Posted by: HOOF

Posted on: 15:56:29 11/12/98

Message:

I haven't gotten irrationally flamed in a while so I figured it was my turn now =)

Actually what I'm posting about is an explanation of the major changes in gunnery in WB with 2.5 and it's effects upon gunnery. People have been complaining about the changes and I don't feel that there is a good understanding of what the changes really are, so here I am.

Before I begin, let me mention a description of WB communications and net lag that I wrote before coming to IMOL. The URL is:

http://www.rdrop.com/users/hoofj/netlag.htm

and in it is a description of comms between front ends and the host, an explanation of net lag, and an analysis of the effects of net lag in the game (including why when you look back at a bogie and see him at range 12 in a 400mph chase he sees you at a much closer range, even range 7 depending on net lag).

Let me sum up the gunnery model before 2.5. Basically each plane had 2 banks of guns (ki43 is an exception). Due to CPU/Bandwidth limitations, each bank was limited to 8 individually tracked "projectiles" per second that can be fired. Thus firing both banks, a maximum of 16 "projectiles" would go out and potentially hit the target.

Obviously no WW2 planes fires exactly 16 "projectiles" per second, so in order to model accurate damage effects from a plane's main guns, each "projectile"'s lethality was adjusted to compensate. Say you have a bank of 4 .50cal guns in the secondary bank. Those .50cals can fire about 48 individual slugs per second (combined). Thus to fit that kind of firepower into 8 "projectiles", each projectile has to be effectively six .50cal rounds in order for the damage to work right. My (now obsolete) plane performance page at

http://www.rdrop.com/users/hoofj

has rounds/"projectile" stats for each of the 2.0 planes, which gives an idea of how much "damage" a single "ping" did for various planes with the old gunnery system.

Since muzzle velocity, gravity drop, and lethality is modelled in Warbirds, this model gave a rather accurate representation of "real life" WW2 prop plane gunnery. However, one critical aspect of "real" gunnery was missing, namely differing Rates Of Fire. Since the number of projectiles a real plane could fire was divvied up amongst 16 "projectiles" per second, this tended to group damage into individual hits, thus a FW190a8's cannon could inflict 6 20mm hits for *each* ping of the cannon every time it hit regardless of the situation. If this was a horizontal stabilizer, this ususally meant a dead plane, regardless of whether the target was flying through the bullet stream at 500mph rolling wildly, or was motionless to the FW's point of view.

In addition, the differences between machine guns and cannon were mitigated because two "projectiles", one from 4x.50cals, one from 1x20mm cannon could do about the same overall "damage" per ping/hit. In "real life" this isn't usually the case, because if the target is moving through the bullet stream when this hit takes place, those four .50cals would get scattered over the target while the 20mm would get concentrated damage when it hit.

In WW2 there was a serious debate whether a lot of high rate-of-fire machine guns were better/worse than few low-rate-of-fire cannon. On the one hand the machine guns gave a much higher chance of scoring a hit statistically in most situations, but the cannon tended to inflict far more critical damage when you did hit, with the machine guns being less effective when the target is moving through the bullet stream (a lot of damage in one spot on a plane is much worse than the same damage scattered all over the plane).

Now on to 2.5. In 2.5 the weapons system was revamped from the ground up and one of the major changes was the gunnery system. With increased CPU power and bandwidth (plus 3+ years of experience with online sims and how to make things efficient), it is now possible to model far more than 16 rounds per second for the main guns without choking the frame rate or modem. We are now able to model every round for recoil, ammo weight, roll intertia changes, and other plane-based physical properties to a much finer degree than before (no more rocking left/right when you fire the P51's guns now, since the recoil is more or less evenly distributed between the wings). However the technology isn't quite there to model every single round from high rate of fire planes, as far as tracking the rounds after they fire (at least without choking the frame rate everytime you touch the trigger). But, we are able to model a much higher rate of fire than before, which allows us for the first time to model a constistant number of rounds per "projectile" between all guns. The .50cal on a P47 hits with the same number of rounds per "ping" as a 30mm Mk108 on a 109K or Me262. A SpitFire Mk 1 hits with as many rounds per "ping" as a bf109f's single pair of 7.9mm. The *big* difference is that different planes can fire different number of "projectiles" per second, in fact we model this per gun, so that the same gun on one plane fires the same RoF as it does on another plane.

