Galactic Starfighter....suggestions to give it wider appeal, short and long term

[LordArtemis]

Made some adjustments to the original OP. May make more edits in the future.

[Ramalina]

You know, if they had a new objective map at some point I wouldn't mind having a KOTOR/KOTOR II style turret or two that you could opt into somehow. Say basically fly your ship into a special TDM style powerup, and it kills your ship and spawns you into the associated turret.

 

Risky proposition, cause the turret would have to be tough but killable, not have a perfect field of fire over the objective(s) and if everyone goes off after other stuff you might be stuck with nothing to do for most of a match, unless they were kind enough to put in an eject/suicide of some sort. Still, sometimes just sitting there and blasting away might be fun.

[QuinMantha]

Still, sometimes just sitting there and blasting away might be fun.

 

It might be about the only way to get a handle on targeting without the immediate threat of crashing. Possible addition to the tutorial?

 

'First, we learn how to aim and shoot. Next, we learn how not to crash. Now we put it all together. Aim, shoot, don't crash. Congratulations! You might actually hit something other than an asteroid in your first match now.'

[LordArtemis]

I think I can understand the difficulty in programming AI. It isn't easy in my experience (it is not my field of expertise). However, they could have and should have, in my view, provided a more rounded experience by adding more "BF2" style features.

 

Here is how it would have been done IMO.

 

First, they have interior ship models, a few simplified ones used during leveling certain storylines. This would have been a good setup for a Capital ship interior, with a connected bay. Then, they open up the other side of the bay to make entry/exit on both sides, and park generic models for player ships in the bay.

 

Now, as far as your on foot combat unit, look at BF2...you had only a few unique appearances, all low poly models based on your class choice, which you chose when you spawned. That would be your appearance on the ground, outside your cockpit.

 

They could have used the 4 legacy pilot sets they released as the base appearance for each "class", and then used the same GSF UI, but with ground abilities, simple ones. Just like in BF2 you would only have a few specials to use.

 

Now, on the ground inside the capital ship you could do the following.....

1) Launch back into the battle in another fighter.

2) If an enemy ship, you could attack NPCs and other players defending the hangar.

3) If an enemy ship, you could attack specific points inside the ship....turret control, weapons control, engines, bridge.

4) If a friendly ship, you could defend the hangar from other players.

5) If a friendly ship, you could man the turrets, or protect the weapons control, engines or bridge from other players.

 

Just like control points in a warzone, controlling points on the ship would count toward taking control of the ship entirely. Once under your factions control, it would fire on enemy players and enemy capital ships.

 

In your fighter, you could chose to engage in normal GSF objectives, which would perhaps be more difficult since you have capital ships firing at each other and you, as well as other players manning turrets on the ships, or you could choose to directly attack the guns and control points of the enemies capital ships from space.

 

So you can attack from without or within. The choice would be yours.

 

This, IMO could have been done using the current framework, and it would have made the feature MUCH more popular IMO, even with the casual crowd.

 

IMO, right now it is only half a feature, or half of what it should be. Not only that, it is too dependent on gear and not enough on skill and depends on keyboard and mouse control because the reticule window (circle) is FAR too large, so large it makes the ships a bit lethargic with respect to running and gunning/turning.

 

Just my slant.

[dancezwithnubz]

imho, the #1 thing that needs to happen to broaden appeal for GSF is cross-server queuing.

 

At a stroke, the effective player base on every server will explode, reducing queue times and allowing for some sort of decent matchmaking process to happen.

 

I've played GSF from early access. Back in the start, we all had the same three ships and we were all nubs.

 

When the f2p and preferred players joined us in the skies, it was a knock-em-down turkey shoot for a couple of weeks while the newcomers climbed their respective learning curves.

 

Today, a new pilot coming in to discover what all the fuss is about just gets blown to smithereens when thrown in with the rest of us. It's not fun for the nubs and, lets be honest, it's not the best fun for us experienced pilots either.

 

I'd also like to see a 'waiting lounge' style free for all. When you queue, you can choose to join a rolling instance of GSF where you get given a random ship, with random components at random upgrade levels and just have at each other (not a TDM, a Deathmatch. Every man for themselves!) while waiting for the queue to pop.

 

A better tutorial system is also badly needed. Preferably you should be able to take a friend into a tutorial space and help them learn to fly before throwing them into a live game.

 

but cross server GSF would fix a lot of the problems people are QQ'ing about.

[arkanone]

Delete Gunships

[Fractalsponge]

Delete Gunships

 

The tears, they are sweet.

