Macrotools

Talk about anything and everything related or unrelated to the FreeOrion project, especially Strategy Games.

Moderators: Oberlus, Oberlus

Message
Author
drek
Designer Emeritus
Posts: 935
Joined: Thu Jun 26, 2003 8:07 am

Macrotools

#1 Post by drek » Fri Jul 02, 2004 10:24 am

Speaking about something as an law without a prove is really bad. I also think there are macro tool with different degree's of complexity. I also wouldn't let some macrotool handle my empire, but some minor things could be handed over. For instance in early game i care about anything but in mid/late game i don't.
Alright. I can't remember the exact quote: "I would have included the proof in the margins, but they are too narrow."

I have nothing to do for the next hour or five, so here's some wide margins:

There are "macrotools" that work. For example, selecting a group of units to issue duplicated orders to.

Then there are bots/viceroys/agents--tools that make decisions for the player. Here's a pale shadow of what I remember from OceanMachine's arguments against the use of macrotools:

1: First piece of evidence is anecdotal.....the lack of precedence. There's no such thing as a viceroy in a real game that does its job properly; feel free offer counter-examples. If teams of full time coders/designers couldn't create useful viceroys, then part time coders/designers probably aren't going succeed either.

Chess is fun. Chess is in-depth. Chess requires no automation. Indeed, it would be suicide to use automation vs. a great player. (unless you keep big blue in your back pocket.) If a game as well understood as chess, as simple as chess, successful automation is still considered a meaty problem to work on.

In a more complicated game (like Moo, Civ, or FO), where strats have not been examined for hundreds of years, I can say with a degree of certainty that the AI will be lackluster, as will the any decision making bots.

2: Often in microheavy games there are optimal or near optimal tactics that can only executed by a thinking, creative mind. For example, a person who micros his workers in Civ3 can achieve a much better result than either the AI or a poor player. Often, seemingly simple decisions (" do I want a mine or irrigation in this square?") require insight into the overall strategy the player is attempting and a great deal of foresight.

I have yet to see an AI with the level of comprehension required for real insight or foresight. Most "AIs" are just overgrown state machines with absolutely no insight.

Eventually, making bunches of tiny decisions becomes boring. So the AI bot is unleashed--and proceeds to perform a less than optimal job.

This is fine in a game vs. non-cheating AIs, since the AIs are straddled with the same piss-poor tools as the player. But in multiplayer, or against a heavily cheating AI, a player must make as many good decisions as possible.

Assuming expert level profiency, we end up with a prisoner game. If Player A gets bored enough to use the less than optimal viceroys, but Player B chooses to slog through microhell, then Player B almost certainly vastly outperforms Player A. The player who gets bored first loses, making the viceroys less than worthless. I won't be able to turn on my bots for fear that you are microing your way to victory.

3: It's tiring to police bots to ensure that they doing the right thing. There will be flaws in the design of bots, since they are unable to account for every possible situation.

Further, often a real player's strategy conflicts with the presumptions made by the AI coder/designer.

Therefore, even if the bot does an excellent job, it still won't be a perfect job. A player is forced to review decisions made by the AI bots, which in the end is almost as much work as simply making the decisions to begin with. When the bot does make some drastic errors, it can be more work to undo the damage.

The solution would be to allow players to assemble their own bots, or be allowed to greatly influnce the bot's actions. Custom build lists are an example of this.

These kinds of activities shouldn't be required to *play a game*.

Sidenote: I'm amazed when people propose macrotools that are teetering close to being Turning Complete. Most people are not computer programmers and have no desire to be computer programmers.

Me whining about custom build lists is perhaps a little silly in a forum full of linux users---but no where near as silly as proposing that players of a 4x game will have to have an understanding of boolean logic, conditional statements, and program flow.

4: If an activity is so boring that we feel the need to automated it, then why include it in a game that being played for fun? This is perhaps the most important reason to avoid macrotools.

People posting ideas to these boards often lay out elaborate schemes in the name of simulation (or worse “strategy” *), then wave the magic wand of automation in order to make the idea playable in mid to late game.

Here’s my own Drekish thing: even if automation is possible, why not shave a few steps off everyone’s day and come up with systems that don’t require automation?

It's true that a sprinkling of the proposed macrotools *might* work as intended (or they might blow up and make the problem 10 times worse. example: the tools that resemble programming languages).

