Tuesday, March 05, 2013

Resolution Versus Radius

Ripard Teg had some posts where he posited that he thinks "something's a little wonky in the actual code that governs how much damage large signature guns should do to smaller targets". He went on to talk about math and tracking and basically got it all wrong as Azual Skoll masterfully demonstrated in a post that could be held up as a shining example of showing someone to be wrong on the internet.

Ripard, being shown the error of his ways, retracted his post soon thereafter:
Earlier this week, I wrote a little post called "Signature" in which I stated my belief that:
I think something's a little wonky in the actual code that governs how much damage large signature guns should do to smaller targets. 
And then I wrote a whole lot of wrong about where I thought this was coming from. EVE blogger and outstanding pilot Azual Skoll took exception and wrote a masterful deconstruction of just how wrong I was and put it on his blog. It's great reading and I suggest you go out and give it a read if you're interested in the topic. 
Now in my own defense -- and not in any way trying to deflect my mistake -- Azual's post is mostly about tracking issues, not damage, and my post was about damage. I fully concede that large guns have a problem tracking smaller targets; I've said in the past that the tracking issue is solvable and indeed I've done my own experiments to solve it. My statement was regarding the damage that large guns do to smaller targets when they actually hit them. Even Azual's data shows this: when the large guns of a Naga actually hit a Cormorant, they were doing equal or better damage than the medium guns of a Ferox. 
What I forgot, and the source of my mistake, is how gun damage is calculated. When a gun does hit a target, it does a minimum of 50% base damage to the target. And damn it damn it damn it, I knew this! But somehow this week I idiotically forgot about it and instead of looking it up, I worked myself into a holy snit over my indignation about something that didn't deserve my indignation. 
So, to summarize, Azual Skoll is completely right and I am completely wrong. I'm really embarrassed, and I apologize. I regret the error, as I say at times like this.

But then he ends his post with this paragraph:
So yeah, the math within the game generating damage on large guns against small targets is working perfectly well. So I'm now going to transfer some small bit of my indignation to the math itself. Should guns automatically do at least 50% base damage when they hit, regardless of the target size they're hitting? I'm thinking "no". What do you all think?
This is what I would like to address.

A while back I was ruminating on tracking, transervsal, etc and learned about the math behind Eve's gunnery mechanics:
My error lies in that I think of the two separately but the formula to determine hits has them combined in the equation. From Evelopedia:

The way I was thinking is that the tracking versus transversal part merely determines if your guns can move fast enough to aim at the target; then assuming you can aim at the target, then the resolution versus radius to determine the dispersion of the shot versus the size of the target area profile.
But its not a two part process, its a one step calculation and turning on a MWD may increase transversal speed but it also increases the radius. So the question is, how much effect does the two changes have?

My instinct prior to that post was that the tracking versus velocity was one calculation, and signature resolution versus radius was a second calculation. In other words, am I fast enough to swing my aim around at the target and is my aim good enough to hit?

But Eve, as I discovered to my public embarrassment in 2011, throws the signature resolution/radius ratio into the tracking calculation. What this means is that ship with ~0 angular velocity negates any effect of the signature ratio in the chance to hit.

Its like saying that I should be able to hit a quarter sitting on a tree with a rifle because the quarter is not moving. This strikes me as wrong.

I think for gameplay balance purposes that large bore guns (ROLEPLAY WARNING: due to inherent nature of attempting to push energy/matter a further distance) should have a built in inaccuracy that is calculated via the weapon's signature resolution and the target's signature radius.

In other words, a frigate with no MWD sitting still at optimal range to a battleship sized weapon should NOT guarantee a hit. I realize this is radically different than what currently happens now and that a lot of balancing would need to make this work but I'm kind of in Ripard's side on this one. Currently I feel that small ships are too easily dealt with by larger ship weapons (mainly  Battleship sized and bigger versus frigates and destroyers) when applied en masse and that giving them more protection would allow for more diverse gameplay options.

That being said, I don't believe Eve is in dire straits with the current gunnery hit calculations.


