Terminal Ballistics · Game Physics

Game Ballistics vs Real Physics

Every cartridge in Escape from Tarkov carries two numbers: penetration and damage. In real terminal ballistics, those two numbers are the same number. Here is where games diverge from physics — and why the divergences are almost always on purpose.

BallisticEngine · Published 23 August 2026

Tarkov's 7.62×39 BP has a penetration value of 47 and a damage value of 58. You can read those as two dials the designers can turn independently — and they are. But a real bullet has no such dials. How deep it drives and how much tissue it destroys are two readings of one event: the kinetic energy it carries and how that energy couples into what it hits. Turn one up and the other moves with it. You cannot buy high penetration and pay for it with low damage, because both are drawn from the same account.

We build a terminal-ballistics engine. Validating it meant spending a lot of time on the gap between what games model and what the physics does, because games are where most people's intuition about penetration actually comes from. This is what we found: where the games are wrong, where they are quietly right, and — the part that matters — why a physics engine and a game engine are solving two different problems.

1. The core abstraction: one number, split in two

A bullet's kinetic energy is the whole budget:

Penetration depth scales with how much of that energy survives to the target and how efficiently the projectile concentrates it — which is governed by sectional density and by whether the bullet stays pointed or deforms. Wound severity scales with how much energy is deposited along the path. Both quantities climb together with retained energy. In physics there is no lever that raises penetration while lowering damage; to change one you must change the projectile — its mass, velocity, or construction — and that moves the other too.

A competitive shooter cannot live with that coupling. If penetration and damage both rose with energy, the highest-energy cartridge would dominate every axis at once — best against armor and best against flesh — and the entire ammo economy would collapse into "buy the biggest round you can afford." So the game breaks the coupling on purpose and exposes two independent sliders. That is what lets Tarkov build a rock-paper-scissors economy: cheap, high-damage rounds that shred unarmored targets but bounce off plates, and expensive armor-piercing rounds that punch through armor while doing modest damage on the other side. The split is not a physics error. It is the mechanism that makes the game a game.

In physics, penetration and damage are two views of one event. In a game, they're two sliders on a balance spreadsheet.

2. A spectrum of games

"Realistic ballistics" is not one thing. A game can simulate the bullet's flight beautifully and then reduce everything at the target to a coin flip, or vice versa. Ranked roughly from most arcade to most simulated:

Two patterns fall out. First, flight and terminal behavior are decoupled: PUBG has real projectile drop but no penetration concept at all, while Ground Branch simulates the bullet from the muzzle but publishes nothing about expansion or fragmentation. Second, nobody in this list models wounding as energy deposition. Every title reduces the target interaction to an integer, a probability, or a flat percentage — because computing residual energy through tissue for every hit in a networked shooter is expensive, and players do not need it to make decisions.

3. What games simplify — or drop

a. Expansion doesn't exist in Tarkov

Real hollow-point expansion is velocity-gated. Below its design velocity a JHP may not open at all — you have paid a premium for FMJ behavior. The FBI's test protocol, set in 1988 after the 1986 Miami shootout, asks for 12–18 inches of penetration in calibrated 10% ordnance gel with roughly 1.5× expansion. A representative modern load — Federal HST 124gr +P — runs about 18.28 inches with a recovered diameter near 0.606 inch. Drive the same bullet slower, or plug its cavity with drywall, and it behaves like ball.

Tarkov models none of this. No round changes diameter; "JHP-like" cartridges are just high-flesh, low-penetration stat blocks with no velocity threshold, because expansion is never simulated in the first place. That is a defensible cut — deformation as a function of impact velocity and medium is costly to compute and would make a round's behavior range-dependent, which complicates balance.

b. Impact angle is ignored for armor

Obliquity is one of the largest real effects there is. The material a bullet must defeat scales with the secant of the impact angle:

Past a critical angle the round stops perforating and starts skipping — which is why police doctrine positions officers at around 30° of obliquity to promote deflection. Tarkov states outright that the angle of impact does not influence penetration chance against body armor; it computes ricochet only for helmets, where a bounced round is the most visible and satisfying. Per-hit angle math against armor geometry for every shot fired online is expensive, and BSG spent that budget where players would notice it.

c. Fragmentation is a coin flip, not a threshold

In reality, fragmentation is a near-deterministic, velocity-gated event. The canonical example is 5.56 M193: it fragments reliably above roughly 2,700 fps and unreliably below about 2,500 fps. The same bullet that shatters at the muzzle stays intact at distance, because it has dropped below the threshold. Tarkov assigns each round a fixed fragmentation chance that does not vary with velocity or range — and, thanks to a well-documented quirk, rounds with a penetration value under about 20 will not fragment against players at all. A static probability is trivially cheap; a velocity threshold would make identical rounds behave differently at 10 and 200 meters, which is realistic but adds another axis to balance.

d. Sectional density does not exist

Sectional density — mass over frontal area — is one of the parameters that actually governs penetration:

All else equal, the projectile with the greater sectional density penetrates deeper — a heavy-for-caliber .308 175gr match bullet sits near SD 0.264 and drives correspondingly deep. Games do not have this concept. Penetration is an abstract integer; a cartridge's mass and diameter are flavor text on the item card, not inputs to an equation. ACE3 is a partial exception, since its ballistic coefficients and Miller stability factor correlate with sectional density, but no mainstream title lets SD drive terminal penetration directly. We pull this thread all the way through in Sectional Density, Explained.

e. Drywall is a binary wall; reality is continuous energy loss

Every common pistol caliber passes through a standard interior wall. Drywall barely resists a centerfire handgun round, and it clogs hollow points — defeating expansion and, perversely, letting the bullet penetrate farther than it would have as a JHP. Even .22 LR from a pistol has been documented punching through roughly ten drywall layers. There is no such thing as a non-penetrable interior wall for a defensive handgun cartridge.

