Editorial - Ultra-Realism in Flight Sims: Myth or Reality?

One of the questions I'm often asked by members of the flight sim community often goes like this: "What is the most realistic sim?"

In my mind, there is no clear cut answer. The definition of "realism" is often interpreted differently from person to person. Is it a matter of visual fidelity, aircraft system depth, "feeling" of flight, using standard operating procedures (SOP), weather effects, air traffic, aircraft component malfunction, damage model, sound effects... or is it something else entirely? More often than not, the encyclopedic debates on various forums end up with a locked thread with everyone throwing arguments in every direction while screaming at each other something like "but MY sim is better than YOUR sim". This time, I'd like to try something different.

A flight sim according to engineers

If you were to create a simulator from scratch, how would you start? Well, you'd likely need to create a list of requirements that need to be met. A requirement can be something as simple as "simulate an aircraft" or as complex as "simulate the left pitot probe pressure input to the flight computer in icing conditions during a blood moon".  Good project managers know how to split requirements in terms that make sense and can be divided into tasks and sub-tasks. These tasks can then be assigned to various engineering teams, which are responsible for specific aspects of a simulation. This creates a need for specialized teams (graphics engine, aircraft systems, user interface, etc.). The simulation of an aircraft itself is typically split based on the reference documentation available in terms of systems and functions. In the civilian world, these are often associated to ATAs (Air Transport Association) chapters/numbers. Think of it as a giant laundry list of systems to "replicate" within a simulation environment. Budgets and schedule estimates are then generated by management and engineering teams, and then these need to be approved by whoever is in charge before the work can begin.

This brings us to the "simulation" aspect. How do you qualify something as realistic or not? How do you draw the line on where the scope of a project begins and ends? The answer is "it depends". For real world pilot training, there are standard certification levels that have their own lists of requirements to be met. These certification levels depend on what kind of training and proficiency level the pilot is expected to accomplish with the training device and what data set is used to "match" aircraft behaviour. These data sets can be flight test data recorded on a real aircraft, wiring diagrams from OEMs (Original Equipment Manufacturers), flight manual excerpts or charts... Aviation certification authorities have their own method of defining realism, which can be simplified as "does it meet pilot training requirements?".

However, simulators accessible to the general public cost a few dozens of dollars... while commercial simulators used for pilot training schools can cost in the millions of dollars. Why is that? The cockpit hardware is one of these reasons, but also the development costs don't involve the same amount of resources. Data isn't "free". Manufacturers know that their IP (Intellectual Property) includes those precious design documents that are needed to create a simulation. Aircraft integrators gathering flight test data isn't cheap either; you're basically asking a whole crew to install a bunch of sensors everywhere in the aircraft and to fly through as much as the flight envelope as possible... for hours and hours. And even then, there are edge cases that the pilot may not want (or have permission) to do since the aircraft may not be certified for these cases. All of these mounting costs make the creation of a realistic simulator extremely prohibitive for small businesses and studios. Therefore, a simulator sold for 60 $ will inevitably have to make assumptions and generic models that simulate the aircraft in a believable manner with the data they have, which is likely incomplete or simply not available due to legal restrictions put in place by companies or governments. A very important thing to note: because something is on the Internet doesn't mean you can use it. Using documents leaked or stolen from manufacturers is a crime and has serious legal consequences. This is why there are whole departments dedicated to obtaining documentation and dealing with various manufacturers.

What is "Ultra-Realism"?

Now... where on earth am I going with all of this? Well... the flight sim world is a complex jungle with tribalistic communities. You'll inevitably run across the new guy who breathes heavily in his microphone as he struggles to find the brakes, the lone wolf who shoots down the entire enemy team in a matter of minutes, the ==\|=Virtual_Squadron_Members=|/==, and other archetypes. One of these clichés is the "milsim" crowd. These people want to push the hobby to the next level. This means flying in an organized fashion, planning missions, flying attack profiles, communicating with brevity terms... basically trying to play more realistically than the average player... and there's nothing wrong with it. It can be fun to roleplay.

