Weapon Spread

UnderwhelmedUnderwhelmed DemoDetective #?! Join Date: 2006-09-19 Member: 58026Members, Constellation
Has UW given any thought as to how weapon spread will be determined? I hope we don't see a truly random spread that makes for somebody shotgunning a skulk dead on only to see most of the pellets dodge the skulk. At the very least I hope pellets/bullets will follow some statistical distribution with maximum density at the center that rewards the players with better aim. Another alternative might be a static distribution, which is also used in some other games.

Comments

  • LazerLazer Join Date: 2003-03-11 Member: 14406Members, Contributor, Constellation, NS2 Playtester
    edited June 2010
    I see what you are saying and although I can't answer, the concern makes sense.

    If instead of a '<!--coloro:#FF0000--><span style="color:#FF0000"><!--/coloro-->random(160) / 10<!--colorc--></span><!--/colorc-->' function returning 0 - 16 (these being made up numbers 0 being dead center 16 being a certain number of max degrees from the origin), it should be something more like '<!--coloro:#FF0000--><span style="color:#FF0000"><!--/coloro-->16 - (rand(40) / 10)^2)<!--colorc--></span><!--/colorc-->'. This will give more of a central bias but still maintain the possibility of odd shots hitting the maximum angle.

    Hopefully this made sense I'm sure any of the coders will know exactly what I'm talking about.

    Also, after the degrees from the origin is calculated it is fine to apply a random number between 0 and 2PI to specify if it is that many degrees up left right down or any angles in between, but I'm sure that part of the code is already done.
  • brownymasterbrownymaster Join Date: 2009-07-11 Member: 68110Members
    in NS they made it so that the spread had a consistent amount in a certain arc then random in another arc. I'm sure they'll do the same, with 1 dead center for sure.
  • Cereal_KillRCereal_KillR Join Date: 2002-10-31 Member: 1837Members
    I don't think the random distribution is going to be our greatest concern for number crunching right now. But rather than using uniform random and toying around with it, unbiased gaussian should be the way to go, it pretty already has everything for it, as it greatly favors the area near the center, yet has some spread. Truncate it to avoid having completely random shots.

    That being said, I have no idea how randomness is achieved in NS1, but it wasn't bad to the point of making me cringe (except the occasional sniper shotgun). Not that big of a problem?
  • LazerLazer Join Date: 2003-03-11 Member: 14406Members, Contributor, Constellation, NS2 Playtester
    edited June 2010
    In NS1 it was a problem. The spread pattern was not random and there was an unofficial cross hair you could download that would show where each pellet will hit. I'm sure some old clanners could pop in and explain this further but I'm sure it was always the same preset pattern. Don't know if this ever got fixed in the latest versions but it WAS an issue for a while.
  • brownymasterbrownymaster Join Date: 2009-07-11 Member: 68110Members
    The consistent spread was an issue mainly because there is no damage falloff in NS and the SG had infinite range (it now can't hit anything past 1024 units i think).
  • Renegade.Renegade. Join Date: 2003-01-15 Member: 12313Members, Constellation
    <!--quoteo(post=1775755:date=Jun 21 2010, 06:17 PM:name=Underwhelmed)--><div class='quotetop'>QUOTE (Underwhelmed @ Jun 21 2010, 06:17 PM) <a href="index.php?act=findpost&pid=1775755"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I hope we don't see a truly random spread that makes for somebody shotgunning a skulk dead on only to see most of the pellets dodge the skulk.?<!--QuoteEnd--></div><!--QuoteEEnd-->
    Huh? If you shotgun a skulk "dead on" from close range (conical) then it's going to take full damage, and if you do so at far range then you deserve to miss it.

