Being good at "strategy"

This really got out of hand. Boiled down, possibly too far:

Strategy: plan
Tactics: execution


Yes it's more complicated than that, but not too much. The complexity and confusion stems mostly from the inescapable facts/realizations that plans change (and part of the overall plan is, most often, to change the plan as new information is discovered) and that successful (or unsuccessful) execution leads to new information (i.e., that far west expansion was successful!! or raided early, etc).


Tactics (execution) can be practiced. US military and in particular it's special forces does this famously.
Strategy is typically theorized and historically relevant samples studied. Again with my clearly biased US examples, West Point can be cited. Since we are playing a simulation that (regular beta balance changes aside) is restarted potentially infinitely with identical technological constraints repeated exactly (never the case in the real world), strategy is perhaps an improperly applied term, and can in this uniquely constrained case be "memorized" from examples pioneered by other players. Since the tactics also suffer from repetition due to scenarios constrained to be unrealistically similar between matches, the two are easily conflated.



Meh. Not sure I'm actually helping here so I'll stop.

Have any of you ever played Advance Wars?
http://en.wikipedia.org/wiki/Advance_Wars

One of my all time fave turn-based tactical/strategy games. I highly recommend it.

I don't agree about that definition for RTS and games in general. I like to frame strategy in "strategic choices" rather than saying that strategy is a long-term plan.
I think that strategic choices can be both long term and short term decisions.
What for me is at the core of a strategic choice is that all the options are, to the player, seemingly viable but have strategical and tactical implications as they are decided.

Yes they can be both, it's relative to the objectives you hope to achieve. However, the center of any strategic plan is the mission itself. How you achieve that mission is tactics.

A true strategy game would be something along the lines of a typical paradox title, eg europa universalis or hearts of iron. PA is too open ended and dynamic at any given moment; the framework for careful, long term planning does not exist beyond the environments in which you wage war, which happen to be very simple maps.

This is a mix of strategy and tactics, but chiefly tactics because the focus is not on running an empire but on fighting battles. The little strategy that there is consists of planning an attack on a specific base with your army, coordinating such preparation with your human ally, or planning a defensive setup for your own base. That is as deep as it gets on that front and then you're back in the tactical realm choosing which units attack which other units.

And speaking of tactics, that is presently a little on the bald side as well, since it chiefly consists of unit and resource spamming due to the size of armies. Every unit is completely expendable. The strongest player is always the one that can spam/zerg the quickest, not the one with the best strategy or tactical decisions.

Where the above can and does change, is in the largest multiplayer games. The more players there are per team, the more planning is required because one spammer cannot carry a team of four or six on his or her own against an equally sized alliance, subsequently cooperation leads to planning which leads to actual tactics and strategy.

You are exactly right! The simple definition of Strategy from the dictionary:

"a plan of action designed to achieve a long-term or overall aim"
"the art of disposing armed forces in order of battle and of organizing operations, especially during contact with an enemy."

I just joined the beta and finding my feet. Right now Im concerned that there isn't a lot of strategy in the game. It all feels very oldskool skirmish. Build as many tanks as quickly as possible and send them all over at once type stuff.

I hope the game evolves away from that. There is so much potential with different planets and putting units into orbit etc. There needs to me more strategy for the player so they need to think how are they going to penetrate the enemy. Not just who can build the fastest.

One quick fix would be to make the defensive buildings much stronger. So even if you herd 1,000 of the same units all in one direction they would all get picked off easily by a couple of defensive towers. That would force the opponent to split their attacks and use special units better. That would also make each game last longer.

Well this is mostly arguing about semantics but isn't the mission always the same in PA? Kill all enemy commanders? You haven't really defined what you mean by mission.

You have to define what you mean by "true strategy". I don't agree that strategy is all about running an empire because then it is only a matter of scale. Like if you would say Risk is more strategy than Chess because you are fighting on the whole world and not just one battlefield. Likewise you could say that PA is bigger scale than Europa Universalis because your are controlling whole planets and not just nations.

I like to emphasize strategic choices when I define strategy and with that definition you can have strategy in the tiniest of fights during the blink of an eye to fighting in the whole galaxy for millennia to come.

I like to think that strategic choices throughout the game is like a series of Rock-Paper-Scissors where the decisions leaves the players with asymmetric options as the game unravels.
I might start with a bomber while you start with a fighter. If your fighter kills my bomber it might give you small edge but if I get some kills with my bomber before it goes down it puts us in different positions where we have different options in the immediate aftermath.
If I go for an early nuke it means I have to dedicate a lot of resources to that which might mean I have a smaller army while you prepare a bomber snipe.
Strategic decisions limits the available options and forces the players into asymmetric positions where they have to figure out their strategic goals and what tactics they should deploy to get ahead and ultimately win.

