Battlefield 4 and the Aim Assist

Guest writer, Hannes Burger, takes aim at the aim-assist in Battlefield 4.

The Journey Begins

A while back some subscribers to my small Youtube channel asked that I do a video discussing the aim-assist system in BF4.

I had the (incredibly misguided) notion that such a video won’t be of much interest, however I coincidentally happened upon a Youtube video apparently showcasing how extremely overpowered the Aim-Assist is in BF4.

The video itself was pretty much a carbon copy of every consecutive, biased video or forum post (God, help me) I had the misfortune of being subjected to thereafter.

It turns out that the topic of Aim-Assist in FPS games is a polarizing one to say the least.

I decided to dig a little deeper and soon I was in that dark, drippy part of the Internet where curiously angry people seem to spend their time – gaming forums.

Excruciatingly I read thread after thread of gamers pissing and moaning about Aim-Assist. The general opinion, it seemed, was that The Dark Lord himself commissioned Evil Game Developers to include the most over powered, damn-near-game-breaking game device since those f@&%!# Hammer Bros. in Super Mario.

1_HAMMER BROS

What surprised me most of all was the apparent sensitivity of these gamers’ senses. I have logged nearly 500 hours since the launch of BF4 and almost 700 on BF3 so I thought after well over a thousand odd hours of playing BF titles that I would have also encountered my cursor “swinging violently towards an enemy” giving me “god-like aim and accuracy” basically allowing me to “hack with the help of Dice” to quote some of my forum dwelling gaming brethren.

So I jumped into the in-game Test Range and I was flabbergasted at how obvious the Aim-Assist actually was, I simply could not fathom how I never noticed it in-game.

I immediately turned it off and jumped into a TDM, expecting to get my ass handed to me so completely I would have to delete my PSN account and burn the game disc – thing is I hardly noticed any difference at all. I played two full map rotations and averaged a 1.81 K/D, which is mere 0.1 below my global average. The only difference I felt was a slightly “looser” feel in extreme close quarters.

This merely compounded my confusion. The disparity between what I noticed on the test range and what I experienced in an actual game was massive.

So what gives?

I finally had the motivation to throw myself into researching exactly how the Aim-Assist works in BF4.

My first goal was to get Aim-Assist data from the actual game code or speak to someone at Dice to get exact values. I contacted EA and although I jumped through all the hoops they presented, I was not terribly surprised when, after a few weeks, they politely declined. In the interim I contacted the creators of symthic.com and this time I was pleasantly surprised when I received a very helpful reply. They sent me a folder from the game code titled “DefaultSoldierAimAssistData.dice”. Pretty much exactly what I was after.

The only problem is that I know less about programming than your average garden chair, which made interpreting the data a tad difficult. In Myffyli’s (one of the Symthic creators) own words “…it can be lots of assuming and guessing…” when it comes to extrapolating workable information from the game code.

Luckily I have the pleasure of knowing quite a few intelligent people with the needed skills and by committing unspeakable acts of bribery they helped me tease some useful information from the code.

2_SYMTHIC CODE

Before I continue let me cover my behind real quick:

Without certain data points and baseline values (to which we don’t have access) we cannot accurately deduce facts from certain parts of the code. Whenever that happened I decided not to speculate too much and rather leave that information out.

There is NO guarantee that the data presented to me by symthic.com is the data actually being used in-game. In fact it seems to be code from the game as it stood on release. The containing folder is labeled “14th Nov”. Regardless of this it does, at the very least, give us an accurate indication of what kinds of Aim-Assist systems Dice built into BF4 and how those systems function.

Some background

After reading the umpteenth nagging forum post I was ready to give up, find a shower, crawl into the fetal position, clutch a pool noodle and cry while the warm water washed away the filth of it all. However, as these things always seem to go, at the last moment I stumbled upon a joint research paper by students from the University of Saskatchewan and the University of Prince Edward Island titled “The Effectiveness (or Lack Thereof) of Aim-Assist techniques in First Person Shooter Games”.

This paper is a treasure trove of viable, unbiased, scientifically tested facts about why and how Aim-Assist is implemented in modern FPS games. I feel it is important to sum up the highlights of this paper and draw some of the information it contains in parallel with my goal.

The first question most people seem to neglect to ask is “Why does Aim-Assist exist?”

