Being good at "strategy"

So there is a lot of talk all the time all over the forums how PA is supposed to be a "strategy" game.
I think that's a horrible thing. Why? Because that word is not very well understood. I might not understand it myself. So let's ask a simple question that tries to reveal what we think of strategy.
I am asking this question without giving you a definition of strategy, use your own definition. You don't need to give the definition directly. Just answer the question. Also note that I am assuming that PA is supposed to be a game that players can learn to play to be good at. That's a trait that a good RTS generally should have, I hope everyone can agree on it.
So the question is:

How does one become good at a 100% strategic game?

My personal answer is:
To be good at a strategic game a player needs to memorize as many game situations and their perfect solutions as possible, including the understanding of the "why is that solution the right one". This is achieved through pregame theorycrafting and lots of focused practice. That way the player can use that knowledge base to form a solution to a given situation in the game. The reason why it is best to memorize situations and their solutions is that it is too complex to "calculate" the best solutions directly ingame.
So a 100% strategic game happens to be all about memorization and repetition.

We were typing at the same time in different topics and we came to similiar conclusions. Great minds think alike haha
I'll copy it here:

I think those are general elements of RTS game (I think it covers it relatively well, cba to find more):
1. Knowledge of units
2. Knowledge of economy
3. Army movement.
4. APM
5. Micromanagament of units
6. Map control
7a Macro scouting
7b Army movement scouting.
8. Adapting to scouting
9. Predicting future happenings

1. You know what composition and what types of units are best for certain situations and moments in the game. Skills involved:
Experience, Experimentation, Observation, Memorization.
or
Copying from pro, Memorization.
2. You know in what order you build buildings and create units. Again
Experience, Experimentation, Observation, Memorization.
or
Copying from pro till certain point in the game, EEO, Memorization.
3. Knowledge where and when to send units.
EEOM or copying and memorization. Trust me, it can be copied.
4. The speed you manage all of your stuff, limited by UI.
This is only marginally copied, you have to learn it yourself.
5. The way you maximize efficiency of your units if you managed to build them in best compositon, in best order and fastest way possible and the best way possible to certain place.
This has 2 parts - Knowledge of how to micro something and then how well you execute it.
6. This is 3 and 2 combined, you take territory, then secure it. Can be copied and memorized.
7. You need to know what your opponent is doing.
You need to remember to do it, then you need to know how to read what you see and then you need to know how to react to it. Memorization.
8. Whatever, Memorization.
9. Can be memorization too, based upon what you know about the game.

To sum up, in theory if you play perfectly, everything except APM and micromanagement is memorization. This is why I got bored with starcraft 2. Most of build orders and reactions to situations are already invented and right now it's mostly about memorization and apm and micro. I like to develop new strategies and new build orders. It's only one time process. Once strategy is invented it is copied and memorized by everyone. The more RTS game is developed, the more it is about memorization and execution and less about smart thinking.

I would say strategy are the points 1 3 6 7 8 9.
Economist is 2 and 6
Tactician is 5 and partially 3.

Starcraft:
Heavy on Tactics, mid-heavy on economy and strategy

Supcom FA:
Relatively heavy on tactics (micro with commander and units) , very heavy on economy, quite heavy on strategy

PA:
mid on tactics (I think it will be high apm demanding in future when we get better), heavy on economy, heavy on strategy. People want to lower tactics even more leading to more memorizing.

How other games coped with the problem of memorization of strategy and economy elements:
1. Chess and to certain extent supcom - Insane number of possibilities in order to make it impossible to memorize them all (This is good but it actually pushes it away in time from what i see from absolutely sick loki's post). Supcom died and at the end it was a bit of memorization. Chess didnt die and it took a couple of hundred of years to have this problem appeared.
2. Starcraft - We introduce a lot of APM and Micro requirements to make it more sporty, less strategy.

Both are not satisfactionary after a couple of years of development of game.
HMMM

Conclusion:
It seems PA is going in the direction in which game will die before we invent all the answers and that's where game leans right now - more planets, a lot of units, million of mex spots, thousands of units. It's relatively decent answer for those who dont like to click fast and micro heavy.

What cola colin previously suggested and I disagreed but after longer consideration this actually might be relatively reasonable solution is FAIR randomization. This would force you to invent strategies on the fly and take away memorization. But this must be fair, you can't have randomization like current when one of guys get random advantage and the game says "deal with it". I have no idea how would fair randomization look like and even then, im not sure whether that would end up in just memorization of possible closest scenarios.