Pardon me for not reading any previous comments. In response to the OP;

My view of the terms-

Strategy - Cumulative goal oriented actions which are done with foresight (planning ahead) to achieve a goal.
Tactics - The actions.

How well your strategy works depends on experience and knowledge; you have to know what you're using to achieve your goal, and how those tools can work.

As mentioned by somebody else, strategy entails reacting to complex changing situations in a complex, often novel way; i.e actual planning. Most RTS games only have one optimal solution toward victory, tied to variables such as APM of competing players and differences between maps.

But how could that be remedied? As long as your meaningful choices revolve around which units to send against which they can always be mathematically analyzed and compared - DPS comparisons, mass efficiency, timings, you name it. The only way to keep enough uncertainty and fuzzines to prevent this would be to introduce random elements with a meaningful impact on your choices. Right now there are random elements - random planets - but they have no meaningful impact.

I admittedly haven't read through the whole seven pages but a lot of people's definitions of tactic/strategy seem to boil down to:

Tactic: Employing an optimal set of actions to a small subset of gamestates that can largely be done through memorization.

Strategy: Tactics, but on a scale so large, the optimal solution is unknowable to the player. And he has to find the most optimal perceived solution on the fly.

The problem with this distinction is that it is based entirely on player capabilities. To once again stress the chess/go analogy, if you've played either game only casually you'll have to plan/stratagize every turn to deduct a good move, if you are a quantum supercomputer all possible outcomes are already known to you and the games is solely a matter of flawless execution

Well that isn't the only type of random elements. I'd say that Rock-Paper-Scissors balance and imperfect information provides a lot of random elements.
If the enemy chose a strategy that counters your strategy you might be able to change your strategy if you scout it in time. However scouting in itself is time consuming and requires resources which could be seen as a deterrent to the execution of your strategy. You have to weigh knowing what the enemy does or predicting what the enemy will actually do.
Scouting mechanics and strategic RPS is a very big element of randomness that you have to take into account when you chose your strategy.
Do you build an anti-nuke just in case, do you sacrifice air scouts to see if and when the enemy makes nukes or can predict when the enemy make a nuke?

While I agree with most of your post I have to ask if Rock-Paper-Scissors can be solved?
There are only a few gamestates in Rock-Paper-Scissors but I'd still say you can't solve the game and it doesn't come down to a matter of flawless execution. In a way you could say that every victory in Rock-Paper-Scissors is flawless execution as there is only one decision that led you to that victory.
While Chess and Go can theoretically be solved it doesn't translate to PA if there exists strategic choices with Rock-Paper-Scissors mechanics.
In PA, you have to make choices based on imperfect information. Flawless execution would mean that you finished your anti-nuke just in time to intercept the enemy nuke without even expending resources to scout the enemy.
Scouting in itself is redundant if you know what the enemy will do. You have to weigh expending resources to scout the enemy strategy versus focusing all your resources on your strategy.
Lets say that we solve PA computationally. We can calculate every outcome of every possible input and variable. Does that mean that there is only 1 strategy that will always win?
No, it doesn't necessarily mean that if there is any Rock-Paper-Scissors component to the game.
If there is a strategy that is considered to be the best strategy because it got the most flexibility and highest win rate it can still lose to counter strategies.
When people talk about metagame they usually refer to this. The strongest strategy is the meta but there can still exist weaker overall strategies that beat the meta strategy.
As PA, Go and even Chess could be considered almost computationally impossible to solve, the thing that makes an RTS with imperfect information more strategic in my opinion is that reading the state of the game itself by scouting have to be part of your strategy. You can play it safe or risky. How much resources can you expend on scouting to avoid nasty surprises?
How much resources can you focus on your strategy without losing to a counter strategy?
There isn't a perfect answer.

While I agree in general I don't think that your example is fitting. Investing in a scout is always better than pouring lots of resources into something you might not even need. This does however not lead to strategic choices - it leads to scouting and subsequently to the application of the correct predetermined counter. Choosing to scout or not to scout is not a strategic choice but a gamble.

