The map consists of two towers connected by a walkway and suspended in orbit above an Earth-like planet. It’s an exquisite setting, albeit with a quarter of a century less computer graphics development. The beauty of the setting is accentuated by the haunting soundtrack.

All of this beauty is, of course, merely a backdrop for the epic Sniper Rifle duels which the map is well-known for. Each tower provides about a dozen spots where you have an unobstructed view of the enemy tower and the route the enemy need to traverse to get to your tower. This makes sniping a guaranteed high-impact strategy: there’s no alternative but to walk onto the “sniper highway” that is the bridge connecting the towers. Of course the challenge is that while the towers give great vantage points, they don’t offer much cover or concealment². So while you’re busy plugging enemies on the far side, they’re busy finding and shooting you, which gives the map its cruel charm.

Today’s post isn’t about the ins and outs of sniper tactics and techniques, mainly because I’m unqualified in this area and can only give second-hand experiences. Suffice to say that camouflage and concealment are just as important as marksmanship for a sniper team.

But I want to focus more on the weapon itself: the UT Sniper Rifle:

There are plenty of these beauties and their and associated ammunition stashed throughout each tower, feeding the fuel of the famous sniper duels. Compared to the rest of the exotic UT arsenal, these seem unremarkable, boring even. They don’t fire bright bolts of plasma or glowing, exploding goo. They don’t cause big explosions. Even the shot’s report is dull and drab in comparison to the noises you get from more fanciful weapons like the Rocket Launcher or the Ripper.

Don’t be fooled: this weapon is perhaps the most unrealistic of the lot. We’ll discuss why in the next section.

…but only by ignoring the laws of physics

Using the Sniper Rifle in Facing Worlds, or any UT map, is very simple: right click to look down the telescopic sight, left click to fire:

This is a gameplay model of sniper rifles (or anything with a scope or sight) which is used in many other shooters. Aim, shoot, and hit (or miss). The game uses a simple model logic called “Hitscan” to determine whether each shot is a hit or not:

This is why bullets seem to cross the map instantly: the game models them that way. This, to state the obvious, is not how real-life bullets work. A real life-logic tree of whether a shot is a hit or a miss would look something like this:

This model was not something that the computer games of the 1990s were able to replicate, especially for rapid fire weapons. Add to this the complexity of multiplayer games, where PCs had to simulate gameplay and communicate and synchronise it across multiple machines over comparatively slow internet connections, and you can see why a simple model of bullet accuracy was preferable.

This began to change not long after the release of UT. Of course, modelling projectiles in video games was nothing new:

By the early 2000s, shooters were coming onto the market which modelled time of flight and bullet drop in a parabolic arc. Operation Flashpoint was a Cold War era West vs. East³ “realistic” shooter which featured large, open maps, actual squad tactics, driveable armoured vehicles, and simulated projectile parabolic arcs from the shooter to the target. Battlefield 1942 was a primarily multiplayer game pitting giant teams of soldiers against each other in large maps where the object was the gain and hold ground. It also featured a wide variety of driveable vehicles⁴ and the aforementioned parabolic modelling of bullet physics. Most of these game engines didn’t actually create a bullet as a physical object and model its travel through the air, rather, the engine did some clever math/maths⁵ to determine if the shot was a hit or miss and then modelled the damage appropriately.

Max Payne (2001) and its 2003 sequel⁶ took firepower physics to a new level, modelling each bullet as an in-game physical object which interacted with other in-game objects (e.g., you, in-game baddies).

Tristan Jung gives an excellent summary of bullets in video games on Medium. As he explains, it’s not a simple case of increased computing power always leading to more complexity. Some of the most popular newer shooters such as the Call of Duty franchise still use Hitscan logic in their game engines:

We’ll discuss the reasons for this in the next section.

Video games take shortcuts, but so does everyone

If we want to understand why game designers moved toward more complex models, one reason (at least with Max Payne and slow-motion bullet time) is for coolness. Slowing down time and making bullets more realistic can make gunfights far more satisfying and believable.

A bigger part of the answer, however, is the aforementioned bigger, more open maps. Take a look at this screenshot from Battlefield 1942:

With bigger maps, the flight time of a projectile becomes a lot more important⁷. Conversely, it really doesn’t matter so much on smaller maps and over shorter distances, because the time lag and parabola effects of bullets at short ranges are negligible. There’s absolutely no point in modelling an effect which the player won’t notice.

