You don''t need a neutralising force, you just need a neutralising strategy. Your implication is that the force on the hill is static whereas the force on the ground is not. So, just amass a larger army. Or use air attacks. Or cut off supply lines. Or just beat them through attrition. Whatever you like. I don''t think the mentioned situations have any correlation to RPS being ''more suitable''.

[ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost ]

I didnt mean for this to be a case for/against rps. I was trying to find flaws in the ideas put forward as to how they would like to see a strategy game build.

G''day!

Some things have been repeated in this thread and they got me thinking. The idea that a game that''s more "realistic" is better than an RPS-based one bothers me. Is it really true that in reality, given a specific tech-period, than no units has an RPS relationship?

You can easily analyze the RTSs mentioned, because many thousands of ''commanders'' wage many ''wars'' each day and publicly contribute that knowledge. The average RTS players has waged many more battles and received far more information about the tech/units available in a week than most real commanders do in a lifetime.

A serious problem with simulating some of the epic real-world battles in an RTS is the ''RT'' part. If I''m sending in a carrier, if the enemy is smart, it will never reach it''s destination. So to protect it, I have to send various types of escorts. The problem is, in reality I would have competant sub-commanders. I may give the general orders, but they are going to carry them out and react to sudden changes. If they get ''out-flanked'' they are not going to follow the orders I gave them, they will react.

Show me an AI in ANY game, in any genre, real-time or not, where I would be able to give detailed unique orders to a sub-commander and have them follow them. I doubt you can. Even if one is created, then the challenge would be to create one that can react to changing situations as best they can, while still keeping my orders in mind. We''ll be waiting a long time for this. It would require that the AI understand the ''spirit'' of my orders.

If you control everything yourself, then it is not possible to handle that many different types of ''realistic'' units and use them effectively. The real-time part of RTS makes that impossible (at least for any but the cream of RTS players).

Don''t get me wrong, I''d love to see things get shaken up. When I get a game like AoE/K, MOO I/II, Civ I/II/II, I don''t read the tech tree. I like the discovery of what my opponent is hitting me with. Realisticly, this goes away after a while because even without any outside references I learn what tech is available. I don''t realisticly see a game coming out that would keep my sense of discovery active indefinitely.

I don''t have any answers, and the few people in this thread who claim to have some, say they don''t want to share. But if we''re not going to try to come up with solid ideas, then what''s the point of a thread like this?

Dauntless''s original post was throwing out an idea and waiting for feedback. I assume constructive feedback, not a lot of hand waving was the goal. So I guess the question is, of all the people who see RPS as bad (or at least want to try to find something ''better'') who has an real idea of how to achieve it?





Stay Casual,

Ken
Drunken Hyena

quote: Original post by DrunkenHyena
Show me an AI in ANY game, in any genre, real-time or not, where I would be able to give detailed unique orders to a sub-commander and have them follow them. I doubt you can. Even if one is created, then the challenge would be to create one that can react to changing situations as best they can, while still keeping my orders in mind. We''ll be waiting a long time for this. It would require that the AI understand the ''spirit'' of my orders.

I don''t think it would take all that much to do at all. And the AI doesn''t need to be able to understand the spirit of your orders, you just have to be given tools that allow you to better specify them.

Ideas (since you called for some):
- allow a variety of formations, such that units guarding other units stay in the correct position relative to the others
- backup orders, to be executed if the original order can no longer be completed (note that this is different from order queuing, yet not mutually exclusive)
- some basic AI levels similar to those seen in Total Annihilation and other games, allowing a coarse auto-response to enemies.
- different amounts of damage taken from hits inflicted at different angles (encourages better use of formation and flanking strategies)

quote: I don''t realisticly see a game coming out that would keep my sense of discovery active indefinitely.

Maybe not, but then Chess has a hell of a lot of depth to it, all with only a handful of units, and the weakest unit can beat the strongest if used correctly. This is something to aim for.

[ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost ]

quote: Original post by Kylotan
I don''t think it would take all that much to do at all. And the AI doesn''t need to be able to understand the spirit of your orders, you just have to be given tools that allow you to better specify them.

Ideas (since you called for some):
- allow a variety of formations, such that units guarding other units stay in the correct position relative to the others
- backup orders, to be executed if the original order can no longer be completed (note that this is different from order queuing, yet not mutually exclusive)
- some basic AI levels similar to those seen in Total Annihilation and other games, allowing a coarse auto-response to enemies.

Good formations would require better tools than I''ve seen thus far. One game (don''t remember which) had a box formation designed to protect ''weak'' troops. Since you had no way of designating ''weakness'' it decided. It put some non-combatents in the front line and weak military in the middle. Good escort for the civilians, eh? Yeah, you lead, we''ll be RIGHT behind you...

