Strategy vs. Tactics

[fizzix]

In a recent thread Nick mentioned that he things angband should focus more on tactics and avoid strategic gameplay. I only partially agree, so I'd like to write down my thoughts.

First a definition for those not familiar with the jargon. When we talk about tactics we talk about decisions in the immediate encounter. Do I attack or do I cast a spell? What buff should I use? Is there more advantageous terrain that I can fight in? Should I retreat/escape? The consequences for these decisions are immediate.

Strategic thinking is more long term. Do I buy the potion of strength or save my gold? Do I try to kill this unique at the expense of some consumables, or should I just avoid it? Which armor should I keep? These decisions have effects that play out over the long term. If you buy the potion you might not have enough money for the wand of teleport other that shows up on your next trip, and so on.

Sometimes the line between strategic and tactical gameplay is blurred. In the heat of the battle you need to decide whether to use your !healing potion or if !CCW is enough. If you use the healing, you might not have it for the next battle. But if you only use !CCW, you might need to use a lot of them, and you might even die.

All roguelikes have tactical gameplay. In some way, it's really the bread and butter of the genre. However, there's a lot more variety in the strategic gameplay. Typically, roguelikes include strategic gameplay in 3 separate axes.

1) character development: choosing what skills and abilities your character will get
2) Equipment/inventory management: deciding what equipment to keep and what to leave behind. I also throw in the use of gold here.
3) Consumables management: deciding when to use powerful items with limited uses. This is the one that tends to bleed over with tactical gameplay the most.

The tricky part about strategic gameplay elements in roguelikes is that the games tend to be long, and you don't want a player to be blithely playing along until they reach a point where they suddenly realize that the decision they made 20 hours ago has caused the game to be unwinnable. Such is the pitfall of strategic gameplay.

Angband pretty much has nothing with regard to character development after the character selection screen. You don't really "build" a character in angband. The inventory management is also not too difficult unless you're playing ironman games. It excels the most with consumables management, and this is where the most interesting long term decisions tend to lie in Angband.

I'd like to compare some other games approaches to this. The other roguelikes I'm most familiar with are Tome v4 (Tales of Maj'Eyal), DCSS and Sil. So I'll use those.

Tome v4 is probably the most pathological of them all. It completely eliminates consumables management. Furthermore equipment/inventory management is almost completely unimportant. You can pretty much carry whatever you need, and the permanence of levels means you can always just leave something for later if it turns out you can't carry it. That leaves character development, and it certainly has a ton of that. You probably can spend more time trying to figure out what skills to take in the game than any other strategic decision. Also it means that choosing a bad path is going to doom your character, since winning often requires exploiting some powerful synergies in the skill tree. Tome v4 winds up being a completely different roguelike from others in that there is no bleed over between tactical and strategic decisions. Every battle can be isolated and after it is over everything resets.

DCSS has character development, although it's not anywhere as intricate as Tome's, and is much more forgiving as far as path choice. It completely disregards inventory management, since again, you can always stash things for later. But it does have consumable management. And this is even more pronounced than in Angband, since the number of any consumable is finite (albeit unknown).

Sil includes all three axes of development. A skill tree that is extremely difficult to master (the hardest part of the game for me, tbh). Consumables that are finite as in DCSS, and the inventory difficulties of an ironman angband game. The only compensation is that games of Sil tend to be a lot shorter, so the burn from a bad strategic decision is less painful.

Ok, so where does that leave us in Angband? I think not having a character development skill path is very reasonable. It's probably a good design choice for the game for a couple reasons. It's very new player friendly (the skill tree in Sil is a bit overwhelming, forget about Tome). And this is often where you are most likely to screw yourself early in a way that doesn't become apparent until much much later.

That leaves the other two axes, inventory and consumable management. Personally, I think the latter one should be the main area of strategic thinking in the game. Deciding when and how to use powerful items is key. Part of the design difficulty here comes with figuring out how to give the player these options, and we have a lot of room for improvement.

Inventory management is secondary, but perhaps it shouldn't be wholly abandoned. These are difficult choices and in many cases they may be interesting.

Now here's the major problem. If inventory management is a key feature of the strategic element, then it often reduces both the tactical richness *and* the consumable strategy. The first is easy to understand. If you can't afford to carry the scroll of dispel evil, then you'll never be presented with a decision of whether to use it or not. (for this, pretend that dispel evil is actually a fairly useful scroll.) The decision is forced, because you don't have the scroll. On the other hand, if you eliminate inventory management altogether, you are left with situations where you very well might always have the perfect answer for any situation. That's not good either.

The solutions I envision tend to have an inventory management problem that allows more diversity but less bulk. So you might be limited in how many !CCW you can carry, but you can carry a whole host of other types of potions. In my ideal game this also involves removing these consumables from the town altogether, since having a town that supplies an infinite amount of !CCW or arrows means that they're going to crowd out any other weaker but sometimes more situationally powerful options. I'm not sure what this looks like, but it probably involves setting up scenarios of interesting choices, and then figuring out how to produce an inventory problem where those scenarios present themselves.