The ramifications for this are huge. Planes with lots of machine guns tend to have hosepipes now, firing many rounds into the air. Anyone who has taken statistics knows that if the probability of hitting a target is x%, then firing 2x the number of "projectiles" as before will yield approximately twice the hit probability (for low probabilities). Basically the planes with 8x.303s or 6x.50s are going to find it *much* easier to score hits than before (especially at range), because now you're throwing out several times the number of individually tracked projectiles as before. This is the #1 cause of the perception of easier long range gunnery. In 2.0 if you hosed away at range 9 and got 1 hit every 3 seconds, now you might get 3-4 hits or more. Naturally the machine-gun planes benefit far more because the cannon planes don't throw out as many rounds per second and are much closer to 2.0 as far as number of "projectiles" modelled.

Now for the flip side of this coin. It is much easier to hit with high Rate Of Fire planes, but the overall damage per second is exactly the same as in 2.0. What this means is each "hit" does less damage. For planes like the FW, or the P47, which had high bullet/"projectile" counts in 2.0, people will notice a huge decrease in damage/ping. In 2.0, an a8 can kill with 2-3 pings of cannon. Now it could take many more pings to do the same damage. In addition, the lack of concentration means that if successive hits don't hit the same location, you'll scatter damage around the target plane reducing the effectiveness of the hits. Thus shooting wildly gyriating planes can take quite a bit of time to down, where before a lucky ping or two would kill him.

To sum up, the new gunnery model gives a sense that it's easier to hit targets, especially at long range, however it also gives a sense that the guns are much less effective, unless you land concentrated bursts on a specific part of the target plane. Remember, however, that firepower/sec is exactly the same. Each plane has exactly the same damage potential per second as it had before. It's just divvied up amongst more "projectiles" than it was before. Rates of Fire and Numbers of Guns for planes now becomes an even bigger difference between the planes of Warbirds.

BTW, each Ping is still heard as a "Ping". People in the BoB will find themselves hearing a wall of Pings when hit by a spit/hurri now where before they'd hear only a few (when they die). Being attacked by an Me262 you'll only hear a few (maybe even just one) before being mortally wounded.

Warbirds is also the first sim to actually track more projectiles than you "see". You see a tracer for each 4-5 rounds "fired". We model other rounds "in between" these tracers as well, this is why you light up the target when you hit much more than the tracers would seem to imply. The statistics are 5 rounds per tracer for .30cal weapons, and 4 rounds per tracer for larger calibre weapons. These are historically common tracer/round counts. Again I wish to reiterate that we actually throw out more "projectiles" than this, while only display a fraction of them.

The ramifications of these changes are these:

People will find themselves able to hit at long range much easier. However their damage potential at range will be statistically the same.

People who tend to get the bulk of their kills at long range or vs maneuvering targets or when targets pass through their gunsight quickly will find that it's now much harder to score kills.

People who rely on close range, tracking/steady targets, and concentrate their fire in one spot will notice little change.

Planes will appear to light up much more on machine-gun equipped planes when you get a good solid hit. These same planes won't do much if you light em up with the same number of "pings" as in 2.0 since you are hitting with much less damage per "ping" than you would in 2.0. Most planes require many more hits than before to do the same damage, and require much steadier shooting to do the same damage.

I hope this clears up some questions people may have on 2.5's gunnery model.

Posted by: Dekker *401 Rams*

Message:

Do the bullets persist by time or by distance Hoof?

Posted by: HOOF

Message:

: Do the bullets persist by time or by distance Hoof?

Both. I don't know offhand how far out they fly, but I do know that the .30cal weapons have a shorter distance they fly before expiring.

