are RTS games popular in Korea sufficiently complex to benefit from computer assisted planning?

this thread should be taken as a continuation of sorts of my previous thread on the topic of making improved UI tools for gamers. Except, here the emphasis is not on making a bot to handle the clickfest but rather on making better analytical tools for managing an army.

Suppose we make a tool that will look at the game world through the same “window” that the player users, without trying to move the window to explore other parts of the map. As it looks, it will use some basic image recognition to pick up on things like

  • training of a unit has begun or ended
  • amount of resources in stockpile has changed
  • there are so many units in this area, by type
  • battle led to death of X of my units, Y of enemy units and Z of enemy units ran away
  • health of my units is so-and-so, by unit (especially if this is easily continually visible on screen)

These various tidbits of info would then be integrated in a user interface separate from the game itself and presented to the user, if he wishes to examine it. So the user would be able to look at graphs of economic production and unit training, evaluate the size and health of his various forces, better keep track of losses inflicted on everybody in the past battles and so on. So the player would be able to spend less time depending on unreliable short-term memory and instead make decisions based on some hard, if uncertain, numbers.

Perhaps quite significantly, the element of time, as measured in seconds or minutes, will be explicitly (rather than as vague hunches or faint memories of when he may or may not have looked at the clock last time) introduced into the the player’s mental calculation process.

Are the build-and-destroy games that they play in Korea sufficiently complex for this approach? Have people tried doing things like that?

On the other hand, could it be that the amount of micromanaging that is demanded by these popular games such that the player simply has no time to stop, look at the map, think about strategy and possibly consult the stats?

Speaking from experience with Starcraft 2, most of that information is either available in the in-game interface, the post-game scores and graphs, and the replay of the game. Also, some of that information (such as enemy units killed) are something that you should be paying attention to yourself, and displaying that information would undermine the fog of war (which is why they’re displayed after the game).

achtung! Wi 'rsind playin for ze vin hier. Ein Kriegspiel, viele many tools, komplett pwnage!

I know what stuff is and is not displayed, and what the player is “supposed” to do. The question is not what is supposed to happen, but rather what can be done to win more frequently.

If this leads to widespread adoption and a minor revolution in RTS UI, so much the better. The game will become deeper and more interesting.

Starcraft is played at such a high speed that there probably is not a huge demand for significantly more information being displayed during the actual games. Is there a demand for better post-game analysis? Yes, and the community has already created free software that can parse .MPQ files of game replays.

Now, if you want to give the player more in-game information, I would imagine that is forbidden by Blizzard, and people using such software would have their accounts suspended. For competitive gamers I doubt they’d want to take that risk. Blizzard additionally has admitted to installing software on end user machines that analyze processes running in memory, meaning they can and do ban people for using software that runs entirely outside of their games if they see it running in memory (this is how some people get banned from their WoW MMORPG when Blizzard’s software sees certain malicious botting programs running in memory.)

I imagine it would help, but wouldn’t it also be borderline cheating?

Since most RTS games utilize (to some degree) a rock-paper-scissors system of unit counters, something as simple as “You need more Unit X to counter the enemy’s 42 Unit Ys” could be helpful. Or even “You could use 5 more extractors at resource Z”. Other examples: “Last seen” unit indicators on your map that show where your scouts found enemies even after they leave the area; spending advice (based on seen/projected enemy unit composition versus your current resources and resources available to you at your bases); a color-gradient for your minimap based on when the areas were last scouted; blah blah blah… and then if you have all that parsing available to you, there’s nothing stopping you from using the same tech to download all of an opponent’s replays and automatically analyzing them prior to fighting them (Based on a statistical analysis of your enemy’s previous build orders and unit compositions on this map, you should probably build X and then Y and send them to location Z).

Straying into IMHO territory for a sec, I think many of these games are DELIBERATELY tactical/micro-management oriented. Every decision you simplify/abstract for the user is one less thing they can get better at. These games aren’t meant to be as efficient as possible or you’d just throw in a bunch of AI or customizable rules to automate the routine processes (kinda like the Stronghold series of castle RTSes with things like auto-buy/sell thresholds at the markets).

Besides, you don’t have to look at hypothetical Korean games; games like World of Warcraft already have lots of UI mods and their fairness (or lack thereof) is the source of endless debate.

Martin Hyde, thanks for the Starcraft pointer. So tempo is too high for strategy in your view.

Are there any other popular games there that are slower? Or is Starcraft basically what everybody plays?

