Anti-dive and anti-squat: side-view geometry earns its keep

You can stop a car diving under braking without a single extra newton of spring — by tilting the arms in side view. Here's where the percentage comes from, and why 100% is almost always the wrong answer.

Every chapter so far has lived in the front view. Time to walk round the car and look at it from the curb, because the side view controls something the front view can’t touch: what happens when you brake or accelerate.

Hit the brakes and two separate things happen. They get conflated constantly, and the whole chapter falls apart if you don’t keep them apart.

One: load transfers forward. How much is fixed by four numbers and nothing else:

\Delta W = \frac{m \cdot a \cdot h_{\text{cg}}}{L}

Mass, deceleration, CG height, wheelbase. Geometry has no vote in this. None. Your front tires get that much extra load however the car is built.

Two: the nose goes down. This one is not a consequence of the first. It’s a consequence of how that load gets into the chassis — and that is entirely a geometry question.

Anti-squat is the same idea at a driven rear axle under acceleration, with the sign flipped.

The nose drops because the spring had to shorten

A spring can only produce force by changing length. That’s the definition: $F = k \cdot x$ . If the front spring has to carry an extra $\Delta W$ , it has no choice but to compress by $\Delta W / k$ in order to generate it.

That compression is the dive. The nose isn’t dropping because load transferred; it’s dropping because the spring had to shorten to react the load that transferred.

So the question becomes narrow and answerable: can anything else push up on the front of the chassis at the same time, so the spring has less left to carry? If something can, the spring compresses less and the nose drops less — at identical tire loads, because the load transfer itself has already been settled by mass, deceleration, CG height and wheelbase, and nothing in the suspension gets a vote.

Something can. It’s the braking force itself, pointed slightly upward.

What tilting does: it aims the braking force

Braking puts a rearward force at the contact patch — call it $F_b$ . That force has to reach the chassis, and the path is the control arms.

An arm pinned at both ends is a two-force member: it can only carry load along its own length. So if an arm lies horizontal, the force it passes to the chassis is horizontal, and none of it is available to hold the nose up. Tilt the same arm in side view and the force it carries is now tilted too — same magnitude, different direction, and a share of it now points up.

That is the whole mechanism, and it is worth being precise about what it is not. It has nothing to do with the arms being stiff. The arms are exactly as rigid at 0% anti-dive as at 100% — rigidity doesn’t change between the two cases, so it can’t be what distinguishes them. What changes is the direction the load path points, and therefore how the braking force resolves. Anti-dive is a projection, not a stiffness.

(Rigidity does earn a footnote, but as an assumption rather than a mechanism: the construction below treats arms and bushings as inextensible. Real rubber deflects under exactly these loads, which is why a car’s measured anti-dive comes in below its drawn value.)

Two arms, one angle

With two arms you have two directions, so which one is the angle?

Start with what braking does to the pair. The rearward force acts at the road, well below both ball joints, so it doesn’t just push the upright back — it tips it. The upper joint is driven forward and the lower one rearward, which loads one arm in tension and the other in compression.

That tipping couple is the part most descriptions stop at, and on its own it produces no anti-dive at all. Two equal and opposite arm forces resolve to equal and opposite vertical components, which cancel. What lifts the nose is the net rearward force the two arms share between them — the resultant, not the couple.

So the angle you want is the angle of that resultant, and geometry hands it to you: the resultant of two link forces must pass through the point where their axes intersect. That point is the side-view instant center, and the line from the contact patch to it is the line the combined force acts along. This is the whole reason the IC construction exists — it’s the shortcut for adding up two arms’ worth of projection without resolving each one separately.

Two consequences fall straight out:

Arms horizontal. Both carry purely horizontal load, so their resultant is horizontal too. The IC sits infinitely far away on the horizon, the line to it is flat, and nothing points up. The spring carries every newton of $\Delta W$ , compresses, and the nose dives. This is the strut preset, at 0%.
Arms tilted fore/aft. Their resultant now leans, and the line from the contact patch to the IC is inclined at some angle $\theta$ . The linkage still has to react the same horizontal $F_b$ — but it can only do so along that inclined line, so the force it actually carries is

F_{\text{link}} = \frac{F_b}{\cos\theta} \qquad\text{with a vertical component of}\qquad F_b \cdot \tan\theta

That vertical component pushes up on the front of the chassis. It is the anti-dive force, and it appears for free — you didn’t add a spring, you tilted a bracket.