A question: if strategy is about remembering solutions to problems, who came up with the solutions?

When I think of strategy two scenarios come to my mind.

The first is chess. Dancing the pieces around trying to get the right approach and leg up on your opponent. This is a top down big picture 10 moves ahead of your opponent kinda strategy. It involves thinking ahead and being prepared for all the different things that could come my way.

The other thing I think of is back when I played paintball all the time. I think of this as more individual strategy. How can I gain the upper hand on my opponents on the other side of the clearing. Do I lay down suppressing fire and let my buddies take care of it, do I make quick brazen moves, or slow calculated sneak moves. Do I stay where I am, do I move a little, or do I fall back and make a large flanking move.

Strategy is all about options and choosing which option is best for the situation at hand and adapting to what comes my way.

A good strategy game presents players with many options so that no game is unique.

Easy example being chess. Another nice example is the strategy of fleet combat in Eve Online. There are untold fits, fleet compositions, and approach strategies. While different engagements may have similar situations, like when jumping through a gate, each situation is completely different at the same time.

This applies to PA in many different ways. The randomized planet generation is a great foundation. It means that while each match often contains similarities, each match is still completely different.

Do I use bots? Vehicles? Air? Which planet should I go to? What type of bots should I use? Do I need to be aggressive or defensive and build up a large force first?

A good player of any strategic game, be it chess, PA, SupCom, or whatever else, must be intimately knowledgeable of every unit and possibility in the game and must adapt to any situation that comes their way.

Kinda rambled there for a bit.

--

colin.

I disagree that a good strategy game is all about memorization and repetition. Memorization and repetition is a factor of becoming good at a strategy game. You must memorize what units/pieces/whatever are, what they can do, and what their strengths are – and doing so requires repeatedly playing the game over and over.

But a good strategy game is all about having multiple options at the player's disposal and using each and every one to their maximum effectiveness in every different situation.

So a good strategy game must have options and create diversity from game to game.

Right now we have 3 discourses going on the forums,
1. We debate how much luck (randomness) should be involved in the game, I think there should be none.
2. We also debate on how much tactic(micro, APM) should be in game and many people advocate for having less of it, I don't agree, tactic if is not heavy is great addition to the game.

3. And we also have debate on how we differentiate and reward the guy who invents strategies from the guy who just copies and memorizes them (and that's most likely inevitable).
I wish we had good solid solution for point 3, this would make Planetary Annihilation absolutely a best RTS ever. I personally have no idea, but I believe current level of randomization isn't good solution.

Right now the game is very interesting, because the constant balance changes give a lot of people plenty of room for creativity. Ultimately, RTS games reach a stable balance and the community learns a collection of strategies that can be applied. There is still a different kind of creativity there, but much less than we get to see right now. I hope for PA to become a game where there is no limit to the ways different known strategies can be combined.

Random maps go a long way to helping extend the length of the creativity phase in the life of this game. Games like Starcraft end up requiring the player to memorize a precise gameplan for each map. SupCom is better, but there are still certain strategies that you see used on each map. Random maps mean that you can memorize strategies that are good for each part of the map, but you have to try new combinations of them each time.

I completely disagree. The moment a game is solved in the way you describe, it ceases to be a strategy game.

Strategy games are about making strategic decisions. As opposed to rote memorization, mechanical skill, aesthetic experience, plot, character development, etc. There are lots of things games can be "about" and strategy games are about decision-making in the context of gaining advantage. In order to be a deep strategy game, those decisions need to be difficult, and the board state needs to be complex.

I consider Go to be a much better strategy game than Chess because chess is becoming increasingly close to being functionally solved. And there is no decision-making in a solved game whatsoever, only execution. Which, due to the low mechanical difficulty of physically moving pieces in chess, essentially means rote memorization is the fundamental skill in chess instead of being able to assess a situation and create a plan to maximize your advantage.


Strategy games are NOT about remembering solutions to problems. They are about FIGURING OUT the optimal next move. They are about creating plans on the fly, and about constantly reassessing and improvising plans within plans. Previous experience obviously helps inform your thought process about what is important in the current situation. But there must be lots of factors to weigh in each choice, and lots of significant choices that affect the game state that players make over the course of the game.

Memorizing optimal moves is not strategy. And following a recipe to beat a solved game doth not Sun Tzu make.

