aren’t explicit rules for throwing fiery oil flasks in Knave, but I think it would be fair to say that oil on the ground creates “Fire” and a direct connection creates “On Fire”. That means that thrown oil is 6x (2d6 vs 1d6, direct damage is 3x) more effective per round that using a sword. Forget swords, we’re just throwing molotovs at everything. It also means that thrown oil does as much damage as a sneaking power attack.

Spells

Spells that scale with the caster is a really cool idea; and one that makes a lot more fictional sense to modern audiences. BX’s spell slots and spells of particular level follow the fiction in Vance’s Dying Earth, but many modern players are much more familiar with game of thrones, harry potter, dr strange, wheel of time, brandon sanderson, etc, where as the caster gets better at doing magic, the same spells become more powerful.

The list of spells is an absolute goldmine for out-of-the-box application, but takes away the standard toolkit entirely. Here’s a short list of stuff that’s missing (but is likely to come up in normal D&D play):

• Creating/purifying food/water.

• Dealing with poison/curses/disease/death/petrification

• Dealing with extreme elements

• Identifying magic objects as magical

• Reliable single-target damage

• Reliable aoe damage

How do 9th level characters journey across the elemental plane of fire to the city of brass? In BX, they have resist fire, create water, and create food. If someone contracts a disease (like from a giant rat), how do they cure it? If someone gets cursed, how is the curse lifted? How do you bring back dead characters? If a character is petrified by a medusa or cockatrice, how do they return to flesh?

BX intentionally places the magical solution to these problems near-in-level to when characters can expect to have to start contending with them. I understand that the GM can handwave and say “oh you find a sage that removes your curse, all you have to do is…”, but this feels immensely unsatisfying to me.

In terms of player behavior, this makes poison/curses/diseases/petrification/death significantly more risky, and so players are likely to try to avoid anything that causes those like the plague. Mysterious fountains often have roughly equal liklihood (from a player’s perspective) to either provide a boon or ailment. This is often worth the risk in BX, because the player can fall back to having their ailment cured; the calculus changes in Knave.

Relic Magic

I’m going to ignore this section because I straight up don’t understand it after reading it 5+ times. If someone can give me a concrete example in the comments, that would be sweet.

edit 2024-09-19: A a reader came though! See Trev's comment.

Alchemy

Potions allow the drinker to produce a single significant magical effect for one turn.

Brewing a potion requires a fire, a cauldron, ingredients, and four hours. The player describes the effect and of the potion they are trying to make, along with the ingredients to be used (usually monster parts).

If the GM approves the effect, the PC makes an INT check. The PC adds +5 to the check if they spend an additional watch brewing. On a success, the potion is created. The parts are lost no matter the outcome. If the PC succeeds at the check by 10 or more, they have created a recipe for that potion and no longer need to make a brewing check to make it when using those exact ingredients.

Harvesting a monster part requires one 10-minute turn and the proper tools. Harvested body parts or plants take up at least one slot, due to the packaging and fluids they must be preserved in.

I think I get it. The players kill 10 Giant Rats and then spend 10 collective turns (with undefined proper tools - does a knife work?) to harvest 10 Giant Rat Livers, which each take up 1 slot. They put these on the ground. The bust out their cauldron (how many slots does a cauldron take?), and start a fire.

The player explains to the GM that they want to use the Rat Livers to create a potion of elasticity (which lets them squeeze through extremely tight spaces) like how a rat is able to do. They either negotiate or the GM just agrees, exhausted. The brewer has 5 INT and will take the full 8 hours, so they roll 1d20+10 vs DC 16 (default). They normally succeed on a 6-15 (50%) and succeed by at least 10 on a 16+ (25%). The fail the other 25% of the time.

After making 4 potions, it’s likely that they know how to make potions of elasticity from rat livers without rolling, so they do so in 4 hours going forward, to make ~7 total potions of elasticity. They can do this at the cost of 4 hours for every rat they kill going forward.

The design ultimately boils down to “player proposes a material cost and effect of the potion, the system specifies time-cost and probability, GM vetos”. The trouble here is that the player has no incentive to propose anything but the most effective potions using the most available resources. If the GM vetos it, all that happens is that the player raises their bid and tries again: they either describe a slightly less powerful potion or slightly costlier materials. They do this until either it isn’t worth it, or the GM agrees.