But I wonder if it's lazy design. Considering the infinite possibilities there should be a game design for a Moo-ish game that's:

a: fun
b: in-depth strategically
c: got plenty of "flavor"
d: in no dire need of automation

It's a challenge, a metagame: the trick is to invent a game that fulfills those four requirements. I know it's the realm of possibility--question is whether or not we've got the collective desire and creativity to invent it.

Cause really: what's better? A game meeting those four requirements, or a game that forces you into either micohell or automation.

(note: an unofficial challenge. No one's ever officially ruled out automation.)

My Jerry Springer-esque Final Word:

Some really fun games: Chess, Super Mario brothers, Advanced Wars on GBA, Europa Universalis, Axis and Allies, Magic the Gathering.

Not a single one of em require that an AI or macro tool holds the player’s hand, playing half the game for him.


* if it's strategic, then how is a dumb robot going perform the task as well as a human?
Last edited by drek on Fri Jul 02, 2004 10:52 am, edited 1 time in total.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12641
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#2 Post by Geoff the Medio » Fri Jul 02, 2004 10:46 am

Tyreth wrote:Regarding macro tools, here is why I don't like them:
1. A macro tool is built to automate repetitive tasks
2. Often, a macro tool is not possible to tweak to a state where it always makes the best 'decisions'
Therefore,
3. When a macro tool will not make the best decision, you would gain more advantage by not using it
Therefore,
4. When you play a multiplayer game, you will need to not use macro tools in these situations in order to remain competitive with other players
Therefore,
5. When a macro tool cannot make the best situation, you again have the problem of repetitive tasks needing to be done by hand

For those situatations where macro tools can make the best decisions:
1. A macro tool is built to automate repetitive tasks
2. Sometimes, a macro tool can be tweaked to always make the best 'decision'
Therefore,
3. The best decision can be managed by a simple tool

If a simple tool can manage the best decision, then why give the choice in the first place? When a macro tool cannot choose best all the time, then we'd prefer to not use it. If it can be trusted to be good all the time, then why give the decision to the player in the first place? We may as well remove these decisions and give the players real choices that don't slow down their playing of the game.
(seemed a good idea to repost this here... I hope you don't mind)

drek
Designer Emeritus
Posts: 935
Joined: Thu Jun 26, 2003 8:07 am

#3 Post by drek » Fri Jul 02, 2004 10:53 am

Thanks Geoff.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12641
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#4 Post by Geoff the Medio » Fri Jul 02, 2004 11:08 am

The main objections to Macro Tools seem to be based around the idea that they can't always make good decisions for you.

This is not what a macro tool does (in my conception). A macro tool makes it easier for the player to impliment and adjust the decisions s/he makes.

Unlike the sterotypical governor in SMAC or viceroys in MoO3, the goal of the macrotool would be to simplify and expedite the execution of what would otherwise be repetative tasks for the player. This would be accomplished through, basically, a well-implimented user interface -> the "Macro Tool". The player would be able to issue a single command that affects multiple similar in-game objects. The player would be able to add or remove objects from the scope(s) of the command. The command, if presistant, would notify or query the player (via sitrep) when appropriate. The command, if persistent, could be automatically assigned to and/or cancelled from objects according to conditions inherent to the command, or that the player specifies (not as complicated as it sounds).

Control groups in an RTS are a closer analogue to the use of macro tools than a viceroy or civ/smac governor.

To reiterate: Macro Tools do not make decisions for the player. Macro tools aide the player to impliment his/her decisions, by eliminating the need to give similar commands repeatedly.

drek
Designer Emeritus
Posts: 935
Joined: Thu Jun 26, 2003 8:07 am

#5 Post by drek » Fri Jul 02, 2004 11:50 am

repetative tasks for the player
If the task is repetative, we should be looking at whether or the task is required at all.

In some cases, the answer is an obvious yes. Pathfinding, rally points, fleets to contain ships are all examples of automation and/or macrotools that allieviate repetative tasks associated with moving ships around the galaxy map.

We could abstract ships to the point that they are not units on the galaxy map. The military units in one my favorite games (King of Dragon Pass) don't appear on any world map, for example.

But that would (a) fly in the face of what most people expect from a moo game (b) greatly diminish military strategies based on the geography of the map. So, grumbling, we have to have deal with some automation--and accept that even though the pathfinding is already quite nice, there will be moments when it offers the less than optimal path.