Backup orders would be a good idea. I don''t recall the TA auto-response, but I seem to recall it being something like the standard RTS aggression modes, but with more detail. Could be good as a backup to the backup (When all else fails...) and/or as a ''operating disposition'' while doing the orders. ''Try to be stealthy'' or ''Try to minimize personal losses'' or ''Complete objectives at all costs''.

If not turn-based, this would require a very different pace than the standard RTS (observation, not criticism). Of course ''real'' battles can easily have more time spent on the planning than the actually combat.

So you''d need to be able to spend time carefully planning and assigning formations (which of course you could save your favourites to be easily called up again).

Has anyone played Robosport? You could prep your armies with all the detail discussed above, and then hit ''Go'' then either then entire battle would take place based on your orders, or a ''phase'' of the battle would take place and then freeze while both sides gave new orders.

Alternately, after hitting ''Go'', it would play out in Real time, fully interactive, and then you could try to manage a real-time battle that was pre-planned in detail.

quote:
- different amounts of damage taken from hits inflicted at different angles (encourages better use of formation and flanking strategies)

That would be good. Makes logical sense too. An infantryman should be better able to defend himself from a straight on attack, then from an attack from his rear.

That could also affect unit morale. Getting hit from behind being rather demoralizing.

quote:
I don''t realisticly see a game coming out that would keep my sense of discovery active indefinitely.
Maybe not, but then Chess has a hell of a lot of depth to it, all with only a handful of units, and the weakest unit can beat the strongest if used correctly. This is something to aim for.

There is a lot of strategic depth in chess, but it''s a little too abstract for my taste. Depending on the game, it doesn''t always make sense for the weakest unit to be able to defeat the most powerful. In a game like StarCraft where power levels are so extreme, a single marine should not be able to take out a Battlecruiser.
But I agree with the basic principle. A well led army should have a chance against a poorly led, but numerically superior, army.

Stay Casual,

Ken
Drunken Hyena

How many have tried Firaxis Games'' Alpha Centari? They don''t have pre-defined units. You get technologies and build your own. For the casual gamer, it will suggest units, but you have total control of what you build. You want to throw the biggest weapon, best armour and cool specials on an anti-grav chassis? Okay, but it''ll cost ya. Your opponent could easily crank out a design of a cheap unit with low armour, but a big weapon and take it out for a fraction of the cost.

A system like that would address a number of the issues presented.


Stay Casual,

Ken
Drunken Hyena

Combining the ideas of my previous posts with Kylotan''s comments:

Picture a campaign map that''s vaguely Risk like. Each region is pre-defined and either player can view a detailed map of it at any time (though not seeing any units/fortifications).

Each side has a home base. This where the economy and war-building efforts happen in a turn-based manner. Economy is important , but not micromanaged in the heat of battle.

Research gets you new technologies, which you can use to design new units. Spies can get you info about the fortifications/units in a region.

If you choose to assault region A, you plan it all out in detail as mentioned above. The defending player (possibly with information from his spies) has defensive orders given to all of his troops. When both sides are ready, a real-time combat ensues. Unless you intervene your units will follow your orders to the best of their ability.

They''ll use their backup orders when necessary (and alert you that they are doing so) all the while acting on their ''disposition'' which is kind of like the ''spirit'' of their mission.

Having detailed orders means that less intervention is required, which reduces the benefit of ''mousing'' skills.

Of course there would be the possibility of 1-shot scenarios, which would be like the above without the long-term economic issues.

Propaganda could come into play as well, adding to your troops morale, and/or reducing enemy morale.

I didn''t copy all of my or Kylotan''s idea down here since, they''re just above this.

Comments?


Stay Casual,

Ken
Drunken Hyena

quote: Original post by DrunkenHyena
Good formations would require better tools than I''ve seen thus far. One game (don''t remember which) had a box formation designed to protect ''weak'' troops. Since you had no way of designating ''weakness'' it decided. It put some non-combatents in the front line and weak military in the middle. Good escort for the civilians, eh? Yeah, you lead, we''ll be RIGHT behind you...

Sounds like Myth Flocking behaviour could be used to improve formations. It could also be used to keep certain units in a given position relative to the unit they''re following. It may well take better tools, so let''s design them

quote: Backup orders would be a good idea. I don''t recall the TA auto-response, but I seem to recall it being something like the standard RTS aggression modes, but with more detail. Could be good as a backup to the backup (When all else fails...) and/or as a ''operating disposition'' while doing the orders. ''Try to be stealthy'' or ''Try to minimize personal losses'' or ''Complete objectives at all costs''.

Yeah... there is basic stuff such as "return fire/don''t return fire", "evade fire/stand ground/roam to defend/hunt to defend" and so on that can customise AI responses. These are mainly useful for ensuring that units defend their positions effectively. For attacking, you could set a ''Flee Percentage'', which tells the unit to withdraw if its damage exceeds a given percentage. This could instead apply to a group rather than a single unit. In fact, making some options apply to whole groups instead of having to be done on a per-unit basis would make a lot of things easier. Being able to specify ''attack from flanks'' for example, might be nice for your faster units, and when the fast units are given the same target as the slow units, they will attack in the different way. (Easy enough to do - instead of pathing directly to the enemy, path to a point near the enemy but perpendicular to the line between the origin and the enemy, then from there approach the enemy from the side.)

quote: There is a lot of strategic depth in chess, but it''s a little too abstract for my taste. Depending on the game, it doesn''t always make sense for the weakest unit to be able to defeat the most powerful.

