Interesting Read From New TF2 Interview
Coaleh
Not Pepseh Join Date: 2003-10-11 Member: 21602Members, Constellation, Reinforced - Shadow
Not Pepseh Join Date: 2003-10-11 Member: 21602Members, Constellation, Reinforced - Shadow
<div class="IPBDescription">It's rather relevant to NS2</div>1up did a really good interview of the Orange Box and in it the asked the TF2 guys some really good questions about gameplay. It's pretty relevant to NS2's development I would say. Here's a few quotes:
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->RW: What we found as we worked on TF2 over the years and went through all these iterations is that depth doesn't demand increased complexity. TF2 is a sucker punch -- it only looks simple. We're already seeing large-scale skill differences.<!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->GFW: Is that luck or skill?
RW: It's skill affecting luck. You can't control the Gaussian distribution of fire when you pull the trigger, but skill is involved in centering that Gaussian distribution on an opponent's head. The idea that there shouldn't be any luck in an FPS is silly.<!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->Adding randomness increases the range of possible tactical skill options. Sometimes, the crazy thing works. If a game is 100 percent deterministic, then only one tactic works. The more randomness you add, the wider the range of options becomes.<!--QuoteEnd--></div><!--QuoteEEnd-->
Full Article: <a href="http://www.1up.com/do/feature?cId=3165930" target="_blank">here.</a>
TF2 section: <a href="http://www.1up.com/do/feature?pager.offset=5&cId=3165930" target="_blank">here.</a>
Although it does advocate that simpler is better, that's not the point I'm trying to apply to NS. What I am trying to say is that things like adding randomness to the game helps it, so not just one tactic is the way to go. How many times have you played NS and had every seasoned player there screaming for things to be done ONE WAY. It was pretty true that often only one tactic would be the best to use in a given situation, but I believe that's something that just makes the game stale and harder for new players.
I say that with a pinch of salt though, because there is often more than one way to do things, but for NS2 its something to keep in mind.
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->RW: What we found as we worked on TF2 over the years and went through all these iterations is that depth doesn't demand increased complexity. TF2 is a sucker punch -- it only looks simple. We're already seeing large-scale skill differences.<!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->GFW: Is that luck or skill?
RW: It's skill affecting luck. You can't control the Gaussian distribution of fire when you pull the trigger, but skill is involved in centering that Gaussian distribution on an opponent's head. The idea that there shouldn't be any luck in an FPS is silly.<!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->Adding randomness increases the range of possible tactical skill options. Sometimes, the crazy thing works. If a game is 100 percent deterministic, then only one tactic works. The more randomness you add, the wider the range of options becomes.<!--QuoteEnd--></div><!--QuoteEEnd-->
Full Article: <a href="http://www.1up.com/do/feature?cId=3165930" target="_blank">here.</a>
TF2 section: <a href="http://www.1up.com/do/feature?pager.offset=5&cId=3165930" target="_blank">here.</a>
Although it does advocate that simpler is better, that's not the point I'm trying to apply to NS. What I am trying to say is that things like adding randomness to the game helps it, so not just one tactic is the way to go. How many times have you played NS and had every seasoned player there screaming for things to be done ONE WAY. It was pretty true that often only one tactic would be the best to use in a given situation, but I believe that's something that just makes the game stale and harder for new players.
I say that with a pinch of salt though, because there is often more than one way to do things, but for NS2 its something to keep in mind.
Comments
Starcraft has nothing random, still an massive number of tactics work. It's all about predicting enemy decisions, fooling him, countering him, countering his counters and so on.
Let's pick an example. Ns_tanith, 6on6. Marines decide to push heavily with 4 rines, 3 sgs through double. Armslab on base. --> The smart kharaa team doesn't fight the nearly impossible fight, but goes from sat side to base. --> The comm expects the baserush, so he gets some mines there. --> The aliens have a gorge with them and it blows the mines up. --> The rines quickly take out an alien rt, but are forced to build a field obs and beacon. --> Aliens spread out, 1 recaps the lost node and the rest go take out marine res, trying to find out what the marine gameplay plan is. No randomness involved, just a bit of different starting moves that lead to new situations and some more tactical calls.
The aliens could have gone for the marine nodes, try to fight the shotguns, quickly getting some off-the-wall sc or dc. The comm could have had obs at base, but he decided to get fast upgrades. Alien build selection, early lerks, more rt:s being dropped, skulk seeing the shotguns. No randomness involved. At any moment an individual effort could upset the gameplay flow and cause the situation go upside down, for example the remaining marine succesfully fending off the baserush.
I'm not saying ns has enough of tactical versatility in all situations, but it doesn't have to be randomness making the tactics viable. The experienced players deciding to do one thing is often because the whole gameplay has led to a situation like that.
And btw, they just mentioned the gun spread as an example. NS has reg and gun spread causing some randomness.
