Accuracy straw poll

The place to discuss balance changes for future versions of the game.
(Master releases & 3.X)

Which accuracy system do you want?

Random to-hit roll, no physics
0
No votes
Gaussian to-hit roll, no physics
7
28%
Physics-based only, no random roll
8
32%
Projectiles always hit (the Starcraft option)
3
12%
Do not change anything / I do not understand this poll / No opinion
7
28%
 
Total votes: 25

Per
Warzone 2100 Team Member
Warzone 2100 Team Member
Posts: 3777
Joined: 03 Aug 2006, 19:39

Re: Accuracy straw poll

Post by Per » 24 Mar 2013, 01:21

Deus Siddis wrote:If the answer turns out to be roughly yes then I think we are voting for the same thing here--
I did some testing, with #3989, and the anwer is "roughly yes" so far. I found it near impossible to intentionally or unintentionally dodge short or medium range projectiles. I only saw one or two shots fire over a droid when it moved over very uneven ground. Other than that all misses seemed to be overkills.

Deus Siddis
Trained
Trained
Posts: 235
Joined: 18 Aug 2007, 06:58

Re: Accuracy straw poll

Post by Deus Siddis » 24 Mar 2013, 02:02

Per wrote:
Deus Siddis wrote:If the answer turns out to be roughly yes then I think we are voting for the same thing here--
I did some testing, with #3989, and the anwer is "roughly yes" so far. I found it near impossible to intentionally or unintentionally dodge short or medium range projectiles. I only saw one or two shots fire over a droid when it moved over very uneven ground. Other than that all misses seemed to be overkills.
Excellent, it sounds like this could be the solution then.
Iluvalar wrote: The accuracy/distance function won't be continuous. There will be NO solution to your problem. No matter how much hard we want it. No matter how much people work on it. Keep neglecting facts lol.
And it seems Iluvalar agrees that we should physically simulate direct fire shots without any trajectory deviation based on distance or accuracy.

User avatar
Iluvalar
Regular
Regular
Posts: 1819
Joined: 02 Oct 2010, 18:44

Re: Accuracy straw poll

Post by Iluvalar » 24 Mar 2013, 05:14

No, I disagree with that as well. As it would mess with the weapons rate of fire and I don't want that.
Heretic 2.3 improver and proud of it.

User avatar
Shadow Wolf TJC
Regular
Regular
Posts: 1047
Joined: 16 Apr 2011, 05:12
Location: Raleigh, NC

Re: Accuracy straw poll

Post by Shadow Wolf TJC » 24 Mar 2013, 11:23

I'd personally go with the Supreme Commander option, which would be closest to a physics-based system, though with some degree of randomness thrown in.

I'd personally want to replace the shortHit and longHit variables in weapons.ini with a maxOffsetAngle variable, which would be the MAX angle of deviation from the target a projectile COULD travel. For example, say that a direct fire weapon with a maxOffsetAngle of 60 degrees (which I doubt anyone would want to assign to a weapon :lol2: ) fires at a target located 32 tiles away. Using some trigonometry (specifically, the law of sines, where a/sin A = b/sin B = c/sin C), and assuming that the projectile isn't homing, the projectile would always land within 32 tiles away from its intended target (since all 3 angles calculated are 60 degrees, hence all sides are the same length), would have a 50% chance of landing within 16.56 tiles of the target (since 16.56/sin 30 = 32/sin 75), and would have a 25% chance of landing within 8.35 tiles of the target (since 8.35/sin 15 = 32/sin 82.5). :lecture:

Edit: The game would calculate the direction that the projectile should go upon creation as something like this:

Code: Select all

var r = 100 - (Math.random() * 201);
var finalAngle = targetAngle + (r * maxOffsetAngle);
where targetAngle is the angle from the attacking object to the target, and finalAngle is the angle at which the projectile will actually travel at (which would almost certainly never be equal to targetAngle, unless the weapon's maxOffsetAngle value was 0).

