The "I liked the combat better when it sucked" phenomenon

[Lyric Suite]

Real time is more complex than turn based because you have less time to make decisions.

 

[Damned Registrations]

Uh, no, starcraft is more complex because you make more decision during the game. If you played chess at the pace you play starcraft at it'd be a speedchess game with a 1 minute clock (and the computer would ANNIHILATE you.) You make more decisions in the first 5 minutes of starcraft than in an entire chess game.

Complexity in a game is about decisions, not how many years old it is.

 

[Johannes]

Starcraft has more complex rules than chess, definitely. But that doesn't automatically imply Starcraft is strategically more complex than Chess.

Amount of decisions per game or per minute doesn't really tell us anything - does Chess become a more complex game depending on the timer you're playing on? Is it more complex to do 10 moves in 10 minutes, or ponder a single move for 10 minutes? If you don't want to get stuck with Chess examples, think about say, Go or Poker. You do a pretty low amount of absolute decisions in those games too, but humans beat computers regardless.

Otoh the most complex game rule-wise can easily be very banal on the strategical level.

 

[Damned Registrations]

It's not about the decisions you make, but could make. There's a limit obviously; tic tac toe on an infinite board offers infinite possible decisions but they're functionally meaningless beyond their relation to eachother. Essentially, you can discard any decisions you know are suboptimal. Taking all your workers in starcraft and running them into a corner and waiting to die isn't a real decision for example. But neither are most of the moves one can make in chess.

But the thing is, we've already established what all the suboptimal moves in chess are, or at least enough of them to ensure a computer always wins. We can't establish those moves in starcraft because moves that appear suboptimal, like moving a worker to a random point on the map and building structures, might actually be the winning strategy under certain circumstances. Chess is far far more transparent in this regard, largely because it's a game of perfect information- both players can see the entire board. Without that information, the decisions become far more complex, because strategies like bluffing arise. There are no bluffs in chess, no deceptions. No metagame. It's just a more convoluted version of tic tac toe. If we were smarter, it'd be equally as dull, because we'd all know the optimal moves and foresee the end of the game before it even started. You can't say the same for starcraft, because without seeing the whole board, an element of chance is present. Assymetrical games are even more complex; something like DotA may as well be Calvinball; trying to teach a computer all the interactions of the rules would be tantamount to constructing a wonder of the world. The drafting phase alone is absurdly complex. If we ever have computers that can beat us at a game like that, it'll be because they truly have a superior intelligence, and they taught themselves.

 

[Abelian]

But the thing is, we've already established what all the suboptimal moves in chess are, or at least enough of them to ensure a computer always wins.

The problem with the chess analogy is that it's not really an example of artificial intelligence, but rather of humans hard-coding certain behavior and rules they have learned over the centuries, and then use the computer's number crunching ability to calculate millions of potential moves per second. I'm not contradicting your post, since you said "we've already established" (that is, humans) and computers currently are much better at playing chess; I'm just adding some background information.

The suboptimal moves were a hard thing to identify in computer chess, since the computer doesn't have the human ability to look at a chessboard and immediately prune off silly moves (pruning is used, but it had to be programmed in). True, a computer will always notice all immediate attacks and will always notice easy captures, but there were some problems in teaching computers to take the initiative when the human opponent played defensively, or making strategic sacrifices.

Basically, computers were excellent at tactics, but it was difficult to program long-term strategy, since each additional move increased the possible number of positions exponentially. Plus, there is always the risk of the horizon effect, where the computer made a move that it deemed sound based on the number of analyzed future moves (plies), but it turns out that the situation would turn sour a couple of moves after the initial projection.

When Deep Blue defeated Gary Kasparov in 1997, it had to use a pre-programmed opening repertoire. In fact, in the 90's it was suggested players use unusual, suboptimal openings in order to make the computer ignore its openings database. Of course, some programs use machine learning to actually learn new techniques/behaviors, such as Blonide24, which played checkers with human players and learned to improve its game from them.

 

[RandomAccount]

B...b...b...but... chess doesn't have Attacks of Opportunity... it must be shit

Of course it has. That is what the en passant move is for.

That only effects Pawns, no other piece, and only effects a pawn that makes their starting move, and that the starting move must be their two squares move, and that this move must pass an opposing pawn and no other piece. The opposing pawn then has an option to fully take the pawn, as in kill it, not just take a swipe at it. This is optional, it is not a forced occurrence. The rule is more about the adding of the new rule of moving a pawn two squares without hindering existing rules than adding a new AoO rule.

