Simulation: How much is too much?

Fair enough, I guess I was only thinking of gun bullets when typing that. All my points still stand.

I had a few other topics in mind, such as with rugged terrain and steep hills and such. TotalA made a big deal of it, at least. But beach ball warfare seems to have stuck.

Tooltips are great, but simply looking at a planet should tell pretty much everything you need to know about it. That's more of a UI thing.

Sounds easy enough... but how far can an atmosphere reach? Not the orbital layer I hope, or a lot of machinery could find itself suddenly crashing down. Accounting for complex flight parameters will make aim an incredible challenge to deal with. Not every world necessarily has an atmosphere, so it's not an end-all solution to artillery.

The simulation exists to make the game interesting or fun. Emergent gameplay is interesting but it can make the game too random or harder to balance.

Some would say a game like TA already goes too far with the simulation compared to say StarCraft or most other RTS games that aren't heavily sim based.

Anything it's an interesting question.

I'm curious to know how the concerns that Total Annihilation had too much simulation came about, and if they were even around when Total Annihilation was still the only game available; and also how much of it isn't just users going on forums to complain because they can, or doesn't amount to "this game doesn't feel like other RTS games on the market".

Let's forget about TA for a second and just talk about anything that adds randomness to a game.

Pro players of any sport don't like random events because it cuts down on the edge they have over the non-pro. It's fairly easy to tweak a rule set of a game to add a bit more randomization which would then make something like a tournament have more randomness which then cuts down on the ability of a pro to get a higher win percentage.

So if someone wants a very hard core reproducible experience staying away from simulation and doing things in a more predictable manner makes the pros happy. Take a look at how predictable the average StarCraft match is for example.

Of course simulation doesn't always mean more randomness but in practice it does make things less predictable in most cases which can potentially cause luck to be a bigger factor. Notice how most tournament style games are played on very specific maps for example.

While I can certainly see why pros would not want to have randomness, I think that it's good for the pro-scene. Not because the pros will enjoy it, but the spectators will. Breaking up predictability will make watching these games more enjoyable, which means that in the long run the game will live longer (which is good for the pros)

Most of the games that people really enjoy playing have at least a decent amount of unpredictability and often at least some amount of randomness as well. I think it's one of the reasons Starcraft II isn't as much fun to watch as it could be. If you've watched a dozen games, you've seen the season.
From that point out, it's just more of the same. And so is playing. It just starts feeling like a chore really fast, and I think the predictability and lack of random events are a major part of that.

In the end, nobody cares if the pros are happy. The pros are only a small group and they cost money, they don't earn it. You want the spectators to be happy. They're the ones paying money to watch your games. (And besides, if the pros are unhappy they'll leave and whoever was slightly worse just gets promoted to the position. As long as there are players, some of them will be the best, and they can be the "pros". And as long as they're entertaining to see in action, nobody will care how good they are.)

Doesn't the unpredictability of a simulation tend to normalize with large numbers (IE statistics). Therefore, with the exception of early game strats or attacks that can 1-hit a commander, PA shouldn't suffer to the degree that starcraft would.

Long post, bear with me.

Almost all the best games, competitive or not, have random elements to them, but it has to be restricted to the scope of what still makes it fun. If your pros don't enjoy how the mechanics, including the randomness, are interacting, you end up with no pros. Too much randomness means no mastery - or at least less mastery. Less mastery means fewer people willing to spend the time to go pro, which means fewer people in general who will maintain interest in that game as it ages.