Two side views of the same corner. Flat arms never meet, so there is no instant center and no anti-effect — that is the strut preset, at 0%. Tilt them fore/aft and they meet at a point; the angle from the contact patch up to it is what the anti-dive formula measures.

Build anti-dive and re-run a stop — strut preset →

Where the percentage comes from

Anti-dive percentage is those two quantities divided by each other: how much the linkage lifts, against how much the spring would otherwise have had to swallow.

\text{anti-dive \%} = \frac{F_b \cdot \tan\theta}{\Delta W} \times 100

Substitute both. The braking force at this axle is the total times that axle’s share of the braking, $F_b = \beta \cdot m \cdot a$ , and $\Delta W = m \cdot a \cdot h_{\text{cg}} / L$ . The $m \cdot a$ cancels top and bottom:

\text{anti-dive \%} = \tan\theta \cdot \frac{L}{h_{\text{cg}}} \cdot \beta \times 100

The cancellation is worth a moment. Deceleration dropped out entirely, which means anti-dive percentage is a property of the car, not of how hard you are stopping. A 30% anti-dive front is 30% in a gentle stop and 30% in an emergency one. What changes is how many newtons 30% represents.

So 100% anti-dive means the linkage’s upward reaction exactly equals the download the spring would have taken. The spring sees no change in load, does not compress, and the nose does not move at all. 50% means the springs handle half of it. 0% means they handle everything.

The strut layout in the app reads 0% anti-dive, for a straightforward reason: its arms have no fore/aft inclination at all, so $\theta$ is zero and so is $\tan\theta$ . The advisor says so directly — “the arms have no fore/aft inclination in side view, so the suspension neither resists nor adds pitch.”

That makes it a clean blank sheet. Everything below is something you build in yourself and then measure, which is a better way to learn this one than reading a number off a preset.

Yes, you can exceed 100% — and you’ll do it by accident

Drag a bracket around in the app and you will see numbers well past 100. That is not a bug, and it is the clearest way to understand what the percentage actually measures.

Above 100%, the linkage pushes up harder than the load transfer pushes down. The net effect at the front is upward, so under braking the nose rises. The car pitches backwards as you slow, which is as unpleasant as it sounds and completely wrong for road feel.

Worse than the feel is what it does to the tire. Over-100% geometry jacks the suspension into droop under braking, and the moment the wheel unloads the jacking force falls away, the suspension drops back, the wheel re-loads, and it jacks again. That cycle is wheel hop under braking — a well-known signature of too much anti-dive, and a fast way to lose the braking you were trying to improve.

It is easy to hit, because the multiplier in front of $\tan\theta$ is large. On this car $L / h_{\text{cg}} = 2467 / 470 = 5.25$ , and with the front doing about 65% of the braking that is $5.25 \times 0.65 = 3.41$ . Every degree of side-view inclination is worth a lot of percent:

Side-view angle θ	Anti-dive
5.0°	30%
8.3°	50%
16.3°	100% — the nose stops moving
23.7°	150%
30.4°	200% — the nose lifts

Sixteen degrees is not a dramatic-looking tilt on screen, which is exactly why it sneaks up on you. If the app reads 250%, your arms are steeper in side view than you think. (It stops reporting at ±500%: a side-view IC almost directly above the contact patch sends $\tan\theta$ toward infinity, and that is a degenerate layout rather than a meaningful number.)

Note what a low CG does to that multiplier, too. Halve $h_{\text{cg}}$ and you double the anti-dive from the same geometry — a low car needs less arm inclination for the same percentage, because there was less pitch moment to fight in the first place.

