Changing Rapid Fire

[vicky_molokh]

Originally Posted by Tyneras

I would rather have the correct results rise organically out of the rules than to have to put band aids named "common sense" over the problem. Makes making GURPS Software much easier.

I now look at my temporarily suspended GURPS Spaceship Battles executables in anticipating horror. I never considered what a large system of cases and exceptions is in demand.

[Ulzgoroth]

Originally Posted by Tyneras

I disagree. A bigger circle over a target means that there is more empty space for it to miss.

It depends on the relative contributions of missing due to your cone of dispersion being wider than the target, and missing due to your cone of dispersion not intersecting the target.

If you always hold a perfect bead, you want a circle smaller than the target but don't care how much. If your aimpoint drifts, but always stays within the target, you want a circle smaller so that the edges of the cone don't go outside the target. If you can't count on your aimpoint being over the target at all, though, a larger dispersion may result in getting more hits.

Just like a shotgun, it can be easier to score some hits because the shots are spread out rather than all going to one place.

[vicky_molokh]

Originally Posted by Ulzgoroth

It depends on the relative contributions of missing due to your cone of dispersion being wider than the target, and missing due to your cone of dispersion not intersecting the target.

If you always hold a perfect bead, you want a circle smaller than the target but don't care how much. If your aimpoint drifts, but always stays within the target, you want a circle smaller so that the edges of the cone don't go outside the target. If you can't count on your aimpoint being over the target at all, though, a larger dispersion may result in getting more hits.

Just like a shotgun, it can be easier to score some hits because the shots are spread out rather than all going to one place.

This made me wonder: if given the range of minimal and maximal angular offset due to recoil (per shot), and assuming that the direction of the angular shift is random, how does one calculate the size of the cone. It is not simple multiplication of average shift by number of shots, because there is a chance that the shift will return a successive shot closer to the point of interest.

[DouglasCole]

Originally Posted by Tyneras

Here is my logic about the whole system.

Think of the target as a 2d silhouette.

Your ranged weapon puts a circle (expanding the farther away from the barrel it gets) of bullets/lasers/whatever out.

FYI, from the not small amount of work I've done on GURPS ballistics stuff, this tracks.

Your Rcl (scatter) defines how big that circle is. A Rcl 1 laser has a tiny circle, a Rcl 3 weapon has a much bigger circle.

Ironically, this does not.[*] The circle your weapon fires into is best represented by the Acc number. If you see the optional rule in Tactical Shooting called "Minutes of Angle," I found that GURPS Acc translates VERY effectively into MoA. It's certainly good enough for a roleplaying game, but what I'm saying is that it's, in a way, better than that. The Acc numbers map very well to how big a circle various weapons will fire into at 100 yards.

Your skill with the weapon dictates how well you put that circle over the target. A success means the silhouette is inside the circle to some degree, a failure means the silhouette is outside the circle. Moving and/or distant targets make putting/keeping the circle over the target harder. Dodging, travel time, etc. must all be accounted for if you want to get that deep into detail.

If this logic holds, then bigger targets should both be easier to hit, but if the target is big enough, reduce effective recoil. If the silhouette I'm firing at is so large that my scatter circle is smaller than the silhouette, then if I succeed at shooting it (accounting for movement speed, target fails to successfully dodge, travel time accounted for, etc.) then every bullet/beam should hit, no matter how many I fired.

I approached this once, in 3e, by calculating with a spreadsheet the fall of each shot. The Acc of a weapon translated into an inherent circle limit for the weapon, while the accuracy of the SHOOTER Root-Sum-Squared into that circle another amount of scatter. I would let each shot deviate with that error, with the aim-point moving based on some funky combination of the 3e Recoil statistic and the shooter's ST.

There certainly might be a way, though it might involve a table lookup, to see what percentage of shots fired in a burst SHOULD hit (raw stats: your dispersion is 50% bigger than your target; you expect half your bullets to hit). Then you'd have a roll to determine how many DO hit. Close range and big targets would move the "floor" for that number up, pure luck can probably deal with exceeding the average by a large amount.

It's a non-trivial problem, as the many pages of discussion here indicate!


[*]Edit: Sorry, "this does not track for a weapon clamped in a vise." is closer to the truth. The higher the Rcl stat is, the higher your skill probably has to be to keep the aim-point where you expect it to be (if it was on-target) or from wandering on follow-on shots (if it was not).

But . . . the inherent "circle" the weapon fires into is the Acc. Where your aimpoint goes after you use your skill to sight in and fire, that's probably Rcl.

[Ulzgoroth]

Originally Posted by vicky_molokh

This made me wonder: if given the range of minimal and maximal angular offset due to recoil (per shot), and assuming that the direction of the angular shift is random, how does one calculate the size of the cone. It is not simple multiplication of average shift by number of shots, because there is a chance that the shift will return a successive shot closer to the point of interest.

I don't think the model behind that is accurate. The gun is shooting a spread pattern, but it doesn't follow that it's jittering around in response to the recoil from each round. If nothing else, over the course of a second the shooter is surely having some effect on the aimpoint, not just letting it drift freely.

Some guns, of course, consistently rise due to recoil. I don't think that's applicable to the support-weapon types we're mostly thinking about here though.

