Flare Timing: Mastering the Last Fifty Feet
You've read the advice. Reduce power over the threshold, look down the runway, flare gently, hold it off. You've done all of it, and the landings are still wrong. Sometimes a slam that bounces the camera, sometimes a balloon followed by a drop, sometimes a touchdown so gentle and so far down the runway that you would have gone off the end of anything shorter. And nobody has told you what a good landing actually measures, so you can't even tell how far off you are.
The uncomfortable starting point is that smoothness is an output. It falls out of an approach flown at the right speed, in trim, on the right path, followed by a flare timed by height rather than by feel. Trying to produce it in the last three seconds, from whatever state you arrive in, is why the same advice keeps failing. Working backwards from the touchdown is the only order that makes sense of it, because every number that decides the landing was set before the flare began.
The number nobody shows real pilots
Touchdown quality has an exact number. Vertical speed at the moment the wheels touch, in feet per minute. Real pilots almost never see it, because only aircraft with flight-data recorders and a monitoring programme get told. Your simulator knows it on every landing, which is one of the few areas where a desk pilot has better feedback than a real student.
Rough bands, for orientation rather than gospel.
| Touchdown rate | What it is |
|---|---|
| Under ~100 fpm | The "greaser". Feels like nothing. |
| ~100 to 250 fpm | A normal, good landing. What airline pilots actually produce day to day. |
| ~250 to 400 fpm | Firm. Noticeable, unremarkable, sometimes deliberate. |
| Higher | Hard. In the real world, above a type-specific threshold it triggers a maintenance inspection. |
Two calibration points stop these bands being misread. First, transport-category landing gear is designed around a descent rate of 10 feet per second, 600 fpm, at the design landing weight. That's a certification design case, not a target and not a damage line, but it tells you that a 300 fpm touchdown is nowhere near the structure's limits. "Firm" and "hard" are different words for a reason. Second, an autoland typically puts the aircraft down noticeably firmer than the community's idea of a good landing, entirely on purpose. A positive touchdown spins the wheels up promptly and gets the ground spoilers and autobrake working.
So the target is not "as close to zero as possible". A good landing touches down around the first 1,000 to 1,500 feet, the aiming point, which sits comfortably inside the touchdown zone proper, conventionally the first 3,000 feet or the first third of the runway, whichever is less. Add the right speed, a rate somewhere in the normal band, and the centreline, and that's the whole definition. Hold that definition, because the last section of this article comes back to why the community's favourite number is the wrong one to chase.
What the flare actually does, mechanically
On a 3° glideslope your sink rate in feet per minute is roughly your groundspeed in knots times five (5.3, strictly). A Cessna 172 at 65 kt groundspeed is coming down at about 345 fpm. An A320 at 140 kt groundspeed is coming down at about 740 fpm. That rule is worth more than any table because it works for whatever you fly, and it makes two things obvious immediately. A headwind lowers your groundspeed and therefore the sink you must arrest, which is why everyone's best landings happen into wind. And the entire job of the flare is to take that 350 to 750 fpm and reduce it to something in the normal band, in roughly the last five seconds of the flight.
The flare does it by trading energy. Raising the nose increases angle of attack, which increases lift, which reduces sink, and it pays for that lift with airspeed. The trade only works while you still have speed above the stall. Arrive fast and you have too much to trade. The aircraft refuses to come down and floats. Arrive slow and you have nothing to trade. You pull, nothing happens, and the ground arrives at the sink rate you brought with you. This single mechanism explains most of the folk advice you've already read, and it's why approach speed, not flare technique, is the first thing to check when landings go wrong.
Two more pieces of mechanism, both usually left out of tutorials.