Here are some examples:
5 !CCW or 1 !healing (assume you currently have 5 !CCW and 1 !healing already)
5? Phase door or 2? teleportation (assume you have 3 !Phase and a staff of teleportation with 3 charges)
1! restore life levels or 2! resist heat and cold (assume you cannot get !rll from town and you have no other source of temp rfire/rcold)
1 rod of TO vs 1 wand of TO with 7 charges (assume you have neither)

etc.

Personally, I think the game would be richer if these choices were presented to the player.

I have more to write on this, but my lunch break is over and I have work meetings to go to.

[Nick]

Thanks for the explanation. I only have time for a brief reply too; I think I can sum up my position roughly as I think strategic gameplay should be emergent rather than planned for. Don't hold me to that though.

[Derakon]

Common sources of strategic gameplay:

* What items do I carry with me? And what do I equip? Even in games that have unlimited storage, typically the character has sharply limited space for the things they can access at a moment's notice. And I can't think of any games that have unlimited equipment space.

* How can I manipulate the setting of this engagement to my advantage? In Angband this is things like luring enemies away from each other, digging corridors, etc, application of buffs before the fight starts, and deciding whether or not the fight should have a ranged combat section.

* Acquiring and using knowledge about the environment. This is not very strategic in Angband because detection is spammable, accurate, and specific, but in some games you may have to make educated guesses based on vague hints (what exactly does a slithering sound indicate?) and/or incomplete information (you know there's a dragon around, but not what kind of dragon).

* Deciding whether or not to fight something. Getting nervous and deciding to flee the level is a strategic decision, for example.

[Pete Mack]

Deciding whether or not to fight a monster is tactics. Deciding whether to be on a level with monsters you can't fight is strategy.

[bio_hazard]

I think I have a broader view of strategies, but I suppose there are long-term and short term strategies, as well as tactics which are the indvidual tools used to carry those out. Sometimes strategies can be a prioritized list of goals

Long term strategy (Dive):
While preventing one-shot death opportunities,
Improve character equipment by exploration and not-too risky fighting
and Improve character stats by exploration and not-too risky fighting
by descending to levels at which these improvements are most likely and at which lethal situations are not too common

Manage consumables to ensure the most reliable escape possible and favor healing, eventually shifting to save healing and battleground managment items for the final fight, rather than attempting to find items to deal with every possible situation


Shorter term strategies would be like
Don't engage in a fight that is likely to cost more than half of your heals, or with equipment-damaging monsters that are likely to get more than a couple of elemental shots at you.
or
Don't engage in summoners if they are likely to be able to summon more than once

[PowerDiver]

Deciding whether or not to fight a monster is tactics. Deciding whether to be on a level with monsters you can't fight is strategy.

That may or may not be true in 4.0 where I read something about chokepoints, but it wasn't true in 3.0.

Deciding whether to commit consumables and/or turncount to take out a given high-value monster was definitely a strategic decision.

Once you've decided to fight, how you fight is mostly tactics, although strategic issues might cause you to abandon a winnable fight in the middle.

[fizzix]

There are clearly both strategic and tactical decisions involved in a given battle. The rationing of consumables, along with the decision to fight monsters that can damage or destroy gear has a strategic element. Limiting how many consumables you are forced to use in a given battle and reducing the chance of destroyed gear involves tactical gameplay. Figuring out exactly where things lie on the scale is not terribly important. All that really matters is whether the choice is interesting.

I wanted to bring up an alternative idea to the "inventory" problem. Here's a TL;DR. Inventory management produces interesting strategic choices but is incredibly annoying. Let's provide the strategic choices without the annoyance.

As I mentioned above, managing inventory is a method for strategic decision making. It's also incredibly annoying. It's really hard to find a game where people point to inventory management as a major enjoyable element. We probably don't want to go the route of giving items complex volumes and weights and essentially force the player to solve the knapsack problem at every juncture. Nevertheless, removing the inventory altogether would clearly remove a point of strategic decision making in the game.

Furthermore, as I mentioned above, the key element of inventory decisions is forcing a choice between two somewhat balanced options. I gave some examples above, like a choice between 5 !CCW and 1 !healing when your inventory already has 5 !CCW and 1!healing. Is it more important to have 10 !CCW and 1! healing or 5 !CCW and 2!healing? That's a good inventory choice, and one that has never been possible in any versions of angband to date (because potions of like kind always stacked.)

For a possible alternative, I want to look to Brogue. I have very little experience with Brogue, so I might get this wrong, but in Brogue (I think) there are treasure rooms, where essentially you pick between 1 out of several items. After which the rest of the items are lost. This is just an adaptation of the inventory problem, just without all the knapsack sorting. If the main interesting part of inventory is choosing between valid alternatives, then I think it might be worthwhile extracting that part out of the inventory altogether and then simplifying the inventory (to be infinite slots or whatever). It would certainly take some balance to get things correct, and we'd need to deal with infinite resupplying shops which is clearly a problem. But it's worth thinking about.

[bio_hazard]

There are clearly both strategic and tactical decisions involved in a given battle. The rationing of consumables, along with the decision to fight monsters that can damage or destroy gear has a strategic element. Limiting how many consumables you are forced to use in a given battle and reducing the chance of destroyed gear involves tactical gameplay. Figuring out exactly where things lie on the scale is not terribly important. All that really matters is whether the choice is interesting.

