Tag Archives: combat

Pahvelorn initiative

Back in June, I discussed potential rules for what happens when magic-users wear armor. I ended up settling on a d6-based individual initiative system, which makes armor increase the chances of spells being disturbed by penalizing initiative. I have been using this for the past three or four months now, and it has worked out well, though we still sometimes forget to enforce the declare spells step. From here on out though: if the initiative die has been thrown, but no spell has been declared, no spell will be cast! This is, of course, true for NPCs as well.

The procedure is as follows.

  1. Declare any spells
  2. Each player rolls 1d6 (including the referee)
  3. Subtract 1 for a dexterity of 13+
  4. Add armor penalty (see below)
  5. Referee counts up (so lower is better), and players act in turn
  • Armor category: heavy (plate) = 3, medium (chain) = 2, light (leather) = 1
  • Armor skill: fighter = 3, cleric = 2, thief = 1, magic-user = 0
  • Armor penalty = armor category – armor skill, minimum 0

That is a complicated way of saying that if you are wearing more armor than your class uses normally, you take an initiative penalty equal to the difference.

Magic-users or clerics casting a spell must chant and gesture, and are thus natural targets. If the character is disturbed before the spell resolves, the spell is interrupted, and a saving throw must be made or the spell is lost.

All retainers (or NPCs being controlled by a player) act on that player’s initiative number, and most of the time all hostiles act on the referee’s initiative number (though occasionally the referee may use multiple initiative dice for different groups of NPCs).

Armor penalty also applies to physical ability checks, physical saving throws, thief abilities, and other similar rolls. For example, even outside of combat, a magic-user wearing heavy armor will take a -3 penalty to attack rolls, strength checks, saving throws versus dragon breath, and so forth.

If a thief “wins initiative” (that is, acts before all hostiles), the option to hide in shadows while in melee is available (requiring the standard thief skill check). This allows thieves to then either retreat without chance of pursuit or take actions in future rounds with surprise (including backstabs). Thieves may always attempt to hide in shadows when not in melee.

Any character that has not yet taken a turn and is not casting a spell may attempt to intercept an attack directed at another character, as long as the action makes sense in terms of fictional positioning. Fighters may attempt one intercept per round without sacrificing their standard action. Retainers that have been directed to defend may need to pass a morale check before they will attempt an intercept, depending on the situation. Succeeding in an intercept attempt requires making an attack roll and hitting a better AC than the attacker (this is essentially a contested attack roll).

The formal hide in shadows rule is new. The intercept rule has been active for a while, but is one that we tend to forget about. It is worth remembering though, because it makes fighters and hired bodyguards more useful.

Degree of success as damage

Men & Magic, page 19

ATTACK MATRIX 1.: MEN ATTACKING (Men & Magic, p19)

A simple house rule idea that just occurred to me, for D&D and similar games: damage from a successful attack = 1 + attack roll degree of success. The 1 is necessary so that some damage is still done when the attack roll succeeds exactly (otherwise, you are essentially applying a -1 penalty to the attack roll).

Some benefits:

  • One roll rather than two.
  • Makes clear the true nature of the attack roll (expected damage is the important thing).
  • Higher level fighters do increasing damage in a pleasing way.
  • The attack roll has many seeming degrees of success.

Expected damage numbers are provided below, for fighters of levels 1, 4, 7, and 10 against ACs of 9 and 2 using ATTACK MATRIX 1 from Men & Magic, page 19 (B/X uses identical numbers other than for zero level people; see the Expert Rules page X26). These numbers are averaged over all die possibilities, including misses, and so are expected damage per round. Comparisons with expected numbers from weapons dealing 1d6 (average: 3.5) damage are provided in parentheses.