Ground effect becomes significant within roughly one wingspan of the surface. A 172's span is about 36 feet, so it enters ground effect at almost exactly flare height and gets a free cushion of reduced drag and reduced sink right when it needs one, which is why light aircraft can be held off and landed near the stall. A 747-400's span is about 211 feet, so it has effectively been in ground effect for the whole final stage of the approach and gets no sudden cushion. This, not tradition, is why jets are flown on with a small pitch change, roughly 2 to 3°, rather than held off. Holding a jet off GA-style bleeds speed with the nose rising, and in long-body types the tail is closer to the pavement than the sight picture suggests. Learn one aircraft's flare and its numbers before switching, because the technique doesn't transfer between classes.
Trim is the hidden variable. In trim, the flare is a small refinement to an aircraft already in equilibrium, a couple of millimetres of stick. Out of trim, you're holding a constant force all the way down final, and the flare becomes a large, badly judged movement stacked on top of a load you're already carrying. A striking proportion of "I can't judge the flare" complaints are trim complaints wearing a disguise. If you're holding pressure at 500 feet, fix that before you practise anything else.
In jets, the thrust cut is part of the flare, not a separate action. Engines take seconds to spool down from approach thrust, and underslung engines pitch the nose down as thrust decays, so closing the levers passing roughly 20 to 30 feet is itself a pitch input you must anticipate. And the landing is not over at touchdown. The nose is lowered flying, by relaxing back pressure, not by releasing the stick and letting it drop. A soft mains touchdown followed by a slammed nosewheel is a bad landing that started well.
Why a monitor makes all of this harder
Real pilots flare on two cues, and a desk provides neither.
The first is peripheral visual flow. In the last twenty feet, height and sink are judged not from the aim point but from the rate at which the runway edges and surrounding ground stream past the corners of your vision. This is also why "look at the far end of the runway" works. It's not superstition. It stops you staring at the aim point (which you will otherwise fly into) and hands the height judgement to your peripheral vision, which is the part actually equipped to make it. A single monitor at default field of view shows a fraction of what human vision covers, so the runway-widening, ground-rush cue arrives late and weak. This is why sim pilots systematically flare late. Widening the FOV or using a head-tracker restores some of the cue, and VR restores much more, but every change to FOV or camera position changes the sight picture your flare is calibrated to. Whatever you choose, fix it and stop touching it. The same goes for eyepoint. MSFS's default camera in many aircraft doesn't sit at the real design eye position, and moving the camera up, down or forward silently changes your perceived flare height. A wrong eyepoint teaches a consistent wrong flare, which is at least trainable. A wandering eyepoint teaches nothing.
The second missing cue is your body. There's no vestibular sense of sink, no seat pressure telling you the descent is arrested, and no control loading. In a real aircraft the elevator goes progressively mushy as speed decays, a built-in warning that the flare is running out of energy. A spring-centred desk stick feels identical at VRef and one knot above the stall, which is why a desk float so often ends in a surprise drop. Nothing warned you the trade was exhausted. The substitutes are explicit rather than felt. The radio-altimeter callouts in airliners exist precisely to time the flare ("thirty" is your cue in most jets), and in GA aircraft without callouts you flare at a pre-decided height, typically starting the round-out at 10 to 20 feet, judged against a sight picture you have deliberately kept constant.
A short-travel spring stick compresses the entire elevator range into a few centimetres, so the flare is a two-millimetre input, which is what sensitivity curves are for. But a curve that flattens the centre steepens the edges, so the same curve that smooths your flare makes a balloon recovery twitchier. Curve setup is its own subject. The point for this article is that whatever curve you use must, like the camera, stay fixed while you calibrate.
Finally, frame rate is a flight-control input. A stutter at 50 feet, common over dense photogrammetry airports in MSFS 2020 especially, breaks the visual flow at exactly the moment you need it, and a landing that felt mistimed may have been a rendering problem. For landing practice, a lower frame rate that never dips beats a higher one that fluctuates.