Economic systems are another thing entirely. First off, each moo game had an economic model utterly inconsistent with it's sister games--so expectations are all over the map. Second off, I believe that many of the types of decisions people enjoy in Civ/Moo2 games can be replicated on an empire-wide scale via the rare wonder-esque buildings. In a sense, your entire empire is one big city. Finally, flavor can be maintained on each indivdual planet (even enhanced compared to moo2/3) through the use of creative specials and events.

When someone proposes an economic system that requires some sort of automation to be playable in mid to late game, I wonder: if a task is repeative and easy to automate, is it really something fit for a human player/interstellar emporer to be worried about in the first place?

vishnou00
Space Kraken
Posts: 157
Joined: Tue May 25, 2004 3:15 am

#6 Post by vishnou00 » Fri Jul 02, 2004 11:51 am

This is a reply to Tyreth quote.

"Why don't like"
1. A macrotool can be used to allow you to make higher level decision, like converging fleets in different locations to one location. I think we all agree that such tools are needed.

2. A macrotool shouldn't make decision, it should only reduce the number of steps necessary to "implement" that decision. In an ideal macrotool, you input your analysis of the situation when it isn't trivial (assessing risk of attack, strategic role...) and the tool make the actions that are direct consequences of this input. I've never seen anything like this attempted in a 4X game, it seems as soon as a designer thinks helper, he thinks it's because the player don't want to understand the game and tries to take the analysis/decision away from him. I want the macrotools to complement an otherwise poor UI, it isn't because I don't care, it's because entering my decision takes minutes of clicking through panels.
So, the macrotool doesn't make decision.
Therefore,
3. He is always optimal within the limits of its functionnality.
Therefore,
4. You can trust him in very competitive game, as in only enables you to play quicker, be more time efficient.
Therefore,
5. You can put all you attention at playing creatively, handling exceptions and general strategy.


If a macro tool if presented a trivial situation, it will work independantly of the player within the limit of triviality. It will control a simple situation with a lot of elements (like supply routes) in the confines of the ordinary situation (peaceful, intra-empire trade route). However, as soon as the situation exit the domain of triviality (a disrupted trade route), it's up to the player to manage the situation. Notice that it will never be boring, because the player will always be handling exceptions, usually involving a challenger, whether player based (ennemy empire) or AI based (pirate or space monster).

If the element of gameplay is wholly abstracted (as in pooling of ressources abstracting supply routes), the basic abstraction only represent the trivial situation, with no problems. To provide interesting gameplay, it has to emulate these behaviour with "extensions" to the abstracted model, and you end-up taking the interesting gameplay away from the player, as you only allow them to indirectly control (through the model) the abstracted gameplay elements (the non trivial situation of the emulated behaviour).

vishnou00
Space Kraken
Posts: 157
Joined: Tue May 25, 2004 3:15 am

#7 Post by vishnou00 » Fri Jul 02, 2004 12:02 pm

drek wrote:When someone proposes an economic system that requires some sort of automation to be playable in mid to late game, I wonder: if a task is repeative and easy to automate, is it really something fit for a human player/interstellar emporer to be worried about in the first place?
The "human player/interstellar emporer" [sic] have to worry about the problems breaking the monotony. He watches the automated cogs working perfectly when everything goes fine, but when somebody throws a rock in the cogs, he dives in and fix what has to be fixed! To be able to fix anything, he's better to understand how it works when everything is well.

vishnou00
Space Kraken
Posts: 157
Joined: Tue May 25, 2004 3:15 am

Re: Macrotools

#8 Post by vishnou00 » Fri Jul 02, 2004 12:06 pm

drek wrote:My Jerry Springer-esque Final Word:

Some really fun games: Chess, Super Mario brothers, Advanced Wars on GBA, Europa Universalis, Axis and Allies, Magic the Gathering.

Not a single one of em require that an AI or macro tool holds the player’s hand, playing half the game for him.
For one, some of them aren't strategy game (Super Mario brothers, Magic the Gathering) and many of them don't pretend to have the epic scale of a 4X game (Chess, Super Mario brothers, Advanced Wars on GBA, Axis and Allies, Magic the Gathering). I don't know much about EU, but while it seem to have interesting diplo, does it tries to be epic?