A modern FPS game’s success is dependent on one thing – player enjoyment. However, the modern FPS game has a steep learning curve and therefore a wide range of skill levels is present in any given multiplayer match. This can lead to both novices and very experienced players not enjoying the game as much. Good players want to be challenged and novice players don’t want to feel completely useless all the time.

The single most critical component of FPS games is aiming. Aiming speed and accuracy is also often the main differentiator between experts and novices.

Before we look at aiming a bit further lets take a very brief swing by Fitts’s Law. In short this law is a model for how human beings interact with computers in terms of “pointing” to a target on a computer screen using a pointing device.

The exact layout and workings of Fitts’s Law is beyond the scope of this article. All you really need to know is that very smart people have used Fitts’s Law to determine the following:

The time is takes to point at a target on screen is determined by how far away the target is and what size it is. The smaller the distance to the target and the bigger that target is the easier and faster it is to point at it using a control surface.

Lastly, using Fitts’s Law it has been proven (quite obviously I may add) that a computer mouse is probably the most intuitive input device you can use to point (aim) at a target on a computer screen, while console gamepads are quite the opposite.

Basically this means that PC players can overcome the skill range and learning curve mentioned above much faster than someone using a gamepad.

Lets imagine we have two people who have to navigate a very challenging test track in a car, but neither of them has ever driven a car before.

We give one person an automatic and the other a manual. It is fair to assume that the person driving the automatic will have an easier task as they can spend more time mastering the steering, throttle, brakes, etc. simply because they don’t have the added burden of also learning how to manually shift through the gears.

This might not the best image to use, but it does crudely illustrate the difference between novice PC gamers and novice console gamers’ ability to overcome the in-game learning curve. Not having to spend as much time first mastering the control surface (the mouse) gives the novice PC gamer extra time to start learning other game mechanics, like flanking, reloading etc.

Even after mastering a gamepad there exists one more crucial difference. An analog stick offers a limited, very small range of motion while the length of your arm and the size of your mouse pad are the only limits to a mouse’s range of motion. Imagine putting your mouse in a box barely wider that your hand and you can start picturing the limitations of an analog stick. Add to this the fact that you can only use your thumb to control an analog stick, as opposed to you entire hand, arm and wrist as with a mouse and we start to see why the learning curve on consoles are generally far greater than on PC.

3_CONTROLLER RANGE           

Console players therefore have to trade off sensitivity against accuracy. You can either turn and move fast, but overshoot your targets all the time or you can tone down the sensitivity allowing more precise aiming, but not be able to respond quickly to threats behind or to the sides.

It is for these reasons that Aim-Assist systems were implemented to console FPS games in the first place. Aim-Assist is meant to act as a “smoothing factor” allowing console gamers to overcome the inherent difficulties associated with using analog sticks for aiming.

Types of Aim-Assist

Various types of Aim-Assist techniques exist and have been present in a wide variety of games for years. The five main Aim-Assist techniques are Target Lock, Bullet Magnetism, Area Cursor, Sticky Targets and Gravity. It is incidentally these five techniques the above-mentioned students tested in their study.

Target Lock moves the player’s crosshairs to a designated hitbox on the target. A button on the controller or mouse activates the lock. The player’s character movement is manipulated to make them face the nearest target. This technique is very noticeable and not used in modern FPS titles. Those of you who have played titles like GTA and Red Dead Redemption will be familiar with this technique.

Bullet Magnetism “bends” a bullet towards a target if it is within the activation range. Interestingly this technique is active only after a player has fired the shot and is therefore hardly noticeable. Games like Halo and Gears of War 3 implemented similar techniques.

Area Cursor has a similar end effect to that of Bullet Magnetism in the sense that it essentially increases the target’s width and is only applied to traveling bullets. In essence the Area Cursor method increases the bullet’s size making it easier to hit nearby targets even if your aim was slightly off.

Sticky Targets (also called Reticule Magnetism) is present in many console FPS titles including Modern Warfare, COD and the BF series. This technique changes the control-to-display ratio when the crosshairs are over a target giving the effect that the target is sticky. Once you are over the target and the stickiness kicks in, it is easier to make small adjustments for a more accurate shot as the stickiness effectively lowers your sensitivity while it is active.

Gravity gives each target an attractive force that results in your crosshairs being dragged toward the target once you are within the specified range. The speed of this attraction varies from a slow movement to a near instant “snapping”. Console versions of games like COD and BF implement some form of this technique.