The landing is decided at 1,000 feet
The stabilised-approach concept comes from airline flight-data monitoring, where the correlation is blunt. Unstable approaches produce a hugely disproportionate share of landing incidents, long landings and go-arounds that should have happened but didn't. The convention (a convention, not a regulation) is that by 1,000 feet in instrument conditions (500 in visual conditions) the aircraft should be in landing configuration and in a steady state. Five checks capture it. Gear down, landing flap set, sink rate under control, on glideslope, on centreline. The stabilised approach in depth is its own article. What matters here is why it predicts the touchdown.
Speed is the clearest thread. VRef, traditionally 1.3 times the stall speed in landing configuration, computed differently under modern transport certification but serving the same purpose, is the margin the flare's energy trade is designed around. Cross the threshold 10 kt fast and nothing feels wrong. The approach was comfortable, the picture looked normal. But kinetic energy scales with the square of speed, so those ten quiet knots buy you hundreds of feet of extra float, transport-category guidance puts it at several hundred feet to over a thousand, depending on type, and then present you with a choice between a deep touchdown and forcing the aircraft on before it's ready to land. Threshold speed against VRef is the single best leading indicator of landing quality, and if yours is consistently high, no amount of flare practice will help, because the flare is being asked to dispose of energy it has nowhere to put.
Weight belongs in this section too, because most simmers never vary it. An A320 at max landing weight and one nearly empty need different VRef speeds, different flare timing and a different thrust-cut point. If you always fly at the default payload you have calibrated to one weight without knowing it, which is fine, until the day you load the aircraft properly and conclude your flare has mysteriously broken.
The five causes of hard landings
Every hard landing leaves evidence, and the fix depends on which failure it was. These five cover nearly all of them.
1. The late flare
What it feels like. The ground came up faster than expected and you were still pulling when the wheels hit. On a monitor this is the default failure, because the visual cue that triggers a real-world flare arrives late and weak on a narrow field of view. The fix is to stop flaring by sight and start flaring by height. Use the radio-altimeter callouts in jets, a committed round-out height in GA, and keep the camera fixed so the sight picture at that height stays learnable.
2. The balloon and the drop
What it feels like. The runway fell away, the aircraft climbed a few feet, then dropped on from a height with the speed already spent. The cause is too much back pressure while there's still excess energy, often driven by aim-point fixation or an eyepoint that reads low. The critical part is the recovery, because balloon and float are the same error at different energies and both have the same correct response. Stop pulling, hold the attitude, and wait. Pushing forward converts a small balloon into a nosewheel-first arrival. If the balloon is large or the speed is gone, go around. Go-around decision-making is its own article, but the short version is that it's always available and almost never taken.
3. Too fast over the threshold
What it feels like, deceptively, is a soft touchdown, 2,000 feet down the runway. Or, if you get impatient and push it on, a flat, firm arrival with the aircraft not yet finished flying, which is where bounces come from. The trap is diagnosis. This presents as a flare problem, and "flaring more gently" makes it worse. Check the speed crossing the threshold against VRef before you change anything about your flare. Gust additives are a real thing, half the gust factor is the usual rule, but they're additives to a correct VRef, not a licence for a permanent ten knots of comfort margin.
4. Out of trim on final
What it feels like. Every landing is different, and the flare is a lottery. If you're holding constant stick pressure down final, the flare input is stacked on a load you're already carrying, and judging a small movement on top of a large one is beyond anyone's consistency. Trim until the aircraft holds the approach attitude hands-off, and the flare shrinks back to the small refinement it should be. On a desk, with no control forces to feel, this takes deliberate discipline. Watch for the aircraft drifting off attitude when you momentarily relax, and trim that out.
5. Disconnecting the autopilot at minimums
What it feels like. Beautiful approaches and terrible landings, and confusion about why. Clicking off at 200 feet gives your hands about fifteen seconds to calibrate to an aircraft they haven't been flying, with no force feedback to speed up the handover. The approach was stable but you were not part of it. The fix costs nothing in a simulator. Disconnect at 1,500 feet, or 3,000, and fly the aircraft to the runway yourself. Your landings will get worse for a week and then better permanently.