I disagree. Scouting gets more and more expensive as enemy AA and fighter coverage increases with time. If you want to scout the nuke being built you have to scout continuously which will eventually outweigh the cost of an anti-nuke.
A preemptive anti-nuke on the other hand might be scouted and the enemy might not make a nuke just because of that.
Actually I think that gamble and risk is largely what strategy is about. It is largely about weighing risk versus benefit and trying to predict the enemy actions.
Also there is an alternative to scouting and that is forcing the enemy to act. Forcing
Instead of spending resources on scouts you can put that into offensive units so that enemy can't build that nuke in peace but have to defend and the extra resources might be the difference that makes you able to push through and destroy that nuke.

I am still not sold on the idea that scouting might become prohibitively expensive at any given point in time - at least if you include radar and all the other options.
The concept of forcing is a nice fit, though, on which I agree absolutely. But then again the two concepts will most often occur together to varying degrees.

Later in the game nukes can be created extremely quickly. The window to scout the nuke before it launches gets smaller and smaller as the income grows. If you have a very dense base where a nuke hit would cause critical damage, making a preemptive anti-nuke might be a good option rather than sacrificing scouts constantly or you can do like me where I spread out my base so much that a nuke hit won't do critical damage.

Yes. If you use Doxes to scout, you not only gain information but also threatens enemy expansion.

I think that a huge part of what sets PA and other RTS's apart from games like Chess and many others is the imperfect information aspect of it. It makes it much more like a blend of Chess and Poker, and I am actually surprised that nobody has brought up the poker analogy yet.

We know Poker played at a medium to high level becomes a fascinating and deep psychological game, and I don't see PA as played by competent players as being much different -- it is in fact much more nuanced in the sense that it isn't all-or-nothing when it comes to revealing important information, because rather than a set of 2 or 5 always-hidden cards (which is just a few bits of information) we have one or two orders of magnitude more strategically-relevant information that can be partially uncovered using scouts, with this scouting action itself having various counters available for deployment...

It's a much more perfect blend in describing the dynamics of real conflict -- at least as well as our science-fiction lore here can describe -- than the games of Chess and Poker do by themselves, and it just seems natural to me that the brand of "strategy" as described by Sun Tzu is completely in line with how to approach problems in PA, its parameters being for waging war between humans on earthly terrain notwithstanding.

PA is such a better game than Chess in my view because it is simply absurd to place arbitrary limits. What says that I have exactly this many pawns to start with which just so happens to be what you also have to start with? Everything just makes more sense. If I have more resources, I can gather more forces and build more megatons of warheads. Come try and stop me. If I want to go to a spot in between two other spots, I can just go there. Why the hell not?

A lot of Chess's complexity arises out of the discrete fractal-like behavior of the structure of the game state itself, and by moving from a discrete 8x8 board to a real spherical geometry and using ballistic trajectory modeling may erase some of that potential for "not being able to kill some particular piece because my knight is one spot away", but I just never liked that to begin with because it made no sense!

Now, there certainly are situations where time drags and outcomes are seemingly determined by which party pumps out more tanks, or better yet, who can pump out the T2 air faster, but I really do see this as just a balance issue, because a couple of well placed laser towers can always stop a land army, and we do now have the AA flak that can counter T2 air. But I definitely think that in high level play this sort of situation won't arise because the only way to get there is by having a crappy strategy: You should always be attacking, always probing your enemy for weak points, and never letting up until their com eats the dust. When the game is not being buggy and when the balance kinks are ironed out, the only reason you should find yourself in a boring stalemate situation is if you put yourself in it or if your enemy is equally lackluster.

"Real time strategy"
A game where strategy is made up on the fly based but using past experience to help give you a bit of a foundation for said strategy. Real time my freinds.. Real time- it's as simple as that it seems.

try

try actung panzer i say the samething to everyone.

Someone else said regarding him. Paraphrased, his play style is said to be ruthlessly accurate, like a computer.

But that's not "strategy", it still relies on the brute force effect.

Do it then. Put your money where your mouth is.


I don't believe in the concept of underdevelopment, it's silly. It's not like you choose when your pieces spawn. They're always on the board, so they are always developed, so they'll always be having an impact. He only uses the Queen defensively in that game. It's an example of a non-offensive Queen game, and then there's just a short step and a jump to a non-defensive Queen game.

That being said, however useful/not useful a Queen is, it does underline my initial point. Turn based video games such as XCOM and CivV up the difficulty by improving both the "computer"'s quantity of units and quality of units, and the human is still able to beat them.

Essentially my argument is tactics = perfect information, strategy = imperfect information, and by giving you perfect information from the beginning, there isn't be any strategy in the game because opportunity costs are eroded for decisions.