The question of what Blizzard can or cannot detect seems a bit OT here. They can’t catch everything. They cannot stop me from pointing a video camera onto the screen and processing the data on another machine. Or something along those lines. I also think that such approaches are a lot less of a departure from the spirit of fair game than any existing bot or hack. After all, if I were a genius, I could remember all this stuff without software, just in my mind (obviously that still wouldn’t necessarily make me the greatest player or anything). And if I cannot do that, why not come up with some software tools to help me. Everybody else can follow my advice and use the “der Kriegspielgeneralstabinstrument” too.

what are those “World of Warcraft” games? I always thought that WoW is role playing and not RTS.

How did people make those UI mods? Did they hack into the Blizzard app? Or did Blizzard release the server communication protocol and the programmers built their own implementation of Warcraft world rules?

another interesting (and more obviously cheating) implication might be better communication between allies. Right now I am not sure if you can even see through allies’ eyes in all cases. You definitely don’t get reports about what is happening in allied economy and so on, and you are too busy with your stuff anyway.

But using this sort of monitoring tool you could automatically continually transmit reports about your forces and about enemy forces you encountered and fought to the allies, not just tell them stuff in chat.

World of Warcraft IS a role-playing game, but it’s still incredibly complex and still benefits from UI aids. My point is that more information could certainly help, but it’s debatable whether it makes the game more enjoyable or just easier (the two are different, sometimes even opposed, concepts). It certainly makes them unfair to those without the aids.

Different UI mods use different technologies for the parsing, and I won’t get into that because it’s (like you said) off-topic.

Shared vision, resource sharing, unit gifting, shared control, and allied warning messages (though typically not about their economy) have all been implemented in various RTS games. I’m sure more could help, but is it a good idea for gameplay?

No, it’s absolutely not done on the server communication level. The WoW user interface has an API which is exposed to third party developers. UI mods run client side and thus can only really handle presentation and processing of information that is sent to the client. They do not interface with the actual communication protocols that send/receive packets from client to game server.

In WoW many in game events are reflected in a text based log which is available to the player, due to the fact the log spams at sometimes dozens or hundreds of events per second, players will not typically read it while playing. Automated tools can read the log and process data, events, and et cetera from it, though. This is entirely allowed by the game’s ToS, and Blizzard has even added true event flags into the game that newer addons can read and process (instead of having to derive events based on certain text being seen in the log itself.)

Blizzard has an API that can be used in Lua to add widgets to their interface, they range of the simple (iterate through every item in your inventory and sell all the junk) to a great deal more complex (determine optimal gear for your class, calculate threat, determine things like average mob damage or drop rates). All mods are, in theory, officially accepted (though not endorsed) by Blizzard by proxy of “if we don’t like something it’s not much trouble to break it for you by removing access to a method.” So you can poll certain things in the server, but not others. For instance, I don’t think you can generally inspect the gear of an enemy player, even with mods. Things that have “too much automation” are discouraged, and there have been cases where Blizzard has simply said “stop using it or else until we can fix it.” But those cases are rather extreme and rare.

And if you had the innate ability to move your fingers really fast, you could use your high APM to micro really well, so why not just plug in a bot to press the hotkeys? These arguments rarely have to do with saying that a human is capable of such a feat, so therefore we should automate it. Every feat in Starcraft is within human capability, Blizzard draws the line because they feel that’s where the correct balance between skill and automation is. (I’m not trying to put down your thought experiment, just stating what I think is the general view on this subject, I’d love to see a lot of these things in an RTS some day).

That’s not ENTIRELY accurate, not all data processed by mods is sent directly to the player. There exists data that is never sent to the player if not explicitly asked for (such as a complete list of all quests completed), but nonetheless can be queried by various API methods for mod consumption. But generally, yes, you’re right.

how about if we consider Starcraft on lower skills level. With millions of people playing, maybe there is a large group of people with rating less than X who play a lot slower and worse than the competitive guys. Could these n00bs be a receptive audience for the hypothetical analysis tools I have described?

Or is Starcraft just inherently fast even for weak players?

I think your ideas would definitely have a place in sandbox training maps. Since you can access most of this data via the world editor (and the stuff you can’t is a few algorithms away), I can’t imagine it would be too much of a hassle to make a simple UI extension for a given map that displays all the data. It would be a useful beginners tool to get them thinking about stuff. Even pro players use “gym” maps, where they hone their skills. While it may not be in the spirit of fair play to make a third party app that helps during competitive ladders, I can DEFINITELY see an interested party if you integrate it into some practice maps that people can learn on (and you can alter Blizzard’s own maps too, so you could even add the training interface to ladder maps, they just wouldn’t be on when someone went to play in league).

I would recommend to the OP that he watch some Starcraft 2 “let’s play” videos on Youtube.

Check out for some very entertaining coverage of professional matches.

You will see that in Replay mode, much of the information you mentioned such as units lost, income, etc is made available, with charts comparing the data for each player.