But also random maps cause random imbalance in favour of certain player

Figuring out the next optimal move is one time thing. Once you figure it out then you figure next ones. After 15000 games played you have decent idea of most of optimal moves. That's when memorization comes(it comes actually a lot earlier).

These kind of discussions always end up with me wondering what is the purpose of life...

Those are extremely important discussions that adress the most core mechanics of games and it's great that we have discuss about this and I hope uberent joins or at least has the solutions for them because solving or not solving such issues results in good or bad end product. I believe Blizzard didn't do too good job. Uberent has serious chance to make better product than sc2.

Strongly agree.

Random maps create diversity and force players to adapt.

Example. I watched a SupCom video of an 4v4 match on Seton's Clutch. Every player knew at the start what their role was and what they were going to be building, before they even saw the map. They knew where their opponent was going to be be and what they were going doing before the map loaded.

That is not the pinnacle of strategy.

The greatest strategic minds are all about adapting and creating new forms of combat. That's why the colonials had a difficult time fighting the native Americans – they wouldn't adapt their style of combat. That's why Napoleon nearly became the emperor of the world – he adapted and created a new strategy that would beat his opponents. That's why the age of the castle ended because new technology changed warfare – those that didn't adapt, fell. Etc etc etc.

Memorizing a certain set of "moves" is stale, boring, and easily defeated by someone who is willing to adapt.

Good strategists, be they generals, foot soldiers, or sports coaches, adapt. They don't use the same move repeatedly. They change it up. They create new strategies. They adapt on the fly if their original strategy wasn't working in execution.

While the randomness of planets creates a danger of having an unbalanced map (which is why there's an argument for symmetrical planets, though I think Uber has done an incredible job with planet generation), the randomness of maps is what separates the good strategy players from the people who read about a good strategy online and copy the steps.

If your game is complex enough to require 15000 plays to fully understand it, then you've done a damn good job.

• Learn the rules so you know the limits
• Practice until the basic choices are second nature, allowing you to concentrate on the complex
• Learn the basic strategies and how to implement them
• Learn the traditional counters and look for innovative alternatives
• Assume your opponents have studied your tactics, introduce new things sporadically to mix things up
• Study your opponent. Try to determine what strategies he will use, how he will react to your actions, and how he thinks you will react to his actions

In the context of games with variable starting positions (eg PA with randomised maps), memorised strategies (and your ability to implement them) will still be important, just not as important as in a fixed starting position game like chess. Recognising when to favour strategy x over strategy y in situation z on-the-fly, that will be more important for variable starting position games.

TL;DR: Learn to adapt what you know to the situation in front of you. Better than the other guy.

Random maps need to be really really good for this to work out. We're still far away from that point.
On a site note I played both games, SC2 and SupCom:FA on a halfway decent level. To get decent in SC2 I had to learn 1 build for each matchup. So 1 vs Z, 1 vs T and 1 vs P. Maps were mostly not that important. In the later phases of me playing SC2 I maybe had like one or two extra builds for "extreme" maps, but generally builds in SC2 were matchup centric, not map centric. That resulted in less builds to learn. In SupCom:FA I had a build for each of the 16 ranked maps. Those builds were vastly different from each other and included things like "reclaim rock at location X at minute Y". So really really map specific.

But what is the best way to figure out the optimal next move? To my understanding the best way is to already know it or at least know the optimal move to a somewhat similar scenario, so your reasoning about what to do has not to start at 0. So that means you need to memorize a lot of stuff and that's it.

Adapting is basing your actions on knowledge of whats best for you. The more game is developed, the more adaptations revolve around remembering and not inventing. A year, two, 5 maybe, yeah inventing has upper hand but then you invent most of things and then what?

I've played 15k games in sc2 and I am nowhere near close to claim i fully understand it. What I claim though is that most of my knowledge (but also other pro players) revolves around memorizing things. You have to memorize before you invent, because if you dont know basics how are u supposed to invent. And basics in sc2 are actually so hard that like 10 players (koreans) actually invent stuff.
Right now in PA you don't have to memorize much because its new game. But in future you will have to memorize even more because of how compex the game is. Unless we keep it randomized as it is now but then we will have obvious imbalances like we do have now

Here is example of imbalance I'm talking about
You and your opponent chose exactly the same early game build order. You spawn near your opponent but you sent your initial dox in wrong direction and you expand around your main base. Your opponent on the other hand sent his 1st dox in your direction and he sniped your initial 2 engineers. You are very behind and coming back from such situation is almost impossible if you play against someone really competent (let's say Godde for example)
Also Mex positioning etc.