I'd also want to add a flightTurnRate variable in weapons.ini, which would control how fast homing projectiles can turn to face their intended targets. This would ensure that even homing weapons would have a chance of missing their targets, and would also allow for the creation of weapons that would be more accurate over long ranges than short ranges, since the farther away the target is, the more time the projectile has to adjust its trajectory to guarantee a direct hit. :3

By the way, why are people proposing accuracy systems where missed shots ALWAYS land within a mere couple tiles of their targets? What's wrong with projectiles (especially long-range indirect fire projectiles) landing over 10 tiles away from their intended targets? :stare:
Creator of Warzone 2100: Contingency!
Founder of Wikizone 2100: http://wikizone2100.wikia.com/wiki/Wikizone_2100

raycast
Trained
Trained
Posts: 131
Joined: 12 Sep 2012, 19:16

Re: Accuracy straw poll

Post by raycast » 24 Mar 2013, 14:12

While I do agree that high-speed low-range shots could be simplified - also with respect to damage prediction (I have been considering to add this to #3748) - I don't think this is an answer to "all your problems". The reason is that with these shots, you will barely notice the difference.

However, e.g. flamers are not fast traveling. And since they also rely on area damage from burning, IIRC, this approach won't help here! Nor will it solve our imprecise concept of what "accuracy" should be. Because, we are actually not talking the same things here... :lecture:

Face it, we are talking about very different problems here:
  • Gameplay: obviously we want something that is intuitive for the players and fun.
  • Balancing: we want to have a simple parameter "accuracy" in our weapons definition to control weapons balancing
  • Implementation: we need something that is fast to implement. True gaussian distributions require more randomness than we probably want to use; something like 1/x^2 is probably good enough.
  • Explanation: we'd all love to be able to tell from the stats files, which unit is the most powerful (but actually, it's part of the fun that it is not that obvious as damage-per-second)
A lot of complaints and discussions here would be obsolete, if people would try to provide a patch to the real code instead of theoretical improvements. Because for example, shots and projectiles are initialized by source and destination; not direction. Which essentially says: it's more expensive to vary angle than varying target destination. Of course you could rewrite the code, but this will likely produce lots of fresh bugs. Go ahead and try it. But other than that, try to work with the codebase that is proven working. :lecture:

Here's a rather radical proposal:

Let's break the accuracy parameter into two parts:
  • Average deviation from target at short and long range (we may want to have weapons where this is linear with range and artillery weapons where it probably isn't!)
  • Chance to do damage on impact.
The latter is probably what most people interpret "accuracy" as. The chance to actually do damage.

Don't get me started with physics... see: if you shoot a MG at a tank, some shots may actually be deflected and not do actual damage. It's not as if every shot tears apart an equally tiny bit.

Maybe these two parameters are more helpful to achieve the desired gameplay behavior, and offer the desired interpretability. The first gives how many shots will be on-target, the latter how many of the hits actually do damage.

Plus, this variant is straightforward to implement. One parameter is used when firing the projectile, the other on impact.

User avatar
Shadow Wolf TJC
Regular
Regular
Posts: 1047
Joined: 16 Apr 2011, 05:12
Location: Raleigh, NC

Re: Accuracy straw poll

Post by Shadow Wolf TJC » 24 Mar 2013, 16:09

raycast wrote:Let's break the accuracy parameter into two parts:
  • Average deviation from target at short and long range (we may want to have weapons where this is linear with range and artillery weapons where it probably isn't!)
  • Chance to do damage on impact.
The latter is probably what most people interpret "accuracy" as. The chance to actually do damage.
Personally, I'd interpret accuracy as the chance to land a "direct hit" on the target. Many weapons in Warzone 2100, including Light Cannons, Mortars, and Railguns, do more damage on a direct hit than they do through splash damage.