A UI mod to make this info available during live gameplay would be a fairly serious cheat and frowned upon in multiplayer play.

ok, so awldune seems to disagree with Martin Hyde - the former thinks that there is real benefit in providing the stats info in-game (making it serious cheating) while the latter says, meh, the game is too fast for that to make a difference. I have no experience specifically with Starcraft (so I am reasoning based on AoE analogy) and I also never played against human players so it’s hard for me to have a meaningful opinion.

My WAG would be that the improvement from what I am proposing would be more of an “aftermarket modification to pimp out my car to make me extra happy driver” than “tanks rolling through the Ardennes, enemy pwned” variety. And, I say, that’s no small matter - people play RTS for fun, so if doing what I am proposing really makes it more fun and intellectually stimulating to the player, then why not. If I want to feel like von Moltke even while still playing on the level of Samsonov, then why not obtain the system that would help me do that.

Of course, the above reasoning can easily break down if the “system” required is actually pretty expensive. What I said about using video camera may be dumb (unneeded signal transformation equals poor quality), but getting the video feed to a device pretending to be second monitor for analysis may be no easy matter given off-the-shelf equipment. I am guessing here that Starcraft will not actually provide nearly as much data as WoW does for the UI mods community.

Then again, if it turns out that I am just ignorant of the sheer power of this proposed UI mod, then hey, even better. After the first wave of pwnage everybody else will get the mod (if it’s easy enough to setup, at least) and gameplay will enter a new era, without Blizzard doing anything.

Now if the system is expensive and if it provides major benefit, then and only then, yeah, I agree that we have a moral dilemma. That would basically mean that richer kids will buy it, poorer ones will not and there will be no way for people from the old-fashioned group to verify that the opponent is “cross my heart not using it”.

incidentally, instead off “second monitor” probably the thing to do here would be to use a monitor splitter that connects to the hypothetical gadget that transforms the data stream into the movie and then feeds it to the computer used for analysis.

So the hardware costs are lining up to be quite nontrivial, even assuming that the monitor-to-video gadget is manufactured and sold for a cost similar to that of an LCD monitor.

If it proves to be expensive but really, really cool maybe this is something that could be offered at gamer cyber cafes. The cafe could even enforce a rule that you play only against others who use the same setup.

“Step right up, folks! Why go to the lame Kim’s cyber cafe, when down here at Lee’s cyber cafe we have got the wonderful new way of playing Starcraft?! Come and learn all about it, and bring all your friends for the ride.”

Having such stat information in-game would likely have some utility at a minimum, but might reduce the level of enjoyment for some players. The thing is, being able to parse the raw data quickly is, at least for some people, part of the skill and the fun of the game. If you remove this aspect of the game by aggregating the raw data into an easily-readable format in the UI, you remove this skill as requirement to play the game effectively. Which may make it more fun for some people, but may make it less fun for others.

For example, one of the Starcraft 2 patches introduced a visible game clock; it would display the number of minutes and seconds that had elapsed since the game began. “Hardcore” players complained that this diminished the amount of game skill required, as it made it easier to defend against certain types of timed attacks (e.g., if the attacking player followed certain build paths, the defender could reasonably estimate when the first attack would land, reducing the need for scouting one’s opponent slightly). It reduced the fun for these players by removing the skill requirement that players mentally keep track of time in the game. Of course, the same effect could be achieved by putting a stopwatch next to your computer, but I digress.

You can certainly build tools to parse and process data efficiently, assist in decision-making, and so forth, but then you risk alienating the people who play the game because they actually enjoy doing all these things manually; that is the fun for these players. So then, to ensure that the game is still fun for this group of players, you need to develop a mechanism so that only similarly-tooled players complete against one another.

Another aspect of this is that tooling-up then becomes an arms race as people strive to build better and better tools. Again, this is fun for some, but diminishes the fun for others.

I’m still not entirely sure I understand what you are proposing. If your tool only gathers information that is available onscreen to the player, then there is not a huge utility there.

If your tool provides info that is not available (opponent’s income, opponent’s army composition, etc) then that is a significant advantage.

When you talk about a camera looking at the monitor I’m not sure I understand that. It should be possible to snoop the video directly without involving a camera.

he’s trying to circumvent the rules, like getting a computer advisor while playing chess, and at the same time limiting that advantage, only allowing the advisor to present possible moves instead of choosing one for him. as far as i can tell.

awldune, the tool should monitor only what is visible to the player through the viewport. It’s not about cheating, it’s about making you smarter when dealing with the info you are supposed to have anyway. So no, it would not be a massive improvement, like I pointed out above. But it could (hypothetically) also be an enjoyable improvement on the user interface and overall experience. Like I said above, “feel like von Moltke, keep playing like Samsonov”.

ETA re shijinn : I think that chess is a bad analogy because computers rule chess but suck at RTS. We could not make a good bot player for RTS even if we tried.