Fighter Level Damage Versus AC 9 Damage Versus AC 2
1  3.3 (1.925)  0.5 (0.7)
4  4.55 (2.275)  1.05 (1.05)
7  6.8 (2.8)  3.3 (1.925)
10  8.55 (3.15)  4.55 (2.275)

You will see that in general, this shifts the damage potential up for most situations (all, in fact, other than first level versus AC 2, at least of those data points shown in the table). ACs 9 and 2 were chosen because they encompasses the full OD&D range from unarmored to plate & shield. Results are independent of other bonuses, which will just raise or lower the expected numbers for all schemes. I imagine the numbers would remain somewhat similar if using a simplification such as hit dice as attack bonus. The same numbers obtain for the other classes, though the level ranges are different (a 9th level cleric hits as a 7th level fighter, for example).

The expected values are close enough that this adjustment will obviously not break the game, though it might shift the dynamics slightly. One could also cap damage at 6 for slightly more restrained damage results (just for comparison, with that modification numbers versus AC 9 become: 2.55, 3.15, 4.05, 4.65 for levels 1, 4, 7, 10 respectively and have a much smaller standard deviation).

The spreadsheets that I used to calculate these numbers can be found here:

JRPG Basic combat positioning

Black Mage

Black Mage (personal sketch)

Here is an abstract system for managing combat options that hopefully provides trade-offs regarding risk and effectiveness and interesting tactical choices. I think this system has intuitive guidelines for attempting things like breaking off from combat, or setting up a surprise attack, which are often somewhat hard to handle when not using tools like grids or other cumbersome procedures. These are only part of the combat rules. The initiative and turn taking procedures are still to come.

Many early JRPGs make a distinction between frontline and rear combatants. For now, I don’t think this distinction deserves separate positioning rules, as it can be handled by the intercept rules (a character that is protected by an interception is much like a “rear” combatant). It may be worthwhile to add more depth to reach weapons, though.

I may formalize some of the terms that are currently handled more descriptively. For example, I was thinking of calling casting spells or firing missile weapons while engaged in melee a perilous action, but I also don’t want to fall into the trap of legalism.

Rules for cover at ranged position will probably also be added later (which may end up just being a simple AC bonus, as is probably familiar from other games).


Tactical Positioning

The relationship of combatants to each other in battle is managed by abstract positions. These postions determine the combat options available and restrict movement in certain ways.

Combat positions include ranged, melee, and concealed.

The lists of actions provided should not be considered comprehensive. Anything may be attempted. The referee should adjudicate the outcome using the action examples given as a guideline and call for ability checks, contests, or saving throws as necessary. Particular class abilities or skills may provide additional options, such as a thief’s ability to use the steal skill. See the relevant ability description for details.

In general, attempting anything other than attacking in melee should probably be subject to a saving throw to avoid damage (following the example of firing a missile weapon while in melee), though more latitude is reasonable for combatants in ranged position.

Ranged

Combats begin with hostile participants in ranged position relative to each other.

Ranged attacks may be targeted against specific ranged enemies, but targets are determined randomly if firing into a melee.

Characters in ranged position are drawn into melee if attacked by a melee combatant using a melee attack. Not all characters in ranged position are necessarily subject to melee attacks, however. For example, a combatant on a balcony above a room, firing arrows down into the room, is not subject to most melee attacks, and thus cannot be drawn into the melee barring special circumstances.

Common ranged actions:

  • Use a ranged weapon against a specific target not in melee
  • Fire into melee (determine target randomly)
  • Charge into melee and attack
  • Hold action in preparation for an intercept
  • Flee current combat, assuming there is an escape route
  • Attempt to hide using the stealth skill

Melee

Melee includes all characters in a limited, abstract space attempting to physically harm each other. Exact positioning is not tracked. Ebb and flow is assumed as combatants jockey for advantage and defend themselves.

A melee attack against a character outside of melee that is not intercepted draws the target into melee, whether or not the attack was successful. Characters in melee may retreat from the melee to ranged position as an action.

Area effects, such as some spells, target entire melees, which includes all combatants participating in the melee.

