Game Design Review - On Guns P1

This is the first in a series of articles which I plan to post irregularly, perhaps once or twice a month, specifically relating to game design mechanics.

One of the first things I noticed about the FPSC: Reloaded engine was the initial guns it came with suffered from the usual host of faults.

Now the thing is,  games have had these faults for a long time.  I generally attribute these faults to developers who simply don't have the hands on experience of USING the actual firearms in question. For context, I'm a long time shooting enthusiast - your typical American Gun nut as it were.  I shot black powder competitively for over 15 years and did reasonably well, all told.

I've shot, firsthand, many of the guns used in video games (at least common ones).  And generally have shot at least the caliber or type/make/model of the gun in question.  I've read ream upon ream of ballistic data, penetration info, wound channel characteristics, stopping power and other items.

So let's start with the first and most glaring issue:

CROSSHAIRS

People - in general - don't have a crosshair built into their eyes.  Now I know this might come as a shock to you, but in the beginning - games didn't use crosshairs.  So it's my preference to remove it from my games.  This is not unprecedented, as games such as wolfenstein,

Now games generally tend to provide a crosshair as a holdover from that day when games only had hipfire but lacked any sort of system to use sights.  Nowadays games increase their complexity and realism.  They include accuracy mechanics, recoil, etc.  I'd say if you use a crosshair, a simple dot reticle provides a simple point of aim versus a full blown crosshair while not requiring a suspension of realism.

This brings us to our second issue:

ACCURACY

Much of the reason a crosshair exists is so people can hipfire and hit the target.  Most games include reduced accuracy for hipfire as a sort of artifical skill penalty.

Unfortunately over the years, this has become less and less understood and now people model the accuracy of one gun versus another.  Allow me to educate you on this:

Most guns are inherently accurate.  That's right.  Accurate.  While not amazing, most (non defective) guns retain sufficient tolerance that 'bad accuracy' means hitting something the size of a 7 inch plate.
Compare that to good accuracy, which would be within say a two inch circle.  Or excellent, where the shots overlap each other.

Rifles, in particular, are measured in terms of 'minute of angle' or MOA.  A minute of angle is 1/60th of a degree.  Consider that the average AR-15 is around 7 MOA.  This is 30% more accurate (seems like a lot, right?) Than your average AK-47, which is around 10 MOA.

We're talking the difference of 3/60th's of a degree here.  So within 100 meters/yards you're not going to see a significant difference in terms of video game accuracy.

Here's a real world example.   This is the shot pattern at ONE HUNDRED YARDS for an AR-15.

Courtesy of thetruthaboutguns.net

Consider the above - those thin lines represent a quarter of an inch. AT ONE HUNDRED YARDS. Using fairly generic hunting ammo.

Now compare that to a standard AK pattern:

From a shooter's forum, name withheld.  The above shows two separate groups of approx 3-4 inches spread.

There are literally THOUSANDS of 100 yard or meter patterns on the web for BOTH rifles that mirror the example pictures I took.  Now bear in mind this is at 100 yards.  Which looks like this:

Courtesy of thrumylens.org

So literally the difference of about two inches variance at one hundred yards, often on iron sights.

Compared to video games, which portrays accuracy like this:

I don't know what movie this is from, but I'd watch it.

The primary factors in missed shots (in the real world) are user skill (did they jerk the trigger, have a bad flinch, etc), technical malfunction (is the barrel loose, is it causing shots to tumble), and of course ... recoil.  Recoil is the physical action a gun takes after being fired.  Many guns use the energy from the bullet to drive a simple bolt assembly backwards to load another round and prepare the firing pin for another shot.  This is called a 'blowback' style gun.  The actual action of moving parts is a significant portion of where recoil comes from.  AR-15's use a lighter form called direct gas impingement whereas AK-47's are piston driven.  The AK has heavier, bulkier moving parts so you get marginally more recoil.  Recoil is also affected by the round itself.  More powder, bigger bullet, etc = more felt recoil.