  1. I dunno, I would think that as there is no wind or atmospheric resistance, and negligible gravity to compensate for, a computer controlled gun on an advanced spaceship would have no problem hitting an unmoving quarter regardless of the size of the shell. Modern tank guns that have to fire on the move in bumpy terrain in an atmosphere are already frighteningly accurate.

    1. Scifi technobabble debate engaged!
      If the weapons have to make compensations for the massive energies involved, and thus impacts accuracy over distance (hence why the weapons have optimal range) then its not surprising they have accuracy issues that they continually have to adjust for as calibrations are made ineffective due to wear, material expansion, stress, etc.

      PLUS, the entire cluster is immersed in a large dust cloud (hence why ships slow down without thrust) and this dust cloud's eddies and currents can also affect accuracy of larger weapons who interact with it more violently.


    2. This comment has been removed by the author.

    3. First off, just because it makes me laugh:


      As to your response, that nicely explains why projectiles have a short optimal and a long falloff, as the result is similar, however, if you note in game projectiles are the "physical" weapons, lasers are energy weapons, and hybrids are, well, a hybrid of the two, the optimal/falloff ratio follows right along. Projectiles have to do the most "compensating", hybrids less so, and lasers just basically hit out to the point where the beam is no longer focused enough to do damage. Therefore, your ideas are valid, but are already part of the basic design of the weapons.

  2. Anonymous4:20 pm

    Eh, I didn't do the math, but I noticed it and I've used this to my advantage. Its why I put an artillery howitzer on my Pilgrim when grinding low end missions - I had plenty of time between when they start moving towards me and when they actually reach me and begin circling to burn them down or at least chew threw most of their EHP. Then the drones cleaned up whatever was left.

    And that leads me to what I would argue is the easy way to fix this - AI needs to be programmed for evasive maneuvers when they are closing with targets. As for human pilots...well, they can be taught and trained. Sometimes.

  3. I think this is a difficult question. Generally speaking, a lone battleship is already pretty vulnerable to smaller ships. Like many of Eve's problems (ECM, supers) it only becomes a problem when you scale it up - putting multiple BS on the field makes it much easier to deal with small ships, since at least one is likely to have low angular at any given time. Like all of these problems, it's very difficult to find a solution that's balanced on one scale (e.g. 50 v 50 fleets) and not horrendously overpowered or underpowered on another (500 man fleets or solo for example).

    In some ways the fact that the tracking problem is fundamentally solvable (via teamwork, fitting, smart flying, etc) is a good thing - it adds a challenge that turret users need to overcome, and thus adds depth. A straight penalty for shooting smaller targets (like you get with missiles) arguably makes for more boring gameplay. The principle that smart flying and support fleets can help turrets perform well against small targets is exactly where we get things like small gang kiting, webbing support in BS gangs, and so on.

    Looking at missiles is probably not a bad place to start a discussion on this one. Missiles are currently implemented so that (within reason) you will never deal anything like full damage to a smaller target. You can still help the situation via painting and webbing, but only to a point. When was the last time you saw a fleet using large missiles - I'm not sure I've *ever* seen one! (other than bombers that is, which have the advantage of always picking their fights)

    Part of the problem there is that when you make a larger ship always ineffective against a smaller ship, you break on the complex rock paper scissors systems that Eve is based on. For example if a BS gang could never deal full damage to a HAC gang, I expect you'd quickly find that nobody PVPed in BS before and everyone started PVPing in HACs - to do otherwise would just be to open yourself to an easy counter.

    Is the current system perfect? No. But it does promote depth and complexity, and I don't feel that large guns or large ships are currently overpowered compared to their smaller counterparts. Tier 3 BCs themselves arguably are, but that's a conversation for another day.

  4. Anonymous5:56 pm

    I spent a good half hour plugging different values into the tracking equation and having wolfram alpha plot it. One thing I discovered is that it's much easier to get under someone's guns than to zip around at range and outtrack them.

