Terminal Ballistics · Barriers

Can Your Car Actually Stop a Bullet?

Every action movie shows the hero ducking behind a car door as rounds spark off it. The physics says that's a death sentence — a door doesn't stop a 9mm, and a controlled test shows a hollow-point that passes through one actually penetrates deeper. A car hides you. It doesn't protect you.

BallisticEngine · Published

"Get behind the car" is one of the most repeated pieces of screen choreography there is, and it has quietly taught a couple of generations that sheet metal is armor. It isn't. We build a terminal-ballistics engine, so instead of arguing from the movies we'll argue from the test data — controlled gel work, forensic glass studies, and the classic penetration series — and sort every part of a car into the only two categories that matter: the parts that stop bullets, and the parts that only stop eyes.

1. Cover versus concealment

Two words get used interchangeably and shouldn't. Concealment hides you — it breaks line of sight but does nothing to an incoming bullet. Cover physically stops the round. Almost the entire visible body of a car is concealment, and most people's intuition badly overrates it.

Crucially, cover is threat-specific. A barrier that reliably stops a 9mm is cover against that round, but if the same barrier is punched through by 7.62×39, it's only concealment against a rifle. Cover is almost always also concealment; the reverse is rarely true. Games blur this line completely — we got into how shooters model (and mismodel) protection in Game Ballistics vs Real Physics — but in the physical world the distinction is the whole subject.

Cover stops bullets. Concealment only stops eyes. A car is almost all concealment.

2. The door

Start with the movie shot itself. A car door is not a plate — it's two thin sheets of steel with an air gap between them. The outer skin runs about 0.65–0.90 mm and the inner panel about 0.75–1.20 mm; in between sit a window regulator, a speaker, some wiring and foam, and one genuinely thick part — the side-impact beam — which is a narrow target you mostly won't hit.

In 2023, writing for USCCA, Joel Nadler ran the cleanest controlled test of this that exists. A Glock 19X firing Speer Gold Dot 124-grain hollow-points, from 10 feet, straight into a 2012 Chevy Cruze driver's door backed by a block of Clear Ballistics 10% synthetic gel — with a chronograph on every round and a bare-gel baseline for comparison. (Synthetic gel depths aren't directly comparable to FBI-protocol ordnance gelatin, but both conditions used the same medium — so the door/no-door difference stands on its own.) The results are worth stating precisely:

USCCA (Nadler, 2023) — Speer Gold Dot 124gr 9mm JHP, with and without a car door in front of the gel
ConditionImpact velocityExpansionGel penetration
Bare gel (baseline)1,075–1,115 fps0.357″ → 0.5″, intact12–15″
Through the car door~samenone — only slight warping5.5–17.5″ (avg deeper)

Every one of the ten rounds punched through both metal panels. In the gel, none of them expanded — and on average the barely-deformed rounds drove farther than the baseline that never hit a door. On top of that, every shot threw fragments of the door's own steel into the gel, 0.5 to 5.5 inches deep — over 70 percent of them lodging within the first 2 inches. Nadler's conclusion was blunt: a car door "does not provide cover," and "may even be worse than nothing in some ways." The classic Box O' Truth "Buick" series found the same thing years earlier — every 9mm and .40 ball round fired into a driver's door passed into the passenger compartment. A single car door offers no protection against handgun rounds.

The door doesn't just fail to stop the bullet. It converts your hollow-point into an FMJ — and adds shrapnel.

3. The glass

Automotive glass comes in two kinds that behave nothing alike. Side and rear windows are tempered glass (~4 mm), designed to shatter into harmless pebbles. Ballistically they're inconsequential — anything breaks them without touching the bullet, and a .22 will sail through both side windows without slowing. Pure concealment.

The windshield is laminated safety glass — two glass plies bonded to a PVB interlayer (~5–6 mm), required by regulation not to shatter — and it's the one piece of glass that fights back. Glass is brittle but extremely hard, and its fracture edges shear the copper jacket off the lead core, so bullets deform and shed energy; from inside a car, ordinary 9mm loads have been observed nearly disintegrating, peppering the target with non-lethal fragments (bonded designs like Gold Dot hold together better). The windshield's steep rake also deflects rounds: forensic testing finds handgun bullets fired from outside deflect downward about 3° on average, and at a shallow enough angle they simply ricochet — one study logged a clean bounce from a 9mm at a 5° incident angle. Direction depends on the load and the angle, so treat it as "unpredictable but real," not a rule you can aim by. Near perpendicular, though, a defensive round still gets through.

