May 31, 2009

Stream of Consciousness About Roles

So, D&D 4E has Leaders, Strikers, Defenders and Controllers, which are all combat roles. Earlier editions had Fighter, Healer (cleric), Skill Monkey (rogue/thief) and Holder Of Win Buttons (wizard). This is an interesting shift and I'd like to expand on it. First, what the heck am I talking about:

  • Fighter and Healer - An combat system with hit points has two obvious "roles": The guy who deals damage to the enemy (4E striker) and the guy who heals damage dealt to his allies (4E leader). 4E expands on this by adding a guy who inflicts status effects (controller). You could argue that a defender heals indirectly, because healing powers are more effective on him (they heal a percentage of his max HP, and he gets a higher total than others).
  • Skill Monkeys were taken out of 4E - or rather, everyone is a bit of skill monkey since everyone gets a similar amount of skills.
  • Holder of Win Buttons is the odd guy out. A Wizard in earlier editions get a limited number of spells that can solve problems when the party's fighting tactics or puzzle-solving skills fail. Sleep wins combats at low level, Fireball wins them at higher ones, Knock sorts out pesky locked doors if the thief is out of commission, and then there are all the divination rituals for when the party is really stumped. Of course, at high level, the Wizard get too many win buttons and goes from "useful party member" to "party leader with three pets".


Anyway, there are two oddities with the old roles, at least as they appear in D&D 3.

  • Nonstandard classes don't fill the roles well (and don't come with instructions).
    • The Druid is a worse healer than the Cleric - the same healing spells are higher level for him, and he gets no spontaneous casting of cure spells - but gets some win buttons to compensate.
    • The Bard (and Monk) can't fill in for a Rogue - at least not in the trap disarming department.
    • The Ranger, Paladin and Barbarian do well enough at replacing the Fighter.
    • The Sorcerer can blast better than the Wizard, but is out of luck when he needs an obscure utility spell.

  • Splitting of spotlight. First the rogue disarms a trap and unlocks a door. Then the fighter rushes in to kill everything, while the cleric stands by to heal him. Then more traps - the thief gets the spotlight again. Meanwhile, the Wizard is standing by waiting for the best moment to step in and press one of his "win buttons". This is a design that's worked for 30 years, but not one I agree with, and plenty of other game systems have changed it.


Of course, D&D 4E missed an opportunity too. Everyone has a combat role, and is relatively competent in skill challenges. However, there is little differentiation in what skills characters have. There is some, in that Arcane characters are likely to have Arcana, Divine ones probably have Religion, Primal ones have Nature and Martial ones get... heck if I know. Still, the roles make it easy to get a good spread of combat roles when characters are rolled up, while skills still have to be doled out manually.

I don't have a better solution. The problem with delineated skill-roles is that the obvious division - Talker, Mechanic, Athlete and Living Encyclopedia - split the spotlight again. The Talker rules social encounters while the Athlete and Mechanic disarm traps. The best is probably to focus on the skills that your stats work with - a diverse group should cover most of the stats, and that provides a natural division of skills. Watch out if everyone plays Arcane classes, though - you'll rock at intelligence-based skills and be mediocre at the others.

May 30, 2009

Circuitous Dungeons Part II

(Part I here.)

These two could also work:



Circuitous Dungeons?

Future project: Populate these two "dungeon maps":



May 24, 2009

On Roguelike-Platformers

Is that even a computer game genre? Platformers where the game utterly hates you, does not provide save states or extra lives, procedurally generates the levels and - in the case of Spelunky - has extensive interaction with the level environment. Well, if there isn't, I'm making it up now.

Spelunky (downloadable) by Derek Yu takes you, in the guise of a chibi Indiana Jones, through four different "worlds" with randomly generated levels. And it's deadly. You get a life meter with three hearts, but that's not worth much once you realize that giant boulders, bombs and certain monsters instakill you anyway. And then you restart from the beginning of the game, because Spelunky hates you.

Yet it's still fun, because the levels are new in every playthrough (though with familiar building blocks after the twentieth playthrough or so) and a game takes about 10 minutes so you aren't actually losing that much progress when you die. Much like Rogue, except as a platformer.

It also borrows the concept that everything is useful (more prominent in Nethack than Rogue), so eventually you will be picking rocks and skulls off the ground to trigger motion-sensitive arrow traps, and then tossing the arrows at enemies to kill them. More evil players do that with human corpses instead...

Tower Of Greed (flash game) isn't nearly as complex as Spelunky, but it still has a certain roguelike appeal. You jump up a downwards-scrolling level like in all those other downwards-scrollers. The twist is that you are supposed to gather gems along the way, and then you can leave the tower (through doors along the tower) when you feel you have enough gems. If you die before leaving the tower, you don't get recorded in the highscore table, and (more painfully) your achievments aren't saved. You have to quit willingly to have the game count.

Of course, there are "achievements" for both how many floors you can survive and for how many gems you gather - I haven't quite gotten to 100 floors yet.

Simplistic, yes, but the controls are close to perfect (unlike the feeling of steering a paraplegic slug I get from many other scrolldown platformers) and the levels are somewhat procedurally generated (though with big building blocks, so you end up learning how to pass certain floor designs anyway).

And much like Spelunky, a game takes about 5 minutes, so failure doesn't lose you that much game time. This... has gotten me hooked in a way no other game in this genre has before. Good job.

I'm keeping an eye out for other games in this genre.

