A turn-based tactics system I would like to play

[J1M]

It makes sense that for a game where you control just one character, that to maintain interest additional complexity is a path towards holding player interest. "Builds", stacking passives, multiple attacks that are all slightly different, and so on.

I have grown less patient with these things for an entire party over the years. It stands to reason that if a D&D 3.5 character is the sweet spot for complexity when someone has one character to manage, that 6 of the same complexity blows way past that optimal window.

I'd like to see a tactics game that pushes complexity of an individual unit closer to that of Chess than 3.5. I propose a system that would have a character with "slots" for "actions" that are based on what they target. For example, "self", "empty space", "ally", "enemy".

Each turn a character moves and then targets a grid cell. Input is snappy. There are no context menus. The choice is about what the character should accomplish, not optimizing for a 10% better outcome by comparing similar options. Based on the target, the player uses an action like "shoot arrow" or "heal".

I predict some will bristle at this being "too much like a board game". Maybe it is. Or maybe it could lead to lessons that can speed up turn-based games without sacrificing depth. There's only so much tweaking animation speed can do in that regard.

One option for increasing depth/complexity while maintaining the means of rapid input would be to add another dimension to the "slots". For each target type, allow additional actions to be assigned based on the state of the battle. For example, "stationary" and "bloodied". This would allow for assignment of 3 different actions that target "enemy". A standard attack, a stronger attack when the character chooses not to move, and a desperate attack only available when below 50% health. Perhaps extra strong actions can only be assigned to a bloodied slot. Or perhaps a cleric would equip a buff to "self" and a heal to "bloodied+self".

Additional depth could be added with more dimensions ("movement+weather+target") or by allowing players to assign a priority order for what should be chosen when the conditions for more than one action are met.

Personally, I think something would be lost if action assignment becomes too complicated, so I would target about a dozen slots for assignment per character with at least a third of them locked for the first hour of the game.

[NotAI]

One way is using randomness to simplify.

The player accumulates more types of actions via a skill tree that unlocks them, but they are used automatically for the party or squad.

Like in chess, the player can just move Spearman 3 to fight Tree Demon A, and the spearman would randomly pick one action out of a growing list. To make this work, like in card games, inserting multiples of an action into that list should be allowed. By unlocking multiple times like in some clicker games.

The good: much faster pacing.

The bad: this does mean outcome has RNG.

[Norfleet]

One way is using randomness to simplify.

The player accumulates more types of actions via a skill tree that unlocks them, but they are used automatically for the party or squad.

Like in chess, the player can just move Spearman 3 to fight Tree Demon A, and the spearman would randomly pick one action out of a growing list. To make this work, like in card games, inserting multiples of an action into that list should be allowed. By unlocking multiple times like in some clicker games.

This seems like you'd want to reduce the list down to one action. Because if you have multiple choices of action, some actions will be below average options. So you improve the output by removing them. You repeat this process until there are no below-average actions remaining, which generally means there's only one action.

[Stack of Turtles]

One way is using randomness to simplify.

The player accumulates more types of actions via a skill tree that unlocks them, but they are used automatically for the party or squad.

Like in chess, the player can just move Spearman 3 to fight Tree Demon A, and the spearman would randomly pick one action out of a growing list. To make this work, like in card games, inserting multiples of an action into that list should be allowed. By unlocking multiple times like in some clicker games.

The good: much faster pacing.

The bad: this does mean outcome has RNG.

This sounds really dumb, partly because of what Norfleet said, but also because your units should generally not be ********. Why shouldn't they just pick the best thing to do? It's one thing if there's a chance of a misfire depending on skill, but it just shouldn't be the case that Spearman 3 goes up to fight Tree Demon A and rolls "answer nature's call" as an action, when that only works on fire demons.

[TKVNC]

Sounds like Dragon Age Origins

[rusty_shackleford]

It makes sense that for a game where you control just one character, that to maintain interest additional complexity is a path towards holding player interest. "Builds", stacking passives, multiple attacks that are all slightly different, and so on.