4. The engine block

Now the exception. The engine block is the one part of a modern car that reliably stops bullets — not just the block casting (5–15 mm of iron or aluminum) but the whole stack in front of it: sheet metal, radiator, fan, and accessories. The Box O' Truth's dedicated engine-block test lays out where the line falls:

Box O' Truth #54 — rounds vs a cast-iron engine block (1974 Dodge 360)
RoundResult on the block
.357 Magnum (158gr JHP)could not penetrate — a lead smear
.44 Magnum (240gr JHP)hole in a thin part of the block, did not reach the cylinder
5.56 M193 ball & M855 green-tippenetrated the block wall, but not the cylinder
12 ga slugbusted the block, did not reach the cylinder
.30-06 M2 ballhole in the block wall, but not the cylinder
.30-06 M2 APthrough the block wall AND the cylinder

Read that against the door: the same 5.56 that treats a car door as though it isn't there is stopped short of the cylinder by the engine — even ordinary .30-06 ball stops there — and it takes a purpose-built armor-piercing round to drive all the way through. That's the difference between real cover and everything else on the vehicle.

The engine block is the only part of a modern car that qualifies as cover.

5. The cover rating table

Putting the whole car together — what each part is made of, and which of the two categories it belongs in:

What each part of a car does against common rounds. "Stops" means reliably defeats; a car is otherwise concealment.
ComponentMaterialStops 9mm?Stops 5.56?Cover or concealment
Door0.7 mm + air + 0.9 mm steelNoNoConcealment
Side windowtempered ~4 mmNoNoConcealment
Windshieldlaminated ~5–6 mmPartial (angle-dependent)NoConcealment (mostly)
Floor pansingle ~1.0–1.2 mm sheetNoNoConcealment
Wheel / hubsteel rim + brake + suspensionLikelyMaybeCover (partial)
Engine block5–15 mm iron/aluminum + fluidsYesYes (not AP)Cover

The B-pillars and rockers (1.2–2.5 mm ultra-high-strength steel) are the only other parts with real steel behind them, and even those are narrow and unreliable against anything but a handgun. Everything a person instinctively hides behind — door, glass, trunk, roof — is concealment. Coverage matters here the same way it does for a plate carrier that leaves your sides open, a point we made in How Body Armor Actually Works.

6. The counterintuitive part: why a hollow-point goes deeper

The single most surprising result in the data deserves its own explanation, because it inverts what most people assume. A hollow-point is engineered to expand — to peel open into a mushroom that dumps energy and limits over-penetration. That expansion is triggered by fluid or tissue filling and pressurizing the cavity. A thin steel panel does something different: it slams the cavity shut and strips or deforms the jacket before the bullet ever reaches soft tissue, defeating the expansion mechanism. What crosses the air gap and punches through the inner panel is no longer a hollow-point behaving like a hollow-point — it's a slug behaving like full metal jacket, driving straight and deep.

CAR DOOR — A HOLLOW-POINT ARRIVES AS FMJ outer panel ~0.7 mm air gap inner panel ~0.9 mm intact JHP expansion defeated exits unexpanded + door fragments → deeper
The two-panel door defeats the hollow-point before it reaches tissue: the outer skin plugs the cavity and strips the jacket, so the round crosses the gap and inner panel as an FMJ-like slug — and drags secondary fragments of door steel along with it.

This is the same clogging mechanism that turns a hollow-point into a deep penetrator after drywall, which we walk through in JHP Fails After Drywall — and it's exactly the kind of layered problem our engine is built to model. A single "will it penetrate?" answer misses it; what you want is the residual velocity and behavior at each layer. Impact velocity also sets whether a hollow-point had any chance of opening in the first place, which ties back to What Barrel Length Actually Changes.

7. Four myths, checked against the data

"Duck behind the car door." The Hollywood staple, and the data says it's worse than nothing — the door won't stop a handgun round and may hand you a deeper wound plus steel fragments. Even the engine block, real cover as it is, was described by the Box O' Truth as "a gigantic bullet magnet": in actual gunfights, people who break away from a shot-up vehicle and go prone tend to fare better than those fixated behind it. The Bonnie and Clyde car took 150+ rounds and 112 penetrations — most of them likely rifle fire — and an old, heavy Ford was no protection.