 

[Cadmus]

The op is stupid. Fallout combat doesnt suck at all. This is sone kind of codexian counter culture edginess thats popular here for some reason. Every combat ever sucks according to the academja here, be it tb, rtwp or action combat. dos has an amazing combat system yet i still read retarded one liners saying its shit. Fuck off already. I the same with mmx, the combat was great, sometimes even difficult. Being different from the old mms is a whole another issue. Aod combat is wasted on the 1 man party, in my opinion, but it looked good the last time i played it. For some reason the ie games get a pass here when their combat is the clunkiest shit ever. It doesnt bore me but saying its better thab dos or fallout is totally delusional.

 

[Ulminati]

D:OS combat is amazing. D:OS encounter design after you hit a certain point in the game, less so. People just have a hard time distinguishing between the two.

 

[Nuclear Explosion]

The op is stupid. Fallout combat doesnt suck at all.

I wouldn't say that Fallout's combat sucks but it's definitely not good. The death animations are awesome but the combat is too simplistic, easy, and poorly balanced.

 

[bonescraper]

Change the thread title to "The "I don't like PE combat" phenomenon".

 

[Roguey]

D:OS encounter design after you hit a certain point in the game, less so.

This is caused in part by systematic problems.

 

[Ulminati]

Change the thread title to "The "I like Infinitron better when he shuts up" phenomenon".

Fix'd

D:OS encounter design after you hit a certain point in the game, less so.

This is caused in part by systematic problems.

The 'broken' system that allows you to get creative is what makes the game great to begin with. The problem is that once the player has all the tools of their disposal, they require enemies that are a lot harder and equally abusive of the system to remain challenged. In a perfectly balanced system, this wouldn't be the case. instead of "Awesome early, r0fleasy late" of DoS, you have "Balanshitboring all the time" of PoE.

 

[Grunker]

The 'broken' system that allows you to get creative

How does D:OS' system "allow you to be creative"?

It's easily one of the most simplistic and basic systems I've played with. Not there's anything wrong with that, I liked D:OS and they clearly spent their energy elsewhere, but calling the system creative is a huge stretch.

EDIT: Or are you talking about the combat system and not the character system?

 

[Roguey]

Here we go again, comparing a completed game to a small chunk of an unfinished one.

Looking forward to D:OS's rebalancing.

 

[Ulminati]

Grunker -- The combat system, obviously. The thread is called "I liked the combat", not "I liked the character generation"

 

[Johannes]

The 'broken' system that allows you to get creative is what makes the game great to begin with. The problem is that once the player has all the tools of their disposal, they require enemies that are a lot harder and equally abusive of the system to remain challenged. In a perfectly balanced system, this wouldn't be the case.

How so? If a system is such that it needs to throw hard and abusive enemies at you to remain challenging, doesn't mean it'd be any less balanced than another kind of system. That's how I prefer my balancing, too.
Disclaimer: I haven't played D:OS, so just speaking in general.

 

[Grunker]

Grunker -- The combat system, obviously.

My mistake. I liked the combat system in D:OS. I do think, however, that it got stale because it began copy-pasting its abilities and environmental interactions, and not because of encounter design. The 100th time I lid oil on fire or used the same slightly upgraded spell, I began being bored.

 

[ERYFKRAD]

D:OS encounter design after you hit a certain point in the game, less so.

This is caused in part by systematic problems.

HAVE YOU SODDING PLAYED THE GAME YET?

I mean, I suppose them patches should be released by now?

 

[Roguey]

HAVE YOU SODDING PLAYED THE GAME YET?

I mean, I suppose them patches should be released by now?

They actually have not.

Should be soon though. Maybe within the next two weeks?

 

[St. Toxic]

My personal favorite is phase based combat in the vein of Frozen Synapse, where you have to mix guessing your opponent's next move with covering as many eventualities as possible to survive and minimize casualties, but without the real-time aspect of having to micro a bunch of crap which inevitably comes down to reflexes. I don't know if it's particularly complex or not, but eh, that almost seems irrelevant to the discussion at hand.

So, for any other basic system and specifically regarding rpg's in general, I try to keep my opinions on combat limited: character skills as the deciding factor of successful actions, no pointless grind in pursuing the main quest-line, mandatory pointless grind in pursuing optional combat encounters (i.e exterminating towns should take ages), no hp-bloat for either player or enemy characters (any indication of sustained damage should really be enough to signify an actual injury, who needs to keep track of scratches?) and that's about it really. Don't care about single or party, class specific abilities, hard or soft counters, cover no cover -- I mean, of course I care, but there's no general statement I could make about a combat system based on the inclusion or exclusion of this tripe that would signify that it's "combat -- done right", because so much relies on encounter design, motivation for engaging in combat in the first place and other factors that systemic shortcomings seem almost irrelevant in comparison.

Having said that, I slogged through Arcanum with guns and avoiding combat as much as possible and what got me to take a 1 year break from finishing it probably wasn't the combat as much as me having a hard time remembering what I was supposed to do and why I should care; think I got lost in all the world-spanning side-quests which seemed so much more important than pursuing some elven prophecy. Soulbringer is yet another classic, where building combo strings around the weaknesses of enemies and watching your character flail around added a layer of pointless complexity but made me enjoy the game a lot more.

