Opinion: Microsoft and Asobo Studios Don’t Understand Flight Simulators
I’m not a fan of Microsoft Flight Simulator 2020 as a flight simulator. It has many shortcomings in the realism of its flight model and systems, and fails to live up to the benchmark X-Plane 11 set six years ago.
I think, since the days of Flight Simulator X, Microsoft has forgotten about what makes a good flight simulator, and FS2020’s developer, Asobo Studios, doesn’t know what one is. This isn’t about the simulation realism, it’s about the systemic flaws in the gameplay experience that show a lack of understanding of the fundamentals of the experience of flight simulation.
A flight simulator is as much about escapism as it is about realism. It’s about being able to get a diversity of experiences you can’t (or are unlikely to be able to) get in the real world. How many people have flown the Red Bull Air Race, dead-sticked an airliner, water-bombed a forest fire, and landed the Space Shuttle? I’d wager there isn’t a single person who has done all of those things for real. But in a simulator, you can — and should — do all of these things. Not only are these experiences fun, they’re eye-opening. And in a flight simulator, they should just be a few clicks away.
Unfortunately, I don’t think FS2020 comes close to capturing that diversity of flight simulation experience.
Hold on, I hear you say, haven’t I previously praised FS2020's incredible open-world visuals that make you feel like you’re really there? Yes, I have. But scenery is alone isn’t escapism in a flight simulator. Think about it this way: scenery is the garnish to a dish, the dish being the flying itself.
A flight simulator’s escapism is also about the flying experience — not just the places you fly over. I want to feel the grit of flying VFR into a no-go-around back-country strip, and the geek-out over a CAT III autoland in a modern airliner — I want these things to evoke distinct feelings in me. I’ve never done these things in real life, and I’m not qualified to either, but the reason I love flight simulators is that I can feel what it would be like to be able to.
Unfortunately, FS2020 falls well short of providing genuinely diverse piloting sensations. I think this is down to three things: its misguided and misunderstood default aircraft selection; and its inexplicable choice of gameplay challenges; and its misconceptualisation of flight instruction.
Aircraft Selection
The standard edition of FS2020 comes with 20 aircraft, increasing to 25 in the deluxe, and 30 in the super-deluxe edition.
Before getting into the issues of the aircraft selection, I want to say something about these tiers. I‘m far from against payware, but what Microsoft have done with their buying options feels dodgy. The deluxe and super-deluxe editions are basically $30 payware bundles of aircraft and airports, which seems ridiculously expensive for “default-ware”. I only own the standard edition, so I don’t have actual experience of these aircraft, but as far as I understand, their modelling quality is the same as the aircraft in the standard edition.
In my view, the point of “real” payware is that you pay (usually $30+) to get one highly detailed, highly accurate aircraft that gives far more depth to the simulation experience than default aircraft provide.
For example, the XP11 Flight Factor 757–200 (in the version I bought) costs $60. It’s a lot of money, but you get a specially designed 757 FMS (not XP11’s default X1000 system), fully configurable ground support vehicles, a unique systems failure model, a customisable avionics suite, and a virtual cockpit where you can interact with every button and switch. You can also change the wing-flex and the amount the passenger blinds move up and down.
I think payware is about “more simulation” not “more planes”. By locking certain aircraft with the same depth of flight/systems modelling into paid tiers, Microsoft are gouging. If they offered “extra-detailed” planes in the higher editions — like a study-level C172, for example — I wouldn’t have an issue. But simply offering “more” of the same modelling looks to me like a transparent money-grab.
Also, word of advice on the editions: there is no easy way either on Steam (the platform I use) or Microsoft’s own website to upgrade the version you own. So, if you do want the “OEM payware”, just buy the version with the plane(s) you want.
Digression over. Back to the planes themselves.
The FS2020 “hangar” is intended to cover a wide variety of types, from ultralights and GA, to transport category. Notably however, it lack a glider — found in FSX and XP11. XP11 also includes a fast-jet, the F-4. Now, I will admit the F-4 isn’t the best plane to fly as most of its systems haven’t been modelled properly, but the ASK-21 is a fun and distinct addition, and one I can’t help but think would take full advantage of FS2020’s fantastic atmospheric modelling.