No, it is not. Figuring out the next move for this current situation is a one-time thing. But this one situation may never have come up before, and may never come up again.

Casual players of Chess will marvel at how many possible boards there are because of the number of pieces and the number of squares on the board. Looking at the board is a strategic exercise for a casual chess player, because the player has to assess the current situation and form a plan. General principles and information from past experience will definitely influence the player's planning.

For a grandmaster, only the very late game is a strategic exercise. The early and midgame of chess have become completely mapped out, and for a sufficiently sophisticated chess player, the early game is nothing more than rote memorization. Novel positions are (at least to current chess theory) impossible, or known to be derivative of inferior positions and therefore bad.


This is why the discussion of depth vs complexity in strategy games is so important, even if it isn't for other games. Deep games have a huge variety of possible moves and possible game states. But in order to create a huge number of possibilities, you have to make the game more complex by adding more pieces, more types of pieces, more mechanics, and so on.

Games that are too simple are too easy to solve. Tic-Tac-Toe is too easy to solve, and therefore completely not interesting from a strategic perspective. Also, games that are too shallow are just boring (in strategic terms). If there's no real thought process necessary to figure out the best move, even if the game is not solved it isn't strategic.

But games that are too complex impose a high burden on the player in order to play them effectively. Games like Dominions 4, with hundreds of types of units, hundreds of items, thousands of spells, are very deep. But extremely complex games achieve it by throwing an encyclopedia at the player. In these types of games it takes a very, very long time to learn all the information necessary just to even begin to start making strategic decisions. The amount of information needed in order to adequately assess your tremendous list of available options, and the amount of information on the board you must process in order to apply that knowledge to the current situation, is huge. For a turn based (even play-by-email) game like Dominions, time is not necessarily a problem. And their hardcore audience is prepared to read (literally) hundreds of pages of spell effects and creature summons that they can choose between.

PA is a real time strategy game. And because the game is real time, complexity becomes more 'expensive' because time and player attention is so valuable. That means the game must be simpler, and cannot afford to go full grognard. That makes it harder to have a huge number of possible game states and player options.

Therefore, in order to get a very deep game, you have to be cleverer. Tight, focused design that leverages simple mechanics to create a huge space of possible board states and player choice is necessary. The best, most enduring strategy games, such as Chess and Go, are definitive proof that it is possible to make an incredibly deep game using very simple pieces. You just have to have the right rules to create very diverse game states and player choices, without defining too many rules so the game becomes overcomplicated.

I don't like it when people say that chess is close to "functionally solved" even though the game is perfectly fine and exciting for 99.99% of the players and there are still close to 50% decisive games even at the grandmaster level and nobody has remembered all opening theory except for some prodigies that have studied the game since they were seven.

In any case, typically a strategy game requires enough depth to the game so that there is always a counter strategy to be discovered for the dominant strategy. In chess this is obviously the case because you can always play different openings and enter new territory, in dota/lol this is the case because there are so many different hero/item set-ups, but for an rts with few units it's tricky.

I consider what Cola_Colin described to be tactics. Soldier are taught tactics. This is how you breach a door properly. This is how you clear a room properly. This is how you patrol a dangerous street properly. Etc etc. These are all things that soldiers drill over and over and memorize, because for the most part, the optimal solution has been learned through past experience by other soldiers. They drill it over and over so that it becomes second nature, there is no thought, only action, all thought was performed beforehand during training.

Generals are not trained the same way. They examine the strategies and tactics employed by past generals and look for common faults and clever maneuvers that were used. When the time comes to implement their own strategy, it is not second nature that they perform without thought, the situation is carefully examined and all possibilities are carefully scrutinized. In the real world, wars are not fought often enough for strategies to be memorized and perfected. Every single war, new technology is present, and the theater of war can sometimes vary drastically from previous ones. Your strategy must be implemented on the fly, not copied from the pages of history.

I would argue that PA requires both tactics and strategy. There are some things you must memorize and perform the same every time, but there is much room for strategic improvisation too.

This is what random map generation tries to solve.. a completely new map for every match. Of course they are very similar in practice, but that is the idea.

In short I think developing a strategy is about adapting to a dynamic system by being creative. What is being creative? sigh.. lol.

I don't think that will happen in PA or any other RTS game. We well see similar situations all the time. We do now and we will see them even more in the future once the game has stabilized balance.