I'll use one of the simplest gameplay mechanics I can think of as an example: Bejeweled. I am, arguably, a fairly high ranking player, because on a good game of Bejeweled Blitz, I can score in the 1.1 to 1.5 million range. My average score though is closer to 500k. The reason for this is because the entire system is predicated by a random seed and therefore random behavior. So as good I as I know I am capable of being in that game, I would never consider participating in a tournament for it unless I had a set of pre-determined seeds I could practice in advance - which in this case, defeats one of the core aspects about what makes the game challenging. It why you don't see a huge "pro" gaming interest in Bejeweled Blitz because it's too random to be reliably fun to watch. I've seen some of the best players around simply get a crappy seed and get a score at 10% of their normal high score. On any given set of 10 games, you'll see the very best players cancel out of probably half of them because they recognize the patterns within 15 seconds that the seed is not a viable seed. That's no fun for the pro, and not really even fun for the viewers, hence why you don't see livestreams Bej Blitz matches (at least not that I ever heard of).

In that case, it's very much by design, and overall, the community and the game doesn't suffer from it, because they recognize that it's a casual game and thats it's bread and butter comes from casual players looking to zen out and compete with their buddies on facebook.

The best RTS's are still, at the end of the day, about mastery - of systems, maps, tactics, strategy. Some randomness is expected and even desired, but too much means consistent mastery is impossible, and therefore, not really a good game. I think you can find a balance that is less predictable than Starcraft, but still predictable enough to allow for mastery of the systems to really get people competing hard.

And don't undersell the value of pros in any given game. If they're not having fun, you don't have pros for long. No pro's, and people start to lose large scale interest in the game. I'm not saying this should drive all your game design decisions, but it's foolhardy to not take into account the appeal of watching the top 5% play has to the other 95%. I can cite examples in almost every genre of how valuable a vibrant pro circuit can be to the overall health of the game. And not many (any?) cases where it hurt. So yah, keeping pros happy is important, but also doesn't mean it runs counter to keeping all the players happy.

Well put Neutrino.

Simulation rather than 'fudges' definitely allow for more refined approaches and novel strategies. It also improves the chances of those 'wow' moments when something you do creates a totally unexpected (but logical, now you've watched it happen) moment.

Those moments are great, and find their way onto youtube. Much more satisfying than the opposite, where your artillery 'should' do this according to your experience of the simulation, but someone's 'if' statement makes your projectile turn into dust at altitude x just 'because'.

Also I think learning about nuances of a consistent, less random world gives more of a sense of mastery for the player than rote learning the if's and but's introduced arbitrarily during development.

Simulation does not equal randomness, and it's not like you can't predict simulation. All it can really do is open up different situations.

The impression I get is people don't want to use situational awareness, and want to only master a few predictable set of orders, hence my statement on people wanting a hard-numbers and dice-roll RTS game.

I do agree though that randomness can detract from a game. Like randomly imposed 'environmental' weather hazards.

So, not to sound like a smartass but.. you'd think that these advanced robots would've figured out a way to adjust the artillery propellant on the fly?
Not to say that in a low gravity, a higher firing arc with slow enough muzzle velocity does true wonders... As opposed to the traditional style of Berthas, Vulcans etc.
Maybe something more like the UEF small artillery in SupCom2?
Also like the suggestion of atmosphere, would ensure that no shots go on accidental orbits.

Now, we all chipped in for something new and awesome..
I say bring on the good sim and take best advantage of it.
Prepare your bazookas you Worms, next stop's Kerbin!
(I'm almost certain someone will create a map of that pretty little rock..)

i was not pleased to see the comments on how PA should be more like Starcraft, leave starcraft to be starcraft and don't try to copy them in any way. Reality Pump tried to bring in some starcraft stylizations into their game Earth 2160, a game that failed to live up to all of the expectations of its predecessors, Earth 2140 (2 expansions), Earth 2150 (2 standalone expansions), and WW3 Black Gold (prequel to the Earth Series). The only thing that Earth 2160 lived up to was quality of Engine which IMO no RTS engine since has replicated it in quality much less surpassed (its 8 years old).

So Please DO NOT TRY TO RECREATE ANYTHING OF STARCRAFT IN PA keep to TA and SupCom (maybe and less of it than TA).