Ranged weapon attacks may not be made against specific targets in a melee. Instead, the target is determined randomly and then resolved as normal (attack roll and so forth). This abstraction represents the chaos of battle. Random targeting does not apply to ranged spells with individual targets, however. For example, a black mage may target a specific melee combatant with the shock spell.

Characters that use ranged attacks or cast spells while in melee must succeed in a dexterity saving throw or take 1d6 points of damage.

Common melee actions:

  • Make a melee attack against an enemy in the same melee
  • Engage someone at range to draw them into the melee
  • Retreat from the melee to ranged position
  • Make a ranged attack or cast a spell (this involves extra danger)

Concealed

Concealed characters may take an action with surprise and may not be the target of individua effects. Concealed characters may, however, still be affected by some area effects, depending on the nature of the effect and how the character is hiding.

Characters at ranged position may attempt to hide. This requires a stealth check. If the check is successful, the character becomes concealed.

Concealment is not always an option. This is dictated by the environment.

Some effects allow the detection of concealed characters (such as spells of the heightened senses of some creatures).

Fleeing

Fleeing from combat is only possible from ranged position. Characters in melee must first retreat to ranged position (this is an action). When in ranged position, a character may spend an action to leave the combat, assuming there is an escape route. Any character at range is drawn into melee if subject to a melee attack (whether or not the attack hits). Melee attacks may be intercepted by other characters or effects, allowing retreat. See pursuit for handling situations where enemies attempt to give chase.

Multiple Melees

Most of the time, a single melee area is sufficient to represent an armed struggle. However, there are cases which require the consideration of multiple melees, such as an adventuring party being attacked from both sides in a hallway. Large open spaces may also sometimes demand the use of multiple melee zones which could potentially merge and divide based on game world circumstances. The melee/ranged abstraction is meant to structure combat in a way that logically represents the chaos and risk of armed struggle, and may be adjusted on an ad hoc basis as needed by the referee.

JRPG Basic Starting Weapons

In JRPGs, weapon upgrades are a big part of advancement. To facilitate that aspect of gameplay, characters start with only a common weapon, and must find or purchase better weapons. I have created a town inventory system (which will be a future post) to go along with treasure tables so that upgrades can be bought and sold according to core rules as well as discovered (though randomness is still involved). Acquiring the most powerful gear will still require adventuring.

I’ve attached some properties to these weapons, but that might end up being overly complicated for this project. In the final game, I may go for something less involved. Maybe just 1d6 for one handed and 1d8 for two-handed common weapons.


Common Weapons

  1. Bamboo pole (1d6, reach)
  2. Cudgel (1d6)
  3. Hunting bow (1d8, ranged, two-handed)
  4. Knife (1d6, throwable)
  5. Scythe (1d8, two-handed, unreliable-1)
  6. Sling (1d6, ranged)
  7. Smith’s hammer (1d6)
  8. Quarterstaff (1d6, two-handed, parry/melee)
  9. Wood axe (1d8, two-handed, unbalanced)
  10. Wooden sword (1d6, quickdraw, riposte)

Bamboo pole (1d6, reach)

Though perhaps the most primitive of spears, and lacking the weight for proper throwing, bamboo poles can still be wickedly sharp. Also useful for poking things from a distance.

Cudgel (1d6)

This heavy oaken club probably once belonged to a town guard, merchant convoyer, or barkeep. What, you are expecting some added ability? It’s just a heavy stick! Being made of wood, it probably floats though…

Smith’s Hammer (1d6)

Good for smashing heads or boxes (+1 to force skill checks).

Hunting Bow (1d8, ranged, two-handed)

This simple hunting bow should be unstrung when not in use and takes one round to string.

Knife (1d6, throwable)

Common tools of daily life, knives also make good weapons of last resort.

Scythe (1d8, two-handed, unreliable-1)

Not the most reliable weapon for combat, the scythe is nonetheless terrifying to behold and can inflict gruesome wounds, though it requires two hands for proper use. Some black mages carry scythes purely for the visual impact.