The lack of these types, while disappointing, is not necessarily a problem in itself. Default XP11 has no aerobatic aircraft, twin-jets, or high-performance turbines. Having many different types is not a prerequisite for a good simulation experience —it’s far more important to make each plane “count”. Sadly, FS2020 doesn’t do this. The seemingly plentiful selection of aircraft is illusory. There are far too many of the same flying experiences in this game.
Consider the light GA selection — by far the most well-stocked part of the hangar. You’ve got (in the standard version) the C152, C172, DA40NG, CC19–180 XCub, Ex Zlin Savage Cub, Bonanza G68, Robin DR400, Aveko VL-3, and Icon A5. Some of these fall into light-sport or ultralight classifications, but they’re all aircraft that can be flown by any SEP-rated PPL or RPL pilot. They’re designed and certified primarily (in some cases exclusively) for VFR operations. Curiously though, only the C152, Ex Zlin Savage Cub, and Robin DR400 are devoid of any glass instrumentation (PFDs or MFDs).
I’m not a fan of glass displays for VFR. For one, they’re complicated and require training to use effectively. For another, and more importantly, they stop you doing the one thing that is fundamental to VFR — looking out of the window.
Using navigation systems on a glass MFD requires a lot of “head down” time. The fiddly preparation of waypoints is usually done and dusted on the ground, but any adjustments made to your GPS route in the air (a diversion for example) will require you to take your eyes away from the real horizon to input changes into at least one sub-menu. You’re also given a lot more to worry about if you use a glass MFD. You have real-time readouts of groundspeed, wind correction angle, and ETA. If you use a paper map, you can only work these things out at discrete intervals (every 10 miles, say)— or at your navigation waypoints. In between, you can’t check these things, so you don’t worry about them. You focus instead on scanning for traffic, maintaining the same “picture”, and enjoying the view.
Using a map does require you to look down, but there’s a fundamental difference. With a map, you have to think about “what the diversion looks like” outside of the window. With a GPS and moving map, you don’t — the waypoints just appears on your display instantly and you can just “follow the line”. Following a line inherently entails that you’re looking (at the minimum) at your MFD. You might even be cross-checking with the HSI on your PFD — which means you’re not looking out the window. Sure, moving maps and HSIs are great aids to safe VFR flight, but their existence encourages you to rely less on what’s outside, and more on what’s inside.
It’s not even about discipline. Actually, in my view, having a glass cockpit there and not using its information is poor airmanship. You have all of these tools there to make your flying safer, so you should use them. The problem is that VFR flying is about using external visual references to fly. Glass cockpits shift the emphasis in VFR flying to validating your performance. The usual cross-checks of altitude, speed, and heading are already time when you’re not looking out of the window. Checking your navigation too means you must be looking outside even less.
More fundamentally, I think navigating VFR by GPS is against the spirit of VFR flying. Unless you’ve taken the time to manually input the exact coordinates of VFR waypoints, you’re probably only approximating your route by using airports or radio navaids. This is problem because, chances are, these aren’t very good visual references. VFR references on VNCs or VTAs are marked because they are obvious. Yes, an airport can be a obvious, but is it really a good idea to be flyingoverhead the airspace of an operating airport? Navaids like VORs are visible, but they’re often hard to see until you’re right on top of them. NDBs are even harder to spot. I’ve flown past the AP NDB in Vancouver probably two-dozen times, and I’ve still not seen it. By contrast, things like lakes and towns are good visual references, but unless you take the time to manually input the coordinates (if you’re able to), they won’t be in your FMS database as it’s designed for IFR.
So, navigating by GPS when flying VFR turns looking outside into a cross-check for what’s on your MFD. This is completely the wrong way around!
