A system for forum games?

[Dominicius]

I think most of us have tried play by bost at least once or twice. Enough at least to realize that D&D does not really support such a format.

First of all you have the per round structure that makes a single combat drag on for up to a week as you wait for people to get online and post and then all the AoO's and immediate actions slow it down even more.

If someone were to make a system for a play by post format, obviously those things would need to go. The per round structure would probably need to be made a lot broader so that a lot more things can happen in a single post. I would even go so far as to say one would need to diminish the reliance on the DM to run an encounter but such a thing is hard to imagine.

What do you guys think such a system would look like? Would kind of design decisions would you make?

[ishy]

What kind of game are you going for? Would something like all actions for a round are defined and only after it the round happens be an example?

Do you want to keep round by round?
What are people supposed to be able to do in a round?
What kind of limits do you want certain systems (like spellcasting) to have?

If just AoO bother you, you could always turn it into smt like: if you provoke you get auto dmg equal to the attackers str score. Or whatever, so you can handle it more quickly.

[hogarth]

First of all you have the per round structure that makes a single combat drag on for up to a week as you wait for people to get online and post and then all the AoO's and immediate actions slow it down even more.

If someone were to make a system for a play by post format, obviously those things would need to go.

Either that, or you just accept that combats are slow.

One thing that I've learned over the years playing PbPs -- it's much more efficient to have one quadruple-sized combat than four single-sized combats. And dungeon crawling is for the birds; if it takes two days to settle a simple question like "Left or right?", to Hell with doing that x100.

[RobbyPants]

Either that, or you just accept that combats are slow.

One thing that I've learned over the years playing PbPs -- it's much more efficient to have one quadruple-sized combat than four single-sized combats. And dungeon crawling is for the birds; if it takes two days to settle a simple question like "Left or right?", to Hell with doing that x100.

This is totally right about dungeon crawls.

As for slow combats, the things I've seen that can help speed them up are keeping the party size manageable and having the DM just do some of the rolling for passive things like saves and skill checks that just "happen".

Also, if the players post stuff like the DC of their spell, it saves the DM from having to answer a question, which can kill the game for a day or two on its own.

[angelfromanotherpin]

I think an ideal play by forum game would have a combat system where the player posts something like:

Okay, my basic approach is (observational/defensive/opportunistic/aggressive/RIPANDTEAR). My primary focus is (disrupt the wizard/draw attacks from the cave troll/get my hands on the macguffin). My secondary focus is (watch Jeff's back/watch the untrustworthy NPC watching Jeff's back/stab the NPC if nobody's watching). (Insert thoughts for specific ability usage)

(insert flavor text re: going for kneecaps, relevant battle cries, mooning enemy pre-charge, etc.)

...and then the resolution system can take those decision inputs and produce a result. Depending on dramatic relevance, fights could go for ~1-3 rounds of that.

It would have to be a fairly abstract resolution system, but I'm okay with that.

[hogarth]

As for slow combats, the things I've seen that can help speed them up are keeping the party size manageable and having the DM just do some of the rolling for passive things like saves and skill checks that just "happen".

I agree that having the GM make some/all rolls speeds things along a bit. AoOs are another type of rolls that it usually makes sense for the GM to roll instead of waiting for the PC.

Other things I've seen:

- Using "first come, first served" initiative order, or a variant thereof.

- Having a strict time window for posting actions (e.g. if you don't choose an action within 24 hours, the GM will choose for you).

I think an ideal play by forum game would have a combat system where the player posts something like: [post general tactics, etc.]

I've actually played in play-by-post/play-by-email D&D games that worked along those lines. You would post tactics that are as detailed as you like, possibly with a bunch of contingencies. Then the GM would use those posts to run the combat (in one game, it would be the whole combat, and in another game it was only a couple of rounds at a time).

It works okay, but it wasn't ideal; coming up with a bunch of contingencies beforehand is too much like work, and if you don't come up with a bunch of contingencies, it's too much like the GM is playing your PC instead.

[PoliteNewb]

- Having a strict time window for posting actions (e.g. if you don't choose an action within 24 hours, the GM will choose for you).

One DM I've played with online uses this, although the action is often "stand there like a retard".

I think an ideal play by forum game would have a combat system where the player posts something like: [post general tactics, etc.]