Sling (1d6, ranged)

Basically just a cloth cup attached to a pair of cords, the sling is one of the simplest, lightest, and cheapest weapons around. You can often pick up ammo from the ground, though metal sling bullets are more effective (improvised ammo only does 1d4 damage).

Quarterstaff (1d6, two-handed, parry/melee)

Quarterstaves are heavy wooden staves that are often shod in metal at the ends for durability. They excel in defensive maneuvers.

Wood axe (1d8, two-handed, unbalanced-1)

This heavy axe was designed for splitting wood, but can split skulls as well. It requires both hands, and does not have the best balance for use in combat. On a natural 1, swinging the axe has left you overextended or unbalanced, and your effective AC is decreased by 2 until your next combat turn.

Wooden sword (1d6, quickdraw, riposte)

This practice sword is well-carved, and weighted with a core of metal. Though intended for practice and lacking a keen edge, in can still deliver brutal strikes.

AC by class AND level

Armor of Stefan Batory (source)

Armor of Stefan Batory (source)

In my recent post about magic-users and armor, there was one option for doing strict class-based AC (that is, all fighters have AC 3, all magic-users have AC 9, and so forth). One downside of that system is that other than gear with special enchantments it doesn’t allow for much in the way of advancement (which is a big part of what makes D&D work as a game).

Here’s another idea that works with OD&D hit dice to address that concern. Take the combined hit dice value (for example, HD 5+1 = 6) and subtract that from 9 (or add it to 10 if using ascending AC). That is the characters armor class (improved further by one if using a shield). This gives fighters the best AC, but also allows gradual progression. Following the same example, a character with 5+1 HD has an AC of 3 [16]. This is sort of the defensive equivalent of using hit dice as attack bonus. AC should also be capped, depending on the desired campaign power curve (if I was using this for Vaults of Pahvelorn, the best AC from hit dice would be 3, or 2 with a shield, because danger should always remain). Really, the term “AC” here becomes a bit vestigial; it’s really more of a defence stat, but continuing to call it AC probably helps from a UI perspective, given that it works exactly the same as AC.

If you wanted to preserve some mechanical effect from armor, maybe allow it to add a bonus to the death saving throw† (light = +1, medium = +2, and heavy = +3). This further reinforces the idea that HP is a mixture of elan and resolve, and that there are no potentially telling blows until HP have been exhausted. The downside of armor would be an encumbrance style penalty to ability checks, non-death saving throws, and escape rolls.


† Rather than deal with negative HP or have death occur at 0 HP automatically, I allow PCs to make a death save. If this is passed, the character is unconscious. If it is failed, the character is slain. This is one of my favorite house rules, and probably deserves its own post so that it can be linked to directly.

Magic-Users and armor

Brainstorming several different approaches. Goals: should provide interesting trade-offs while respecting the fictional logic.