    Consider a taranis tackling a 425mm AC hurricane with barrage, orbiting at 3500m/s at 20km out. ( http://www.wolframalpha.com/input/?i=y%3D.5%5E(((((1.7)%2F(x*.1))*((125)%2F(81)))%5E2)%2B(((x-3)%2F(18))%5E2))+from+x%3D0+to+x%3D30 ) Even though the ranis, zipping along at 3500 m/s is orbiting at an incredible 17x the tracking speed on the hurricane's guns and (at interceptor skill V) has a sig radius a bit lower than the sig resolution of the medium autocannons, the hurricane still has a 16% chance to hit. A lower skilled interceptors IV has a higher sig that bumps that number up to 24%. A cane that does 500 DPS on paper is still going to be able to deal a respectable 80-100 DPS (from guns alone) to a taranis. That's in ideal circumstances. Factor in the hurricane's own speed, slingshotting and slightly slowing down the transversal as the taranis travels the back end of the loop, and the cane will most likely do a bit more damage.

    Now let's look at that same taranis with close range tackle orbiting in close at 3km with mwd off. It might be webbed, so we'll halve the speed down to 260m/s. Now we've got the ranis outtracking the guns only by about 8.5x. ( http://www.wolframalpha.com/input/?i=y%3D.5%5E(((((.85)%2F(x*.1))*((125)%2F(36)))%5E2)%2B(((x-3)%2F(18))%5E2))+from+x%3D0+to+x%3D30 ) But at 3km, there's no chance the cane will even score the occasional hit.

    Other numbers (even with large artillery shooting an ares at range) follow this same pattern. If you want your small ship to stay safe against a single ship with large guns, it's usually much better to orbit in close than it is to zip around at range.

    1. From a design point of view it's inelegant but I have to agree that gameplay should trump physics.

      The ideal solution might be a gameplay design that was equally fun but which didn't leave a coin equally easy to hit as a barn door.

  5. the signature resolution of a large gun is about 400,
    and the radius of a FF is about 40

    It means that the large shot will be disperse in a circle with radius 400m.

    so even the FF stay still, it is possible for the large turret miss the target

    1. That is what you would think should happen, but the formula is definitely different :)

  6. My understanding:
    The "optimal" value is the distance where the various environmental effects can be either discarded or taken into calculation fully (like gravity, wind pressure, medium density as the ammo leaving the barrel has enough energy to overcome them) with supercomputers during aiming. "Optimal+falloff" is the value, where these effects can be more-or-less taken into account hence the reduced hit ratio.
    A computerized targetting system on the gun has more than enough time to aim during a reload cycle so if a target is stationary in reference to you, it's perfectly sane to assume it will be hit with any gun.

    As for your quarter and gun analogy:
    If your rifle is of a decent quality and your eyes and hands are good enough, you will be able to hit that quarter on the tree (assuming you're in the "optimal" distance of the rifle). As it's proven by snipers over the world dozens of times.

    1. Re quarter: yes, its dependent on my aim, but not dependent on how fast I move my gun.

  7. You wrote:
    "Its like saying that I should be able to hit a quarter sitting on a tree with a rifle because the quarter is not moving."
    So, why would you want to move your gun if the quarter is not moving?

    Maybe I'm not getting what you're trying to say. :-/

    1. I'm trying to say that there is a floor to how accurate my firing is, and that floor is not 100% accurate when the quarter stops moving. IN Eve, that floor is 100% regardless of target size or weapon accuracy.

    2. Anonymous5:31 pm

      That floor is 100%...out to optimal. Once you get into falloff, even if both you and the target are stationary, you'll still start missing.

    3. Weapon accuracy is represented with optimal and falloff, where optimal is 100% and optimal+falloff is 50%. So no, as araziah said, it's not regardless of weapon accuracy.

    4. Thus the problem: shouldn't the radius/resolution ratio impact the optimal/falloff calculations instead of the tracking calculations?

    5. That would have the following odd effects:

      1) Large sig resolution guns would have an easier time (than currently) tracking small targets up to their optimal, but a shortened falloff

      2) Small sig resolution guns would have a harder time tracking larger targets up to their optimal, but a longer falloff

      3) Increasing the sig radius of a target would not make them any easier to hit inside optimal, but would make it so that you could hit them farther away