Regarding the flying experience, this is also a problem. Having a glass cockpit when flying VFR isn’t fun. You become hypnotised by the huge display. Its precision and bounteous information make you chase that 10ft/min of vertical speed or that half-bar deviation on your HSI. It’s unavoidable as a pilot — you want those things to be exactly right — and it fundamentally changes your experience of flying. With an analog panel, I accept that I can only be so precise, and that’s okay. I only glance at my instruments because they’re just a cross-check for what I can see outside. A glass cockpit makes you worry about, and consider, a lot more.
I’m not saying in a everything-was-better-in-my-day voice: “I only want to have analog gauges in the planes I fly VFR because that’s how things should be.” I’m saying that a flight simulator should acknowledge that glass-free cockpits give a particular feel to flying VFR, and one that I feel is more in the spirit of the discipline.
By providing so many GA planes with glass cockpits, FS2020 isn’t providing a good representation of the experience of GA VFR flying. There’s just far too much overlap and not enough distinctness.
If I want to fly “analog” VFR, I can choose between a C152, Cub, or DR400. Why do I need all these choices? I can understand the inclusion of the Cub — it provides a nice contrast in bush-flying experience to the modern, more powerful, glass-equipped XCub. That’s fine, but I’m not really getting a different experience flying a C152 or a DR400, so why do I need both? In the real world they might handle differently, but unfortunately the flight model in FS2020 is too coarse-grained for me to pick up on any meaningful differences beyond just comparing the performance numbers. Just pick one plane — I’d go for the C152 because it’s a ubiquitous trainer.
The same problem appears among the glass-equipped light aircraft. What’s the difference between the C172, DA40NG, and G68? They even have the same type of glass cockpit — the G1000. So, not only do they basically look the same inside, they’re operated in very similar ways.
If it were up to me, I’d choose to have a glass DA40NG for the modern GA experience, and an analog G68 for the “old school” one. They would also double as representative single engine aircraft for IFR GA. Yes, the C172 is the most widely produced single-engine aircraft in history, but it’s essentially a 50s design with updated avionics. The G68 is of a similar ilk, but it has the experiential benefit of being more complicated with retractable gear, and having “more” performance. Flight simulators are about diversity of experience, and including a C172 when you’ve already got a 152 feels like too much of the same thing. You could do what XP11 does, which is include the C172 in analog and glass versions, but FS2020 has decided to lock the analog 172 into the deluxe edition. You could ditch the C152 and replace it with an analog 172, if you really wanted to insist on having one, but my point is that I’m not getting more experiential mileage out of including a C172.
I’d also get rid of either the C208 Caravan or the TBM in the utility category. Personally, I’d kill the Caravan because the TBM is both iconic and “more plane” than the Caravan. In FS2020 both have G1000s, and both have a PT6 up front. What’s the point?
The only experiential contrast that has been done well is the King Air 350i and the CJ4. Both are utility twins, but one is a turboprop and one is a jet. The transport category is okay — the gulf between the A320NEO and the 747 is a bit too big in my view, but they do admittedly span the spectrum of transport.
To be clear, I’m not against having a lot of planes. I’m just against having a lot of the same ones. By getting rid of some of these “redundant” planes, FS2020 could have included an analog light twin like the Duchess to compliment and contrast with the glass DA62, and maybe thrown in a glider like the DG808S or the ASK-21. They could also given us a regional turboprop to fill in a gap in the transport category, or even something “classic” like a DC-3.
A flight simulator can have loads of default planes, but each one should feel different to fly. If I want subtleties and nuances, I‘ll go aftermarket.
In-Game Challenges
FS2020 is obsessed with landings. Every single “challenge” you can find is to do with landing. I put “challenge” in quotation marks not because the landings aren’t challenging, but because they’re completely misguided as piloting challenges.
The Transport Canada PPL and RPL training manual goes to great pains to emphasise that (paraphrasing) “landing is not the end-goal of learning to fly”. And that’s completely right. Landing is one aspect of flying — you can make the softest landings in the world and still be an awful pilot. Yes, it’s important — it’s what gets you back on the ground — but the landing itself is just a few short seconds in the process of getting from A to B.