Why 100% is the wrong answer

If anti-dive is free suspension travel and free pitch control, why doesn’t everything run 100%?

Because the force has to go somewhere. Anti-dive works by feeding braking load through the control arms into the chassis instead of into the spring. At 100%, every newton of that transfer arrives through the bushings. The result is a car that transmits every longitudinal impact straight into the structure and the driver — brake over a bump and it feels like hitting something.

Second, and worse for a driver: dive is information. Pitch is one of the main cues telling you how hard you’re braking and how much grip is left. A car with no dive at all feels numb and disconnected, and drivers routinely brake less confidently in it.

Third, high anti-dive means the geometry is strongly inclined, which couples braking into the wheel’s vertical motion. The suspension gets stiff in exactly the moment it also needs to absorb road inputs — the wheel is fighting two jobs with one linkage.

So the honest answer is that anti-dive is a seasoning, not a main ingredient. 20–40% at the front removes the worst of the nose-dive while leaving the pitch cue intact.

Go look at it

Open the strut layout.

Find Anti-Dive in the GEOMETRY METRICS panel. It reads 0%, and the Setup Advisor explains why: no fore/aft inclination in side view.

Switch to the SIDE VIEW panel. This is the plane the whole chapter lives in, and the arms here are flat.

Build some. Drag the lower arm’s inner pivots so the front and rear bushings sit at different heights — that height difference is the tilt, and the tilt is the mechanism. Watch Anti-Dive climb off zero as you go.

Aim for roughly 30% and stop there. That’s a realistic road-car figure.

Now feel it. Open the Brakes tab and run a stop, noting the front dive. Set anti-dive back to 0% and run the identical stop again. Same braking event, different amount of nose-down — and identical tire loads, for the reason below.

Then overshoot deliberately. Keep tilting past 30% until the readout passes 100%, and run the stop again. The nose now lifts instead of dropping. That is what the percentage is measuring.

What it costs you

Beyond harshness and lost feel, there’s a packaging cost that catches people out: creating anti-dive means the inner pivots sit at different heights fore and aft, which turns a simple arm into a more complex bracket and puts a twisting load into the bushings. On a road car this shows up as bushing wear; on a kit car it shows up as a bracket that cracks.

There’s also a hidden interaction with chapter 6. Tilting the arms fore/aft changes the path the wheel takes, which changes where the tie rod needs to be to avoid bump steer. Adding anti-dive to a car that had good bump steer will usually cost you some of it.

And one genuinely counterintuitive point: anti-dive doesn’t reduce weight transfer. Not by a newton. The load on the front tires under braking is set by mass, deceleration, CG height and wheelbase, and geometry has no vote (that’s chapter 18). All anti-dive changes is the path the force takes and therefore how much the body moves. A car with 100% anti-dive has exactly the same tire loads as one with 0%.

Rules of thumb

Front anti-dive, road car: 20–50%. Enough to calm the nose, not enough to numb it.
Rear anti-squat, road car: 10–30%.
Circuit car: often lower than road cars — 0–20% — because drivers want the pitch cue and the platform is already stiff enough that dive is small.
Drag racing: the exception that proves the rule. Very high anti-squat is used deliberately to plant the rear tires on launch, accepting every downside because the car only has to do one thing.
Rally-raid and desert: modest anti-dive, and for an unusual reason — with 350–600 mm of travel the arm angles sweep through an enormous range, so an anti-dive figure quoted at ride height may be nothing like the value at full compression. As with roll centers, design the envelope rather than the number.

Try this

Build 30% anti-dive into the strut layout. What did it do to the bump-steer figure from chapter 6?
Push it to 100%. Run the brake sim. Is the reduced dive worth what you can infer about the harshness?
Get anti-squat on the rear axle to about 25%. What had to happen to the side-view instant center?

Next: the lever between your spring and your wheel — and the reason a 400 lb/in spring can behave like a 200 lb/in one.

Preset hardpoints are illustrative sketches. Anti-dive figures here assume a 500 mm CG height and a 65/35 brake split, which is what the app assumes too.