    <!--quoteo(post=1775777:date=Jun 21 2010, 10:25 PM:name=Lazer)--><div class='quotetop'>QUOTE (Lazer @ Jun 21 2010, 10:25 PM) <a href="index.php?act=findpost&pid=1775777"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->In NS1 it was a problem. The spread pattern was not random and there was an unofficial cross hair you could download that would show where each pellet will hit. I'm sure some old clanners could pop in and explain this further but I'm sure it was always the same preset pattern. Don't know if this ever got fixed in the latest versions but it WAS an issue for a while.<!--QuoteEnd--></div><!--QuoteEEnd-->
    Yes, the issue was that after a certain range of conical spread, the pattern became fixed, with pellets traveling parallel thereon. So the crosshair displayed the parallel trajectory of each pellet at maximum-spread. The solution to this was (I believe) the team both allowed the pellets to continue outwards in a conical spread and capped their distance.
  • tankefugltankefugl One Script To Rule Them All... Trondheim, Norway Join Date: 2002-11-14 Member: 8641Members, Retired Developer, NS1 Playtester, Constellation, NS2 Playtester, Squad Five Blue
    edited June 2010
    <!--quoteo(post=1775768:date=Jun 22 2010, 04:18 AM:name=Cereal_KillR)--><div class='quotetop'>QUOTE (Cereal_KillR @ Jun 22 2010, 04:18 AM) <a href="index.php?act=findpost&pid=1775768"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I don't think the random distribution is going to be our greatest concern for number crunching right now. But rather than using uniform random and toying around with it, unbiased gaussian should be the way to go, it pretty already has everything for it, as it greatly favors the area near the center, yet has some spread. Truncate it to avoid having completely random shots.

    That being said, I have no idea how randomness is achieved in NS1, but it wasn't bad to the point of making me cringe (except the occasional sniper shotgun). Not that big of a problem?<!--QuoteEnd--></div><!--QuoteEEnd-->

    Actually, a gaussian distribution is used for regular weapons. For the shotgun, it proved too random - sometimes the pellets would all be way too close to the center, other times they all would be so far spread out that they would miss any skulk or fade. We ended up partitioning the spread into three cones and forcing a number of bullets to be within the inner cone, a number of bullets between the inner and outer cone, and then a few stray ones outside the outer one.

    This is what, after feedback, gave the (seemingly) right amount of randomness as well as consistency to it.

    (Of course, this was after it was fixed; a bug caused the random spread to be constant for a few versions.)
  • jamesbchjamesbch Join Date: 2004-04-20 Member: 28035Members, Constellation
    edited June 2010
    Why should we use only the rand function ? It seems logical that a player focusing on a fixed point would have a better aim than a player who just turned his head and shot at the same time ! You should take it into account.
  • tankefugltankefugl One Script To Rule Them All... Trondheim, Norway Join Date: 2002-11-14 Member: 8641Members, Retired Developer, NS1 Playtester, Constellation, NS2 Playtester, Squad Five Blue
    There are many variables one can take into account. In the end, the real qualifier is whether it works in the game's context or not, not whether it's "logical" or not.
  • Cereal_KillRCereal_KillR Join Date: 2002-10-31 Member: 1837Members
    <!--quoteo(post=1775777:date=Jun 22 2010, 04:25 AM:name=Lazer)--><div class='quotetop'>QUOTE (Lazer @ Jun 22 2010, 04:25 AM) <a href="index.php?act=findpost&pid=1775777"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->In NS1 it was a problem. The spread pattern was not random and there was an unofficial cross hair you could download that would show where each pellet will hit. I'm sure some old clanners could pop in and explain this further but I'm sure it was always the same preset pattern. Don't know if this ever got fixed in the latest versions but it WAS an issue for a while.<!--QuoteEnd--></div><!--QuoteEEnd-->


    The shotgun pellets got randomized in 2.0 :D
    Or was it 3.0? I don't recall using the shotgun that often in 1.0x but I remember clearly the perfectly constant shotgun spread.

    In any case, I think that if we're ready to complain about the weapon's spread distribution at game launch, then that means the game's already near perfect.
  • LazerLazer Join Date: 2003-03-11 Member: 14406Members, Contributor, Constellation, NS2 Playtester
    Definitely an issue in 2.0 I think it got fixed by 3.0 (or later?)
  • kuperayekuperaye Join Date: 2003-03-14 Member: 14519Members, Constellation
    <!--quoteo(post=1775818:date=Jun 22 2010, 09:07 AM:name=Lazer)--><div class='quotetop'>QUOTE (Lazer @ Jun 22 2010, 09:07 AM) <a href="index.php?act=findpost&pid=1775818"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Definitely an issue in 2.0 I think it got fixed by 3.0 (or later?)<!--QuoteEnd--></div><!--QuoteEEnd-->


    I think it was one of the later patches for 3.0 but I could be completely mistaken. I do remember using the fixed crosshairs for shotgun pellets.

    What they should do for NS2 is do random spread for pubs but have a cvar that you can turn on for competitive to turn on a fixed spread. I know TF2 does this and it works pretty well. Because it gets annoying to shoot someone point blank and they only take like 20 damage.
Sign In or Register to comment.