By placing all of the emphasis of the challenge of flying on “landing”, FS2020 completely fails to understand what flying is about. And by “landing”, I literally mean “landing”. You’re graded on your vertical speed, touchdown zone, and centreline deviation. There’s absolutely no consideration of how stable your approach was, how suitable your configuration was, and how well you braked. Before I get into that though, I want to highlight that FS2020 doesn’t even understanding the event of landing.
Contrary to popular belief, a smooth touchdown is not always a good touchdown. On a grass strip, a smooth touchdown is essential to prevent the gear digging in, but on concrete, that’s not a consideration. In dry conditions with light wind, a low vertical speed is generally good because it’s comfortable and saves wear on the landing gear. In a crosswind though, you don’t want a smooth touchdown. You want a “positive” touchdown, because it lets you gain directional ground control quickly and it prevents you from being blown off course when you de-crab. You don’t want to put too much lateral load into the gear, but that’s distinct from a low vertical speed. And when a runway is wet, you want a firm touchdown because it prevents hydroplaning and enables you to brake as soon as possible. FS2020’s scoring system doesn’t give any regard for these things — on a wet runway in a strong crosswind, a “greaser” gives you a high score.
In fact, even on a dry runway, it may not be desirable to have a smooth touchdown. In the 737, for example, light touchdowns are not recommended in the POH. This is because low vertical speeds at touchdown can cause the main landing gear to shimmy in their partially loaded condition, resulting in destructive oscillations that can cause the gear to completely fail. Several accidents (such as EI-STD below) have been caused over the years by overly light touchdowns. FS2020 doesn’t have a 737, but according to its scoring system, a good touchdown is an objective standard, independent of considerations of aircraft type and POH.
Now that we’ve covered why FS2020’s touchdown scoring system doesn’t understand anything about landing, let me explain why a touchdown scoring system shows no understanding of anything in aviation else either.
If you select a landing challenge in FS2020, you’re set up on an approach a variable distance and position from the runway. You might be on the downwind or on a straight-in. Let’s take this week’s landing challenge as an example: a crosswind landing into a snowy LFPG in an A320NEO.
First, we start off below the glidepath. And by below, I mean way below. Four reds on the PAPI is definitely go-around territory — especially 5ish miles from the runway. The fact FS2020 doesn’t even set you up on the right glidepath is damning — it shows zero care was taken and also that the developers don’t care about how you get to the touchdown zone.
I’m also fully configured (flaps 3 and gear down) with no Vapp shown. Is flaps 3 the appropriate choice for my landing weight and the weather conditions? Does FS2020 care? No. Moreover, the fact I start off below the glidepath and need to level off to reacquire it throws all of the stabilised approach criteria out of the window. My thrust is all over the place as the autothrust scrambles to provide enough power to level off in landing configuration, and my speed is nowhere because I have no clue where it’s meant to go. This is bad flying, yet the scoring system doesn’t care.
When you eventually reacquire the glidepath, you can continue to bust every rule in the aviation rulebook without it affecting your score. Generally (I say that hesitantly because it is operator, aircraft, and airport dependent), -1000ft/min is considered the maximum permissible sink rate below 1000ft — more conservatively, after leaving whatever altitude you begin your final approach from. Exceeding that is a mandatory go-around. Guess how much FS2020 cares about that? Yup, not at all. The first few times I tried to do the approach with the autothrust engaged, I routinely busted -1200ft/min without any impact to my score and (more concerningly) without the GPWS going off. Eventually I decided to use manual thrust after being completely baffled by how to use pitch-and-power with autothrust, probably violating SOP to do so, and the game didn’t care about this either.
In the rollout, because I hadn’t mapped my throttle quadrant, I didn’t deploy the reversers. Now, reverse thrust isn’t mandatory, but given it was snowing and both reversers were working, it was very poor airmanship to not deploy them. I also wasn’t judged on how well I maintained the centreline after landing, or how long it took me to stop.
So, despite violating SOPs and Flying 101 in every aspect of my approach, thanks to my (completely inappropriate) light touchdown and good aiming, I manged to get a score of 1200000-ish. Yay. Well done me.