"Shoot the gas tank and it explodes." It doesn't. Tests leave a leaking tank that looks like Swiss cheese; there's no cinematic fireball.

"Newer, heavier cars are safer." Backwards. Modern cars use thinner high-strength steel, not thicker mild steel — a 0.8 mm high-strength door skin can match older 1.2 mm mild steel for dent resistance while giving a bullet even less to chew through.

"If the engine block is cover, the car is cover." No — the block is a small, specific mass. Two feet away, the door, fender, and floor are all just concealment again.

8. Try it yourself

The reason a car door fools people is that "does a bullet go through it?" is the wrong question — the useful question is how much energy actually lands, and in what state the bullet arrives, after each layer. That's a sequential-barrier problem: outer steel, then air, then inner steel, then tissue. Model the exact stack and watch the hollow-point arrive unexpanded. Run the door stack in the calculator — 0.7 mm steel, an air gap, 0.9 mm steel, and a target — and see the residual velocity at each step.

9. FAQ

Will a car door stop a bullet?

No. In a controlled gel test, all ten 9mm hollow-points fired into a modern car door passed through both metal panels, and the classic Box O' Truth series found the same with 9mm and .40 ball. A car door is two thin sheets of steel (about 0.7 mm and 0.9 mm) with an air gap — it's concealment, not cover, against handgun rounds, and rifle rounds pass through it easily.

Is the engine block good cover?

Yes — it's the only reliably bullet-stopping part of a modern car. Testing shows a cast-iron block stops a .357 Magnum entirely and halts 5.56 ball and green-tip short of the cylinder; only a purpose-built armor-piercing rifle round like .30-06 M2 AP drives all the way through. The real barrier is the whole front stack: sheet metal, radiator, fan, and the block itself.

Does hollow-point ammo perform better through car doors?

It performs differently, and in a way that's worse for anyone behind the door. A thin steel panel plugs the hollow cavity and strips the jacket before the bullet reaches tissue, defeating the expansion the bullet is designed for. In the USCCA test, none of the hollow-points expanded after passing through the door, and they averaged deeper penetration than identical rounds fired into bare gel — plus they carried fragments of door steel with them.

Can a windshield deflect bullets?

Yes. A windshield is laminated safety glass set at a steep angle, and forensic testing shows handgun rounds fired from outside deflect downward by about 3° on average, with clean ricochets at shallow incident angles. The glass also tends to shear the jacket off the core, deforming the bullet. The exact deflection direction depends on the load and angle, so it's unpredictable — but near-perpendicular, a defensive round still penetrates.

What part of a car actually provides cover?

Realistically only the engine block, and to a lesser degree the wheel and hub area, where a steel rim, brake, and suspension mass sit in the path. Everything else — doors, all glass, trunk, roof, floor pan — is concealment: it hides you but does not reliably stop an incoming bullet.

Methodology & caveats

The car-door figures are from a controlled, chronographed test by Joel Nadler for USCCA (2023, 2012 Chevy Cruze, Speer Gold Dot 124gr) conducted in Clear Ballistics synthetic gel. Absolute penetration depths in synthetic gel are not directly comparable to FBI-protocol 10% ordnance gelatin — but the bare-gel baseline and the through-door series used the same medium under the same conditions, so the comparison between them is internally controlled and is what this article relies on. Corroboration from The Box O' Truth "Buick" series; the engine-block results are from The Box O' Truth #54, fired at a 1974 cast-iron Dodge 360 — a heavy old block, so treat it as a best case (modern thin-wall and aluminum blocks resist less, and the pistol rounds tested were JHP loads). The windshield-deflection figures are from published forensic ballistics research. Credit to those sources for the underlying data.

We deliberately do not quote "feet per second remaining after each panel" — clean per-layer velocity data isn't published, so the gel-penetration results stand in for it (and are exactly what a multi-layer barrier model computes). Buckshot-versus-door behavior is described qualitatively because we found no controlled pellet-by-pellet test. Engine-block wall thickness varies widely by engine (iron versus aluminum), so it's given as a range. Windshield deflection direction is load- and angle-dependent, not a fixed rule.

This article describes what projectiles do to vehicle materials, for general education. It is not tactical, self-defense, or legal advice; it is physics, not positioning.