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12641
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

#9 Post by Geoff the Medio » Fri Jul 02, 2004 12:08 pm

drek wrote:In some cases, the answer is an obvious yes. Pathfinding, rally points, fleets to contain ships are all examples of automation and/or macrotools that allieviate repetative tasks associated with moving ships around the galaxy map.
And this is the sort of thing I've preivously suggested be dealt with via a macrotool. You (drek) and tzlaine have previously been less than enthusiastic about fleshed-out examples, however...

viewtopic.php?p=12413#12413
viewtopic.php?p=12419#12419
So, grumbling, we have to have deal with some automation--and accept that even though the pathfinding is already quite nice, there will be moments when it offers the less than optimal path.
The macrotool could let you mark planets/starlanes as "off limits" or "preferred travel route" for some or all ships. You could also plot a course for several fleets, currently at different locations, with several waypoints to follow (at different times, since they start at different locations). The order would also warn you and ask for what to do if something blocked or otherwise interfered with the plotted course.
Economic systems are another thing entirely.
Agreed. The focus system and limited wonder buildings is a much better system that moo2-style umpteen buildings you need everywhere and which takes a lot of micro to deal with. Most governors/viceroys were meant to deal with this sort of thing, which they are not as well suited for. Strictly followed, or brancing logic development plans might reduce the tedium, but abstraction does so much better.

drek
Designer Emeritus
Posts: 935
Joined: Thu Jun 26, 2003 8:07 am

#10 Post by drek » Fri Jul 02, 2004 12:48 pm

The "human player/interstellar emporer" [sic] have to worry about the problems breaking the monotony. He watches the automated cogs working perfectly when everything goes fine, but when somebody throws a rock in the cogs, he dives in and fix what has to be fixed! To be able to fix anything, he's better to understand how it works when everything is well.
An evocative image. SimGalaxy would be a fun toy.

FO, for better or for worse, isn't SimGalaxy.

And yes, EU2 is of an epic scope. Spartan is another epic game, and a 4x game to boot, wherein the designers made a succesful effort in reducing the micro while maintaining the feel and strategy.
And this is the sort of thing I've preivously suggested be dealt with via a macrotool. You (drek) and tzlaine have previously been less than enthusiastic about fleshed-out examples, however...
I remain less than enthusaistic about the tool as described. Would prefer to worry about it later. Other more pressing issues need to be worked on for the v.3 design doc, and this is issue that should probably handled in v4. (the combat/ships/ship design release).

vishnou00
Space Kraken
Posts: 157
Joined: Tue May 25, 2004 3:15 am

#11 Post by vishnou00 » Fri Jul 02, 2004 1:09 pm

drek wrote:An evocative image. SimGalaxy would be a fun toy.

FO, for better or for worse, isn't SimGalaxy.
You understand that the image represents the economy of an empire attacked by an opposing faction? This is something present in 4X games, not in Sim games.

guiguibaah
Creative Contributor
Posts: 441
Joined: Tue Oct 07, 2003 1:00 am

Sim games

#12 Post by guiguibaah » Fri Jul 02, 2004 10:49 pm

My sim would constantly get attacked by annoying neighbours entering MY house and using MY bathroom. So I did what every self-respecting sim would do. I walled off the offending sim and started a collection of 'sim neighbours'.

They were my precious, precious littles! Oh, the screams, they are of pleasure, not of hunger or pain. Be silent, my little sims, for your day will soon come, and your cell will soon again be empty for another body to fill.
There are three kinds of people in this world - those who can count, and those who can't.

vishnou00
Space Kraken
Posts: 157
Joined: Tue May 25, 2004 3:15 am

#13 Post by vishnou00 » Fri Jul 02, 2004 10:54 pm

Please don't mix up the Sims with the Sim games (SimCity...). One is a doll game and the others are a management toys.

vishnou00
Space Kraken
Posts: 157
Joined: Tue May 25, 2004 3:15 am

#14 Post by vishnou00 » Sat Jul 03, 2004 12:11 am

Now onto drek's first post.

1. lack of precedence : like I said before, I've never seen a macro tool attempted at a large scale, but I can descibe those used in Stars!
-default and custom plantet production queue: there is a default production queue for planet as soon as you colonize them, and you can select/customize other ones (through the diamond menu)
-diamond menu for transport order: since a transport order has many options (4 type of cargo + fuel, load %, load qty, load dunnage, quick load, wait load...) the order may be saved in the diamond menu for quick usage afterward. It doesn't make decision, it only reduce the typical number of steps from 12 to 1.
-repeat orders: takes the just done order and put it at the end of the queue. Makes for patrolling and supply routes.
While imperfect, I don't feel they have pushed the tool to the limit.