Now, with some approaches, you might be able to follow SOP and get a good score. The problem is, FS2020 doesn’t care about whether you did or not. It doesn’t understand or appreciate the value of good airmanship. And in the case of the LFPG approach, it’s impossible to follow SOP, as you start miles below the glidepath in mandatory go-around territory.
Why does FS2020 only care about landings? That’s a good question. FSX didn’t. For all its realism faults (and there were many), FSX had a whole plethora of engaging (and some genuinely challenging) missions that showcased aviation’s diversity. You could fly the World Gliding Championship, dead-stick a 737 to a tiny airfield in the Indian Ocean, and fly hard IMC into Yakutak with a partial panel on a King Air 350i. It was an immersive aviation experience.
I genuinely miss those FSX missions — so much that bought the game again on Steam to relive them. Yes, some of them were more about fighting the infurriating in-game bugs getting the cargo to load into your Grumman Goose, but for the most part they genuinely showcased the breadth of possibilities and experiences in aviation. I could drag race a rocket truck in an Extra 300 and learn to fly IFR across the English Channel in a C172. They actually taught you something about flying too — how to adapt to unexpected changes and to manage emergencies. I also loved the entertaining and anecdote-rich co-pilot and ATC dialogue. My favourite one was when you had to deliver cargo to some remote airfield in the DRC in a battered DC-3, and my co-pilot would not shut up about “the last time he was here”.
The fact that I remember these missions now, more than 10 years after I first did them, tells you everything you need to know. I can barely remember which of the approaches or bush trips I’ve flown in FS2020, despite doing some just the other day. There’s no experience in FS2020’s “challenges”.
XP11 doesn’t hold a candle to FSX in terms of its challenges, but it does provide a genuine experience in doing so. Take the aircraft-carrier landing challenge. Superficially, it seems the same as what you’d find in FS2020 — you just need to land on the deck. But in XP11, you don’t get a score. The “situation” exists to give you a genuine appreciation of how hard it must be to fly a real carrier landing — quantification be damned.
I was flying on a CAVOK day, with light winds, and no pressure of minimum fuel or completing a combat sortie — and I still managed to screw it up most of the time. I can only begin to imagine how tough it must be to land on a rolling deck in marginal visibility or at night.
Flight Training
I thought FS2020’s misunderstanding of piloting challenges was bad. That was until I tried out their “training” lessons. It makes their landing challenges look like masterworks.
In FS2020, a lesson is a box-ticking exercise — much like the modern education system in general, but I won’t get into that. Worse, it doesn’t even make you tick the right boxes.
The first lesson is learning how to control the plane. You would think that’d be about learning to turn, climb, and descend. Yeah, no. It’s about learning to press the right keyboard shortcuts.
Eventually, you do a bit of “turning”, “climbing”, “descending”, and a spot of “throttle management”. I put these in quotation marks because you just need to move your controller (joystick, yoke, etc.) in the right direction, not actually do a maneuver. You could be in a nosedive, but as long as you move your stick up, even for a fraction of a second, that counts as completing the objective. In XP11 and FSX, you’re judged on the duration and precision of your maneuver, and the lesson won’t advance until you fulfill the criteria.
The first lesson doesn’t even get you to learn the keyboard shortcuts properly. As with the controller inputs, it doesn’t matter if you use them properly, you just need to press the button. In the first tutorial task, you have to locate the airport. This means you need to press S on the keyboard, but you don’t need to actually look at the aiport to pass — just a fraction of a second will do. I’d also complain that the S shortcut is completely disorientating in the way it snaps your view, but really, that’s the least of this game’s problems.
More alarmingly, at no point (not even in any of the subsequent lessons) do you learn about the aerodynamic functions of the flight controls, or about the fundamental relationship of pitch, power, and performance.
The problems don’t improve with the other lessons. You’re never told about coordinated turns, about adverse yaw, proper use of trim, or about stall recovery. These are flying absolute basics. Even steering on the ground isn’t discussed. FSX had a whole tutorial all about ground maneuvering in different aircraft — it was infuriatingly tedious, but informative regarding the ground handling model. At no time in FS2020 are you encouraged to understand the game mechanics or how to actually fly.
