Stalls and Turns — Why Pulling Harder Doesn't Turn You
Stall speed rises the moment you roll into bank. Load factor, turn radius and corner speed with the actual numbers — and why slowing down in the valley is the wrong instinct.
“I’m pulling and it won’t turn.” “It fell out of the sky the moment I rolled in.” This is the most common way to fail in this game — and the real aircraft does exactly the same thing.
What follows is what is actually happening, in the flight model’s own numbers. All of them come from the aircraft constants.
1. Bank raises your stall speed
Turns here are not auto-coordinated. They are bank-and-pull, the way the real thing works.
To hold altitude at bank angle φ, the vertical component of lift has to balance weight. That takes more lift than level flight does — it takes load factor n.
n = 1 / cos φ
stall speed Vs(φ) = Vs × √n
At sea level the 1g stall speed is about 240 km/h. By bank angle:
| Bank | Load factor | Stall speed |
|---|---|---|
| 0° | 1.00g | 240 km/h |
| 30° | 1.15g | 258 km/h |
| 45° | 1.41g | 286 km/h |
| 60° | 2.00g | 340 km/h |
Pull there and angle of attack goes into the stall region (past 15°), lift drops rather than rises, and no turn rate appears. That is why pulling harder turns you less.
The stall line on the HUD speed tape moves with bank angle and with altitude. It is Vs(alt)/√|cos φ| shown directly. Nothing flies the aircraft for you, but in exchange the display is honest about the danger. Whenever you are turning, that is the line to watch.
2. The slower you are, the less you turn
“Slow down to be safe” is counterproductive here. Working out the available G and turn radius by speed:
| Speed | Max G available | Best turn radius |
|---|---|---|
| 300 km/h | ≈1.8g | ≈460 m, on the edge of the stall |
| 400 km/h | ≈3.0g | ≈440 m |
| 500 km/h | ≈4.6g | ≈440 m |
| 600 km/h | ≈6.5g | ≈440 m |
| 700 km/h | ≈8.8g | ≈445 m |
| 725 km/h | 9g — the structural limit | ≈450 m |
| 900 km/h | 9g, capped | ≈700 m |
| 1,100 km/h | 9g, capped | ≈1,050 m |
Three things fall out of that.
① 450–600 km/h is the working range. Below it you cannot pull G, and turn radius falls apart quickly. In an F-2, 300 km/h is approach speed — not a speed you maneuver at.
② Inside that band, going faster does not tighten the turn. Minimum radius sits at about 440 m all the way from 400 to 700 km/h and moves less than 2%. Lift (∝V²) and centrifugal force (∝V²) cancel each other, so “faster turns tighter” is not true. What speed does buy is turn rate — how quickly the nose comes round.
③ Past 725 km/h the radius opens up again. That speed is corner speed — where aerodynamic lift first reaches the structural 9 g limit (sea level, reference weight). Above it the available g is pinned at 9, so radius grows with the square of speed: about 700 m at 900 km/h, about 1,050 m at 1,100 km/h. §8 has the details.
Speed therefore has a hard floor, and a penalty on the top side that arrives all at once at corner speed. When in doubt, take the faster option — with 725 km/h as the ceiling on “faster”.
3. At low speed the controls themselves stop working
There is a second trap sitting just short of the stall. Control authority is proportional to dynamic pressure.
authority = (speed[km/h] / 396)²
Halve your speed and you have a quarter of the authority. Control response also carries a time constant of 0.3 seconds.
Down at the bottom of the speed range, “the input does nothing” and “the input arrives late” happen at once. Grabbing at the controls once you are already near the stall is too late — and that is not a problem with your hands.
4. Recovering from a stall
The order matters.
① Roll wings level → ② lower the nose to build speed → ③ re-enter the turn once the speed is back
Trading altitude — potential energy — for speed is the fastest route out. Wings level comes first because as long as bank is in, the lift you need stays high.
The problem is that down low you cannot do this. There is no altitude to give away. So in the valleys the question is never whether you can recover: not getting slow in the first place is the only defense you have.
5. Flying it in the valley
The low-level courses — Kurobe Gorge, the Southern Alps, Kamikōchi and the rest — are meant to be flown along the river. Give up on the straight-line shortcut and follow the terrain.
There is one rule.
Get the speed first, then a steep bank and a firm pull.
Building speed and changing direction in one committed movement is safer than easing around in a shallow bank. A shallow bank spends time without producing turn rate, and the valley wall keeps coming.
In practice it becomes throttle back just before a bend, throttle up on the straights — because too much speed makes the radar ceiling hard to manage.
The low-level ceiling works in AGL, height above the terrain. Level flight alone will push you through it as soon as the ground rises. Heading for a ridge, lower the nose first.