  1. Chance of spell failure. Maybe N in 6 chance, where N is based on the armor heaviness (light = 1, medium = 2, heavy = 3). Problem: the rational course of action is to carry a suit of armor and put it on after all spells have been used. This is lame. Such an approach might even encourage annoying things like taking off armor to cast spells and then putting it back on (yes, this can be balanced with random encounter checks, but still lame).
  2. Wearing armor causes an armor penalty, equal to the type of the armor (where light = 1, medium = 2, and heavy = 3). This penalty applies to all physical checks (attack rolls, saves, ability checks, etc) and works much like the encumbrance penalty (and in fact is cumulative with it). This could be offset by a class-based armor skill (fighter = 3, thief = 1, magic-user = 0). Problem: while this would increase the game cost of magic-users wearing armor slightly (-3 to physical saving throws, escape rolls, and constitution checks associated with drowning are big deals), it doesn’t really impact casting spells at all. Further, one seeming corollary of this system is that zero level characters would probably have armor skill of 0, complicating the common case if consistency is maintained.
  3. Use a roll-to-cast system. I love this, but it’s also very invasive, and probably requires reworking many spells as well to do correctly. I would like a solution that supports the traditional Vancian system.
  4. Maybe magic-users just don’t get as much benefit from armor, but still take all of the downsides? Kind of like this. Magic-users could get AC 8 from leather, AC 7 from chain, and AC 6 from plate. Has promise, but is perhaps too complicated. Also, I don’t much like the corollary that magic-users have less skill with armor than standard zero level civilians. I prefer to think of all the character classes as somewhat competent adventurers; magic-users are not assumed to all be frail academics. I suppose zero level characters could also gain less benefit from armor, but that seems to introduce unattractive complications.
  5. Magic-users can wear armor, but only special enchanted armor (elven chain perhaps). This necessitates some justification for why magic-users can’t cast in normal armor (such as overly scientistic claptrap like how metal interferes with magic, which just invariably leads to subversion of the balance rule through creation of things like wooden or chitin armor). Also, it means that all “fully upgraded” magic-users must aspire to finding a suit of special armor (kind of like how cloaks and rings of protection are so critical in AD&D). Suboptimal.
  6. Learn to stop worrying and love magic-users in armor. I don’t mind this aesthetically. In fact, I quite like sorcerers in armor (picture by Stefan Poag), but it does seem wanting in terms of class balance (greatly decreasing the relative combat power of the fighter, specifically). That’s not the end of the world, especially as I have already improved the to-hit rolls of fighters in my current game.
  7. Provide compelling alternatives that compete with armor (such as robes). Combine with options 6, and maybe a increase the encumbrance cost of armor while not adding a full more physical penalties on top of the general encumbrance penalty. This is close to status quo, with the exception of explicitly legitimating magic-user armor use.
  8. Class-based AC. This has a certain attraction, especially if taken to the logical extreme where you literally say that fighters and clerics just have AC 3 no matter what they wear, thieves have AC 7, and magic-users have AC 9 (aside: clerics really should be the AC 5 class, but whatever). That solves the problem in one sweep, maintains both abstraction and balance, and makes armor only important as a kind of magic item or cosmetic affectation. That last bit is also maybe a problem though, as it does away with consequences of armor for things like drowning.

As you can see, my thoughts are all over the place. Anyone else have any good ideas or suggestions?

Edit: added option 8 based on Guy F.’s comment on Google Plus.

Weapons Revisited

Update: I recommend using the slightly simpler approach to this same basic idea described in the Weapons Quick Reference post.

Image by Piranesi (source)
Image by Piranesi (source)

I started work compiling various blog posts into a Vaults of Pahvelorn Player’s Guide, and the very first thing I decided to look at was my old weapon properties post. This is one of those topics that I think would really benefit from the accessibility of being included in a player’s document, because despite being designed mostly as bonuses, in play we still often forget many of the weapon features. One of my ongoing personal design goals is to make weapon choice just as interesting as spell choice, without relying primarily on variable damage dice (which promotes an overly numerical approach that I find lacks interesting trade-offs).

Of course, rather than just copying the old rules into the document and moving on, I immediately start to significantly revise them. Since I wrote the original version, I have been exposed to Apocalypse World and Dungeon World, which both take approaches to weaponry that have influenced me substantially. I think weapon tags are a wonderfully efficient way to remember the various features, and while yes you need to read the definitions once, it seems like once you know that dangerous-1 means backfire on a natural roll of 1, you are unlikely to ever need to look that up again. I think this version below is much improved over the original, and many of the rules have also been simplified (for example, riposte just automatically deals damage to attackers that miss and roll poorly rather than requiring another attack roll).


Weapon damage is 1d6 by default. Some weapons have additional benefits, as described below.