I've actually played in play-by-post/play-by-email D&D games that worked along those lines. You would post tactics that are as detailed as you like, possibly with a bunch of contingencies. Then the GM would use those posts to run the combat (in one game, it would be the whole combat, and in another game it was only a couple of rounds at a time).

It works okay, but it wasn't ideal; coming up with a bunch of contingencies beforehand is too much like work, and if you don't come up with a bunch of contingencies, it's too much like the GM is playing your PC instead.

I use this in a PbP AD&D campaign I run...it works better if you use a fairly abstract combat system, rather than a fine-grained tactical miniatures game.

My players can literally say stuff as simple as "attack the bugbear, then the goblins" or "shield, then sleep, then magic missile on the cleric", and I can run a couple of rounds based on that. If the situation changes drastically, I put out another call for actions; most of my peeps post daily.

It's certainly slower than face-to-face gaming, of course...we've been in this campaign since August, and thus far have completed two "adventure" scenarios and are deciding on a third from some rumors they've picked up. That's tons faster than another PbP I was involved in, which didn't even get through 1 adventure in roughly the same amount of time.

[Dominicius]

But isn't that putting too much work on the shoulders of the DM? In a way, you are handling the math and decision making behind the NPCs actions and now you also need to handle the math and decision making behind NPC actions.

But yea I agree, dungeon crawls just don't work. The best way I've seen it handled is when the DM gave us the map to the dungeon from the start and allowed us to plan our exploration course ahead of time. Then he just guided us through the rooms.

[hogarth]

But yea I agree, dungeon crawls just don't work. The best way I've seen it handled is when the DM gave us the map to the dungeon from the start and allowed us to plan our exploration course ahead of time. Then he just guided us through the rooms.

The best way I've seen it handled is when the GM says "you walk through a dozen rooms that are not particularly interesting and then you reach room XYZ". Actually, the even better way is to just remove the dozen uninteresting rooms altogether.

[Grek]

A system designed from the ground up for PbP would be alot like Diplomacy in the turn structure and resolution: Everyone submits their actions all at once via PM and them the DM sits down for half an hour and works out what happened.

[BearsAreBrown]

Tangential question: Anyone here every used FantasyGrounds? I'd love to see a web-based version of that which didn't cost an arm and a leg.

[Josh_Kablack]

Some potential options for working with the existing system:

1. Use something like the Amber Diceless general combat options. They'd have to be modified to account for die rolling and limited use powers. But it works like this: in combat you are assumed to be "opportunistic" (defending yourself but taking advantage of easy openings) unless you declare otherwise -- other declarations are on a slider between "cautious" (defending yourself) and "aggressive" (going all out without regard to personal safety) with a couple of exceptions for feigning strength or weakness to intimidate or bait an opponent.

2. Revive the long-dead idea of the "lead player". The guy who's always there, who has nothing else to do, who might be the MC's roommate. All other players discuss their general combat tactics with him and give priorities at the start of each fight - and if they want to go as far as FFXII gambit-style scripts they can - but this guy handles the player side nitty gritty of fight scenes like a pokemaster.

3. Each player has the player to the left (or with the next username in alphabetical order) as an "understudy" who gets to declare moves for their character when they can't post by deadline. So while Orangegar is Abe's character and Usk the Orc is Bob's - when post posts Usk's move, he checks to see if Abe has already posted Orangegar's - and if not he adds Orangegar's move to his posts in the specified understudy font. If Abe manages to post between the time Bob posts and the move deadline, the actual player move declaration renders the understudy move declaration null, but if Abe can't make it by deadline, then Bob's declaration determines Orangegar's move.

[hogarth]

Tangential question: Anyone here every used FantasyGrounds? I'd love to see a web-based version of that which didn't cost an arm and a leg.

Maptools isn't web-based, unfortunately, but the price is right and it's written in Java so it's system-agnostic.

[Ice9]

You'd definitely want to eliminate any out of turn actions, or make them automatic. So:
"When in aggressive stance, any foe that moves away from you gets stabbed." is fine, but 3E-style AoOs are right out.

Also, you'd want to make turn order a simple "one side goes, then the other side" thing.


The basic difference is that individual action resolution is (relatively) instant, even with complex mechanics, but getting feedback is potentially very slow. In many ways, the opposite of face to face gaming.
So for example, rolling 3d100 and consulting a chart for attack rolls is actually not a big problem, but the opponent being able to choose different defense options in response is terrible.