I don’t buy the theory that learning to fly “properly” is too boring for the causal gamer. Every game requires you to understand how its mechanics work — regardless of the genre. An FPS, RPG, and MMO all require you to learn how to operate the game controls and apply them practically. There’s no reason this can’t be engaging.
Take The Outer Worlds, an FPS, as an example. The opening gameplay is all about learning how to jump, sneak, and move around. If you don’t crouch and sneak in the grass, two feral canids will jump out and eat you — you don’t have any weapons so you’re extra tasty. Why can’t the same thing kind of interactive thing be done in FS2020?
Moreover, you buy a flight simulator to fly. This might be news to people at Microsoft and Asobo, but people who buy their product are actually interested in flying in some way. And flying is technical — there’s no getting around that. FSX didn’t shy away from giving people the full gory-story of learning to fly, and I really can’t understand why it’s child FS2020 is so shy of doing the same. In fact, FSX took a two-pronged approach to teaching people how the simulator worked. There were casual, fun mission tutorials where you had to fly through gates — which you could only do if you were doing roughly the right things with the throttle and stick. Then there were the lessons, where you actually had a very brief crash-course (pun not intended) PPL.
XP11 isn’t very keen on its tutorials, but the few it does have are done very well. You actually learn about why you do certain things when you fly. You also get a score at the end based off the quality of your maneuvers — with actually meaningful metrics like how well you tracked the ILS and how on target your airspeed is. I actually like doing some of these tutorials every now and then as a test for how well I “get” the game mechanics and modelling. Flying in a simulator is different to real-world flying, and each simulator is different, so having these tutorials with meaningful benchmarks is useful.
FS2020 doesn’t benchmark you informatively at all. Considering the game’s obsession with scoring landings, it’s completely baffling to me that there isn’t even a scoring system for how well you execute your landings in the tutorials. There is no metric by which you can judge how well you did in any of your lessons — only that you did them.
I had hoped, among other things, that FS2020 would have a whole host of new flying lessons that are far more diverse and challenging than those found in FSX and XP11. It’s been sixteen years since FSX came out, and six for XP11. Imagine how many new metrics you could measure and grade people on! And wouldn’t it be amazing to get the vastly more sophisticated algorithms to suggest improvements to your flying technique? Instead, we’ve been given a tutorial system that’s less engaging and completely non-educational.
Conclusion
Microsoft and Asobo just don’t get flight simulators. This is perplexing to me, as Microsoft developed FSX. From the choice of planes to the misguided challenges and the absence of meaningful tutorials, the whole game feels like it was written by people who never asked themselves why people buy and use personal flight simulators.
To me, aircraft selection is indicative of how well developers understand their simulator. I look at the planes XP11 have chosen, and I can tell that these were planes chosen to give a wide-ranging cross-section of aviation. From the Aerolite 103 ultralight to the Space Shuttle, there’s no overlap in any of the choices. I look at FS2020, and I sigh. The higher “OEM payware” editions do increase the diversity, but there’s still far too much overlap in experiences.
The lack of meaningful challenges and lessons in FS2020 is exasperating — especially when FSX did them so well. That’s not to say FSX was a good flight simulator — it was dreadful on the realism front — but it was a genuinely evocative as an experience. It’s frankly unacceptable that FS2020 completely fails to deliver the quality of immersion and escapism that its predecessor did.
If I had to give a reason for all of these shortcomings, I would, as with realism, blame that third-person view HUD. All of the aircraft issues go away because the planes look different externally. The landing-obsession is harder to explain, but I can see the scenery becomes the more significant component of the experience than the flying, so the different landings are more distinct. The lessons — well, you don’t really need lessons if you’re just using your plane as a glorified cursor to move around the world.
I think Microsoft and Asobo decided the main experiential value in a simulator was the scenery. That’s like a chef deciding that the most important part of their cooking is the garnish. Microsoft and Asobo just don’t understand flight simulators.
I thought that 16 years of technological advancement would make FS2020 experientially immersive in an unparalleled way. Turns out, it is, but for completely the wrong reasons.