[arkanone]

I think I can understand the difficulty in programming AI. It isn't easy in my experience (it is not my field of expertise). However, they could have and should have, in my view, provided a more rounded experience

 

I have a hard time understanding why it would be so difficult to program AI when they did it quite well in X-Wing...that was all the way back in 1993. It's due to a lack of programming talent.

 

But if we're going to play along with Bioware's excuses then at least put in some ground attack missions. That's what the battle of HOTH was...which has been recreated countless times over the years by programmers with MUCH less computing horsepower at their disposal.

[LordArtemis]

I have a hard time understanding why it would be so difficult to program AI when they did it quite well in X-Wing...that was all the way back in 1993. It's due to a lack of programming talent.

 

But if we're going to play along with Bioware's excuses then at least put in some ground attack missions. That's what the battle of HOTH was...which has been recreated countless times over the years by programmers with MUCH less computing horsepower at their disposal.

 

Well, from my limited view, listening to the folks that I work with around the office, the engine determines the ease or difficulty in adding proper AI elements. If the framework does not support easy development in that respect, than it will be a long road every time.

[Zoom_VI]

I have a hard time understanding why it would be so difficult to program AI when they did it quite well in X-Wing...that was all the way back in 1993. It's due to a lack of programming talent.

The problem with AI is not creating them. The problem with AI is that the number of operations that have to be constantly performed every second in order to make a decent AI puts a lot of load on the servers. This is why no mmo in existence has any form of AI.

[EvilTyger]

I wonder...

 

Could they shift some processing power to home users via consoles?

 

Make a console version of Galactic Starfighter, let them host games. Immediate increase of players, easy access to at least that part of the game.

 

Heck, I wouldn't mind firing up the X-Box for a shorter Que and not having to log in to the full game.

[LordArtemis]

I wonder...

 

Could they shift some processing power to home users via consoles?

 

Make a console version of Galactic Starfighter, let them host games. Immediate increase of players, easy access to at least that part of the game.

 

Heck, I wouldn't mind firing up the X-Box for a shorter Que and not having to log in to the full game.

 

I had the idea that they should have tried a PVE version, then released it as a short demo. Nothing fancy, just to get a gauge of how folks would take to it.

 

That could give them an idea if folks would even be interested in that sort of thing.

[Verain]

I wonder...

 

Could they shift some processing power to home users via consoles?

 

 

No.

 

So, to tell you why:

 

1)- Consoles have crap for processing power. They are awful tiny little boxes that accomplish precious little. These days, the only thing they manage to do is ape an old video card. The CPUs on them are anemic, pathetic, worthless. You'd be better off having phones do the work- at least there's enough phones to matter!

 

2)- If you could make the server work get shoved off, you'd already be doing that with PCs, which are more numerous and vastly superior.

 

3)- The tech push on this game is LIGHT. This is a minigame from a coding perspective. It uses all the same logic as the rest of the game, it isn't revolutionary or interesting from a technical perspective. The cool part is that they were able to reuse so much to make such a deep and interesting game. To make the game distributed would be huge. The devs would, if they were serious about this, have a MUCH easier time implementing a dedicated GSF rack, and then making it such that ALL pilots zoned into that, instead of it being handled the way it currently is. That would be a big deal of code, but it's way less than what you said above.

 

4)- Latency is the real killer, and the logic needed to run the game is small compared to other things. It wouldn't make sense to have the game logic anywhere but a server.

 

5)- You can't trust a console. The console would be game genied to give the host wild powers. The damage and kill reports would be spoofed. No reason to walk away from the existing secure model for a crap one.

 

Heck, I wouldn't mind firing up the X-Box for a shorter Que and not having to log in to the full game.

 

A version COULD be made that could launch from the launcher, but there's no way any of this code will ever run on some jalopy of a console.

 

...words...

 

Artemis is a Lady or a Goddess, not a Lord. Why on earth would you name yourself that?

[Kuciwalker]

The problem with AI is not creating them. The problem with AI is that the number of operations that have to be constantly performed every second in order to make a decent AI puts a lot of load on the servers. This is why no mmo in existence has any form of AI.

 

False. Dude, seriously, Moore's law. We had decent spaceflight AI on computers that are far less powerful than modern cell phones. Processing power is not an obstacle here.

[Asbetos]

False. Dude, seriously, Moore's law. We had decent spaceflight AI on computers that are far less powerful than modern cell phones. Processing power is not an obstacle here.

 

Yes, it is. Yes, you had "decent spacefight AI" 20 years ago, BUT:

1) how many polygons did those models have?