This is a nasty subject in Ballistics because of the way bullets work flying through air. Basically there is this concept, called the Ballistic Coefficient, which governs how well a bullet holds its speed over distance, which then affects things like how much it's affected by relative wind and other things. As any long-distance shooter probably knows, spin-stabilized projectiles are less affected by relative wind than fin-stabilized projectiles (like the M1 tanks Sabot round), because the fin stabilized projectiles tend to line up with the wind (which is why it's important on an M1 tank to know which way the wind is blowing when firing at a target :) The reason I bring this up is that bullets tend to do nasty stuff when they slow down into the transonic region (the speeds near the speed of sound), and anything that causes them to fly off the perfect flightpath gets exasperated drastically decreasing accuracy (this is why the .22 round commonly used by kids as their first "rifle" deliberately keeps the bullets subsonic, as increasing muzzle velocity even by 100fps causes an alarming decrease in wind deflection tolerance). In addition the slower the bullet travels, the less damage it does, in fact its kinetic energy is proportional to the square of the speed. Unfortunately we aren't this detailed (yet) in WB, so to "simulate" this phenomenon we give bullets a "lifetime". For example, .303s have a (relatively) poor ballistic coefficient, going from approximately 2500fps to the transonic region (1100) in less than 800 yards. This is the #1 reason in WW2 why they weren't as effective for long range shooting as, say a .50cal or 20mm. Going from 2500fps to 1100fps loses over 80% of the hitting power of the round (based on the Kinetic energy of the round which is where the damaging energy for solid slugs comes from). So even if you hit a guy at 800 yards with .30cal weapons you'll probably watch as the bullets merely bounce off his plane.

50cal and 20mm have much higher ballistic coefficients, in fact if I'm not mistaken they can go 1500 yards or more before slowing down to the transonic region, but even then, a .50cal loses approximately 50% of it's killing power over 800 yards or so (requiring many more hits to do the same damage). Cannon have an advantage over machineguns in that at long range they still have just as much explosive firepower as at close range, though some cannon like the 30mm Mk108 have time fuses to blow up the shell at long range (makes sense, U aren't gonna hit anything no point in dropping explosive shells into the landscape).

One of the "biggest" reasons long range gunnery wasn't too effective in WW2 wasn't just that you couldn't hit at that distance, but because the odds of doing any serious damage is rather low. Only cannons could reliably kill at long range when you hit, but even then very few planes had the ammo load to waste shooting at such long ranges, especially since it a) alerted the other guy you were there, and b) getting in closer guaranteed a kill more often. Maybe a Fw190 would perish with only 20 or so .50cal hits at 150 yards, but at 800 yards it might take 50 hits or more, and good luck getting enough hits on him to do that :)

We actually model this aspect in WB better in 2.5 than in 2.01 because at those ranges the odds of 2 different "hits" landing on different parts of the plane are very high. Unless you get lucky and hit the pilot, you're much less likely to land a concentrated burst at 900 yards than at 200 yards, so it takes many more rounds to chop up a plane at 900 yards. With 2.01's hit packeting you might land 6 .50cals in one spot at 900 yards due to the limits in the numbers of hit packets, but in 2.01 the odds of getting 6 effective .50cals in the same location at that range is very low. Unless you put a heck of a lot of rounds out there :) This does *not* mean that you can't die at 900 yards as many people have complained about :)

As I mentioned in my previous post, due to technical limitations we can't do the Holy Grail of gunnery models, so one of the things we do is bullet lifetimes. We gotta stop tracking the bullet sometime, since tracking 500+ bullets in flight most of them many seconds after firing, would reduce the framerate to unacceptable levels. This is why the bullets "run out" at the distances they do (again I don't have the exact numbers for you unfortunately). I believe we have a good compromise between framerate and letting the bullets live long enough to give a good full range of potential hits.

Posted by: HOOF

Message:

: Therefore in the past, less computor/bandwidth power, less projectiles with more punch/projectile. In the future, higher computor/bandwidth power more projectiles with less punch/projectile = more realism.

: correct?

Sounds right :) One day I hope to see a sim with the holy grail of Gunnery models, modelling every round with ballistic drag, differing trajectories for different round types, bullet dispersion, full kinematic collision model using actual kinetic energy for damage calculations and realistic shell explosion fragmentation modelling, bullet transition through planes including deflection and richochetting off differing material, full mach zone drag and deflection modelling, gun heating, rates-of-fire affected by G's and gun heat, jam model, ammunition feed modeling, charging delays/problems, high alt cold jamming problems, real propellor-arc gun syncing, etc etc etc.

That day will be a glorious day in Flight Sims =) For me Warbirds is a good approximation for now :)