Bear in mind most games portray AK-47's with completely unrealistic recoil.

From combat arms - and this is relatively TAME for a videogame AK-47.

When really... it's more like this.

OH MAN SO MUCH RECOIL HOW DOES HE HOLD ONTO IT

In context of the Reloaded engine, virtually every gun needs to be brought down to a zero or one value for accuracy.  At WORST a two.

Which brings us to the biggest habitual offender.

SHOTGUNS.

Shotguns in video games are consistently the WORST example of a firearm - EVER.  Even 'accurate' games like battlefield tend to be so woefully inept at portraying them.

For history, the shotgun is designed from revolutionary war-era shooting that basically was "If you can only shoot two or three times a minute, shoot lots of pellets in the hope of hitting.".  The shotgun is a very simple concept.  It's a long tube, typically smoothbore (no rifling) but sometimes includes rifling (for slugs - more on that in a minute) which fires whatever is in the cartridge.  You can shoot all kind of crazy, insane shit out of a shotgun.  Go to this youtube channel if you want a real good idea of just how nuts you can get:  https://www.youtube.com/user/taofledermaus

Your TYPICAL video game shotgun is based on using buckshot.
These are typically 33 caliber (a unit of measurement) pellets (smaller than a 9mm pistol bullet!) arranged in a cartridge with generally nine (standard 12 gauge) or sometimes up to thirteen (12 gauge magnum) pellets per shell.

This is what it looks like:

So firstly, at least get the pellet count right.  No meaningful defense round uses anything less than buckshot (unless they're using an oddball or they're one of those doofuses who uses birdshot).

Next, let's look at the accuracy.  Video game accuracy:

From: https://www.youtube.com/watch?v=eFZQ8NL6WhE

Now this doesn't look COMPLETELY bad right?  I mean .. it's about a three foot circle at 15 feet or so.  You can tell by the width of the body of the badguy and the shots that missed going over his shoulder.  Watch the entire vid for context.

But here's the truth.  Shotguns don't spread out THAT much.  They spread out about this much at fifteen feet:

From http://www.theboxotruth.com/the-box-o-truth-42-precision-shooting-with-buckshot/

So as you can see at 5 yards (15 feet) it's a 5 to 6 inch group.  That's literally a 600% over-estimation in group size by even the most tame video game developers.

In reality, most shotguns are effective out to about 35 yards (with buckshot).  I can't find the source but I recall reading that the british had tested it and found that you had a 30% increase in accuracy vs muskets out to this range.  Past that, it was exponentially less effective.

From bearingarms.com

Note - two to three hits in the critical zone(s) at 35 yard range.  So tone down your spread guys :)

Lastly...

Shotguns are not like this:

Yes, that's a real artillery launched nuke.

They don't have THAT kind of power.  Instead, think of it like a small caliber pistol (33 caliber, to be exact) of above average pistol velocity (typically 1300-1600 feet per second for a shotgun, vs 900-1100 for a pistol).  Compare that to say.. an AK-47 which shoots a 32 caliber projectile at 2200-2400 FPS!

So basically they're like shooting someone a small pistol round... up to nine times in one go.

So it looks like this in flesh at around the 10 foot mark:

Hint, it's the fourth one down.

You can clearly see the damage caused by rifles is pretty substantial, just like the shotgun at near point blank range.  So just try to factor that in when you are modeling your games.
I'll save my recommendations for the second part of this article.  Ta-ta for now!

Next time:  Magazines vs Clips, shells vs bullets, casings, slugs, and how to implement into your own games realistic gun configurations!

Technique Evaluation - Ground Clutter and Vegetation #1

In this series, I will be evaluating Gallery photos (freely available on tgc's website) which are highly rated.  These photos will be specifically selected for the purposes of learning or extracting a technique which I feel they exemplify.