2) did they use client-server model that goes over the internet?

3) did those servers have a ton of other stuff to compute at (virtually) the same time? (here I mean the rest of the game, in case it's unclear)

4) Did they use an MMO engine that adds even more computing?

5) Did they use simple collision calculation and simple "this weapon does X damage and drains X weapon power" or were they calculating in Evasion vs. Accuracy, DR, bunch of talents etc. etc. etc.?

I had some more questions in my head, but they seem to have vaporized.

Edited July 9, 2014 by Asbetos

[Kuciwalker]

Yes, it is. Yes, you had "decent spacefight AI" 20 years ago, BUT:

1) how many polygons did those models have?

 

...

 

... ... ...

 

 

 

 

 

...

 

Dude, you have no idea what you are talking about. I'm not going to bother.

 

(The rest of the post was also terrible but not as shockingly awful as item #1.)

[Asbetos]

Dude, you have no idea what you are talking about. I'm not going to bother.

 

(The rest of the post was also terrible but not as shockingly awful as item #1.)

 

Right...

[LordArtemis]

Artemis is a Lady or a Goddess, not a Lord. Why on earth would you name yourself that?

 

Because.....

 

A) Artemis is a goddess....and a starship bridge simulator, a fiber-class pCell mobile experience, a genome browser, an embedded computing organization, a type of horse post, a standard 16 tooth edged gear, a video game sword, an Integrated Portfolio and Project Management company, A character in an animated book series, etc etc etc.

 

B) You are named after a town in France.

 

There are your reasons.

[dancezwithnubz]

Could they shift some processing power to home users via consoles?

 

the major problem with too much happening client side is that the opportunities for hax starts increasing.

 

a lot of game hax rely on fiddling with client-side stuff and fooling the server. i'll pass thanx.

[Verain]

False. Dude, seriously, Moore's law. We had decent spaceflight AI on computers that are far less powerful than modern cell phones. Processing power is not an obstacle here.

 

You are a little bit wrong and mostly right.

 

 

The processing consideration is actually not trivial, because if you were to use what passes for sophisticated AI, you'd have to ensure that the remote and local box are doing the same processing. While the server is king, one of the ways we get away with such a snappy experience is to do some things both locally and remotely. The other piece is that you'd have to demonstrate that the AI code could execute tightly in some time- the script based AI of NPCs in most games right now is very, very, simple. So processing does have SOME consideration...

 

A bigger part is that the gameplay in these games is based around entirely or mostly predictable pieces. Plenty of raid bosses have the power to just wipe everyone out if they properly selected their moves, and DEFINITELY if they properly selected their targets. These characters often have extremely hard hitting moves that need to be responded to appropriately, and couldn't be tuned like that.

 

The final piece of that is this: the gameplay experience would be utterly inconsistent. Plenty of classic games have AIs whose exploitation is legendary. Pacman has patterns for every level that can be memorized to play a perfect game. The patterns are emergent- each ghost makes a decision based on several variables, none of them random, and as such there's a "pattern" to beat each given set of AI parameters. Ms Pacman gets away from this by injecting random things, but a lot of the play is still based on the assumed actions of each of the four enemies. Many, MANY classic and modern games end up with an enemy who makes decisions based on hero actions, and much testing is spent trying to ensure that the enemy doesn't become predictable. So while my raid group may have a really easy time because the AI weighted that he should do X then C then J, your raid group is totally screwed because he doesn't get confused and does X then either A through F then M, which is much harder to deal with- maybe the devs didn't test the raid comp I had, or realize that he becomes predictable in very narrow circumstances.

 

 

And that's the biggest reason- solid computer villain AI is an industry standard but is absolutely absent in MMO NPCs because of consistency of difficulty and experience.

[Altheran]

Because.....

 

B) You are named after a town in France.

 

There are your reasons.

 

Wat ?

[Kuciwalker]

You are a little bit wrong and mostly right.

 

The processing consideration is actually not trivial, because if you were to use what passes for sophisticated AI, you'd have to ensure that the remote and local box are doing the same processing. While the server is king, one of the ways we get away with such a snappy experience is to do some things both locally and remotely. The other piece is that you'd have to demonstrate that the AI code could execute tightly in some time- the script based AI of NPCs in most games right now is very, very, simple. So processing does have SOME consideration...

 

Again, 15 years ago I played games that never demonstrated the slightest bit of AI-related lag. I promise you this is just not an issue.

 

Hell, Freelancer even handled the synchronization issue just fine for online open-world MP.