I wanted to bring up an alternative idea to the "inventory" problem. Here's a TL;DR. Inventory management produces interesting strategic choices but is incredibly annoying. Let's provide the strategic choices without the annoyance.

As I mentioned above, managing inventory is a method for strategic decision making. It's also incredibly annoying. It's really hard to find a game where people point to inventory management as a major enjoyable element. We probably don't want to go the route of giving items complex volumes and weights and essentially force the player to solve the knapsack problem at every juncture. Nevertheless, removing the inventory altogether would clearly remove a point of strategic decision making in the game.

Furthermore, as I mentioned above, the key element of inventory decisions is forcing a choice between two somewhat balanced options. I gave some examples above, like a choice between 5 !CCW and 1 !healing when your inventory already has 5 !CCW and 1!healing. Is it more important to have 10 !CCW and 1! healing or 5 !CCW and 2!healing? That's a good inventory choice, and one that has never been possible in any versions of angband to date (because potions of like kind always stacked.)

For a possible alternative, I want to look to Brogue. I have very little experience with Brogue, so I might get this wrong, but in Brogue (I think) there are treasure rooms, where essentially you pick between 1 out of several items. After which the rest of the items are lost. This is just an adaptation of the inventory problem, just without all the knapsack sorting. If the main interesting part of inventory is choosing between valid alternatives, then I think it might be worthwhile extracting that part out of the inventory altogether and then simplifying the inventory (to be infinite slots or whatever). It would certainly take some balance to get things correct, and we'd need to deal with infinite resupplying shops which is clearly a problem. But it's worth thinking about.

FWIW I think what first got me hooked on Angband was the exploration/discovery aspect. To some extent that hasn't left, which is probably why I really like having the level feelings.

I think the 5CCW = 1 Heal will make the early to mid game more interesting, but could make the late game really hard. Would 4 !Heal then = 1!*heal*? 4*heal* = 1!life? So 1 life would be like one inventory slot. Obviously figuring out these equivalences would be really tricky.

I'm not sure how similar the brogue case is to the knapsack problem. In the knapsack problem, choosing none-of-the-above becomes a more interesting choice as weighed against future options. In the treasure room situation, the choice among the 3 is independent of future treasure rooms, so as long as picking up something doesn't bump into broader inventory trade-offs, you should pretty much always pick up the most useful thing. And technically the items aren't lost, but you need to replace the first item to retrieve one of the other two.

[fizzix]

I think the 5CCW = 1 Heal will make the early to mid game more interesting, but could make the late game really hard. Would 4 !Heal then = 1!*heal*? 4*heal* = 1!life? So 1 life would be like one inventory slot. Obviously figuring out these equivalences would be really tricky.

Honestly, anything that makes the late game more difficult but doesn't cause too much problems in the early game is a huge plus.

I'm not sure how similar the brogue case is to the knapsack problem. In the knapsack problem, choosing none-of-the-above becomes a more interesting choice as weighed against future options. In the treasure room situation, the choice among the 3 is independent of future treasure rooms, so as long as picking up something doesn't bump into broader inventory trade-offs, you should pretty much always pick up the most useful thing. And technically the items aren't lost, but you need to replace the first item to retrieve one of the other two.

The Brogue case isn't the knapsack problem, it's a way to get inventory decisions without making players solve knapsack problems. Knapsack problems aren't fun, unless you're into that sort of thing.

I didn't know about the replacing items, but obviously that's not what we're going for. I was thinking more like instead of chests, you get some item or terrain feature that when you interact with it, you get to choose one of three options.

The tricky part would hopefully be figuring out which of the options is the best one.

[Derakon]

I didn't know about the replacing items, but obviously that's not what we're going for. I was thinking more like instead of chests, you get some item or terrain feature that when you interact with it, you get to choose one of three options.

A room with three chests; when you go to open one, one of the others opens to reveal a kobold. Do you switch which chest you choose?

[Pete Mack]

A 1/2 chance is better than a 1/3 chance. Monty Hall says: Yes!

[fizzix]

A 1/2 chance is better than a 1/3 chance. Monty Hall says: Yes!

What if I want the kobold?

[Estie]

What if I want the kobold?

So you are into THAT sort of thing....

[PowerDiver]

A 1/2 chance is better than a 1/3 chance. Monty Hall says: Yes!

I hope you are joking ...

The only time I've ever seen 1/2 come into this problem using probability correctly was once when I posed it to my dad, who figured I was trying to trick him, and his answer was to "flip a coin to decide whether to switch".

One year when I was a TA, I gave a section once a week in a room that was obviously teaching junior level probability the hour before, given the leftovers on the blackboard. They were always long gone before I got there, until one day, as I approached from far away, I saw the TA sitting at the desk, with his head in his hands, surrounded by half a dozen students yelling at him. From 50 feet away I knew what topic he had just covered. He was so happy when I arrived to kick him out.

[Nick]

What if I want the kobold?

Always wondered who the guy in the beret was...