I have grown less patient with these things for an entire party over the years. It stands to reason that if a D&D 3.5 character is the sweet spot for complexity when someone has one character to manage, that 6 of the same complexity blows way past that optimal window.

I'd like to see a tactics game that pushes complexity of an individual unit closer to that of Chess than 3.5. I propose a system that would have a character with "slots" for "actions" that are based on what they target. For example, "self", "empty space", "ally", "enemy".

Each turn a character moves and then targets a grid cell. Input is snappy. There are no context menus. The choice is about what the character should accomplish, not optimizing for a 10% better outcome by comparing similar options. Based on the target, the player uses an action like "shoot arrow" or "heal".

I predict some will bristle at this being "too much like a board game". Maybe it is. Or maybe it could lead to lessons that can speed up turn-based games without sacrificing depth. There's only so much tweaking animation speed can do in that regard.

One option for increasing depth/complexity while maintaining the means of rapid input would be to add another dimension to the "slots". For each target type, allow additional actions to be assigned based on the state of the battle. For example, "stationary" and "bloodied". This would allow for assignment of 3 different actions that target "enemy". A standard attack, a stronger attack when the character chooses not to move, and a desperate attack only available when below 50% health. Perhaps extra strong actions can only be assigned to a bloodied slot. Or perhaps a cleric would equip a buff to "self" and a heal to "bloodied+self".

Additional depth could be added with more dimensions ("movement+weather+target") or by allowing players to assign a priority order for what should be chosen when the conditions for more than one action are met.

Personally, I think something would be lost if action assignment becomes too complicated, so I would target about a dozen slots for assignment per character with at least a third of them locked for the first hour of the game.

I suspect this is based upon what you learned from trying to make a 4e game?
Seems like an attempt to correct the issues it has with translation to video game.

[NotAI]

"This seems like you'd want to reduce the list down to one action. Because if you have multiple choices of action, some actions will be below average options. So you improve the output by removing them. You repeat this process until there are no below-average actions remaining, which generally means there's only one action."

"This sounds really dumb, partly because of what Norfleet said, but also because your units should generally not be ********. Why shouldn't they just pick the best thing to do? It's one thing if there's a chance of a misfire depending on skill, but it just shouldn't be the case that Spearman 3 goes up to fight Tree Demon A and rolls "answer nature's call" as an action, when that only works on fire demons."

Indeed, it has all the downsides of a collectable card game, where a leading meta strategy is generally having a dozen or more copies of a single best card (the only card you really need is your credit card) in a deck, so that the probability of drawing the best action is maximum. You have multiple decks, based on what you guess your opponent plays. Until an expansion comes out changing the rock paper scissors for the best cards. Whereupon in a deck or two you remove a few instances of the best card and add several of a second card. After a few expansions, change out entirely the old best card for a different best card. Which can become a bore at some point.

Upside wise it's fast, simple, and encourages achievement and exploration by getting multiples of great gear/cards/items (learn from gatcha hooray), and your characters are like most real people: ********. They make mistakes, they feel real: dummmm. Then again, realism might or might not be fun, as has been often rightly said.

One improvement is making sure that there is no single best action, as this is drawn from AI, for problems where there is high entropy in the search space, it is well mixed. Another is adding a contextual satisficing system. So characters never make a move that is worse than some threshold, so you don't need to prune actions from the list, but optionality only helps for those contexts where the usually subpar actions are actually good. This threshold might be tied to character level. Lower level characters are ********, higher level characters are not.

It should be a roguelike, so characters can permanently die and this frees up slots, and one thing the player tries to do is hire or recruit higher level characters to their party directly, not always just grind to level up their novices. Always give players many chances to replace characters, not just fiddle with actions lists.

[Norfleet]

Upside wise it's fast, simple, and encourages achievement and exploration by getting multiples of great gear/cards/items (learn from gatcha hooray), and your characters are like most real people: ********. They make mistakes, they feel real: dummmm. Then again, realism might or might not be fun, as has been often rightly said.

This is why sound engineering involves removing the option to make mistakes. If you force people to turn the machine on by pressing buttons located on opposite sides, they cannot hold their hands in the path of the blades. People can't choose the wrong option if you don't give them that option.