 

A bigger part is that the gameplay in these games is based around entirely or mostly predictable pieces. Plenty of raid bosses have the power to just wipe everyone out if they properly selected their moves, and DEFINITELY if they properly selected their targets. These characters often have extremely hard hitting moves that need to be responded to appropriately, and couldn't be tuned like that.

 

The final piece of that is this: the gameplay experience would be utterly inconsistent. Plenty of classic games have AIs whose exploitation is legendary. Pacman has patterns for every level that can be memorized to play a perfect game. The patterns are emergent- each ghost makes a decision based on several variables, none of them random, and as such there's a "pattern" to beat each given set of AI parameters. Ms Pacman gets away from this by injecting random things, but a lot of the play is still based on the assumed actions of each of the four enemies. Many, MANY classic and modern games end up with an enemy who makes decisions based on hero actions, and much testing is spent trying to ensure that the enemy doesn't become predictable. So while my raid group may have a really easy time because the AI weighted that he should do X then C then J, your raid group is totally screwed because he doesn't get confused and does X then either A through F then M, which is much harder to deal with- maybe the devs didn't test the raid comp I had, or realize that he becomes predictable in very narrow circumstances.

 

And that's the biggest reason- solid computer villain AI is an industry standard but is absolutely absent in MMO NPCs because of consistency of difficulty and experience.

 

Yes, it would not be "MMO-y" and it would be hard to get rid of "AI exploits" - but single-player games deal with this issue all of the time too, and it's just not that big of a deal.

Edited July 9, 2014 by Kuciwalker

[Verain]

Because.....

A) Artemis is a goddess....and a starship bridge simulator, a fiber-class pCell mobile experience, a genome browser, an embedded computing organization, a type of horse post, a standard 16 tooth edged gear, a video game sword, an Integrated Portfolio and Project Management company, A character in an animated book series, etc etc etc.

 

The starship bridge simulator is named after http://en.wikipedia.org/wiki/Artemis

Why? An archer goddess is used as an analogy for the bridge- someone "calling the shots".

It is very definitely named after the goddess.

 

The "mobile experience" (Artemis Networks) is named after http://en.wikipedia.org/wiki/Artemis

A classical name is pretty common in tech and military.

 

The genome browser is named after http://en.wikipedia.org/wiki/Artemis

It is presumably named after her for being a huntress.

 

The embedded experience http://www.artemis-ju.eu/home_page actually has a picture of Artemis, you know, this one:

http://en.wikipedia.org/wiki/Artemis

On their logo.

 

I can't find the horse post. I can find horses named after many gods and goddesses. Lemme assure you, whatever it is that you think this this, it's named after http://en.wikipedia.org/wiki/Artemis

 

The "hunting gear" thing is named after http://en.wikipedia.org/wiki/Artemis because she's a goddess of hunting.

 

http://en.wikipedia.org/wiki/Artemis_(software) is a "a sister product to Apollo". Any of these things are all named after the goddess Artemis, normally for being an archer, a huntress, or any of the other things associated with her.

 

Any character named Artemis in any media is named after http://en.wikipedia.org/wiki/Artemis

 

 

B) You are named after a town in France.

 

My name has nothing to do with your name. However:

 

1)- Verain is one of the archaic spellings of some plants of the Verbena family.

2)- If you google image search Verain, you'll see a flower first off, and a bit later you'll see a WoW character of mine.

 

The town in France ("Saint-Veran") is unrelated and from a different source ( http://en.wikipedia.org/wiki/Saint-V%C3%A9ran ) - it is named after a literal saint who apparently beat up some dragon. http://en.wikipedia.org/wiki/Saint_V%C3%A9ran

 

 

There are your reasons.

 

So you named yourself after a goddess and messed up your title, and when questioned the best you can do is a bad ad hominem. Got it.

[LordArtemis]

So you named yourself after a goddess and messed up your title, and when questioned the best you can do is a bad ad hominem. Got it.

 

I think its fairly obvious you actually didn't get it, but I will make one more attempt.

 

I did not likely choose this name based on a goddess considering the word "lord", which is likely male (naturally not in every case, but the most common use). The reasons I chose that name are my own, but there is no connection with the goddess you reference. It is a mispronunciation of a name of one of our commanders in Iraq that ended up as a nickname, nothing more. Not everything that has a name similar to something or someone from the past is an intended connection.

 

Much like you likely did not choose your name because it is a town in France....I would say your reaction and further explanation presents more evidence to that likelihood..