If we want to keep the each-table-comes-up-with-costs-and-effects design, better is to split that job up. Either have the player propose the cost, then the GM tells them what the potion will do, or have the player propose the effect, then the GM tells them what materials that would take. The book already specifies the time cost and probability, I would prefer some structure around material cost too (see Roleplaying's Fundamental Act - Vincent Baker). Even something like “it takes the monster parts from monsters totaling Y hit dice to create a potion replicating a X level spell effect” would do wonders.

In any case, this is potentially how we reintroduce our gygaxian building blocks, if we need to. The potions can be whatever, and so one can imagine brewing up all the stuff you need to make it to the city of brass, or potions that turn stone to flesh, or cure curses.

Overall, really neat system, though not my cup of tea. I have a massive preference for the game designer designing the magic and potions and whatnot, rather than offloading that to the GM or players. I also think this has massive, undiscussed, implied setting implications (If creating potions is fast, cheap, and effective, are potions commonplace? Do we expect wizards with monster farms churning out potions on an industrial scale?)

Imagine if the book (or a supplement) went through and provided a concrete list of potions each with explicit mechanical effects and directly tied those which monsters (from a popular bestiary) would need to be harvested.

Equipment

Warning: this is where, IMO, things get the most off the rails. I’ll start with what I really appreciate, and then get into the weeds.

Splitting up items into common/uncommon/rare and 1h melee (d6, 1 slot, 50c), 2h melee (d8, 2 slots, 100c), 1h ranged (d4, 1 slot, 60ft, 50c), and 2h ranged (d6, 2 slots, 120ft, 100c) is very good. BX gets weirdly fixated on tiny differences between weapons that, in practice, don’t end up mattering. When there’s crunch for no real reason (other than modeling reality), this is a huge red flag that something needs to change.

Here, we can decide if we want a short sword, longsword, or spear just based on flavor, rather than having to weigh the mechanics against our imagined character.

Coinage: All costs are in coins (c). 10c is the wage for a day of unskilled labor.

We know from earlier that when adapting OSR modules, that 1g = 1c. This one line is a massive re-imagining of the fantasy economy. Compare this to BX. A peasant (unskilled labor) wants 1g to be a mercenary for a month. A foot soldier will fight for 2g/month. Unskilled labor in Knave would want ~300c/month; so they’re demanding 300x higher wages.

Common items: Found in any settlement. Rope, torch, saw, arrow, quiver, etc. (5c).

This implies that a torch is worth half a day of unskilled labor. We also see that a single arrow costs 5c. Say that you loot the Dragon’s Hoard in The Black Wyrm of Brandonsford:

The total gold value here is 12405. In BX, that’ll buy you ~50000 arrows. In Knave, it’ll buy you ~2500 arrows (20x more expensive). In BX, that amount of money is enough to hire ~500 footmen for a year. In Knave, you can buy the unskilled labor from ~3 peasants for a year. Absolutely wild.

Melee weapons: One-handed weapons: d6 damage, 1 slot (50c). Two-handed weapons: d8 damage, two slots (100c).

Compared to BX, where a short sword costs 7g (7x price hike), and a two-handed costs 15g (7x price hike). Gear is significantly more expensive, though not as expensive as labor.

Armor pieces: PCs can wear up to 7 pieces. Each uses one slot and grants 1 AP (max of 7 AP or 18 Armor Class). Shield (100c), Helmet (100c), Gambeson (100c), Mail shirt (200c), Breastplate (500c), Arm plate (500c), Leg plate (500c).

Exactly 7 pieces are listed. I think the implied intention here is that you can’t double up on the same piece. A full suit (which you’re allowed to start with at character creation) would thus cost 1900c. The shield is 10x more expensive than in BX, and full plate is 32x more expensive. Yikes!

Funnily enough, a party of 6 fresh PCs start with more net wealth than the Dragon’s hoard in Brandonford!

Transport: The crew requirements listed are not included in the price.

• Mule, 50 slots (30c)

• War Horse, 80 slots (10,000c)

• Cart, 200 slots (50c)

A single warhorse costs as much as the Dragon’s hoard in Brandonsford!

A Mule costs 30c and Cart costs 50c. So, you have everyone start the game with full plate, then you have one player sell their leg plate for 500c and buy a mule and cart, right? This gives your team 250 slots to work with, and we presumably form a gentleman’s agreement to stop tracking inventory.