That's actually a bit weird talk by a TF2 developer(?) if you ask me. Like chess had only one strategy. And of course there are large scale skill differences, no matter if you've got depth or not. A Quake 3 pro could beat a newbie in an empty room, no matter what and it doesn't mean it had anything to do with depth. Depth is the thing that starcraft is still being played professionally after about 10 years after its publishment and some new tactics are still being created (helped by active renewal of map pools though)and even the professionals are improving.
Random hive locations, some random features in the map are the things I want to see, since the players can see them at the start of the round or scout them and then adapt their strategy. However, you can't adapt your strategy for your enemy getting those 5 crit hits in a row. Going for a risky strategy is also ok, but the risk should be created by the fact that you don't know if your enemy is going to do a move affecting your strategy.
Gaussian spread damage is an acceptable form of randomness in combat, because it's randomness contained within an absolute. You know the area that the spread is restricted to, so you are aware of the limitations of the weapon. Something like critical hits is just a source of frustration, because you don't need to do anything right to get a critical hit. In TF2 you could miss a player and catch a cloaked Spy with a critical hit, instantly killing him. That's not skill, it's fluke that has been facilitated through the design of the game.
If randomness is put into NS2, it neads to be a random element wrapped up in a known quantity. E.g. (assuming Hives and map layout are well-balanced) having a random starting Hive isn't bad because it is a random location from 3 definite, known possibilities. The same is true of something like res-for-kills. If a Skulk were to get 1-3 res from each Marine kill, it would be possible for a Commander to calculate a range of res that Skulk has, and determine how likely it is that that Skulk has enough res to Fade or drop the Hive. This estimate may prompt a skilfully anticipant play by the Commander to move on a Hive very soon after, as or even before it is dropped.
Well articulated point Crispy.
I find that I ultimately become frustrated with randomness in games the more I play them <i>competitively</i>. On the flip-side it can be equally frustrating when somebody is just plainly better than me and pwns my face repeatedly when there is no means for me to beat them with fluke (happens a lot in <a href="http://www.pvkii.com" target="_blank">PVKII</a>).
Games that focus on teamplay tend to have a fair amount of randomness built into their combat, so that one "hardcore" player cannot work as a one man team. At this point I'm confident the NS team know how to and will strike a balance in playtesting.
Unchanging certainity gets stale and unexciting. A sudden crit rocket into that room of camping engineers is always fun, but being wiped out by something there's no reasonable way to defend against... isn't.
In the case of TF2 I'd rather have the crits stay though; emotional highs and lows makes the games more memorable.
Unchanging certainity gets stale and unexciting. A sudden crit rocket into that room of camping engineers is always fun, but being wiped out by something there's no reasonable way to defend against... isn't.
In the case of TF2 I'd rather have the crits stay though; emotional highs and lows makes the games more memorable.<!--QuoteEnd--></div><!--QuoteEEnd-->
You're exactly right. Although the members of the Steam forums will not agree with you on that, and they seem to skim past any post that argues in favor of critical hits. They claim it takes away skill, but you're right in that it makes the game more fun. A set pace in a game doesn't result in a good experience, there needs to be low and high points. I think the addition of a respawn timer to TF2 is another example of that, in addition to the fact that having no respawn timer breaks the game completely. 'Waiting' may be boring, but if the action itself isn't exciting, then a game has failed in its primary purpose, to entertain.
Gaussian spread damage is an acceptable form of randomness in combat, because it's randomness contained within an absolute. You know the area that the spread is restricted to, so you are aware of the limitations of the weapon. Something like critical hits is just a source of frustration, because you don't need to do anything right to get a critical hit. In TF2 you could miss a player and catch a cloaked Spy with a critical hit, instantly killing him. That's not skill, it's fluke that has been facilitated through the design of the game.
If randomness is put into NS2, it neads to be a random element wrapped up in a known quantity. E.g. (assuming Hives and map layout are well-balanced) having a random starting Hive isn't bad because it is a random location from 3 definite, known possibilities. The same is true of something like res-for-kills. If a Skulk were to get 1-3 res from each Marine kill, it would be possible for a Commander to calculate a range of res that Skulk has, and determine how likely it is that that Skulk has enough res to Fade or drop the Hive. This estimate may prompt a skilfully anticipant play by the Commander to move on a Hive very soon after, as or even before it is dropped.<!--QuoteEnd--></div><!--QuoteEEnd-->
I totally agree -- couldn't have said it better myself. I can assure you we won't have anything like critical hits in NS2.
Depth shouldn't be confused with skill differences. Of course depth demands complexity - tic-tac-toe's gameplay level is much shallower than say, Chess or even Checkers because all you have is nine spaces and two sides.
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->Adding randomness increases the range of possible tactical skill options. Sometimes, the crazy thing works. If a game is 100 percent deterministic, then only one tactic works. The more randomness you add, the wider the range of options becomes.<!--QuoteEnd--></div><!--QuoteEEnd-->
Deterministic does not equal predictability. Chess is completely deterministic, yet no win-all strategy has been discovered yet.