 

I am fairly certain, as well, you can not indicate with any kind of accuracy the reasons that folks choose a name. I am also fairly certain that they can and would do that themselves. You can certainly infer that there is a likelihood or not that they have done so, and evidence may exist that would make that more or less likely, but posting the same link over and over does not make conjecture fact, nor does it give any weight to this silly discussion we are having at the moment.

 

My guess would be you assumed I was not aware Artemis is a goddess, that noone had brought this up in the past in this forum....and perhaps you meant this as a feeble attempt at some kind of slight. If that is the case you would be wrong on both counts.

 

So I presented a good point (IMO naturally) that you completely and obviously missed...or perhaps deliberately did so to be combative. I could not speculate as to your motives with any accuracy, you would have to speak to that.

 

It was supposed to be a lighthearted way of pointing out to you that not everyone chooses a name because it has some kind of obvious connection to someone from the past.

 

If you still fail to understand this there is nothing I can do to assist you with this further. So we will have to leave it at that.

 

 

BTW, Saint-Vérain is situated in the Nievre, Burgandy, whereas Saint-Veran (the town you linked) is actually in the Hautes-Alpes region. They are two separate places. It is a common mistake.

 

Also, a standard 16 tooth edged gear is NOT a "hunting gear"...if even any such thing exists. It is a type of gear used in transmissions to dictate speed (in other words, related to speedometer operation). No connection to the Greek goddess whatsoever. It is named after that particular designer. I am fairly certain that the goddess Artemis did not design this gear.

 

In all fairness you may have made this mistake simply because I did not post enough information for you to do a cursory internet search to present information, so that you could appear informed...Perhaps you may wish to do more research in the future (and avoid Wikipedia) to ensure you have not made any mistakes before posting your replies.

Edited July 9, 2014 by LordArtemis

[Nemarus]

On the surface, it seems like it would be fairly tricky to make a competent 3D pilot AI, especially one that can interchangeably react with both the player and other pilot AI's. But thinking back to X-wing and TIE Fighter, I do not recall ever finding the AI easy or unbelievable. Part of this may have been due to the mission structure and heavy scripting.

 

I recall that many missions would involve stopping a wing of TIE Bombers from attacking some objective. The Bombers would fly toward the objective, lock on torpedoes and fire at range. That's pretty simple. The trickier bits came when you started shooting at them. They'd juke and try and make themselves difficult to hit. At first they'd "stay on target", but eventually they'd enter some evasive pattern (which could distract you from the rest of their wingmen). And if you made them severely damaged, they'd even abandon their mission and try to return back to their capital ship.

 

And usually there would be a wing of TIE Fighters or Interceptors that were programmed to attack the player and/or his wingmen. I wonder how they determined who to attack within that target set, and what their conditions were for switching. But even then, that's all just simple scripting.

 

The more interesting questions concern their movement and shooting, moment to moment. Were they always just flying directly toward the target and shooting? Did they try to get behind a target? Did they try to avoid collisions? I know for sure they worked hard to avoid missiles fired at them.

 

As a kid, I found it all pretty natural and believable. I should give it another try and see how that AI holds up now.

 

I can only immediately recall one exploit. In many missions when you destroyed a full wing of fighters (such as all TIE Bombers in the Beta wing), another identical wing would immediately spawn at the enemy Star Destroyer. Some missions had respawn limits, some didn't. The exploit was that if you destroyed all but one of the ships in the wing, but only severely damaged the last one, it would go into retreat mode and begin the long, slow flight back to its capital ship. The new wave would not spawn until the damaged ship had gotten back home. Doing this trick often bought valuable time. You could delay a new Bomber or Interceptor wing from coming, which gave you an opportunity to deal with secondary and bonus objectives.

 

But apart from that (which was an exploit of scripting), I cannot remember any particular exploit cheese you could use to win a dogfight against AI. In fact I remember the most difficult missions were the ones where you had to defeat a lot of enemy fighters who, instead of focusing on objectives, were solely focused on you. In particular, I recall one very difficult mission where you were one TIE Defender against a whole lot of TIE Advanced with advanced concussion missiles. You couldn't use aggro tricks or any other scripting exploits. You just had to outfly them, and on the hard difficulty, it was indeed hard.

 

I also remember X-wing had a training mission where you dueled against progressively more difficult AI pilots, culminating in a fight against the "TOP ACE". I wonder what the differences were between the TOP ACE pilot AI and the lesser AI levels, in terms of why the TOP ACE as so much harder to beat in a dogfight.

Edited July 9, 2014 by Nemarus