Practise like it's a simulator
The strangest habit in flight simulation is training like it costs 300 pounds an hour. Flying a full 45-minute circuit to earn one flare is how real students practise because they have no choice. You do.
Spawn on a five-mile final. Land. Reload the same final. That's thirty flares an hour instead of four, with every variable held constant, same aircraft, same weight, same runway, same weather, which is precisely the condition under which motor skills calibrate fastest. Change nothing until your touchdown rate sits reliably in your chosen band, then add exactly one variable. A 5 kt crosswind, then 10, then gusts, then night. (Crosswind technique itself, crab versus wing-low, decrab timing, is its own article. The ladder is the point here.)
Three drills no real aircraft can afford.
- The no-flare flight. Once, deliberately, fly a stable approach all the way to the runway without flaring. Feeling what 350 to 750 fpm at the surface actually looks like recalibrates every subsequent flare, and it's a demonstration that would write off a real aircraft's day.
- Bounce recovery. Deliberately induce a small bounce and practise the correct response, hold the attitude, a breath of power, re-land or go around, until the instinct to shove forward is gone. This is explicitly forbidden in real training and free at your desk.
- Deliberate firm landings. Put the aircraft down on the 1,000-foot markers at 200 to 300 fpm on purpose. Controlling touchdown point and touchdown rate independently is the actual skill. Softness alone is a side effect you cannot steer.
Use the replay afterwards. From an external side view, your flare height and float distance are visible in a way they never are from the seat, and the gap between what you felt and what happened is usually the whole lesson.
One landing tells you nothing
Here's the part that undoes most self-assessment. Landing skill is a distribution, not a highlight. Everyone has produced one greaser. The measure of skill is the median across twenty landings, the spread around it, and the worst one, and a tightening spread is real improvement even in a week when your best number never moves. Without recorded data you cannot see any of that, and memory is a flatterer. It keeps the greaser and quietly discards the two thumpers either side of it.
The metrics worth tracking together, touchdown rate, touchdown point past the threshold, threshold speed against VRef, and centreline through the rollout. Rate alone rewards floating, while rate plus touchdown point exposes it immediately. This is the logic of airline FOQA programmes applied at a desk, and it's what I built My FS Flights to do automatically. Its landing report records the flare, touchdown rate, threshold speed against VRef, and centreline and glideslope tracking on every flight, its stabilised-approach badge checks the five criteria at 1,000 feet, and the logbook turns single landings into the distribution that actually measures whether you're improving. No manual logging is involved, so the final-only drill above generates twenty data points an hour by itself. Track your aircraft separately, though. Your 172 landings and your A320 landings are different skills with different baselines, and averaging them hides both.
The butter trap
Which leaves the target itself. The community's prize landing, sub-100 fpm, screenshotted, shared, is optimising the wrong variable, and it's worth being clear-eyed about why. Airline flight-data programmes flag long and fast landings far more urgently than firm ones, because a deep touchdown is a runway-excursion precursor and a firm touchdown is a non-event. A 250 fpm landing on the markers at VRef is a better landing, by every standard professional aviation uses, than a 40 fpm landing 3,500 feet down the runway. There are also conditions in which chasing softness is actively wrong technique. On a wet or contaminated runway, a positive, firm touchdown is what gets the wheels spun up and the antiskid and spoilers working, and the simulator will never punish you for getting this wrong because contamination and aquaplaning modelling is one of MSFS's weaker areas. Learn it as knowledge, because the sim won't teach it as consequence.
None of this means softness is worthless. It means softness is the last variable, earned after touchdown point and speed are under control, and cheap whenever it's bought by floating. Fly the approach on speed and in trim, flare at a height rather than a feeling, and put the aircraft down where it belongs. A smooth touchdown then stops being something you chase and starts being something that happens to you, more weeks than not, in the numbers where it belongs.