Personally i hate Starcraft, there was no randomness in it, everything had to be the same in order to succeed, Build X at time Y have enough of it by time z, don't build Structure A which unlocks unit B because that Unit Sucks even though it looks like it can throw down. and don't build defenses because they get roflstomped, by 2 Zelots. there was no room to play the way i wanted to play, no ability to have different strategies.

There needs to be room for many different types of strategy to be successful, no single tactic should never be UP and stupid even turtle, because the more strategies that are viable the more randomness there is but the same basic ability to master each and be successful with however you want to play even if you want to play differently for a little while.

I like the idea of simulation should prove to be fun

I appreciate the point Garatnw made. In my eyes, gameplay in a proper strategic game should not depend on the amount of "randomness" in a game. It should depend on how effective actual the employment of actual strategy is. Going the Starcraft route wouldn't bother me in the slightest if it weren't for the fact that because the in game simulation is so very deterministic, it becomes part of the metagame for high level players, a major part of which is defined by micromanagement.

For a game that is less deterministic in the simulation, such as TA, metagame tactics such as micromanagement become less useful, and the focus is shifted more towards the concentration of higher levels of strategic thinking.

While this means that a few units may be needlessly lost due to a "random" lucky shot or something similar, it shouldn't matter that much to the player with the better strategic play.

This means that a game can be as random as it wants, up to the point that if the majority of random event occurrences have disadvantageous results the overall outcome of a superior strategy would be largely unaffected.

Many turn-based, grand strategy games follow this concept. One good example is the game Unity of Command, in which battles have a rather large random element, and yet a superior strategy will almost always be successful.

BTW nobody should take anything I've said here as implying I want to remove simulation from the game. To me the simulation is one of the big differentiators. I mainly was going through the thought process of why someone would want a more predictable game. StarCraft is not a game I'm terribly interested in emulating.

And no, simulation doesn't necessarily imply randomness. However, most physics sims aren't terribly stable and end up with a significant amount of randomness.

Anyway we draw lines on the simulation all the time. Anything we cook up is still terribly crude compared to the real world, especially from a physics standpoint. As some have pointed out we generally don't even take air resistance into account. The unit "physics" is also a fairly crude simulation.

One thing we are doing is drawing a more clear line between effects/animation and the simulation. For example animation will be driven by the state of the aim node but it won't necessarily use the same skeleton so we'll be able to do more complicated articulation but not pay the cost in the sim, only in the rendering.

Just to clarify my post, since both neutrino and ayceeem mentioned it, what I refer to as "randomness" means autonomous events outside of the player's control. For example, a tank shooting at an opponent in TA may miss because it's shooting a moving target at a long distance, while the tank itself is moving. This is not entirely random, and makes sense in the simulation context of the game itself. However, to a player concentrating on the big picture of hundreds of units, this missed shot might as well have been "random" as there are plenty of instances of tanks in the same situation that would hit their targets.

I don't know how emergent gameplay makes a game random. From the definition on Wikipedia emergent gameplay is just complicated situations derived from simple gameplay mechanics. Whether or not the players or the game developers understand those situations and their gameplay implications is another question.
It is harder to balance because it is something that you do not know if the players will use or something that you didn't take into account when you balanced the game.
Do you mind if I use the game in another way than you intended neutrino?

I don't think that there is a big difference between a random number feeding an event and player input causing random outcomes.
Let's compare a t2 stationary artillery sniping an ACU and a tactical missile launcher trying to snipe a dodging ACU in FA.
The artillery unit have inaccuracy and whether or not it will be able to kill the ACU before it escapes depend on if it hits or not. It is randomness that is out of the players control.
The Tactical Missile Launcher always hit spot on where you target it. If you can predict the movement of the enemy ACU you can hit it but the enemy might change direction and avoid being hit. It is basically a guessing game if you can't see the missile approaching and if the game comes down to guessing, the outcome of the game is basically random.

If a random event has small consequences or happens frequently enough it normalizes. It becomes predictable.