6. How this differs from the real aircraft
Work backwards from this model’s lift coefficient and you get a mass of about 12.7 t — roughly the real aircraft’s landing weight. The 240 km/h stall and the 280 km/h approach were both measured there, and thrust is set against the same figure: the published F110 numbers divided by 12.7 t are exactly what the model carries.
Every number in this section is measured at that 12.7 t. Actual launch weight depends on what is aboard and how much fuel is left, so the Matsushima circuits and the formation missions — which carry two external tanks — fly at 15.1 t, where the stall sits at 262 km/h. Burn fuel off and the aircraft moves back toward the figures below (see The F-2 §4).
| Item | Game | Real aircraft |
|---|---|---|
| Top speed (low altitude) | ≈1,160 km/h (mil) / ≈1,380 km/h (AB) | Mach 1.1–1.2 in AB |
| Thrust-to-weight | 0.61 (mil) / 1.04 (AB) | F110-GE-129 ÷ 12.7 t — the same |
| Climb, 400 m → 3,000 m | 35 s (mil) / 16 s (AB) | matches the analysis at that weight |
| Service ceiling | ≈14,900 m (mil) / ≈17,400 m (AB) | around 18,000 m |
| Stall and turn aerodynamics | true to the aircraft | — |
| Structural g limit | +9 / −3 g | the same as the real FBW |
| Roll | rate command, 180°/s | true to the real FBW |
What caps the top speed is not parasitic drag but transonic wave drag, which is zero below Mach 0.78 — so the band you actually fly in passes through untouched.
7. High up, it is a different world
The atmosphere is the real ISA standard model, and thrust, lift, drag and Mach number all read the same air. Sea level passes through untouched (density ratio 1.000), so none of this reaches carrier landings or low-level ingress. It reaches you when you climb.
| Altitude | Stall speed (true) | Wave-drag wall | Width of the band |
|---|---|---|---|
| 0 m | 240 km/h | 956 km/h | 715 |
| 6,000 m | 327 km/h | 889 km/h | 561 |
| 10,000 m | 414 km/h | 841 km/h | 427 |
| 16,000 m | 654 km/h | 829 km/h | 175 |
In thin air you have to fly faster for the same lift. Meanwhile the speed of sound falls with temperature, so the drag wall comes down to meet you. The higher you go, the more the band you can fly in closes from both sides.
Near the 17,000 m ceiling there is barely more than 100 km/h between the stall and the drag wall — and rolling into bank pushes the stall side up into what is left (same reasoning as §1). What real pilots call the coffin corner falls out of this model with nothing added to produce it.
The ceiling comes out of the physics too. Minimum drag is a constant that does not depend on altitude, while thrust falls away with the thinning air; above the single point where those two cross, no choice of speed will sustain level flight. Accelerating to top speed down low and zooming does not get you past it either.
8. The structural g limit — +9 / −3 g
The FBW in the real F-16 and F-2 will not accept a pull that would break the airframe, and neither will this model: load factor is capped at +9 g and −3 g.
It stops you by tightening the rate at which angle of attack may change, in proportion to the g margin left — not by killing the elevator. The pull therefore settles onto the limit rather than bouncing off it or punching through: a full pull from 1,134 km/h peaks at 8.95 g and sits there for five seconds with no spike. Holding a high-g turn takes very nearly full authority, which is why the command is not faded out at the edge.
Load factor is shown above the airspeed box, and turns amber as you close on the limit. Like the real limiter it is silent — there is no over-g warning because there is no over-g — so this is where "I'm pulling and the nose isn't following" becomes visible.
Below corner speed — 725 km/h at sea level and reference weight — it never engages at all, because the angle-of-attack limit binds first. Carrier landings, air-to-air refuelling, the procedure stages and low-level ingress fly exactly as they did before the limiter existed. It bites only above 725 km/h: pulling up after a full-throttle run down low.
With tanks and missiles aboard, the real aircraft’s limit load is set well below 9 g (see F-2 armament §5), but this model holds 9 g whatever is on the wings. What tanks take away is the g you can generate — the extra weight means less load factor at the same speed — not the structural ceiling.
Both of the limits the real FBW puts on the stick — angle of attack and load factor — are now in place. The flip side is that as long as you keep your speed up, it flies like the real aircraft.
Summary
- Stall speed rises the moment you bank. 286 km/h at 45°, 340 km/h at 60°.
- The HUD stall line moves. In a turn, that is the line to watch.
- Don’t try to turn below 450 km/h. Slower means less turn, not more.
- Turn radius bottoms out around 440 m — and opens up again past 725 km/h, where the 9 g limit takes over.
- Wings level comes first in a recovery. But down low there is no second chance.