Also, I feel that the closer to the target the weapon is, or the bigger the target, the more likely the projectile SHOULD hit the target. An accuracy system based on static accuracy stats for weapons, regardless of range to their targets, just seems unrealistic. :lol2:

By the way, I'm very much against the idea of StarCraft 2's method of having every shot fired from every weapon being guaranteed to hit their targets, as not only would it make accuracy upgrades pointless, it would make it impossible to develop weapons that are deliberately made inaccurate in order to balance out their other attributes. For example, Contingency's version of the Mini-Rocket Pod and Mini-Rocket Array were deliberately made highly inaccurate in order to balance out their higher damage-per-second (compared to cannons) to the point where they're actually inferior to cannons in 1vs1 engagements, but superior in large-scale engagements, since stray shots could hit nearby enemies. If we were to switch to the StarCraft 2 method, then it would be a step backwards in terms of moddability. I'd have to find a new way to make Contingency's Mini-Rocket Pods distinguishable from Light Cannons, since without a way to influence a weapon's accuracy, these highly-inaccurate weapons would suddenly become overpowered highly accurate weapons. :evil:

Also, since we're switching from the old .txt format to the new .ini format, which, although it takes up much more space, actually seems more modular than the older format, I don't see as big of a reason as to why we should be reluctant towards introducing new variables for weapons as before. :roll:
Creator of Warzone 2100: Contingency!
Founder of Wikizone 2100: http://wikizone2100.wikia.com/wiki/Wikizone_2100

User avatar
effigy
Regular
Regular
Posts: 1217
Joined: 22 Jan 2010, 03:21
Contact:

Re: Accuracy straw poll

Post by effigy » 24 Mar 2013, 19:59

I think it's logical that the further away a unit/def from a target the less accurate it should be. No randomness needed, except to determine when a shot will miss. If a shot with splash hits a tile with out a droid/def it's logical to me that anything in the area of effect should take some damage, though, anything close to the area of impact would take more damage than something further away.

This has me thinking about the options in the orders menu for optimum, short and long range. Are these only in use when a unit is told to fire at something? Optimum meaning the range at which a droid has it's highest level of accuracy or what?
This is why some features aren't implemented: http://forums.wz2100.net/viewtopic.php?f=30&t=7490&view=unread#p87241

User avatar
aubergine
Professional
Professional
Posts: 3459
Joined: 10 Oct 2010, 00:58
Contact:

Re: Accuracy straw poll

Post by aubergine » 24 Mar 2013, 20:35

The short/optimum/long range options have been removed in WZ 3.2 branch.
"Dedicated to discovering Warzone artefacts, and sharing them freely for the benefit of the community."
-- https://warzone.atlassian.net/wiki/display/GO

User avatar
Shadow Wolf TJC
Regular
Regular
Posts: 1047
Joined: 16 Apr 2011, 05:12
Location: Raleigh, NC

Re: Accuracy straw poll

Post by Shadow Wolf TJC » 24 Mar 2013, 21:47

effigy wrote:I think it's logical that the further away a unit/def from a target the less accurate it should be.
Quite true if the projectile lacked any ability to home onto its target, though if the projectile can do so, then there would exist a point where, after reaching a certain distance, the chance to hit would actually start to increase since the homing projectile would be getting more time to adjust its trajectory before reaching its target. Eventually, at twice the distance from that point to the weapon (assuming that the projectile's speed never changes), the homing projectile's accuracy would reach 100%. :wink:

By the way, I haven't actually voted yet, though I'm considering it.
Creator of Warzone 2100: Contingency!
Founder of Wikizone 2100: http://wikizone2100.wikia.com/wiki/Wikizone_2100

User avatar
effigy
Regular
Regular
Posts: 1217
Joined: 22 Jan 2010, 03:21
Contact:

Re: Accuracy straw poll

Post by effigy » 24 Mar 2013, 22:28

Shadow Wolf TJC wrote:[...
Eventually, at twice the distance from that point to the weapon (assuming that the projectile's speed never changes), the homing projectile's accuracy would reach 100%. :wink:

By the way, I haven't actually voted yet, though I'm considering it.
Interesting.

I'm one of the no opinion votes, currently, though I seem to be leaning towards physics based.
This is why some features aren't implemented: http://forums.wz2100.net/viewtopic.php?f=30&t=7490&view=unread#p87241

raycast
Trained
Trained
Posts: 131
Joined: 12 Sep 2012, 19:16

Re: Accuracy straw poll

Post by raycast » 24 Mar 2013, 23:51

Shadow Wolf TJC wrote:Personally, I'd interpret accuracy as the chance to land a "direct hit" on the target. Many weapons in Warzone 2100, including Light Cannons, Mortars, and Railguns, do more damage on a direct hit than they do through splash damage.
But the chance to hit as such is rather meaningless with splash damage. In fact, it is quite artificial; driven from a simple dice roll semantics. Why not use the two parameters I suggested?