[Whatever]

While combat can be glacial in PbP, I always found the biggest problem was in noncombat stuff, when there's less of a tendency to wait for people to post. Obviously you don't want every encounter to last days and days because someone might have something to say (it's much better to say "Okay, unless anyone objects, let's do X" than "what does everyone think of X?"), but when two people are online at the same time and each make 5+ posts back and forth, it can leave everyone else in a tough position; you don't want to overwrite any of what they said, but your character might have wanted to say or do something in response to one of their earlier posts.

On a mostly unrelated note, I'd favor giving more narrative agency to the players, rather than leaving everything up to the DM, because you can just play with people who aren't awful. Everything is written down anyway, so the closer you make this to collaborative story telling, the better.

[Prak]

I think an ideal play by forum game would have a combat system where the player posts something like:

Okay, my basic approach is (observational/defensive/opportunistic/aggressive/RIPANDTEAR). My primary focus is (disrupt the wizard/draw attacks from the cave troll/get my hands on the macguffin). My secondary focus is (watch Jeff's back/watch the untrustworthy NPC watching Jeff's back/stab the NPC if nobody's watching). (Insert thoughts for specific ability usage)

(insert flavor text re: going for kneecaps, relevant battle cries, mooning enemy pre-charge, etc.)

...and then the resolution system can take those decision inputs and produce a result. Depending on dramatic relevance, fights could go for ~1-3 rounds of that.

It would have to be a fairly abstract resolution system, but I'm okay with that.

This puts in mind a kind of Pokemon-esque Rock Paper Scissors resolution system. I think that could work decently. Say RIPANDTEAR beats observational, Observational beats Defensive, Defenisive beats Opportunistic, Opportunistic beats Aggressive, and Aggressive beats RIPANDTEAR, or something.

[Grek]

Also, you'd want to make turn order a simple "one side goes, then the other side" thing.

Honestly, no. Let's look at three games that work well in PbP and see how they do it:

1. Diplomacy. In this game, 99% of the actual gameplay is talking tactics with the other players and bluffing the enemy. There's no turn structure; everyone submits their actions and acts simultaneously. A RPG would like something like the WW2 variant where half the players are working together to play the allies while the other half plays the axis. Or, in RPG terms, half the players run the adventuring party and half the players run the dungeon/monsters using a Descent-esqe system. Another important thing to note is that if you fail to submit a turn, the game doesn't grind to a halt. There's a default action that you'll take and everyone else can continue on without you.

2. Mafia. Also known as Werewolf. Again, no turn structure. Everyone spends the Day talking to eachother and jockeying for position until something happens that ends the turn. This easily translates over into a tactical positioning mini-game where everyone secretly draws a hand of cards or rolls a dice at the start of the turn and then bluffs at eachother until someone forces the turn to end. Instead of voting to lynch someone you think is on the other team, you move in to engage someone that you know is on the other team, and who's you hand/roll you think you can beat. Then you go to the Night phase where everyone secretly declares a special action based on their role (read character class) and it all gets resolved simultaneously.

3. Chess. The game has no random elements, aside from the initial assignment of colour. Every move is straightforward to resolve, requires no input from the other side and is 100% deterministic. In RPG terms, most attacks are Save: No, but allow you to take some action (with an opportunity cost) to either prevent or negate the effects.

It's pretty clear that the ideal system for PbP involves submitting actions secretly, simultaneous resolution of every player's actions, making choices based on predicting your opponent's actions and a deterministic ruleset that allows any player to resolve a turn independently of the DM.

[Vebyast]

It's pretty clear that the ideal system for PbP involves submitting actions secretly, simultaneous resolution of every player's actions, making choices based on predicting your opponent's actions and a deterministic ruleset that allows any player to resolve a turn independently of the DM.

Beat me to it. I was going to put up a post deriving the same conclusions from first principles. Basically, you want to never be waiting for a player's input. That means asynchronous actions. The easiest way to do that is simultaneous resolution, with the resolutions occurring as infrequently as possible. The rest are required to make simultaneous resolution possible (deterministic state) and fun (choices based on predicting opponents' hidden information).

On a side note, deterministic state doesn't prevent the game having random elements. The key is that you hide parts of the state from each player. Infodump:

For example, random number generation. At the start of the turn, before any decisions are made, have each player roll a giant pile of d20s. Each publicly publishes some signature feature of their list of numbers (probably a cryptographic hash) along with their actions for the turn. Once all the actions are published, each player publishes their list of numbers and proves that they were generated before decisionmaking by comparing the two signatures. To generate the nth die roll for that turn, take the nth number from every player and combine them somehow (product modulo the die size? Probably not so simple). The system is deterministic and resolves simultaneously, but you still have truly random numbers. The only problem is the overhead; this scheme requires one extra post between decisionmaking and resolution, which could bog the game down pretty significantly.