Predictability is the denominating factor when determining whether something is based on luck or not.
When players play on a map they don't know, they don't know the metagame and the strengths and weaknesses of different strategies. Choosing 1 strategy for another has to be done with less knowledge and a more uncertain prediction. You might just as well say that one player was lucky to chose the "better" strategy or was it skill and intuition?

Let us compare that to a proposed PA tournament.

1. Do the players know what planets/planetary system that will be in the tournament?
If players don't know how the map looks like they can't practice and prepare strategies. The players ability to read the map and predict favourable strategies are more important.
Like Bejewelled players where who understand how to make the best use of their seed as they play it have an advantage, players in PA that can understand the metagame of the map gets an advantage. It comes down to experience, skill and intuition rather than practice and memorisation.

2. Do you play with symmetrical planets/planetary systems or do you play with asymmetrical planets/planetary systems?
Players could get unlucky in Bejewelled and get a bad seed while in PA one player in a 1v1 could get unlucky and get disadvantageous terrain and/or bad resource distribution.
If all the players in Bejewelled get the same seed and if both players in 1v1 PA game play on a symmetrical map it doesn't matter if they haven't seen the map before because they still have the same starting conditions.
If a map is asymmetrical in PA it could be balanced or it might not be obvious how to use that asymmetry to your advantage.
If there hardly is any advantage to be gained from the asymmetry and players have not played on the map before then the asymmetry will most likely be a diminishing factor compared the player who can understand the map and employ the best strategy.

3. Are the planets created or are they randomly generated planets/planetary systems?
A created planet can have an intent and be modified over time to emphasize a certain gameplay or keep several strategies balanced to each other.
Similar effects can be caused by a strict planetary generation algorithm.
This can allow tournament managers to chose what type of gameplay they want to see in their tournament.

Thank you! This guy gets it.

A simulation doesn't mean that gameplay is necessarily more random. It only means that unit interaction has more elements. These features can be broad brush strokes- Bullet physics is a major one. Bullets follow a trajectory based on gravity, and they hit whatever they hit. Projectile spread adds randomness to this system.

Simulation elements can be very nuanced- for example TotalA has a wide range of pathing capabilities over rough terrain, smooth terrain, steep terrain, rocky terrain, and everything in between. Each unit's movement capabilities control how it bumbles arounds terrain, and the shape of the terrain has a powerful impact on how they behave. This can make the movement of any particular set of units seem to be random, as each army moves slightly differently across maps.

Another nuanced feature was how terrain shape effected projectiles. I think any Supcom player can appreciate just how annoying it was for Cybran turrets to have half their range. This happened due to minor changes in terrain height, a feature meant to be more aesthetic than real. The map view largely gave you no idea where the turrets could use their full range and where they couldn't, so the result was a seemingly random outcome in game.

Just remember that players use "Random" to cover a wide variety of gameplay experiences. At the high level, this is usually a discussion of legitimate random elements that were likely coded into the game. At the low to mid levels, "random" can cover any event that seems unusual, counter intuitive, or unexpected. If something weird happens and there's no obvious explanation for it, then it must be random! Whether it ends up good or bad pretty much needs to be determined case by case. Generally, a unit that screws up for no reason is bad, and a unit that succeeds at something new is good. Completely changing how you think about a unit is the coolest outcome.

You've basically summed up what I said, but having your commander killed in one shot isn't a small consequence, and neither does it happen frequently (a max of once per game in fact). It contradicts the previous conclusion, which is why I included it as a caveat. So I'm not sure what your purpose in quoting me was; would you care to clarify?

I agreed to what you said and used that as a part of my argumentation of why the randomness of accuracy and player input can have similar effect on the outcome of the game.

Like bobucles said:

Edit:If a pro player say that the outcome of the game is random it doesn't necessarily mean that it is because of randomness in the simulation or because there are so complicated situations that a human can't understand or predict those situations but it could be because the game is won by guessing.
Rock, paper, scissors is a guessing game. You could argue that you can improve your skill in that game by outsmarting your opponent or learning human psychology but it is still a guessing game.