Yeah, but the other end of the spectrum (Unit A always beats Unit B unless you tell Unit A not to retaliate) is even less satisfying to me. So something in between would be nice.

Example: It might be interesting to see a game where certain tanks (''Tank A'') have heavy armour at the front and light armour at the sides and back. This means that, to be effective, they must line up in tight ranks to reduce the chance of being hit in the side. The opposition might normally be able to use a fast and light tank (''Tank B'') to outmanoeuvre one or two of the above Tank As in order to score a hit on the side, but would have trouble against a line of them unless they can get in behind them. The implication of this, is that one Tank B will beat one Tank A, but 10 Tank Bs will probably lose to 10 Tank As if the Tank As are used effectively. In both cases, the onus is on the player to avoid using their units in a situation where they will come out worse.


[ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost ]

Drunken:

you are absolutely right in how critical AI would be in a gmae of the nature I propose. That''s why I think it will still be a few years before what I''d like to see is possible. The sheer level of decisions that would need to be handed off to sub-commanders is massive and would require a sophisticated level of AI.

A possiblity would be to hand this off to other players like a co-op team based effort which I thought about, but I don''t know how manageable that would be. I actually did toy around with the idea of having teams-based strategy games but I''m still trying to figure out how to implement such a game. Each player would essentially have a rank, and would therefore be able to only order units at his disposal. The overall commander could also manipulate the economic and social problems of his country. In my tabletop project, I have some rough ideas on this, but for a computer version, I really don''t know how to implement it, but I think it''s a neat idea and hope I can come up with something.

I think you are also very right about the "real time" nature of RTS''s. I think they actually go too fast for strategical purposes, and if they could be slowed down a great deal (some games do have this option) it would allow for greater strategical thinking. Personally, i would like to see a hybrid system that incorporates both styles.

However, RPS situation really don''t work in the real world because in the real world they work more like this:

A beats B
B beats C
C beats D
E beats A
F beats B
G is equal with H
I can beat B and D
etc etc

Generally speaking, there is no loop involved. I think a better model is Generalized vs. Specialized units. Generalized units are sort of your utility handy men, they can do several things ok, but no one thing extremely well. Then you have your specialized units which are very good at defeating one kind of unit, but are useless against virtually everything else or have a sever handicap against other units. There are exceptions to these rules however, but for the most part this is how it goes.

Think of the Aircraft carrier again. Here we have a ship that can destroy every other ship in the water with ease....at a long range. But, it in turn is incredibly vulnerable to virtually every other ship out there at close range. It relies on escorts for protection against subs or its own air superiority squadrons for protection against other aircraft. And against battleships or other land vessels, it better stay out of range of its guns, or make sure its fighters destroy it before they do. In other words it doesn''t follow RPS at all. It can destroy anything, and in turn be destroyed by anything.

ditto with infantry really. In small quantities, they are vulnerable, but in large enough groups with it''s specialty teams attached, they become a very potent force. Never fight an infantry unit with unsupported Armor, and even planes have to watch out for SAM teams. But again, it can be destroyed by any other unit.

My argument against RPS is basically this...it''s incredibly limiting. Sure it works, but the depth of strategy becomes very limited. For the umpteenth time I will point out chess as an example. It does not utilize RPS and has an amazing depth of strategy because any one piece can destroy any other piece at any time (save the King versus a queen). This simply isn''t possible in RPS and is the reaon why I think it is an inferior form of balancing design. and here is a game that is not "realistic" as the detractors of my comments have pointed out. For the ones who claim that realism is not fun, chess is the perfect example of a game that does not use RPS, has a much richer and deep strategical mode, and has withstood the test of time as perhaps the greatest strategy game ever.

Let''s try to expand our creativity:

A > B
B > C
C > A

This is the standard R/P/S model. How about some alternatives:

A + B > C
B + D > C
C + D > A

Or how about this:

A > B on Tuesdays
B > A after dark

Or this:

2A > B
B > A

Or this:

5A > 5B
B > C
C > (1 to 4)A

Or some simple support type formats:
4A + B> 6A
A > B

4A + B + C > 10A
A>B, A>C, A>(B+C)

Now, which of these formats looks the most fun to play: All of the above! If all were incorporated into a game, that game would be much more interesting than one by itself.

Now, let''s think about the purpose of R/P/S. It''s to encourage diversity, is it not?
Thus, by definition of its purpose, it should not be adhered to. That is not to say; however, that it should not be used.

R/P/S makes a great starting line. It makes a terrible finish line.