 

[Cowboy Moment]

Damned Registrations, you're wrong. You don't make more meaningful strategic decisions in a Starcraft game than in a chess match. Starcraft has much more complicated rules, and therefore a complicated metagame, where specific knowledge becomes extremely important - it's the same for MOBAs. I'm pretty sure that if the same amount of R&D went into developing Starcraft AI as it did for chess, we'd already have unbeatable ones, especially given how important mechanics are in RTS games, and machines have a huge advantage there. There are already AIs capable of perfect micro with certain units, developed by hobbyists just for fun. If anyone were ever to invest serious money into coding the combined knowledge of 10 years of competitive Starcraft into an AI (not unlike it was done for chess), you'd have your eternal champion. Even training AIs specialized at executing a single build play pretty decently nowadays.

Actual strategic decisions requiring real thought are very rare in a typical Starcraft game - it's much more about execution and multitasking. Sure, there's incomplete information, and bluffs and hidden all-ins/expansions are possible, but with enough multitasking it's often possible to figure things out either way, most players simply don't because they can't micro a scouting worker for an entire game while doing everything else.

 

[Damned Registrations]

But where to allocate your skill in mechanics is part of the strategy. Every moment, deciding the timing of things, where to position precisely, are all decisions to make during the game. And they matter. Scout in the wrong direction by just an inch and you might miss an important building. And as I said, the twitch factor is the only reason AI is even close to legit in starcraft anyways, and even then only at fast speeds. At low speeds (like chess) a human could position units perfectly as well to avoid psi storms or whatever, and then you're left with overall strategy, and the machine has no chance. Besides, 'most players' don't matter. What matters are the best players, and the best starcraft players CAN afford to by constantly scouting, upgrading, moving their units, etc. It becomes much more a matter of tactics and strategy, of which units to build and when, where to spend resources. Same as chess, except a fuckton more complicated. If it were as simple as you're pretending, every single game between any two given races on a similar map would follow the same build orders.

Whats the excuse for losing to AI in a chess game? Forgetting how to execute a particular endgame (AKA specific knowledge)? Not noticing something obvious when you have all the time in the world and it is literally your fucking job to notice? Did it take 10 years of research to teach a computer to beat humans at Reversi? Why was Counterstrike figured out ages ago but not Starcraft if they're both just twitchfests where only execution matters? Civilization is a lot fucking older than 10 years and the AI there still can't beat humans even with MASSIVE amounts of cheating. On a giant free for all where you ought to win less than 50% of the time even with equal skill anyways. You're going to seriously argue Civ isn't strategic or that no investment was made towards it's AI?

ITT a bunch of chess fanboys rationalize why chess is the bestest game that can ever be made and will never be surpassed by anything ever, using a bunch of semantics and bullshit vague wording about what 'strategy' is. 'Hurr durr I only lose at MOBAs because they're not strategic enough! I'd totally win if it were slower, even though I walked into an ambush every 3 seconds when I played and wasted all my gold on shit items.'

 

[Awor Szurkrarz]

The op is stupid. Fallout combat doesnt suck at all.

I wouldn't say that Fallout's combat sucks but it's definitely not good.The death animations are awesome but the combat is too simplistic,easy and poorly balanced.

And too random.

 

[TigerKnee]

Civilization is a lot fucking older than 10 years and the AI there still can't beat humans even with MASSIVE amounts of cheating. On a giant free for all where you ought to win less than 50% of the time even with equal skill anyways. You're going to seriously argue Civ isn't strategic or that no investment was made towards it's AI?

I don't think each Civilization is subjected to the sort of rigorous "multiplayer-style" balancing the way a game like Starcraft would in order while developing its AI. If the AI knew that e.g Slavery + chopping forests in Vanilla Civ4 was the way to go I'm sure it would play a lot better and so on.

It reminds me that Master of Orion's 1 races was considered "balanced" in the past, which is a rather crack-filled statement and while MOO1 is still a really good game, I figure if the game devs can't judge game elements, it's hard to develop an AI when you're under the impression that say, Race X is equally powerful as Race Y and Z when actually they are shit and need specific strategies to win a game.

 

[Damned Registrations]

Well the civs tend to use specific civics according to their culture, rather than what is optimal, admittedly. But there is still a laughably huge gap in basic strategy. There's just too much shit to take into account. Trying to get it to accurately value naval tech (let alone actually making use of it) would be a huge feat alone, requiring tons of factors regarding the land available, resources available, other nations to compete with, availablility of that tech, associated wonders... trying to weight all that shit to get it to come up with a good approximation of it's value would be a nightmare. Repeat for every tech in the game, and then realize that valuing techs properly during a trade is an incredibly small corner of a huge game.