All of these techniques are to some degree negatively influenced by “distractors”, basically other nearby players. Distractors can cause the Aim-Assist to “target” a different enemy than the one you intended to shoot, but more on this later.

Aim-Assist in BF4

As far as we can tell BF4 uses two of these techniques namely Sticky Targets and Gravity, the game code labels these “StickyBox” and “SnapBox”.

The StickyBox system slows down the player’s crosshair when it is over the targets hitbox. What we cannot determine is by exactly how much it slows your cursor down. One possibility seems to be a reduction of 15%, although I would wager that it is slightly more than that based on in-game testing.

The StickyBox size (scale) is indicated to be 50% bigger than the normal hitbox along the x and z-axes and this seems to check out in-game.

4_STICKYBOX

SnapBox is in fact just a very aggressive form of Gravity we discussed above. The “LateralSpeedLimit” for the SnapBox in the game code is set to 1000.0 while the “SnapZoomTime” is set to a value of 0.2.

Although we cannot know with real accuracy what these values refer to it is safe to assume that they refer to crosshair movement in time and therefore indicate a near instant “snapping” of the crosshairs to the hitbox, which is backed up by in-game testing.

The last piece of information we can gather about the SnapBox is that the “MaxToTargetAngle” is set to value of 45.0. This seems to indicate an angle of 45° to be the maximum angle possible to still have your crosshairs snap to the target. My in-game tests seem to indicate an angle closer to 15-20°, but obviously that is impossible to confirm with any precision.

5_SNAPBOX

Various references in the code are made to something Dice calls “Attract”, but we could not bind this with any certainty to an Aim-Assist system in-game. It could merely be a characteristic descriptor for the other two systems.

Once again, I stand to be corrected on any and all of the above deductions/interpretations of the code supplied to me, if anyone has more knowledge or better insights regarding this, please speak up! 

In-game vs. Test Range (why I hardly notice Aim-Assist)

All this information did not really answer my question about the huge disparity I found between what the Aim-Assist seems to do on the Test Range vs. what happened in an actual game.

Once again the saintly research students gave me the key to the answer.

Two of their studies particularly interested me. They tested the performance of Aim-Assist techniques in a “Shooting Range” scenario with stationary targets and very little to no other game elements like lighting effects, distractors etc. As a counterpoint to this study they also tested the same techniques (used by the same players) in a “real game” scenario where the game had moving targets, lighting, sound, other players, etc.

What they found was what I expected all along. Nearly every Aim-Assist technique they used was quite noticeable and even distracting when the test players used them on the shooting range, however the opposite was true for the “real game” study where other game elements are present not only influencing the player’s experience but also influencing how the actual Aim-Assist systems performed.

After all the results were gathered the two “weakest performing” Aim-Assist systems were Sticky Target and Gravity, which is to say these two systems aided the players the least. These are also the systems most prevalent in modern FPS games on console.

This mimics almost exactly my experience in BF4. On the test range (or an empty server with a friend) where your targets are isolated and possibly stationary and with the lack of real gunfights, suppression, screen shake, dynamic effects and so on the Aim-Assist systems’ effects are grossly exaggerated. However when all these elements are present as they are in a real game and with the added “stress” associated with chasing objectives, trying to outsmart your opponents and playing to win, their effects are massively reduced.

One more reason I don’t notice the Aim-Assist in BF4 as much seems to be the fact that I go ADS 99% of the time. This means that I very rarely noticed any “snapping” of my cross hair to targets because I was already ADS when they entered my sites.

My Opinion

So what is my opinion on the Aim-Assist in BF4? Short answer: I’m OKAY with it, kind of.

Even though I still strongly disagree with all the Forum Folk and their incessant, whiny exaggerations I do have a few gripes with the system as it currently stands in BF4.

The StickyBox system works wonderfully and perfectly embodies what Aim-Assist should do in my opinion. Aim-Assist is there to subtly reduce the inherent difficulties of using a controller. It should be as unnoticeable as possible and should not effect your aiming decisions and target acquisition negatively. This way novice players can get acquainted with the aiming mechanics on a controller yet experts do not need to feel like the game is making decisions for them in some scenarios.

The SnapBox system is flawed in a few ways. In the right scenario the “snapping” motion of your crosshairs is noticeable in-game and it detracts from your experience. This can certainly be made much worse with certain play styles, but to be fair if you are constantly and rapidly going from hipfire to ADS back and forth (basically quick-scoping constantly), you are doing it wrong.