Is the intention here that we list out what’s in all 250 of those slots? I’m baffled as to how this is supposed to work in play. Like, at some point we just sort of shake our heads and say “wait, this all just probably fits, right?” and then we’re back to GM fiat which is what we already had with BX.

Cost of living: Use the following table when determining a PC’s lifestyle between adventures. A social level’s monthly cost of living is also a good guideline for bribes and gifts for someone of that class.

Beggars: 90c, Laborers: 300c, Bakers: 600c, Lawyers: 1200c, Knights: 3000c, Barons: 12000c, Dukes: 60000c, Kings: 600000c.

Base level feudal economics aside, this has large mismatches with both the game’s internal logic and also module compatibility. You take home the 12000g hoard from The Black Wyrm, and you find out that a baker is blowing through that in 2 years, just living normally. You accumulate 16000c treasure over bunches and bunch of expeditions (spending 5c per arrow along the way) to discover that a Knight blows through that in half a year.

In term of player behavior, the diegetic reward is much lower. Previously, they were risking their lives, but if successful would be lavishly wealthy and could finance small armies, buy strongholds, form religions, buy the finest clothes and drink the finest drink. Now, these Knaves venture into orc-ridden pits, risk death, and come out with a Baker’s wage.

Buildings

A small castle costs 2,000,000c. A character who reaches level 10 (max) will have earned (not kept) 500,000c. That means that if 4 level 10 (max) characters take every penny they’ve ever earned and spent none of it, they’ll be able to buy a small castle and be completely broke afterward.

You can’t adventure to buy castles in Knave; presumably the GM has to decide to gift you one. Why put prices in there if you literally can’t afford it?

Warfare

This system (prices aside) is very cool. I’m not a fan of the having a CHA check improve power by 50% (if you have 10000 soldiers, vs 15000, a single speech being worth 5000 soldiers feels ridiculous).

The underlying warfare game is wonderful.

You randomly determine how many objectives (0, 1, 2, or 3) each side secures, weighted by their relative army size. Then, for each objective, that side blindly picks a result from the following list:

• Capture: An enemy NPC is captured.

• Loot: An enemy item is looted.

• Slay: Slay 10% of the enemy force.

• Rescue: Cancel a “Capture” result.

• Guard: Cancel a “Loot” result.

• Shield: Cancel a “Slay” result.

So say the PCs win 2 to 1. This whole battle was against Bandits who kidnapped the King’s son. The Bandits would like to choose “Rescue” to prevent the PCs from “Capturing”. They also know that the PCs can choose “Capture” twice, and they’ll get back the kidnapped Prince anyway. They choose “Slay”, to go down swinging. The PCs, rather than risking another battle, choose “Capture” twice to guarantee that they get the Prince. In the end, they’re able to retrieve the prince, and bring back the bandit leader for questioning, but at a heavy blood price.

It resolves quickly, is low crunch, interesting, and creates compelling narrative. A+.

I wish more games used these sort of abstract bluffing games for more systems. Poker is a wonderful, wonderful game and games like netrunner and magic the gathering lean really hard into the yomi.

Downtime

Carousing: Carousing takes up a whole night and costs a PC d10×50c in villages, d10×100c in towns, d10×200c in cities. It also grants the PC XP equal to the amount spent. In the morning, succeed at a CON check or roll a mishap.

First, I think XP-for-spending-money rules are an absolute can of worms. Now, players no longer need to go into dungeons, they just need to acquire money. The king gives you a small castle for rescuing the Prince? Sell it for 2m, carouse for 2m xp.

Find some magic sword? Surely there’s a Knight that’ll pay top dollar (from their generous monthly expenses) to have it, and spend the money on XP.

Now, players really want to sell everything they get their hands on so that they can grow more powerful. Now every purchase also costs them the opportunity cost of XP. Why spend 30c for a nice night in a tavern when you can sleep in the common room and spend the extra change on more XP?