It's not randomness you want. It's variation. The more randomness you add, the less predictable everything becomes, and the less of an effect your actions have on the outcome of a game.
Deterministic does not equal predictability. Chess is completely deterministic, yet no win-all strategy has been discovered yet.
It's not randomness you want. It's variation. The more randomness you add, the less predictable everything becomes, and the less of an effect your actions have on the outcome of a game.<!--QuoteEnd--></div><!--QuoteEEnd-->
I agree that you can make a game have more variation without adding a lot of random elements, but then you're risking making the game for a really niche audience. There's something else that I can't quite articulate yet. I think a TF2 without crits would be less fun but I'm not sure why. I think part of it has to do with the ups and downs of gameplay as suggested before.
Gamers know that this is wrong, but here are some words to munch on anyway.
When you have human players coupled with the uncertainty of knowing what your opponent will do, this creates a huge opportunity for varied tactics, because of the "Tactics Landscape", similar to the "Fitness Landscape" from evolution and complexity. The landscape is ever-changing in response to the population of tactics used, so there can never be a one-tactic-fits-all. Multiple recursion -> complexity.
What is a great example? The RUSH. It works when your opponent is not prepared. Why not prepare for the rush? It gives an advantage if there is no rush. And the advantage is enough to risk it sometimes. However, if preparation is abandoned too often, this changes the landscape, making it more worthwhile to try the rush. But then players adapt again. There is a dynamic optimization in progress, which adds variation to the game, but has nothing to do with artificially imposed Gaussians in the mechanics.
A similar argument can be made on a smaller scale, on knowing when to ZIG and when to ZAG against the soldier with his rocket launcher. Very new players run in a straight line. So new soldiers shoot ahead of them. More skilled players run back and forth. But then more skilled soldiers shoot behind them. But vet players may then sometimes just run straight, making skilled soldiers miss.
So, we have "100 percent 'determinism' " which still allows for great variation. (not actually determinism just not artifically imposed randomness, lets not get into philosophy lol)
That said, it IS possible to enhance the tactical landscape through artificial means. It can make a certain tactic viable when it would not otherwise be. The example here is the TF2 crit. Maybe you do a suicide rush against some defenses which you would defeat if you crit, when normally you would have to go back and take time to heal, also a risk, facing running out of time. And if you kill them you take the point. So a new tactic is available which would not be otherwise.
Of course in NS2, critting a fade/onos high cost unit would be a nightmare. HEADSHOT lol. Also, deterministic damage from melee classes adds to the skill ceiling for them, so they know when to run or when to go in for the last swipe. Randomness would be a no-no there as well.
But I trust the Flay. He has glowies.
It's like how I enjoy being the only one on a server with godmode on, while not liking being on the receiving end.
You should see the discussion on scripts/bunny hopping. You probably wouldn't be paying complements then. The forum members save most of their venom for those topics. <img src="style_emoticons/<#EMO_DIR#>/nerd-fix.gif" style="vertical-align:middle" emoid="::nerdy::" border="0" alt="nerd-fix.gif" />
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->I agree. As much as I enjoy instantgibbing someone with a rocket, I feel it doesn't belong.
It's like how I enjoy being the only one on a server with godmode on, while not liking being on the receiving end.<!--QuoteEnd--></div><!--QuoteEEnd-->
I'm starting to think C-rockets are a special case different from other critical shot weapons. It's harder to land a crit pipe or sticky accurately, especially at range. Rockets are accurate at long range with high damage <i>and</i> decent AoE so they tend toward cheap kills. Demomen get cheap kills too but they don't have the range <i>and</i> they have the disadvantage at close range so their crit advantage is "counterable".
I think the game would be much better competitively if they removed the crit from the rocket. Maybe if the removed the crit from the spy too because if you're a spy and not backstabbing with the knife or using the revolver, you're doing it wrong and a crit back stab is actually worse for the spy.
The only random factor should be the players.
I like how StarCraft is quite binary with a lttle probability throw in, but it still has enormous depth to it. Granted, it's an RTS and not a FPS, but you get my point.
it "belongs" in rts because that's a simplification of actual combat, so instead of calculating that soldier 1 took a swing at soldier 2 and soldier 2 ideally ducked in time but still got nicked on the top of the head, it just adds a random number of 'damage' to simulate that interaction. the same with rpgs/mmos.
but i think it's both good and bad that there's a degree of uncertainty to a game. if you always know that one headshot will kill someone, you'll always be aiming for one headshot. if you always know that three shots will kill someone, 'skill' involves you being able to keep the crosshair on an opponent for three consecutive shots. by not having that level of certainty, you might be frustrated that your 'certain kill' hit-number is a -range- and not a -number-, but it also means you have to invest more into the game. i'm not sure if i'm getting this particular point across.
i still think randomisation is a rather artificial way to add variety to fps gameplay.