Originally Posted by DouglasCole

Ironically, this does not. The circle your weapon fires into is best represented by the Acc number. If you see the optional rule in Tactical Shooting called "Minutes of Angle," I found that GURPS Acc translates VERY effectively into MoA. It's certainly good enough for a roleplaying game, but what I'm saying is that it's, in a way, better than that. The Acc numbers map very well to how big a circle various weapons will fire into at 100 yards.

I think it may be dangerous to not clearly distinguish single/first shot cones of fire from cyclic shooting cones of fire. An M2 Browning can be used as a sniper weapon, but in automatic fire I'm fairly sure it'll lay a cone similar to that of other machineguns.

[apoc527]

How do games like Battlefield 3 or Modern Warfare do it? I believe they use the computer to calculate aim point, movement of the aim point due to recoil, and shot dispersion applied to each bullet fired while the aim point is shifting.

That's not something I want to do in an RPG!

The GSS issue is merely the same symptom of the system's issue with large numbers. I to, however, believe it is meaningless to compare hit percentages with single shots to those of auto fire (I don't think you were doing that either).

As far as GSS is concerned, the mistake there is treating individual shots as a form of Rapid Fire. If you gamed out those attacks with one second rounds, thereby getting many attack rolls, many shots will hit. I just think they went with a quick-and-dirty approach by co-opting the Rapid Fire rules for something that's not true Rapid Fire.

As for mini guns, I don't know how you fix those. Use Suppression Fire mostly, I guess, and make sure to Aim at actual targets you want to hit.

Finally, let me just say that us never want to actually roll damage for the whole minigun burst in the first place!

At any rate, I'm no more convinced now than I was earlier that these rules fail only for certain unusual circumstances but otherwise work fine (excepting the plain misuse of them in GSS of course).

[DouglasCole]

Originally Posted by Ulzgoroth

I think it may be dangerous to not clearly distinguish single/first shot cones of fire from cyclic shooting cones of fire. An M2 Browning can be used as a sniper weapon, but in automatic fire I'm fairly sure it'll lay a cone similar to that of other machineguns.

Yeah, you have to read my addition that I put in. Under perfect clamping conditions, a weapon will still shoot into a cone. With (occasionally deliberate) loose mounting, this cone will widen. Applying Rcl as an effective decrease to Acc (or some function of Rcl) in auto-fire situations would be an interesting approach. Not sure how fruitful it would be.

After TG playtest is done, and I finish my alternate GURPS article on (something), I'll have to dork around with this again.

[Ulzgoroth]

Originally Posted by DouglasCole

Yeah, you have to read my addition that I put in. Under perfect clamping conditions, a weapon will still shoot into a cone. With (occasionally deliberate) loose mounting, this cone will widen. Applying Rcl as an effective decrease to Acc (or some function of Rcl) in auto-fire situations would be an interesting approach. Not sure how fruitful it would be.

After TG playtest is done, and I finish my alternate GURPS article on (something), I'll have to dork around with this again.

My conceptualization was that even if your machine gun is in a clamp, you'd get a different cone from a series of well separated single shots than from a long burst. Though your suggestion of a mounting producing the same effect is interesting.

[apoc527]

Originally Posted by DouglasCole

Yeah, you have to read my addition that I put in. Under perfect clamping conditions, a weapon will still shoot into a cone. With (occasionally deliberate) loose mounting, this cone will widen. Applying Rcl as an effective decrease to Acc (or some function of Rcl) in auto-fire situations would be an interesting approach. Not sure how fruitful it would be.

After TG playtest is done, and I finish my alternate GURPS article on (something), I'll have to dork around with this again.

Correct me if I'm wrong, but aren't many modern guns so inherently accurate (i.e. very small MOA) that the cone isn't going to be very large? I thought I read somewhere that a fully stabilized modern, long-barrelled minigun can shoot hole-in-hole out to fairly impressive ranges. In other words, the dispersion of shots is less than or near the diameter of the bullet.

I think a key point here that has been mentioned, and perhaps unfairly dismissed, is that we need to define the circumstances under which one should actually use the GURPS Rapid Fire rules. If I clamp a minigun to test its accuracy, why would you use the Rapid Fire rules? They clearly don't make sense in that situation.

If I'm on a Blackhawk helicopter and trying to hose down enemy jeeps driving on the road below me with the mounted minigun, perhaps the Rapid Fire rules make a lot more sense and create a reasonably close hit percentage to whatever reality is (do we really know?).

So, while this discussion is hugely interesting, what is the practical, game-level effects that we really need to be worrying about?

And what are the right mechanics for realistic use of an LMG at long ranges? I believe one should use Suppression Fire and probably steer away from directly targeting individuals with LMGs at long range, but perhaps that's a naive viewpoint.

[apoc527]

Originally Posted by Ulzgoroth

My conceptualization was that even if your machine gun is in a clamp, you'd get a different cone from a series of well separated single shots than from a long burst. Though your suggestion of a mounting producing the same effect is interesting.

The recoil has nothing to do with the gun's inherent accuracy. The "circle" of dispersion in a perfectly clamped gun will not vary if the weapon is on single shot or fired on full auto. The main exception arises when a weapon fires from a closed bolt on single shot and an open bolt on full auto. THen you'll get different dispersion amounts for sure.