Our own engine agrees. A 9mm Wilson Bulwark load driven through one standard wall exits at 1058 fps still carrying 302 ft·lbf — the wall absorbed about 8% of its energy. Across four pistol calibers put through the same wall, exit velocities cluster within about 50 fps of one another. Reality is a shallow, continuous energy tax; the drag relationship behind that decay is the same one we derive in How Far Do Bullets Travel Underwater. Tarkov and most games collapse all of that to a single question — does the round's penetration value clear the surface or not — because a binary is cheap to compute and readable at a glance. A player needs to know "can I shoot through this wall," not "how many foot-pounds land on the other side."

4. The clearest inversion: the leg meta

Tarkov has no leg armor. Limbs are unprotected against every round in the game, so a leg shot bypasses the entire armor system — penetration becomes irrelevant, and only flesh damage and fragmentation decide the outcome. The result is the infamous "leg meta": low-penetration, high-damage rounds are optimal into legs. A 9mm RIP, with a penetration value of 2 and a flesh damage of 102, beats a 7.62×39 BP (penetration 47, damage 58) when both are fired into an unarmored leg.

Physics says the opposite. Leg-wound severity is a function of velocity, mass, and expansion or fragmentation — real wound-channel mechanics — not a flat damage integer. And the energy gap is enormous:

The 7.62×39 round carries roughly four times the kinetic energy of the 9mm and is far more likely to fracture the femur or sever the femoral artery. In physics, it does more to a leg on every axis that matters. In Tarkov, it loses to a fragmenting pistol round — because "flesh damage" is a balance stat divorced from energy. This is the cleanest case in any of these games where the logic inverts the physics: the round that wins the leg meta is the one that would do the least to a real leg. And that inversion is correct game design. RIP is a cheap, niche, situational cartridge precisely because it is useless against armor; the leg meta is where it earns a reason to exist. Physics has no balance to protect and so hands the win to the rifle.

The round that wins Tarkov's leg meta is the one that would do the least to a real leg. That isn't a bug — it's the balance working.

5. Armor: a probability roll vs a hard threshold

Tarkov resolves armor as a probabilistic roll. Penetration is checked against the plate's class and its remaining durability, with a rule of thumb that the penetration value needed is about ten times the armor class. A 5.45 BP round (penetration 45) against a Class 4 plate (roughly a 40 threshold) penetrates reliably but not deterministically — and as the plate's durability falls, the penetration chance climbs. Every hit, whether it penetrates or not, degrades that durability by at least one point. The behavior is a smooth curve: armor is a consumable that wears down over a firefight.

Reality, at the soft-armor boundary, is a cliff. Soft body armor rated to NIJ IIIA is defeated by essentially any rifle round as though the armor were not there — the fibers shear and the bullet passes. There is no 60% chance; it is a categorical yes. Against a hard plate the outcome depends on the plate compound, impact velocity, and angle — a genuine Level III plate stops both 5.45 and 7.62×39 ball, while the IIIA soft panel next to it does not even slow them. The defeat criterion is a threshold, not a die roll.

Why model armor as a probability at all? Because durability-as-a-resource is good gameplay — armor becomes a consumable that attrites in a fight — and a smooth curve avoids the feel-bad of a hard binary where a single durability point flips a firefight from survivable to lethal. Reality has no durability-as-gameplay to protect, so it keeps the cliff.

6. What games get right

The critique cuts one way; the credit cuts the other. Several things Tarkov does are genuinely faithful:

It uses a real projectile model, not hitscan. Rounds have muzzle velocity, gravity drop, air-friction deceleration, travel time, and ricochet. They lose both damage and penetration as velocity bleeds off with range — the game even defines an "effective distance" as the range at which a round sheds 25% of its damage and penetration. That is genuine exterior ballistics, and it is the part most games skip.

Durability degradation is physically motivated. Real armor is damaged by impacts, and its protective capacity falls as damage accumulates — multi-hit-limited by design. Having penetrating rounds do slightly less durability damage than non-penetrators is even a reasonable abstraction: a round that punches clean through deposits less of its energy in the plate.

Material destructibility is ordered correctly. In-game, ceramic has the worst destructibility (it shatters) and aramid the best, with UHMWPE and steel between — matching the real world, where ceramic plates are brittle and multi-hit-limited for exactly that reason.

The coarse ordering holds. Rifle rounds defeat armor that stops pistol rounds; bigger, faster cartridges penetrate more. Tarkov gets the first-order physics right. It is the fine structure — angle, expansion, velocity-gated fragmentation, sectional density — that is abstracted away.

Tarkov's bullets fly like real bullets. It's what happens at the target that's been turned into a spreadsheet.

7. Run your own tests

Our engine computes residual velocity and energy through materials from first principles — drag, material properties, and projectile geometry — rather than a lookup table or a probability roll. The wall result from the drywall section is one worked example; here it is alongside the placeholders for the rest of the four-caliber comparison, which need real engine output dropped in.

Pick a cartridge, a barrier, and a thickness, and watch the residual energy come out — no die roll, no penetration integer, just the physics. Run your own tests in the calculator.

8. FAQ