Most of the videos I mentioned at the start were of people playing in this fashion. Not only does it look like they are hopped up on PCP and the intravenous application of Red Bull, but it is a blatant attempt to exaggerate the Aim-Assist systems’ effect.

This would be the equivalent of someone engaging an enemy by firing 1 bullet from their primary then 1 with their secondary then 1 with their primary again and complaining that the switch over time between guns is too long when they eventually lose the fight.

But I digress…

I have two main problems with the SnapBox technique in BF4.

Firstly, for balancing reasons your crosshairs only ever snap to the “torso” or “body” hitbox of an enemy, regardless of where you aimed. This means that expert players who actively attempt to aim for the headshot are negatively affected when the Aim-Assist system moves their crosshairs down to the body.

Secondly, the angle at which the SnapBox system is activated is too large. Not only does the wide angle make the effect very noticeable but it also causes havoc when multiple enemies are in your sights. When their respective hitboxes overlap, the Aim-Assist can mistakenly snap you to an enemy you didn’t intend to engage. This can quickly ruin that beautiful 5 man back rage. 

Suggested Tweaks

So how would I tweak the Aim-Assist in BF4?

I would leave the StickyBox system pretty much untouched. Like I already said I feel this technique is the most subtle and perfectly embodies the original intent and purpose of Aim-Assist.

Regarding the SnapBox technique here are my suggestions:

  1. Reduce the maximum angle of activation down to around 5-10°. This would reduce the noticeability of the system and some of its inherent issues with distractors.
  2. Change it so that the system only activates when the crosshairs are in line with the body when in hip fire mode. This way you can aim for the head and not have the system snap your crosshairs anywhere. It also emphasizes the higher skill level necessary to acquire a headshot.
  3. Gradually reduce, and eventually remove, the SnapBox system as you rank up. Hypothetically you can have full blown SnapBox from ranks 1-20, reduced SnapBox from 21-50, much-reduced SnapBox from 51-80 and none at all from rank 81 and over. This gives more than adequate time for novices to get acquainted with the aiming system in BF4 and has the added benefit of gradually forcing players to acquire the fine motor skills necessary for good aim.
  4. Give players the option to select/deselect the Aim-Assist systems independently.
Conclusion

So now that my hands are cramping and my keyboard is smeared with the blood of my abused fingertips, here are my final thoughts on Aim-Assist in BF4.

I hope I have illustrated why Aim-Assist exists at all and that most developers, Dice very much among them, try and keep Aim-Assist systems as subtle as possible when they implement them.

I agree that Aim-Assist should be present to promote flow and game balance and cannot agree with the vehement complaints from parts of the community. Aim-Assist, in essence, serves the same function as Multiplayer Matchmaking (based on skill levels and stats) you find in other games. Aim-Assist is grouped under “Game Balancing Techniques” and should be viewed as nothing more than that. Something as fundamental to the game as the Class System is an elaborate way of achieving the same goal – balance. Giving players the option to select a class with unique weapons, gadgets and tactics that suit their individual play styles exist largely so novices and experts can play together with a smaller skill range.

With or without Aim-Assist good players will always do better than bad ones. If Aim-Assist was so crazily overpowered as it has been suggested then all console players will have MLG-like stats and console multiplayer matches will be one big kill streak for everyone, like a bunch of Neo’s fighting each other in the Matrix. The reality is that it simply does not play out that way. If everyone has Aim-Assist no one has Aim-Assist…

I can totally understand (and agree) how playing with Aim-Assist off can be seen as a kind of bragging right (I’ve seen at least two clans who claim to play with Aim-Assist off at all times) just like you have bragging rights if you dominate in close quarters with a bolt action rifle and a red dot sight, at least in my opinion.

In fact, I can see a few advantages to certain play styles to have Aim-Assist off, especially the way it works in the current build of the game. For instance, players who routinely pull of big flanks and aim to make big multi-kills, where target prioritization is key, will certainly benefit by not having the Aim-Assist snap them to the wrong target.

The bottom line is that Aim-Assist is necessary, but as with pretty much every device/mechanic present in AAA FPS titles, it is damn near impossible to make it perfect.

I hope this article and the accompanying video serves as a rational, thought out, well researched opinion on the Aim-Assist system in BF4.

Check out more BF4 related content on my channel here.

About Guest Zombie

Here today. Gone tomorrow. Always a zombie...