This amateur art evaluation is as much for my own learning as it is for yours.  And while many of these photos will specifically exhibit multiple characteristics that are useful, I will typically focus on one element of it.   I plan on doing these evaluation posts once or twice a week, so far less frequently than store evaluations, at least for the time being.  In a sense, it's the best I can do at trying to give back to what I feel is already a very giving community.

Today's study is in Ground Clutter, it's uses as a means to improve a scene's depth.

When I first started making levels, I noticed that I had a significant problem in that my levels just felt cheap and amateurish.  These levels, I had hoped, would have a good feel and playthrough.  But even after throwing what I felt was a significant amount of clutter about (trash bags, rocks, debris, and other small scenery items) I was very frustrated to find that it didn't have the level of depth I was looking for.

Now take a look at this beautiful photo.

Author: Kajocat.  Skill level? Over 9000.

Compared to my tripe, this was the work of an old master vs a four year old's crayon drawing on a cardboard box.   Mine was on the cardboard, if that wasn't clear enough.  So let's evaluate this picture in depth for a moment here.   The trail leads you to the center of the screen, as a proper focal point should.  In terms of ground clutter items, there are approximately 15-20 bushes of four different types that I can discern.

These are skillfully placed around other objects, to add TO to those other objects.  Example: The house on the left.  It's got several bushes near it to help it from looking naked in the wild.  The trees also have small scrub bushes, as is often found in nature.   Indeed in my native home here in Pennsylvania, this is a very common sight - medium or large scale trees with small brambles or scrub bushes near the base.  The two pair together very well.  Finishing off the scene is some fairly expert use of vegetation.

The vegetation system in and of itself is possibly the best component of the new FPSC engine.  While the terrain system has a lot of flexibility, it clearly needs a bit of work to be up to snuff.  The vegetation system though is impressive in a lot of respects.  The sprites are animated, easily sizeable and colorable and there's a whole host of options for those who own the vegetation pack (Gold backers, for example).  I've found the stock settings are generally insufficient if you want to give your game depth.  One of the biggest 'uglies' in my work vs this picture was how objects would just touch the ground, leaving a discernable binary look to the graphic.  Like "This is the house" and "this is the ground".  There was no visual blending of any sort.  Then I realized... I could use the vegetation to blur that line.

Bare scene

Scene plus clutter

As you can clearly see, there's a marked difference on these scenes.  One is a simple 'town objects, nothing else.  The other is 'town plus desert terrain plus some clutter objects' scene.  The other is the exact same scene, and about 3 different types of plants, rocks, and scrub bushes

.This time, with plantlife.

The above here is simply the same scene with me spraying wildly with some stock vegetation settings plus a vegetation I felt looked good in the terrain.  While it helps and gives it an arid appearance, it seems very blocky and inconsistent.  It also makes clutter objects actually stand OUT more.  It's not bad though, and definitely an improvement.  It's not, however, quite what I'm looking for.


Now check out the difference when I plug these numbers to the maximum (Which is actually my preference, at this point).  Note, the final picture includes minor graphic tweaking to show you what a 'final' version would look like - example, higher bloom settings, adjusted ambient, etc.

The only thing different here is the size of the vegetation and the aforementioned graphical tweaks.  The vegetation values have gone from quantity/width/height of 50 to quantity/width/height of 100.   It's not perfect, but for what basically amounted to ten minutes of work I've created a rather convincing desert scene with significantly more depth.


The shearing of textures as they collide is much less noticeable and you get a much more professional flair with the addition of the vegetation.  On top of that the vegetation looks REALLY impressive with it's size at maximum.  It's flowing movement is much more noticeable in game.

Bearing in mind that this is merely an example; the proper value would probably lie in the middle between half and full settings for vegetation but these are things you have to experiment with to get the best results.  My goal here was to simply show you a process for improving your own scenes as best as possible.    If this was of utility to you, please let me know; I'd love to see some before/afters of your own.