WeaponProperties
Axemelee, damage: re-roll 1, shield-smasher
Daggermelee, quickdraw, throwable, range-2, concealable, grapple
Macemelee, penetrating-2
Swordmelee, quickdraw, riposte
Two-handed swordmelee, two-handed, damage: 2d6 take highest
Pole armmelee, two-handed, interposing, damage: 2d6 take highest, reach, awkward-2
Quarterstaffmelee, two-handed, parry (melee)
Spearmelee, throwable, range-3, interposing, reach
Javelinmelee, throwable, range-5
Clubmelee
Shieldarmor-1, parry (missile), damage: 1d3
Bowmissile, two-handed, range-7, reload-0
Crossbowmissile, two-handed, penetrating-2, range-6, reload-1
Slingconcealable, missile, insignificant, missile, range-4, reload-0, versatile ammo
Oilbombdangerous-1, flaming, immolating, penetrating-2, range-1, reload-1, unreliable-3
Powderbombarea, damage: 2d6 take highest, dangerous-1, flaming, range-1, unreliable-3
PropertyEffects
Areano attack roll, all enemies in area of effect take damage, save for half
Armor-N+N armor class
Awkward-N-N attack if not used at reach
Concealableeasy to hide in standard clothing (will not be noticed without a search)
Damagedamage inflicted is modified as stated
Dangerous-Nbackfires (damages wielder) on natural rolls of N or less
Flamingdeals fire damage and flammable targets must save or be lit up
Grapplefuture attacks auto-hit if a dexterity/strength contest is won
Immolatingsave or ignite, continuing damage, additional save per round
Insignificantdoes not count as an item for encumbrance purposes
Interposingmelee enemies must save to attack wielder, and on failure take damage
Meleemay only be used when engaged in melee (essentially, range-0)
Missilerequires ammunition
Parry (type)save to deflect one attack per round that hits (limited to type, if given)
Penetrating-N+N attack versus targets with armor
Range-Nweapons of higher range afford a free attack round as enemy closes
Reload-Ntakes N rounds to reload (reload-0 fires every round)
Ripostedeal damage if enemy misses and rolls 5 or less
Shield-smasherdefender with shield must save or have their shield destroyed
Two-handedrequires both hands to use effectively
Quickdrawmay ready and attack in the same round
Unreliable-Ndoes not function on rolls of N or less (overridden by dangerous)
Versatile ammomay use any small hard object (coin, rock) as ammo

Further Notes

  • “Mace” includes warhammer and military pick.
  • Wielding two weapons (where one is not a shield) grants +1 to the attack roll (credit to Philotomy).
  • A flask of oil may be used to coat a weapon and then ignited to give a metal weapon the flaming property (likewise, arrows). This is a reload-1 type operation. On attack rolls of 5 or less, the fire goes out. Such flaming weapons will also go out after one exploration turn (or after combat).
  • Target of a grapple may spend an action to attempt to free themselves (this is another dexterity/strength contest).
  • Ranges: bomb < dagger < spear < sling < javelin < crossbow < bow
  • Ranges are not measures, but are only used in relative comparisons.
  • Shooting or throwing into melee: determine target randomly.
  • Crushing/bludgeoning damage is sometimes important (skeletons, living statues, and so forth), but I decided that this is probably clear enough contextually, and thus doesn’t require a property (Google Plus discussion). I may change my mind on this, though.
  • A strength/dexterity contest means: both contestants roll either a strength or dexterity check (their choice). This is a less than or equal to d20 check, and the one that makes it by the most wins the contest (ties go to the defender).
  • I kind of want to add a great axe and maul (two-handed varieties of the axe and mace), but that would probably necessitate giving the two-handed sword an added benefit, and I haven’t been able to think of anything that I like.

Thanks to Robert G. on Google Plus for suggesting the property name interposing.

For ease of future reference: G+ threads on flaming oil: here and here (and Philotomy).