Which is actually my concern with the current version, as well as my patch. See: a shot with splash damage could not miss a building in the current version, which is more than 1 tile in size, just for example. Because an inaccurate shot meant: 1 tile off.
And with my patch, this still applies to close range. No way to not hit at close range, making machine guns and flamers even more powerful.
Shadow Wolf TJC wrote: Also, I feel that the closer to the target the weapon is, or the bigger the target, the more likely the projectile SHOULD hit the target. An accuracy system based on static accuracy stats for weapons, regardless of range to their targets, just seems unrealistic. :lol2:
Incorrect. At close range, a mortar will have to shoot a high angle, making it actually quite inaccurate. But of course, at the end of its range, it will be inaccurate, too.

Which is why I suggested to add back *multiple* parameters. You could then specify that a machine gun will have 0 error at close range, and 2 tiles at maximum range without upgrades; whereas a mortar will maybe have 2 tiles both at minimum and maximum range. I don't see why a mortar should scale inaccuracy from 0 to 10 tiles at maximum range, just to get a realistic 1 tile at the range where it is most effective (because of shot traveling time, it will usually miss at maximum range anyway).

User avatar
Shadow Wolf TJC
Regular
Regular
Posts: 1047
Joined: 16 Apr 2011, 05:12
Location: Raleigh, NC

Re: Accuracy straw poll

Post by Shadow Wolf TJC » 25 Mar 2013, 08:18

raycast wrote:Incorrect. At close range, a mortar will have to shoot a high angle, making it actually quite inaccurate. But of course, at the end of its range, it will be inaccurate, too.
When you're aiming straight up, it'll take a long time for a mortar shell to fall back down, but in the end, it'll still be more accurate at shorter ranges, since the area where the shell could land is still smaller. Of course, given how long the shell stays in the air, units that can move would likely have enough time to move out of the way, so in a sense, mortars ARE inaccurate at shorter ranges, but ONLY against anything that can actually move. :wink:
Creator of Warzone 2100: Contingency!
Founder of Wikizone 2100: http://wikizone2100.wikia.com/wiki/Wikizone_2100

raycast
Trained
Trained
Posts: 131
Joined: 12 Sep 2012, 19:16

Re: Accuracy straw poll

Post by raycast » 25 Mar 2013, 23:32

Actually, if you shoot straight up with an error of ~5°, and if you shoot at 45° with an error of ~5°, the results will be quite similar, if you shoot up to the same height.

User avatar
War2070
Rookie
Rookie
Posts: 25
Joined: 06 Mar 2012, 04:42

Re: Accuracy straw poll

Post by War2070 » 02 Apr 2013, 07:11

I know you guys like to look like you know what your doing. Your changing it so much it looks like its headed to be crap and you won't be able to remember what you did to make it like that. You guys made a mess of everything from what Warzone 2100 version 1.1 was. To me you you all look like a bunch of kids trying to learn programming on everyone's backs. The weapons in 1.1 were dependable and fired just the way it was supposed to. Now if you can make the weapons dependable I have no problem with that.

Per
Warzone 2100 Team Member
Warzone 2100 Team Member
Posts: 3777
Joined: 03 Aug 2006, 19:39

Re: Accuracy straw poll

Post by Per » 02 Apr 2013, 10:26

War2070 wrote:To me you you all look like a bunch of kids trying to learn programming on everyone's backs.
It never ceases to find it so strange that anyone in their sane mind could seek out an all-volunteer, unpaid effort to produce stuff you can download and use for free, and say things like that. You know that you can still grab the original Warzone and use that instead, right? Or some other version or fork of the game? We are not depriving you of anything - we are giving you more options, for free, and you are attacking us personally.

Didn't anyone teach you manners? :roll:

Post Reply