No matter how much resources you throw at a simulation, it will only ever be an approximation to reality. Ammunition designers will employ teams of designers and supercomputers to create a simulation of the real-life fluid dynamics⁸ affecting a bullet in flight, and still fall short of reality:

No computer game will ever have the level of sophistication to model shockwaves, fluid dynamic effects, spin⁹, the Coriolis effect, and the Magnus effect. Don’t know what these are? Don’t worry, you’re not alone. Snipers and artillery gunners have heard of these things, but don’t need to worry about them every time they take a shot. This is because they have tables of data which approximate the real world sufficiently to the task in hand. Need to know what ammunition to use? There’s a table for that. What charge should the projectile be, and what bearing and elevation should we set the gun to¹⁰? There’s another table for that. What if it’s really warm or cold: does the change in air density make a difference? Sure does, and there’s another table for that. There are lots of tables (it’s all the one page in the links above), and they account for most reasonable scenarios within a reasonable degree of precision, but they do not represent the “true” physics governing the motion and explosion of the artillery round.

A similar (but much simpler) table is below for a normal assault rifle round (.223″ Remington¹¹, courtesy of and transcribed from Ammo.com). This shows roughly how fast the bullet is travelling at various distances, how long it takes to get there¹², and how far the bullet drops in that time:

A few things you might notice from the above:

• The bullet falls a lot once it starts slowing down
• It also seems to rise before it falls—wait, what’s that about?
• It takes nearly a full second to travel half a kilometre

Here’s a though experiment: imagine you fire a bullet on a horizontal axis and, at the same instant, drop an identical bullet from the same height. Which one hits the ground first?

The semi-surprising answer is that both hit the ground at exactly the same time. The fact that the fired bullet is travelling at nearly a kilometre per second horizontally doesn’t have a bearing on how gravity accelerates it in the vertical direction¹⁴. But when we’re teaching soldiers to shoot in the real world, we use a very simplified model. Rather than learn about the parabolic trajectory of the round, they learn to assume a flat trajectory and no bullet drop at the “zeroed” range, which in our case was 300 m¹⁵. For any distance closer than this, the bullet was said to “rise”, so the estimated point of impact at 100 m was 10 cm above the point of aim. That’s exactly what you see in the table above too, although they have assumed zeroing at 200 yards (metres).

We also teach soldiers to fire at the centre of mass of the target, i.e. the enemy soldier’s midsection. This means that a bullet “rise” of 10 cm would hit the upper chest or neck, still a probable lethal shot (ignore the focus on “headshots” you see in videogames and films: real soldiers go for the centre of mass.). This simplified version of reality allows us to account for the real-world effects of gravity with the minimum impact on a soldier’s perception of how his/her rifle works:

Games which use hitscan mechanics might be taking a shortcut, but it’s one for which can forgive them. The effects aren’t likely to affect realism too much, and we take these shortcuts in real life as well.

Conclusion: All models are wrong; some models are useful

All models are wrong, but some models are useful

—George Box (statistician, 1976)

The above quote is one of my favourite from my engineering university days. It applies to every branch of the sciences, and failure to appreciate its truth is one of the leading causes of ignorance. It means that no matter how hard we try to simulate the infinite complexity of the real world, we will fall short. If we accept this limitation, however, and are aware of the limitations of our model, then we might get to do something useful with it.

Hitscan mechanics in shooter games is a perfect example of this: it’s a model which works just fine in most cases and allows us to program games with very simple mathematical rules under the hood. Unreal Tournament is no exception: it’s hard to notice the limitations of its hitscan mechanics in what is for the most part a corridor shooter, with fast-paced, high-octane, close-range engagements. Enemies are rarely far enough away for bullet projectile physics to become important. Facing Worlds is an exception: the instantaneous flight time from rifle muzzle to enemy head is quite noticeable. This works, however, because it’s such a balanced map where the enemy has the very same advantages as you do.

The Sniper Rifle in UT is one of the most unrealistic weapons in the entire game, but its unrealism is subtle. Like everything UT-related, we will happily park our beloved but often pesky physics to one side for a few minutes while we grab a [magic] Sniper Rifle and revel in the low-gravity, high mortality asteroid that is Facing Worlds.

That’s all for this week folks, and I hope you enjoyed it. Please let me know what you think in the comments below. And don’t forget to subscribe using the link below if you haven’t done so already. Finally, please share this on your social media feeds of choice. Thanks for reading, and I’ll see you next week.

Featured Image: The making of Facing Worlds, Unreal Tournament’s most popular map, Kyle Hoekstra, Rock Paper Shotgun