May 17, 2009

Summon Monster V in D&D4

Summoning in D&D 3 is a tricky thing. Done right, it allows one NPC to multiply its effectiveness by spamming summoning spells and letting the beasts fight for him. Sure, most summonings (notable Summon Monster/Summon Nature's Ally) summon creatures weaker than the caster, but Gate, Lesser Planar Ally and so on can make a dedicated summoner really nasty. And if the player manage to disrupt the casting, or dispel the summoned creature, the fight just became a cakewalk (or it's balanced now and would have been a potential TPK with the summon still around).

And let's not get into monsters summoning other monsters. A D&D 3 Marilith has a 20% chance of summoning another Marilith and making the fight twice as hard. How many mariliths do you account for when figuring how tough the fight will be? (XP is easy - summoned stuff don't give XP when killed.)

This is interesting for any edition, but D&D4 focuses on balance, and having Schrödinger's encounter be potentially twice as hard as planned is outside the design paradigm for 4E.

Fortunately, D&D4 provides examples of summoning monsters, though it's just one and a half in the whole Monster Manual. The half is the Berbalang, which summons a duplicate of itself by spend one-quarter of its HP. More interesting is the Pit Fiend, who has an encounter power that summons either 8 legion devils, 2 war devils, or half of each. However, they're weak compared to the Pit Fiend himself - if he's level appropriate, the summons are cannon fodder. Which is probably as it should be.

I had another idea. Either make the summoner a solo and let it summon minions as a recharging power (the summoner should probably be a solo), or give the summoner a higher XP cost to make up for a one-time summoning power. (The Pit Fiend does not do this - it follows the guidelines for statting an elite almost perfectly.)

If I was going to make a Marut Blademaster (Level 21) with the ability to summon his debtors, I'd double his XP value (from 3200 to 6400) and give him a minor action encounter power to summon a set bunch of monsters worth 3200 XP. Say, give him these choices:
  • 4 Angel of Valor Legionnaires (L21 minions - how did these get in the Marut's debt?)
  • 4 Legion Devil Legionnaires (L21 minions - our Marut gets around.)
  • 1 Ghaele of Winter (L21 standard monster - okay, he outsmarted a noble Eladrin?)


There you go. Two monsters masked as one dude summoning another. Quadrupling the XP budget could obviously make one weak caster with the ability to summon up three times his own worth in opponents once the PC:s barge in.

May 10, 2009

XP for GP in 4E

...or how to go old-school and reward characters getting gold by any means rather than rewarding fighting. If that's your kind of thing.

In the old style of gaming, characters got most of their XP for bringing home loot (at a 1:1 gold:XP ratio) and just a little for killing monsters and rarely any for doing quests (except that quests could give monetary awards). Rewarding what you want the game to be about is a good thing, so if you want characters to be creative, sneaky and conniving, rather than show tactical prowess, you might want to go back to that scheme. Fortunately, WOTC makes that very simple in their own Dungeon Master's Guide.

Simply look at the treasure parcel tables. Yes, it's four pages of tables, but you want the headers that tell you how much gold 5 PC:s are expected to find during one level. Once your group finds this, they level up. Scale as needed if you don't have exactly five players.

(Alternatively, the expected gold is twice the cost of a magic item of the party's level, listed in the Player's Handbook.)

Magic items would still have to be scattered around the adventures. I'd spread them about liberally with some to spare - the sky won't fall if the group has one item too many or too few. One may want to make sure the group at least has the expected +2 to attacks and defenses as they leave the heroic tier (and +4 when going epic). Maybe even require such equipment as a prerequisite to going to level 11/21 or adding those items as rewards for the quests taking them to the next tier.

Taken together, this plan would promote acting like an explorer rather than a tactical fighter. It's up to you if that's a good thing.

May 03, 2009

Item-based Skills, or How I Reinvented Scrap

I was pondering a roguelike. Many roguelikes have you "build" your character over time. Even Nethack has skills, even if they're not vital in the case of weapon skills (but magic skills can make the difference between being able to cast a spell reliably and not). Incursion is based on the d20 SRD and has build options accordingly. I don't know how much ADOM's skills affect gameplay, but they're there too. Of course, most 7 Day Roguelikes don't have skills, as implementing a decent skill system takes time and detracts from the design goal of the game.

Anyway, I don't fancy skill systems. They lock you into one specific way to play the character (range vs melee vs magic vs sneak vs charm etc), and can serve as a trap for players who haven't played enough - if a Nethack player doesn't know what to spend skills on, he might waste skill slots on weapons as a wizard, or on bad weapons as a fighting class. And Nethack doesn't even deign to tell you that there is such a thing as skill slots.

So I was thinking of a system where "skills" are more fluid. Either allow generous retraining - you have X dynamic skill points that can be shuffled among skills - or base your abilities entirely on loot. An itemless character would just have the basic abilities of an adventurer - comparable to the protagonist of Rogue, but carrying a book of spells would instantly give him magical abilities - and then various implements could refine him further like in D&D4 (where staff, wand and orb wizards get different special abilities). The "wizard" could shed his magic items and grab a set of thief equipment or fighter armament instead, or even mix and match - a robe may boost spells, but real armor gives better protection in melee which is still interesting if you are carrying a Staff Of Becoming Ground Zero Of A Fireblast.

This would be relatively balanced with harsh carrying capacity rules (to prevent people from "multiclassing" too much by carrying extra items) and generous item drops (to encourage changing your setup without it taking half the game). Angband-style levels where levels vanish when you leave them would prevent extensive stashes.

Alas, Scrap already does this, but with the player being a robot scavenging robot parts. Nice game, try it.