At a high level, I think the player behavior it drives is really unhealthy (penny pinching to blow it all on alcohol over and over). In terms of specific implementation, I find the consequences very frustrating:

1 You made a public fool of yourself.

2 Take d3 direct damage from a fight.

3 Pay d100c due to fines.

4 You are engaged to be married.

5 Lose d1000c from gambling.

6 Groupies follow you everywhere.

7 You’ve made an enemy.

8 You have an ugly, prominent tattoo.

9 Hangover: take -5 on all tests today.

10 You joined a local faction (p. 50).

11 Robbed: Lose all remaining coin.

12 You wake up in prison.

13 The building is on fire!

14 You’re expected to complete a mission (p. 51) due to your boasts.

15 A duel is scheduled for the next dawn.

16 You signed a shady contract.

17 A stranger’s corpse is on the floor.

18 A faction hates you (p. 50).

19 All your belongings have been stolen.

20 You meet a new companion who wants to join your party.

These are player facing. The player, knowing what the results could be, takes measures to hedge against these risks (they bury the money they aren’t using to carouse, so they can’t get #11 robbed).

Many of the downsides are comedy-of-errors inconveniences, and improvisational hell for the GM. What happens when someone makes a public fool of themself, mechanically? Do they have lower reaction rolls? What happens when someone gets engaged? Can they just call it off? What happens when someone wakes up in prison? Do I need to invent a penal system and improvise a prison break now? What happens when you’re “expected” to complete a mission? Can you just not do it? How do burning buildings work? What happens when you sign a shady contract? Can you just not fulfill it? What are you contractually obligated to do? What happens when a stranger’s corpse is on the floor? Whose corpse is it?

Keep in mind, that at low levels of play (when this is most significant), con checks likely have a 25-35% chance of success. It’s also likely that each player will want to carouse if one does, so we’re talking ~3 rolls on the failure table; 3 things to improvise every time the players are back in town with some money.

I’ll level with you here; I just want to play my modules! If the players do some carousing in Brandonsford in between when they loot the barrow and when they plan to set out for the goblin hideout, now there’s player-facing rules that can shoe-horn me into a get-out-of-jail scene.

Gambling: The player wagers some money (up to a house limit), then the GM rolls a d6. The player then makes a choice; either bow out and forfeit half of their wager or try to roll higher on a d6 than the GM’s roll. If they roll and succeed, they double their money. If they fail, they lose their entire wager.

This is mathematically elegant. Here’s the analysis:

So players should bet if the GMs roll a 4 or lower (a 1/3rd chance to win 2x your money is better than losing half of your money) and walk if the GM rolls a 5 or 6. The GM is equally likely to roll any particular number, so with optimal play the player wins loses ~6c for every 100c they bet.

This means that gambling is not a magic money creator (just like in real life), but that if you’re risk-tolerant it can be a fun and not-stupid thing to do. Awesome!

Recruiting

Hirelings: 300c per month. AC 11, HP 3, LVL 1, ATK punch (d2), MOV 40’, MRL 4. 10 item slots.

Hirelings (such as porters or torchbearers) perform unskilled labor and avoid danger at all costs. d10 are available in a village, 3d10 in a town, and 10d10 in a city. Reroll every month.

I have a (well documented) love for availability numbers. I would have love to seen availability numbers like these listed in the equipment section as well :)

So, a hireling wants 300c per month, which is the exact same as our 10c per day for unskilled labor number in the equipment section, that’s internally consistent.

I’m hung up on this line: “avoid danger at all costs”. Do they enter dungeons? Those seem dangerous. They have a morale of 4, implying that they would hypothetically pass morale checks 1/6th of the time, but if they avoid danger at all costs, would they ever make a check? Like, the things that normally cause morale checks are inherently dangerous, and those things are “avoided at all costs”, so what’s the deal?

Yet, the text specifies that these folks are torchbearers, who normally are the ones carrying the torch to provide light in a dungeon. Carrying a torch in a dark dungeon sounds dangerous to me, so I’m confused.

As far as I can tell, the playable reading of this is that “avoids danger at all costs” is a exaggerated description for “Morale 4”. They will accompany adventurers into dungeons, and don’t just automatically flee when a fight breaks out, but if a morale check ever does happen (like if a PC dies), they flee 5/6ths of the time.

Companions are generated like a PC, do not make morale checks, and will fight to the death for their employer. They take a half share of any treasure and XP found and can gain levels over time. They will only follow PCs of a higher level. A PC’s CHA sets the maximum number of companions that will ever follow the PC over the course of their life.

There’s a couple of key changes to retainers here, relative to BX.

First, we massively buff their loyalty and morale. BX retainers have morale (unspecified) and loyalty based on 7 + CHA (of the character hiring them). Knave retainers are perfectly loyal and totally fearless (for reasons unknown).