Obviously you'd want automate this scheme somewhat (I built an implementation that did it over email), but it's possible, and you can tweak it to make it faster. For example, whisper the number lists to the DM in the post where you make your actions, and give the DM enough privileges in the thread that they can go through and remove the whisper tags once everybody's made their decisions. Alternatively, just resolve after 24 hours using all of the numbers that have been posted so far; if someone doesn't post their numbers, they don't contribute.

[...You Lost Me]

Vebyast, every time you write something, my eyes glaze over. What is modulo?

[Vebyast]

Vebyast, every time you write something, my eyes glaze over.

Erp, sorry about that. I have some bad habits.

What is modulo?

"a modulo b" is a short, clear way to write "the remainder when a is divided by b". It fits into sentences slightly more easily.

[jadagul]

Vebyast, every time you write something, my eyes glaze over.

Erp, sorry about that. I have some bad habits.

What is modulo?

"a modulo b" is a short, clear way to write "the remainder when a is divided by b". It fits into sentences slightly more easily.

But you'd wanna add the results rather than multiplying them--otherwise you get something really nonrandom, and a rather high chance of a 0 or 20 (which increases with the number of dice rolled).

[Vebyast]

Right, my bad. Addition it is.

The method itself doesn't really matter. The point is that you can have simultaneous resolution without the game having to be as deterministic as chess. The trick is to treat randomness as a secretly-submitted action and to make your results depend on other people's randomness (to prevent cheating). You choose your strategy publicly and submit randomness secretly, everybody reveals their randomness at once, and the turn proceeds deterministically from there.

Hmm. Now that I think about it, that's kind of what poker does.

[hogarth]

As for slow combats, the things I've seen that can help speed them up are keeping the party size manageable and having the DM just do some of the rolling for passive things like saves and skill checks that just "happen".

One more comment -- I actually find very little benefit in terms of speed for having a small party vs. a large party, assuming you're not too strict about posting in initiative order. In play-by-post games, your turn will usually be as fast as the slowest player; adding more fast players doesn't change that.

In fact, I think larger parties are more beneficial for the survival of a PbP. A game with six people posting every day just seems livelier and has more momentum than a game with three people posting every day.

[RobbyPants]

The method itself doesn't really matter. The point is that you can have simultaneous resolution without the game having to be as deterministic as chess. The trick is to treat randomness as a secretly-submitted action and to make your results depend on other people's randomness (to prevent cheating). You choose your strategy publicly and submit randomness secretly, everybody reveals their randomness at once, and the turn proceeds deterministically from there.

Why do the players even have to randomly generate encrypted rolls? Why can't they just post actions and have the DM do all the rolls? It seems like it'd satisfy both parts of asynchronous decisions and decisions made before knowing the outcome.

As for slow combats, the things I've seen that can help speed them up are keeping the party size manageable and having the DM just do some of the rolling for passive things like saves and skill checks that just "happen".

One more comment -- I actually find very little benefit in terms of speed for having a small party vs. a large party, assuming you're not too strict about posting in initiative order. In play-by-post games, your turn will usually be as fast as the slowest player; adding more fast players doesn't change that.

In fact, I think larger parties are more beneficial for the survival of a PbP. A game with six people posting every day just seems livelier and has more momentum than a game with three people posting every day.

Also, with people suggestion asynchronous input, the size of the party would also have a lot less impact in terms of getting player input. There'd be more work for the DM, but that would be a lot smaller than the days you're forced to wait if running initiative by the book.

[Vebyast]

Why do the players even have to randomly generate encrypted rolls? Why can't they just post actions and have the DM do all the rolls? It seems like it'd satisfy both parts of asynchronous decisions and decisions made before knowing the outcome.

Either way, you're blocking once after actions are declared. With secure multiparty rolls, you're waiting for at least one person to post rolls so you have randomness. With DM-rolls-everything, you're waiting for the DM in particular to post his rolls. That is far more prone to failure. It's also insecure, and that's bad; I've seen PbPs disintegrate when the players accused the DM of dice fixing.