Chess is not applicable as it has very little in common with 4X games. I could very well tell you that macro exist and is a huge advantage in fighter games (such as King of Fighters and Last Blade) where you enter commands for special moves as quick manipulation of joystick and buttons. This is because it reduces the (execution) barrier between your thoughts and the game. The differences with chess:
-doesn't have eXploration, a basis of 4X gaming
-doesn't have eXpansion, as you don't create new units (pawn promotion isn't creation)
-doesn't have eXploitation, as there is no production economy
-doesn't have eXtermination, the goal is to checkmate the king. Usually when you "exterminate" every opponent's pieces, you are on the verge of drawing the game, which is only good for the loser.
-it's a perfect information game.
-has very coarse granularity. 64 cases and 16 pieces aren't "epic".

2. As I said before, a player should supply creativity and analysis to the tool and the tools do the chore of implementing it in the game. For example, in a civ game, you could decide only once that a square should have irrigation, then road, then super-irrigation, then railroad (when available). A worker assigned to work in that group of square would develop that square as you want, without requiring further attention from you.

I don't want AI trying to have insight, I just want them to execute idea that too 30 seconds to have, 2 minutes to translate to the tools but that would take me 1 minute of attention every turn.

This may be the reason why experienced developer/coder don't do good macro tools: they don't try to. They just staple the AI somewhere in the UI without further thought for inputting player's decision. (Well, SMAC attempted to let you configure the AI, but instead of presenting you with options related to player activity, they provided the parameters of the AI, so you just end up configuring the piss-poor AI.)

3. You have to police the tools if they overstep their duty: executing orders. As soon as they try to make decision, you can't trust them and have to police them. If they are strictly following your instructions, you only have to police your inputs and the game, which you already do because you're playing the game.

This is why the tools must be simple yet powerful, so the player can really know the limits of the tools, recognise when a limit is reached and input further instructions. It doesn't have to be very complex to be powerful: it doesn't require branching, as branching implies a decision. But with just grouping, sequence of orders, one line if (as in "if not already done, do it") and simple loops (as in repeat always), you can transform micromanagement into decision making.

These kind of activities are already done by players, only that they do it in their head and tend to forget what they were doing with what. That's why when I play a slow strategy game (PBEM with more than 2 hours per turn) I use some kind of notepad to avoid forgetting my plans.
drek wrote:proposing that players of a 4x game will have to have an understanding of boolean logic, conditional statements, and program flow.
These are very simple, everyday activities. Everytime you say "and" and "or", you are doing boolean logic. Everytime you say "if" you are doing conditional statements. Program flow (as basic as we need) can be experienced in various grade school activity, such as music (repeat bars), early math (multplication is often presented as addition done multiple time, the same thing for exponential and multiplications). These are things people already know. We aren't asking people to master pointer and bitwise operations! Presenting it in an intuitive fashion is only a UI problem.

4. Repetitive activity are not worth the player's attention as long as they aren't disrupted. Every activity must be justified by its gameplay consequences, so if it is something that just "grow" without it being significantly disruptable, it shouldn't exist. Players' empires exist as long as there are a threat to that empire: when there is no threat anymore, it's game over, nobody cares about the remaining empire. It is possible that player continue to play even after the game is over, but they aren't not playing anymore, they are just toying.

I'm not FOR micromanagement, but I don't think microphobia/distrust of macrotools should get in the way of good gameplay mechanics.

Hexxium
Space Floater
Posts: 41
Joined: Tue Apr 06, 2004 8:40 pm

#15 Post by Hexxium » Sat Jul 03, 2004 2:43 am

I agree that AI automation, like viceroys, will most probably fail. But in my opinion that's not what a "macro tool" is. For me, a macro tool is a tool that makes it easier for the player to apply his strategy, for example issue a comand to all free feets, all farming focus worlds, all shipyards, etc.

Automation that doesn't involve AI, like build lists, also fits in this category. I think it's a good thing in general.

Micromanagement is that it is fun in some situations. For example, in most games micro is fun in the beginning, sometimes also for important ("key") worlds, or when there's some kind of "crisis" (blockade, food shortage, etc.).

Removing micro definately removes some interesting game aspects / strategies. It's important to keep that in mind - there just have to be other things to make up for this, and it will be fine.
Macro tools are an easy way to keep the common "interesting situations" and strategies to deal with them, while providing a way to skip the tedious tasks in "not-so-interesting situations" ;)

Also, from what Drek said about micromanagement / micro tools, he concluded that it's better to completely remove the need to micro. Don't forget the other possible conclusion: Make micromanagement more interesting.

Post Reply