Second, we make it so that retainers must be lower level than the hiring PC, rather than less than or equal to. This is substantial nerf to level 1 PCs; they have to survive to level 2 before they can pick up their hired fighting help.

Third, a PC can only ever have as many companions as their CHA, in their lifetime. Huge nerf; makes hiring a companion very risky and losing a companion very costly. I’m relatively sure the play here is to max out your companion usages, level up the companions that you hire, then retire that PC and start playing one of the companions (who still has all of their slots). The companion should only be a level behind.

I also think that the mechanic harms the suspension of disbelief. I have 2 CHA, hire two people, and they both get gobbled by a dragon or some such. Then I travel halfway around the world to a totally new town where no one knows who I am, and find that I’m totally, universally incapable of hiring more companions.

I think the design intent here is to make hiring and losing a companion feel more significant, but now I think the pendulum has swung too far.

Designer’s Commentary

Careers are a quick way of giving players a background and a starting loadout without having to do a lot of shopping.

Here’s the text: “You gain those careers’ items, as well as any of the following that you can carry: 3d6×10 coins, 2 rations, a 50’ rope, 2 torches, any armor pieces or weapons (p. 38) and a quiver of 20 arrows. If the PC has any points in INT, they may receive a random spell book (pp. 22-25) for each point.”

So we’re still shopping. We’re getting an average of 110c of spending money, which is ~22 common items. That’s shopping!

Like, say that you roll antiquarian (5) and orator (65). You get an old coin, flag, lore book, 100 marbles, bullhorn, wax tablet, 120 coins (to spend on whatever), 2 rations, rope, 2 torches, bow, sword, shield, helmet, gambeson, mail shirt, breastplate, arm plate, and leg plage. Then you spend your money and choose ~11 of those things.

This is hardly solving the problem of not shopping!

Note that although PCs can start the game with expensive armor pieces if they wish, the resale value of those pieces would be very small, since armor has to be precisely tailored to a PC.

Why bother saying “very small” when you can specify something useful? What’s the general resale value of used goods, and how much worse is the resale value of leg plates? 50%? 10%?

I’ve removed the Copper/Silver/Electrum/Gold/Platinum system that many old-school RPGs use since that requires the GM to do math.

That math still needs to be done =/ Pop quiz! I’m running Silveraxe and the player’s loot this room, how much coin do they receive?

More troubling - how many slots does it take up?

Note that an aspiring overlord could use the construction rules to build their own dungeon.

Not following - is the implication here that players will ever have enough money to do that?

The ability to purchase expensive things like buildings, ships, and so on will likely not come from collecting piles of coins from dungeons but from finding unique treasures that can be traded to rich NPCs.

Do those treasures have the same 1g = 1c exchange rate as coins, right? If not, the exchange rate for treasure needs to be listed as well. If so, the PCs still don’t have anywhere near enough wealth to buy buildings and ships.

Players may find that with enough coin they can hire large numbers of followers. This is fine! Hirelings run away at the first sign of trouble and mercenaries are too smart to enter dungeons at all.

Here, Ben doubles down on the “avoids danger at all costs” remark, noting that hirelings “run away at the first sign of trouble”. Then what good are they? There’s going to be tons of signs of trouble on every expedition to a dungeon, and you’ll get signs of trouble in basically every room. Why do hirelings exist?

Conclusions

Knave feels, on a mechanics level, like a blueprint or a storyboard. You can’t live in a blueprint, but you can use a blueprint to build a house that you can live in, it just takes a lot of labor, knowledge and skill.

You can use a storyboard to plan and execute an actual story, and some people can look at a storyboard and get roughly the same satisfaction (because they’re so familiar with the art of story creation) as reading an actual story.

Knave (unlike BX) feels the same way to me; it isn’t an actual, stand-alone game that can play OSR modules. It doesn’t bother to define things like what melee combat are, and doesn’t have a bestiary or magic item list. I need other, actually complete and self-contained OSR books to use Knave. I find that frustrating.

On the upside, the GM tools are fantastic, and (like WWN) are where the real value lies. Lists of tools, fabrics, names, ingredients, symbols, magic schools, elements, forms, activities, mechanisms, room details, rooms, locations, signs, and careers. I think Ben poured a lot of love and research into these and it shows.