One improvement is making sure that there is no single best action, as this is drawn from AI, for problems where there is high entropy in the search space, it is well mixed.

I don't see how that'd work: If there are two actions, A and B, then one of these (A) is better, which means the other is worse. Otherwise they'd be the same. Therefore, your goal is to remove B from the options so that it cannot be chosen as it is subpar.

So now that we understand the optimal number of actions for a piece is basically "1", you've reduced things so each unit performs a single action. Congratulations, you're reinvented Chess.

[NotAI]

Loss is calculated for the path in the whole space when you don't directly control the system. You are trying to minimize total loss, not loss in any particular case. This is because we've let the character decide what to do to make it simpler when handling a party of say a dozen characters.

A is better in cases X and Y, B is better in cases U and V.

X,Y,U,V are well mixed in the space explored by the system.

You bias the randomness such that as you keep playing, total score increases at the highest rate, if you do not know in advance which case you will encounter and must select your actions in advance, not in detail in the moment.

[Norfleet]

Loss is calculated for the path in the whole space when you don't directly control the system. You are trying to minimize total loss, not loss in any particular case. This is because we've let the character decide what to do to make it simpler when handling a party of say a dozen characters.

Yes, but rather than having characters that randomly do the most dumbass things imagineable, we build characters that do one, predictable thing. Then we control which thing will occur by using our control of which characters go where and act on what to decide what happens. Rather than giving two characters both Fire and Ice attacks, and then having no control over which will be used, which is highly detrimental, I'd just equip one with ONLY the fire attack and the other with ONLY the ice attack, and be confident that when I move the character of my choice in to attack, he will use the correct attack because he only has one. So this proposal would simply reduce down to "characters have one attack, and I control which one is used through positioning and choice of actor".

[NotAI]

Godzilla appears out of the ocean with a roar.

Knight: Tremaine:
(a) Run away (Defense: +50)
(b) Attack with sword (DMG: +200)
(c) Attack with bow (DMG: +100)
(d) Attack with turnip (DMG: +5)
(e) Cast ice magic (DMG: +150)

Sometimes you just have to let characters attack with the turnip they picked up. Player didn't sell it yet and it's on him; turnip attack does 5 dmg when Golden Magic Sword of Thunder and Lightning does 200 and hangs on the char's hip.

When Tremaine levels up he will stop occasionally attacking Godzilla with inventory crap a turnip, skillet, sack of corn, pebble, rubber eraser.

On the other hand, if a shopkeeper gives you too high a price, throw a turnip in the shop to get lower price. ("Hey! You are messing up my shop! Here, here, have your lower price, just go away, stupid barbarian...")

If you fight a Yak-Shirak, Yak-Shirak is weak to turnips and this is a good move.

[J1M]

It makes sense that for a game where you control just one character, that to maintain interest additional complexity is a path towards holding player interest. "Builds", stacking passives, multiple attacks that are all slightly different, and so on.

I have grown less patient with these things for an entire party over the years. It stands to reason that if a D&D 3.5 character is the sweet spot for complexity when someone has one character to manage, that 6 of the same complexity blows way past that optimal window.

I'd like to see a tactics game that pushes complexity of an individual unit closer to that of Chess than 3.5. I propose a system that would have a character with "slots" for "actions" that are based on what they target. For example, "self", "empty space", "ally", "enemy".

Each turn a character moves and then targets a grid cell. Input is snappy. There are no context menus. The choice is about what the character should accomplish, not optimizing for a 10% better outcome by comparing similar options. Based on the target, the player uses an action like "shoot arrow" or "heal".

I predict some will bristle at this being "too much like a board game". Maybe it is. Or maybe it could lead to lessons that can speed up turn-based games without sacrificing depth. There's only so much tweaking animation speed can do in that regard.