Another stunt system

Jeremy D. recently posted this simple and elegant stunt system based on the DCC deed die but intended for use with traditional rules. The basic idea is that attack rolls are made with d16 + class hit die (assuming B/X style variable hit dice), and a pre-declared stunt is successful if the hit die comes up 4 or greater and the attack roll is high enough to hit. This gives fighters (d8 hit die) roughly a 62% (4 or higher on a d8) chance of pulling off a stunt, given a high enough modified attack roll. Magic-users (d4 hit die) have a 25% chance (4 or higher on a d4). What’s the trade-off? If the stunt fails, the attack misses, even if the number would have been high enough to hit had a stunt not been attempted.

The system is clean, but does require Zocchi dice, which is a downside. However, thinking about the trade-off gave me an idea for another system based on a similar principle. The basic idea is to gamble on two independent dice both coming up high. So why not make stunts require success on two attack roles rather than one? This would still represent a single action, but would be similar to roll twice, take the lowest (since both need to hit). The essential dynamic of fighters being most able to benefit from stunts would be preserved, because (assuming the same level) fighters will have the best chance of hitting. This also seems like it would be easy to communicate to players: just make two attack rolls; no new mechanics would need to be introduced.

Interception

To throw yourself in the path of an attack directed toward another character, make an attack roll. If this intercept roll hits an armor class as good as the attack roll being intercepted, the interceptor becomes the new target of the attack and moves between the attacker and the original target. The decision to intercept must be made prior to the attack roll.

Fighters may perform one intercept reaction per combat turn. Characters of other classes may only perform an intercept if they hold their action. Retainers directed to intercept attacks may be required to pass a morale check.

I want to add a sentence about how intercepts can only be attempted if they make sense logically, or are supported by the fiction (or whatever), but don’t have quite the proper language down yet.

Hit dice as attack bonus

Costume design for the Opera "Prince Igor" by Alexander Borodin

Ivan Bilibin, costume design (source)

When I recently played in Evan’s Uz campaign, he had hit dice do double duty as attack bonus. Uz is based on Swords & Wizardry WhiteBox, which, like OD&D, only uses the d6 for hit dice (modulating the difference between classes using bonuses as described here). The elegance of this approach impressed me. Only one number is required for both hit points and offensive ability. And it generalizes to monsters, though you might want to cap the bonus depending on the campaign power curve (max +10 seems pretty reasonable to me).

I don’t remember exactly how he did it, but this is how I might do something similar. Fighters find their attack bonus by adding their hit dice expression together. For example, a fighter with HD 5+1 has a +6 attack bonus. All other classes ignore the bonus part of the hit dice expression and just use the base HD. So, a cleric with 4+1 hit dice attacks with +4. Tougher classes always have more hit dice than weaker classes, which is also how attack bonuses should work.

This is what the attack bonuses would look like given the recently posted rationalized hit dice progression (hit dice are in parentheses):

LEVEL HIGH MEDIUM LOW
1 +2 (1d6+1) +1 (1d6) +1 (1d6)
2 +3 (2d6+1) +2 (2d6) +1 (1d6+1)
3 +4 (3d6+1) +2 (2d6+1) +2 (2d6)
4 +5 (4d6+1) +3 (3d6) +2 (2d6+1)
5 +6 (5d6+1) +4 (4d6) +3 (3d6)
6 +7 (6d6+1) +4 (4d6+1) +3 (3d6+1)
7 +8 (7d6+1) +5 (5d6) +4 (4d6)
8 +9 (8d6+1) +6 (6d6) +4 (4d6+1)
9 +10 (9d6+1) +6 (6d6+1) +5 (5d6)
10 +11 (10d6+1) +7 (7d6) +5 (5d6+1)

Another way to look at this rule would be that all classes use base hit dice as attack bonus, but fighters get an additional +1. Or, one could just use the additive hit dice for all classes, which makes things simpler (no special case for the fighter) at the cost of decreasing the relative power of the fighter slightly.

Using this system necessitates running with ascending AC. I have actually been considering switching to ascending AC off and on for a while now anyways. This attack bonus system is probably simpler and easier to understand than my attack ranks system, and it does away with another table (usually a good thing).