But... there is a specific school of thought that goes even further. That's the "Ultra-Realism" bunch. They want to emulate what it's like to be fighter pilot, to a point where they pretend that what they do translates directly into the real thing. Some of their dogmas include (but are not limited to):

• Anyone who does not use as a primary source of information the real NATOPS/NAVAIR/FCOM/obscure-document-he-got-from-wikileaks is garbage.
• Anyone who does not say a 9-line correctly is a shitbird.
• Having questions means you haven't spent enough time RTFM'ing (reading the f***ing manual).
• Anyone who does not perform SOPs (standard operating procedures) in any given situation doesn't know anything, and is therefore worthless.
• You're not playing a game. This is basically like real life.

I'm sure you've run across that one person at some point. It's often difficult to understand this mindset since it often involves long-winded discussions with verbal diarrheas that go on for pages and pages, throwing giant chunks of manual excerpts left and right. Since this is an editorial, I can only share my own opinion based on my own experiences on developing and using simulators on both the lower and higher end of the realism scale.

Get to the point already!

Yes, yes, I'm getting there. In my opinion, I think "ultra-realism" is a delusion that is closer to myth than reality. Flying with a 30-inch monitor and a gaming chair isn't the same as pulling Gs in an aircraft. There is a physical aspect to airplanes that cannot be replicated due to mechanical limitations that are present even in high-end simulators. There are a number of limitations that sims accessible to the general public have that cannot be overcome by sheer willpower.

Money has a big part in it. Most of the systems are approximations by design since the cost of having all the data required to make it "as real as it gets" is made deliberately prohibitive by IP holders. There is a reason why flight sim studios are small in scale and have relatively slim profitability margins. Licensing costs can be very high, and finding the right person for a system can be more difficult than meets the eye since it involves a high degree of specialization in both physics and programming. And even then, you need a useable platform to work with... which brings us to my second point.

Technical limitations are often seen as an easy cop-out, but it's often more complicated than that. "Technical debt" is a concept in software development that reflects the implied cost of additional rework caused by choosing an easy (limited) solution now instead of using a better approach that would take longer. Most flight sim engines are decades-old and are sometimes limited by the way they were initially designed. Working with older codebases is often quite challenging since it can become gradually more difficult to re-write certain features without having to re-write heavily integrated core elements of the engine... a bit like playing a game of Jenga. Most flight sim developers do not have the deep pockets of triple-A video game studios to re-write an engine from scratch, so this idea that you can just shake a magic wand and say "let's do it all over again, but better" isn't common practice for developers who have accumulated years and years of technical debt.

This means that certain solutions, while "more realistic" in nature, sometimes need to be simplified in order to avoid an excessive performance cost. As an example, modern aircraft distribute information with avionics protocols. As a safety measure, redundancy requirements have controllers distribute data across multiple channels, "duplicating" the amount of information transmitted. This can have a significant cost in performance/computing time. One of the biggest challenges of simulating an aircraft is how to simulate a component breaking in the aircraft. What if an engine ingests a bird? What if an aileron is jammed? What if a missile explodes nearby and shrapnel clips a part of your wing? What if certain avionics buses are failed but others are not? What if you have a gradual engine performance degradation? How will the aircraft behave if exceeding rated G limits during flight or exceeding sink rate limitations during touchdown? How can we even evaluate if the malfunction modelling is accurate or not? Flight safety is no joke; the list of possible malfunctions is mind-boggling, and real pilots are trained to perform adequate corrective actions to either mitigate the problems or save the aircraft in cases of catastrophic failures. Modern planes have many automated system control loops that perform some of these corrective actions automatically, but the ultimate equalizer is the man or woman behind the stick or yoke. Part of the job of being a pilot is to know how to act when things don't go according to plan. This is one of the primary training goals of a simulator, and in my humble opinion one of the reasons why "ultra-realism" isn't a thing for simulators accessible to the general public. The development cost to meet "ultra-realism" standards vastly exceeds what average consumers like you and I can afford.

This doesn't mean that flight sim developers can't get close to this imaginary realism line, or that they can't simulate certain aspects of a plane within good tolerance margins... but you cannot have something that isn't somewhat simplified either for gameplay purposes or for budgetary reasons. In the end, pushing the limits of an engine is an uphill battle that is mainly driven by the passion and dedication of flight sim developers more than monetary incentives. There is no criteria set in stone that will determine if something is "real" enough; that is up to the consumer to decide based on what he is willing to pay... and it is also up to the developer to decide based on what is feasible within his means and ability.

So, those were my two cents. If you have anything to add or comments, please share them with us below!

Cheers everyone!

Chuck