One option for increasing depth/complexity while maintaining the means of rapid input would be to add another dimension to the "slots". For each target type, allow additional actions to be assigned based on the state of the battle. For example, "stationary" and "bloodied". This would allow for assignment of 3 different actions that target "enemy". A standard attack, a stronger attack when the character chooses not to move, and a desperate attack only available when below 50% health. Perhaps extra strong actions can only be assigned to a bloodied slot. Or perhaps a cleric would equip a buff to "self" and a heal to "bloodied+self".

Additional depth could be added with more dimensions ("movement+weather+target") or by allowing players to assign a priority order for what should be chosen when the conditions for more than one action are met.

Personally, I think something would be lost if action assignment becomes too complicated, so I would target about a dozen slots for assignment per character with at least a third of them locked for the first hour of the game.

I suspect this is based upon what you learned from trying to make a 4e game?
Seems like an attempt to correct the issues it has with translation to video game.

More of an alternate idea I decided not to explore when I chose to work on the 4e game. I guess you could say both were a search for a turn-based framework better than what the industry has coalesced around.

[Norfleet]

Godzilla appears out of the ocean with a roar.

Sometimes you just have to let characters attack with the turnip they picked up. Player didn't sell it yet and it's on him; turnip attack does 5 dmg when Golden Magic Sword of Thunder and Lightning does 200 and hangs on the char's hip.

That, or you don't install turnip-throwing ability on your swordsmans, and instead have a guy whose job it is to pick up turnips that you do not send to engage the enemy. You see how play here involves specifically trying to eliminate options for misbehavior. In this regard, it's very similar to the Sims "No X" whack-a-mole.

[Stack of Turtles]

"This seems like you'd want to reduce the list down to one action. Because if you have multiple choices of action, some actions will be below average options. So you improve the output by removing them. You repeat this process until there are no below-average actions remaining, which generally means there's only one action."

"This sounds really dumb, partly because of what Norfleet said, but also because your units should generally not be ********. Why shouldn't they just pick the best thing to do? It's one thing if there's a chance of a misfire depending on skill, but it just shouldn't be the case that Spearman 3 goes up to fight Tree Demon A and rolls "answer nature's call" as an action, when that only works on fire demons."

Indeed, it has all the downsides of a collectable card game, where a leading meta strategy is generally having a dozen or more copies of a single best card (the only card you really need is your credit card) in a deck, so that the probability of drawing the best action is maximum. You have multiple decks, based on what you guess your opponent plays. Until an expansion comes out changing the rock paper scissors for the best cards. Whereupon in a deck or two you remove a few instances of the best card and add several of a second card. After a few expansions, change out entirely the old best card for a different best card. Which can become a bore at some point.

Upside wise it's fast, simple, and encourages achievement and exploration by getting multiples of great gear/cards/items (learn from gatcha hooray), and your characters are like most real people: ********. They make mistakes, they feel real: dummmm. Then again, realism might or might not be fun, as has been often rightly said.

One improvement is making sure that there is no single best action, as this is drawn from AI, for problems where there is high entropy in the search space, it is well mixed. Another is adding a contextual satisficing system. So characters never make a move that is worse than some threshold, so you don't need to prune actions from the list, but optionality only helps for those contexts where the usually subpar actions are actually good. This threshold might be tied to character level. Lower level characters are ********, higher level characters are not.

It should be a roguelike, so characters can permanently die and this frees up slots, and one thing the player tries to do is hire or recruit higher level characters to their party directly, not always just grind to level up their novices. Always give players many chances to replace characters, not just fiddle with actions lists.

This is why only pathological autists play dickbuilders, yeah.

[NotAI]

dickbuilders, yeah.

Woah.

I just got this idea. "Dickbuilder" is a potentially REALLY good title; why doesn't that exist?

That's the next Minecraft!

I mean an RPG where your character builds his ****. (Her ****. Their ****. Its ****. Zer ****. Xer ****.)

That's it.

Cloud doesn't wanna save the world. He just wants to BUILD his ****. I need to work on this.

Gameplay: You freely build your ****. You can level it up, or equip stat boosting equipment on it. You fight with it.

Genius.

THAT **** WILL SELL!

Rozen Garten Saga class ****!

It's the game that best fits our time. The game of our era.



Except if....Steam doesn't allow titles like that....I will check.